Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Interactions between cytochrome P450 genetic polymorphisms and plasma organochlorines in non-Hodgkin… Dhalla, Anar 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2014_september_dhalla_anar.pdf [ 3.1MB ]
Metadata
JSON: 24-1.0107219.json
JSON-LD: 24-1.0107219-ld.json
RDF/XML (Pretty): 24-1.0107219-rdf.xml
RDF/JSON: 24-1.0107219-rdf.json
Turtle: 24-1.0107219-turtle.txt
N-Triples: 24-1.0107219-rdf-ntriples.txt
Original Record: 24-1.0107219-source.json
Full Text
24-1.0107219-fulltext.txt
Citation
24-1.0107219.ris

Full Text

i   INTERACTIONS BETWEEN CYTOCHROME P450 GENETIC POLYMORPHISMS AND PLASMA ORGANOCHLORINES IN NON-HODGKIN LYMPHOMA  by  Anar Dhalla  B.Sc., Simon Fraser University, 2008   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE  in  The Faculty of Graduate and Postdoctoral Studies  (Health Care and Epidemiology)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August 2014   © Anar Dhalla, 2014   ii Abstract  Following World War II, the production of chlorine-containing organic molecules known as organochlorines (OCs) became commonplace. Although heightened regulation has since occurred in Canada, OCs continue to pose health and environmental concern s due to their persistence and ongoing use elsewhere.  To date, the largest study assessing the association between plasma OCs and non-Hodgkin lymphoma (NHL) was conducted in British Columbia (BC) between 2000 and 2004. Eight hundred twenty-eight newly diagnosed NHL cases were ascertained from the BC Cancer Registry and 848 population controls were randomly obtained from the BC Ministry of Health Client Registry. Significant associations were observed between NHL and 14 individual organochlorines.  Because the effects of OCs on NHL may be modified by an individual’s genetic makeup, gene-environment (GxE) interactions between OCs and cytochrome P450 (CYP) genes were examined here, using data from the subset of participants of European ethnic origin (“Europeans”) in the BC study. CYPs were chosen because they are involved in the metabolism of chemicals, including some OCs.  First, the effects of the OCs were re-evaluated in Europeans only, as this was the analytic group in subsequent genetic analyses. Significant trends were noted for polychlorinated biphenyls (PCBs) 153, 180, 187, summed PCBs (total, dioxin-like, non-dioxin-like), beta-hexachlorocyclohexane (β-HCCH), hexachlorobenzene (HCB), mirex, trans-nonachlor, and oxychlordane. Significance was maintained after controlling for multiple testing for PCB-187, total summed PCBs, β-HCCH, HCB, trans-nonachlor, and oxychlordane. No significant trends were found for PCB-28, 99, 105, 118, 138, 156, 170, 183, cis-nonachlor, p, p’-DDT, or p, p’-DDE. Secondly, 129 single nucleotide polymorphisms (SNPs) in 18 CYP genes were selected for study. Significant trends were noted for rs743572 (CYP17A1) and rs1322179 (CYP2C19). Significance was not maintained after controlling for multiple testing. Two iii SNPs, rs9332197 and rs10509679 (CYP2C9), in the same gene cluster as CYP2C19 were of borderline non-significance.   A significant GxE interaction was found between rs743572 and mirex, even after controlling for multiple testing. The increased risk of NHL conferred by higher mirex levels is lessened in minor allele homozygotes of rs743572. As CYP17A1 is involved in steroid metabolism, these results suggest the involvement of hormonal modulation in NHL risk.  iv Preface  This thesis uses data from a study on risk factors for non-Hodgkin lymphoma undertaken by the team of Dr. John J. Spinelli at the British Columbia Cancer Research Centre (BCCRC). Research was approved by the University of British Columbia - British Columbia Cancer Agency Research Ethics Board [certificate #H98-61471 Genetic and Environmental Risk Factors for Non-Hodgkin's Lymphoma]. Study design, including selection of genes, genetic variants and organochlorines, and data collection were conducted by this team prior to Anar Dhalla’s involvement. A. Dhalla carried out the literature review in Chapter 2. A. Dhalla also carried out the quality control procedures described in 3.3.6.2, the statistical analyses described in Section 3.4 and the occupational coding described in Section 3.4.1.4. A. Dhalla carried out almost all of the analyses in 3.7 and all the analyses in 4.4 and Chapter 6. Exceptions to this were the quality control processes for cleaning the Illumina genotyping data and the organochlorine data and the preliminary quality control processes for the Taqman genotyping data described in 3.7. More specifically, the quality control processes for the Illumina genotyping data were carried out by the team at BCCRC under the direction of Dr. Denise Daley at St. Paul’s Hospital in Vancouver. A. Dhalla’s contribution was limited to observing this process. A. Dhalla carried out subsequent quality control processes for the Taqman genotyping data using the methodology developed for the Illumina assay.  The statistical methods used in this thesis were largely adapted from previously published articles and theses. The quality control processes and statistical methods are described in these works:  Spinelli JJ, Ng CH, Weber JP, Connors JM, Gascoyne RD, Lai AS, et al. Organochlorines and risk of non-Hodgkin lymphoma. Int J Cancer 2007; Dec 15;121(12):2767-75.  v Ng CH, Janoo-Gilani R, Sipahimalani P, Gallagher RP, Gascoyne RD, Connors JM, et al. Interaction between organochlorines and the AHR gene, and risk of non-Hodgkin lymphoma. Cancer Causes Control 2010; Jan;21(1):11-22.  Ng C. Plasma organochlorines, interaction between the Aryl hydrocarbon receptor gene and organochlorines, and risk of Non-Hodgkin lymphoma 2007; 05.  Schuetz JM, Daley D, Graham J, Berry BR, Gallagher RP, Connors JM, et al. Genetic variation in cell death genes and risk of non-Hodgkin lymphoma. PLoS One 2012;7(2):e31560.  Schuetz JM. Genetic variation in lymphocyte life and death genes and risk of non-Hodgkin lymphomas 2011; 08.     vi Table of contents Abstract ........................................................................................................................... ii Preface............................................................................................................................ iv Table of contents ............................................................................................................ vi List of tables ..................................................................................................................... ix List of figures ................................................................................................................... xi List of symbols ................................................................................................................ xii List of abbreviations ....................................................................................................... xiii Acknowledgements ........................................................................................................ xvi Dedication..................................................................................................................... xvii Chapter 1 Introduction ................................................................................................. 1 1.1 Rationale ............................................................................................................ 1 1.2 Thesis objectives ................................................................................................. 4 1.3 Outline ................................................................................................................ 8 Chapter 2 Background .................................................................................................. 9 2.1 Background on NHL ............................................................................................. 9 2.1.1 Cancer .......................................................................................................... 9 2.1.2 Non-Hodgkin lymphoma ............................................................................. 11 2.1.2.1 Classification ........................................................................................ 11 2.1.2.2 Epidemiology of NHL ............................................................................ 12 2.1.2.3 NHL risk factors .................................................................................... 14 2.2 Background on organochlorines ........................................................................ 17 2.2.1 Organochlorines ......................................................................................... 17 2.2.2 Literature review of associations between OCs and NHL ............................. 25 2.2.2.1 Parent study (British Columbia non-Hodgkin lymphoma case-control study)  ............................................................................................................ 28 2.3 Background on CYPs .......................................................................................... 30 2.3.1 Cytochrome P450 genes and protein products ............................................ 30 2.3.2 Literature review of associations between  CYPs and NHL ............................ 33 2.4 Chapter synopsis ............................................................................................... 37 Chapter 3 Study design and methods .......................................................................... 39 3.1 Study design ..................................................................................................... 39 3.1.1 Study population ........................................................................................ 39 3.1.2 Cases .......................................................................................................... 40 3.1.3 Controls...................................................................................................... 45 3.2 Organochlorine measurement methods ............................................................. 46 vii 3.2.1 Organochlorine measurement .................................................................... 46 3.2.2 Organochlorine quality control ................................................................... 46 3.2.3 Organochlorine lipid-adjustment ................................................................ 46 3.3 Genetic measurement methods......................................................................... 47 3.3.1 Gene selection ............................................................................................ 47 3.3.2 TagSNP selection ........................................................................................ 48 3.3.3 DNA extraction and whole genome amplification ........................................ 48 3.3.4 Sequencing ................................................................................................. 49 3.3.4.1 Taqman ................................................................................................ 49 3.3.5 Genotyping ................................................................................................. 49 3.3.5.1 Illumina................................................................................................ 49 3.3.5.2 Taqman ................................................................................................ 50 3.3.6 Genetic quality control ............................................................................... 50 3.3.6.1 Illumina................................................................................................ 50 3.3.6.2 Taqman ................................................................................................ 52 3.4 Statistical analyses ............................................................................................ 54 3.4.1 Organochlorine analyses ............................................................................. 54 3.4.1.1 Descriptive statistics ............................................................................ 54 3.4.1.2 Organochlorine categorization ............................................................. 55 3.4.1.3 Descriptive organochlorine analysis...................................................... 55 3.4.1.4 Occupational analyses .......................................................................... 56 3.4.1.5 Organochlorine main effects ................................................................ 60 3.4.2 Genetic analyses ......................................................................................... 63 3.4.2.1 Descriptive statistics ............................................................................ 63 3.4.2.2 Genetic main effects ............................................................................ 63 3.4.3 Gene-environment interaction analyses ...................................................... 68 3.4.3.1 Gene-environment models ................................................................... 68 3.4.3.2 Multiple comparisons ........................................................................... 69 3.5 Power calculations ............................................................................................ 69 3.6 Ethical considerations ....................................................................................... 72 3.7 Chapter synopsis ............................................................................................... 72 Chapter 4 Results ....................................................................................................... 74 4.1 Organochlorine analyses ................................................................................... 74 viii 4.1.1 Organochlorine quality control ................................................................... 74 4.1.2 Descriptive statistics ................................................................................... 74 4.1.3 Descriptive organochlorine analysis ............................................................ 77 4.1.4 Occupational analyses ................................................................................ 95 4.1.5 Organochlorine main effects ..................................................................... 100 4.1.5.1 Confounder selection ......................................................................... 100 4.1.5.2 Final organochlorine models in NHL ................................................... 102 4.1.5.3 Interactions with covariates ............................................................... 106 4.1.5.4 NHL subtypes ..................................................................................... 106 4.2 Genetic analyses ............................................................................................. 108 4.2.1 Genetic quality control ............................................................................. 108 4.2.1.1 Illumina.............................................................................................. 108 4.2.1.2 Taqman .............................................................................................. 109 4.2.2 Descriptive statistics ................................................................................. 110 4.2.3 Genetic main effects ................................................................................. 113 4.2.3.1 Genetic models in NHL ....................................................................... 113 4.2.3.2 Interactions with covariates ............................................................... 117 4.2.3.3 NHL subtypes ..................................................................................... 117 4.3 Gene-environment interaction analyses .......................................................... 119 4.4 Chapter synopsis ............................................................................................. 122 Chapter 5 Discussion ................................................................................................ 123 Chapter 6 Conclusion ................................................................................................ 136 6.1 Summary of results ......................................................................................... 136 6.2 Strengths and limitations ................................................................................ 136 6.3 Implications .................................................................................................... 140 6.4 Future directions ............................................................................................. 141 Bibliography ................................................................................................................. 142 Appendices .................................................................................................................. 174 Appendix A Search strategy for literature review of associations between CYPs and NHL  .......................................................................................................... 174 Appendix B Supplementary organochlorine tables ................................................ 178 Appendix C Supplementary genetics tables ........................................................... 191   ix List of tables  Table 3.1  Genetic models .............................................................................................. 66 Table 3.2  Minimum detectable odds ratios for main effect of OCs and risk of NHL ......... 70 Table 3.3  Minimum detectable odds ratios for main effect of SNPs and risk of NHL ........ 70 Table 3.4  Minimum detectable odds ratios for gene-environment interactions .............. 71 Table 4.1  Characteristics of European cases and controls with organochlorine measurements included in analyses [frequency (percentage)] ........................................ 75 Table 4.2  Organochlorine measurements above the detection limit in blood plasma ...... 78 Table 4.3  Spearman rho for PCBs/ summed PCB measures in European controls ............ 79 Table 4.4  Spearman rho for pesticides/ pesticide metabolites in European controls  ....... 80 Table 4.5  European lipid-adjusted organochlorine concentrations (μg/kg) by case/ control group and p-values for the Wilcoxon rank-sum tests ....................................................... 81 Table 4.6  European lipid-adjusted organochlorine concentrations (μg/kg) by age group and p-values for the Wilcoxon rank-sum tests ................................................................ 83 Table 4.7  European lipid-adjusted organochlorine concentrations (μg/kg) by sex and p-values for the Wilcoxon rank-sum tests .......................................................................... 85 Table 4.8  European lipid-adjusted organochlorine concentrations (μg/kg) by region and p-values for the Wilcoxon rank-sum tests .......................................................................... 87 Table 4.9  European lipid-adjusted organochlorine concentrations (μg/kg) by body mass index group and p-values for the Wilcoxon rank-sum tests ............................................. 89 Table 4.10  European lipid-adjusted organochlorine concentrations (μg/kg) by education group and p-values for the Wilcoxon rank-sum tests ....................................................... 91 Table 4.11  European lipid-adjusted organochlorine concentrations (μg/kg) by fam ily history of NHL and p-values for the Wilcoxon rank-sum tests .......................................... 93 Table 4.12  Farming history and 1st usual occupations of European cases and controls with organochlorine measurements included in analyses [frequency (percentage)] ................ 96 Table 4.13  European lipid-adjusted organochlorine concentrations (μg/kg) by occupational OC exposure group and p-values for the Wilcoxon rank-sum tests .............. 98 Table 4.14  Confounders included in final logistic regression models of the associations between organochlorines and NHL in Europeans based on the change-in-estimate criterion ....................................................................................................................... 101 Table 4.15  Lipid-adjusted organochlorine associations with non-Hodgkin lymphoma ... 103 Table 4.16  Associations between PCB-99 and NHL subtypes ......................................... 107 Table 4.17  Characteristics of European cases and controls with genotyping measurements included in analyses [frequency (percentage)] .............................................................. 111 Table 4.18  Various genetic models for CYP SNPs selected for gene-environment interaction analyses ..................................................................................................... 116 Table 4.19  Additive associations for CYP SNPs with heterogeneity between NHL subtypes .................................................................................................................................... 118 Table 4.20  Likelihood ratio test p-values for CYP SNPs and organochlorines tested for gene-environment interactions .................................................................................... 120 x Table 4.21  Odds ratio estimates for interaction between rs743572 (CYP17A1) and mirex in Europeans .................................................................................................................... 121 Table B.1  Organochlorine half-lives ............................................................................. 178 Table B.2  Possible organochlorine exposed occupational groups with corresponding Standard Occupational Classification codes (1980) and number of European participants with organochlorine measurements included in analyses reporting as 1st usual occupation .................................................................................................................................... 181 Table B.3  Organochlorine concentrations (μg/L plasma) by sex in European controls with organochlorine measurements included in analyses ..................................................... 183 Table B.4  Organochlorine concentrations (μg/L plasma) by sex in European cases with organochlorine measurements included in analyses ..................................................... 185 Table B.5  Lipid-adjusted organochlorine concentrations (μg/kg) by sex in European controls with organochlorine measurements included in analyses ................................ 187 Table B.6  Lipid-adjusted organochlorine concentrations (μg/kg) by sex in European cases with organochlorine measurements included in analyses .............................................. 189 Table C.1  Codominant associations between CYP SNPs and NHL with additive p-trends (dominant associations shown when minor allele homozygotes <10) ............................ 191 Table C.2  Codominant associations between CYP SNPs and NHL ................................... 203 Table C.3  Additive associations between CYP SNPs and NHL ......................................... 215 Table C.4  Dominant associations between CYP SNPs and NHL ...................................... 219 Table C.5  Recessive associations between CYP SNPs and NHL....................................... 223 Table C.6  Dominant associations with additive p-trends between rs743535 (CYP2E1) and NHL subtypes ............................................................................................................... 227 Table C.7  Codominant associations with additive p-trends between rs915906 (CYP2E1) and NHL subtypes......................................................................................................... 228 Table C.8  Codominant associations with additive p-trends between rs1007219 (CYP7B1) and NHL subtypes......................................................................................................... 229 Table C.9  Codominant associations with additive p-trends between rs2470152 (CYP19A1) and NHL subtypes......................................................................................................... 230 Table C.10  Codominant associations with additive p-trends between rs4774584 (CYP19A1) and NHL subtypes ........................................................................................ 231 Table C.11  Codominant associations with additive p-trends between rs7172156 (CYP19A1) and NHL subtypes ........................................................................................ 232 Table C.12  Codominant associations with additive p-trends between rs12911554 (CYP19A1) and NHL subtypes ........................................................................................ 233  xi List of figures  Figure 3.1  Participant selection for parent NHL study and subsequent organochlorine analyses ......................................................................................................................... 42 Figure 3.2  Participant selection for Illumina genetic analyses ......................................... 43 Figure 3.3  Participant selection for Taqman genetic analyses ......................................... 44 Figure 4.1  Linkage disequilibrium plot of pairwise r2 values displayed as percentages between CYP SNPs analyzed on chromosome 10. .......................................................... 114    xii List of symbols  α:   Alpha β:   Beta δ:   Delta γ:   Gamma ε:   Epsilon   xiii List of abbreviations  A:   Adenine AHR:   Aryl Hydrocarbon Receptor AIM:   Ancestry Informative Marker AR:   Androgen Receptor AT:   Ataxia Telangiectasia  B-cell:   B-lymphocyte B-CLL:   B-cell Chronic Lymphocytic Leukemia  BaP:   Benzo[a]pyrene BC:   British Columbia BCCA:   British Columbia Cancer Agency BCCRC:  British Columbia Cancer Research Centre BMI:   Body Mass Index BMS:   Mean square between samples bp:  Base Pair bw:   Body Weight C:   Cytosine CEPA:   Canadian Environmental Protection Act CEPH:   Centre d'Etude du Polymorphisme Humain CEU:  A Centre d'Etude du Polymorphisme Humain population comprised of Utah residents with ancestry from northern and western Europe CI:   Confidence Interval CIE:   Change-in-Estimate CLL:   Chronic Lymphocytic Leukemia CLUE 1:  Campaign Against Cancer and Stroke  CNV:   Copy Number Variant COMT:  Catechol-O-transferase CRD:   Capital Regional District; Greater Victoria CV:   Coefficient of Variation CYP:   Cytochrome P450 DDA:   2,2-bis(p-chlorophenyl)acetic acid DDE:   1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene DDT:   1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane DF:   Degrees of Freedom DL:   Detection Limit DLBCL:  Diffuse Large B-cell Lymphoma DNA:   Deoxyribonucleic Acid  EBV-EA:  Epstein-Barr Virus Early Antigen ESR1:   Estrogen Receptor 1 FL:   Follicular Lymphoma G:   Guanine GxE:   Gene-Environment xiv GM:   Geometric Mean GVRD:  Greater Vancouver Regional District H. pylori:  Helicobacter pylori HCB:   Hexachlorobenzene HCCH:  Hexachlorocyclohexane HHDN:  Hexachlorohexahydromethanonaphthalene HHV-8:  Human Herpes Virus-8 HIV:   Human Immunodeficiency Virus HWE:   Hardy-Weinberg Equilibrium IARC:    International Agency for Research on Cancer IBS:   Identity by State ICC:   Intraclass Correlation Coefficient Inf:  Infinity ISCO:   International Standard Classification of Occupation IUPAC:  International Union of Pure and Applied Chemistry KFSHRC:  King Faisal Specialist Hospital and Research Centre LD:   Linkage Disequilibrium LPL:   Lymphoplasmacytic Lymphoma LRT:   Likelihood Ratio Test MAF:   Minor Allele Frequency MALT:  Mucosa-Associated Lymphoid Tissue Max.:   Maximum MCL:   Mantle Cell Lymphoma MDS:   Multidimensional Scaling MF:   Mycosis Fungoides Min.:   Minimum MSP:   Medical Services Plan MZL:    Marginal Zone B-cell Lymphoma NA:   Not Applicable NAT:   N-acetyltransferase n.e.c.:   Not Elsewhere Classified NHL:   Non-Hodgkin Lymphoma NHS:   Nurses’ Health Study NK:   Natural Killer NOS:   Not Otherwise Specified NR3C1:  Nuclear Receptor Subfamily 3, Group C, Member 1 (glucocorticoid receptor) o:   Ortho OC:   Organochlorine ODC1:  Ornithine Decarboxylase 1 OR:   Odds Ratio p:   Para p-het:   Statistical significance of heterogeneity test p-trend:  Statistical significance of trend PAH:   Polycyclic Aromatic Hydrocarbon xv PCB:   Polychlorinated Biphenyl PCDF:   Polychlorinated Dibenzofuran PCR:   Polymerase Chain Reaction PEL:   Primary Effusion Lymphoma PHAH:  Polyhalogenated Aromatic Hydrocarbon POPs:  Persistent Organic Pollutants PRL:   Prolactin PTCL:   Peripheral T-cell Lymphoma QC:   Quality Control Ref:  Reference RNA:   Ribonucleic Acid ROM:   Reactive Oxygenated Intermediate SD:   Standard Deviation SLL:   Small Lymphocytic Lymphoma SNP:   Single Nucleotide Polymorphism SOC:   Standard Occupational Classification SS:   Sézary Syndrome T:   Thymine T-cell:   T-lymphocyte TAT:   Tyrosine Aminotransferase TDE:  1,1-dichloro-2,2-bis-(p-chlorophenyl) ethane UBC:   University of British Columbia U.S.:   United States USEPA:  United States Environmental Protection Agency WAS:   Wiskott-Aldrich Syndrome WGA:   Whole Genome Amplified WHO:   World Health Organization WMS:   Mean square within a quality control pair WT:   Wild-type XLP:   X-linked Lymphoproliferative XME:   Xenobiotic-Metabolizing Enzyme Xp:   Xth Percentile   xvi Acknowledgements  This thesis would not have been possible without a number of people. First, I would like to thank my thesis supervisor, Dr. John J. Spinelli, and my thesis committee members, Dr. Kay Teschke and Dr. Angela Brooks-Wilson, for answering my many questions, motivating me to finish and bearing with me for far too long.  Secondly, I must thank the many friendly faces at the British Columba Cancer Research Centre (BCCRC). The support of everyone in Cancer Control Research has been invaluable. I could not have finished without the research and personal support of research project coordinator, Agnes Lai, the statistical guidance of Zenaida Abanto or the  chats with Treena McDonald, Daryl Armstrong, Donna Kan, Barb Jamieson, Eunice Rousseau, Maria Andrews, Magali Coustalin, Colene Bentley, and Charlene Gunn.  I have also been incredibly lucky to go through this process with fellow grad students, Derrick Lee, Niki Garzia and Yang Zhang. Without question, I am greatly indebted to Dr. Johanna Schuetz, who not only was my source of information for all things genetic, but was also a huge source of inspiration when life was particularly challenging. I have also appreciated the laboratory support of Rozmin Janoo-Gilani and Stephen Leach. Special thanks to the ladies at the BCCRC library! Furthermore, the financial support of the Canadian Institutes of Health Research, the BC Cancer Agency and the University of British Columbia (UBC) must be acknowledged. Thank you to Hillary and Paulo at UBC. Finally, I would like to thank my family and the friends I have not yet mentioned. There are too many individuals to mention whose kindness has been priceless and not forgotten. Here, I will acknowledge Janine Jongbloed, who kept me relatively sane throughout the journey. I could not have done this without the unconditional support of Mum, Aalisha, Nahid, and Nani. You are my kindred spirits! xvii Dedication  To all the inspiring women in my life.    1  Chapter 1 Introduction  1.1 Rationale  Non-Hodgkin lymphoma (NHL) is the fifthi most common form of cancer in Canada.[1] The incidence rate of NHL increased by approximately 50% between the 1970s and the 1990s, but has since stabilized.[2] Although the etiology of NHL remains unclear, many risk factors have been demonstrated, including infectious agents, immune deficiency and various chemical exposures.[3] The increase and subsequent plateau in NHL incidence has been suggested to be partially due to the widespread use and subsequent ban on and heightened regulation of many organochlorines (OCs).[4]  Organochlorines are chlorine-containing organic molecules for which large-scale production began after World War II. Organochlorines have raised much public health and environmental concern, as many are persistent organic pollutants (POPs). POPs are toxic and highly stable carbon-based chemicals that are distributed throughout the environment and bioaccumulate ii.[6] The latter may be attributed to the lipophilicity iii of many OCs.[7] Organochlorines can be classified according to whether or not they are used as pesticides.[8] Polychlorinated biphenyls (PCBs) are synthetic non-pesticides that were produced for use in electrical equipment, as well as for other products, until their ban in most countries during the 1970s.[8, 9] Commercial PCBs are mixtures of one or more of the 209 different “congeners”, defined by the number and position of the chlorine atoms on their two benzene rings.[8, 10] These congeners can also be classified according to                                                   i Tied with bladder cancer and excluding non-melanoma skin cancers  ii Bioaccumulate: The accumulation of a substance from the outside environment in an organism’s tissue.[5]  iii Lipophilic: soluble in lipids 2 whether they are dioxin-like (i.e. their toxicity is mediated by the aryl hydrocarbon receptor, AHR) or non-dioxin-like.[8]  Numerous studies have investigated the potential link between NHL and organochlorine pesticides and PCBs.[11-14] Positive associations have been reported by some studies, while no associations have been found by others. This may be attributed to differences in the specific chemicals analyzed, as well as to methodological limitations,  such as insufficient power and/or imprecise exposure assessment.[12, 13]  A recent study by Engel et al. noted mostly significantiv exposure-response trends for NHL and lipid-corrected, prediagnostic blood levels of PCB-118, 138 and 153, using data collected between the 1970s and 1990 from three prospective cohorts: the Campaign Against Cancer and Stroke (CLUE 1) cohort from Maryland (74 cases and 147 controls), the American multistate Nurses’ Health Study (NHS) cohort (30 cases and 78 controls) and the community-based Janus cohort from Norway (190 cases and 190 controls).[15] Obvious trends were not found for p, p’-DDE.  The CLUE 1 study also found no significant trends between NHL and prediagnostic serum levels of dieldrin, β-HCCH, HCB, trans-nonachlor, oxychlordane, heptachlor, heptachlor epoxide, summed chlordane/ heptachlor-related compounds (trans-nonachlor, oxychlordane, heptachlor, heptachlor epoxide), or summed 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane (DDT)/ 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene (DDE) (o, p’-DDT, p, p’-DDT, o, p’-DDE, p, p’-DDE).[16] A significant trend was found for summed PCBs (28 congeners).[17]  In a subset of 100 cases and 100 controls from a four-centre U.S. population-based case-control study conducted between 1998 and 2000, De Roos et al. did not find significant trends between NHL and the lipid-corrected plasma levels of the PCB congeners measured (PCB-74, 99, 118, 126, 138/ 158, 146, 153, 169, 170, 183, 187), nor with summed noncoplanar PCBs, dieldrin, β-HCCH, trans-nonachlor, oxychlordane, heptachlor epoxide, 1-chloro-4-[2,2,2-trichloro-1-(4-chlorophenyl)ethyl]benzene (p, p’-                                                  iv Significance set at alpha 0.05, unless otherwise indicated  3 DDT), or 1-chloro-4-[2,2-dichloro-1-(4-chlorophenyl)ethenyl]benzene (p, p’-DDE; a metabolic of p, p’-DDT).[18] Nevertheless, significant trends with PCBs 156, 180 and 194 were found. Significant trends between NHL and PCB-180 and p, p’-DDE were found using carpet dust from 603 cases and 443 controls from the same parent study as the previous.[19] No significant trends were found for PCB-105, 138, 153, 170, alpha-chlordane (α-chlordane), gamma-chlordane (γ-chlordane), or p, p’-DDT. However, risk was significantly increased when total PCBs were categorized into detectable and undetectable groups. Brauner et al. used a case-cohort design within a prospective Danish cohort (239 cases and 245 controls).[20] Participants provided prediagnostic adipose tissue at enrolment, which occurred between 1993 and 1997. Incidence rate ratios were examined for 18 lipid-adjusted OCs: PCBs 99, 118, 138, 153, 156, 170, 180, 183, 187, 201, dieldrin, β-HCCH, HCB, cis-nonachlor, trans-nonachlor, oxychlordane, p, p´-DDT, p, p´-DDE, summed PCBs, summed immunotoxic PCBs (138, 153, 180), summed chlordanes (cis-nonachlor, trans-nonachlor, oxychlordane), and summed DDT/ DDE, A significant dose-response trend was only noted for p, p'-DDT.  Between 2000 and 2004, Dr. John J. Spinelli and his team at the BC Cancer Research Centre (BCCRC) conducted a large population-based case-control study in BC, hereafter referred to as the “parent study”.[13, 14] Refer to Section 2.2.2.1 for additional details on the parent study. Not only was this the largest study to date assessing associations between plasma OC levels and NHL, but a variety of other potential NHL risk factors were also researched. Organochlorines were measured in the plasma of approximately half of the participants (422 cases and 460 controls). These OCs included 14 PCB congeners (dioxin-like PCBs: 105, 118, 156; non-dioxin-like PCBs: 28, 52, 99, 101, 128, 138, 153, 170, 180, 183, 187) and 11 pesticides/ pesticide metabolites (aldrin, β-HCCH, HCB, mirex, α-chlordanev, γ-chlordanevi, cis-nonachlor, trans-nonachlor, oxychlordane, p, p’-DDT, p, p’-DDE).                                                   v cis-chlordane  vi trans-chlordane 4 Associations between these OCs and NHL were assessed in participants of all ethnicities. Significantly increased odds of NHL in the highest compared to the lowest OC categories and significant trends were found for total summed PCBs, summed dioxin-like PCBs and summed non-dioxin-like PCBs. These findings can be attributed to significantly increased odds of NHL in the highest compared to the lowest OC categories and significant trends for the dioxin-like congeners, PCB-118 and 156, and the non-dioxin-like congeners, PCB-153, 170, 180, and 187. Significance was maintained even after adjustment for multiple testing. PCB-99 and 138 showed significant trends, but not significantly increased odds in the highest compared to the lowest OC categories. The trend for PCB-138 also remained significant after adjustment for multiple testing.  Significant trends and significantly increased odds of NHL in the highest compared to the lowest OC categories were also found for the following pesticide analytes: β-HCCH, HCB, mirex, trans-nonachlor, and oxychlordane. p, p’-DDE showed only a significant trend. All associations, except for β-HCCH and mirex, remained significant after adjustment for multiple testing. No significant trends or significantly increased odds of NHL in the highest compared to the lowest OC categories were found for PCB-28, 105, 183, cis-nonachlor, or p, p’-DDT.   1.2 Thesis objectives  As the risk conferred by environmental exposures, such as organochlorines, can be modified by an individual’s genetic makeup, and vice-versa, gene-environment (GxE) analyses are of great interest.[21] The effects of some exposures may be more detectable in genetically susceptible individuals and the effects of some genotypes may be more phenotypically expressed in the context of specific exposures.[22] In general, gene-environment interaction analyses may guide cancer control efforts in targeting groups that are at higher risk for disease due to greater environmental exposure and/or the presence of certain genotypes. Furthermore, gene-environment studies may be able to help to elucidate the biological mechanisms of carcinogenesis and to develop disease avoidance strategies, through the 5 identification of disease pathways that may be impacted by specific susceptibility/ resistance genes and environmental exposures.  For these reasons, the research team looked at interactions between organochlorines and genetic point variants known as single nucleotide polymorphismsvii (SNPs) in the AHR gene, which were found to be significantly associated with NHL.[25] AHR is a transcription factor that can act by many mechanisms.[26] AHR can be activated by various ligands, including some organochlorines.[27, 28] This causes AHR to translocate from the cytosol into the nucleus and to form a complex with aryl hydrocarbon receptor nuclear translocator.[26] This complex then interacts with the xenobioticviii response elements of target genes, such as xenobiotic-metabolizing genes and genes involved with proliferation. Five AHR SNPs were studied, X1-459G/A (rs7796976), IVS1+4640G/A (rs17722841), IVS1-3946G/A (rs2282885), X2+132T/C (rs17779352), and X10+1661G/A (rs2066853).  A significant trend was found between the tagSNPix IVS1+4640G/A and risk of NHL, before correcting for multiple testing. Gene-environment interactions were found between the IVS1+4640G/A and PCB-118, oxychlordane and trans-nonachlor in participants of European ethnic origin only (hereafter referred to as “Europeans”) and all ethnicities combined. Risk of NHL was increased in the highest OC exposure compared to the lowest OC exposure in the major allele homozygous group. This increased risk was not seen in the heterozygous/ minor allele homozygous group. However, these interactions                                                   vii SNPs are single base genetic variants in which more than one allele or version of the DNA building blocks (A (Adenine), Thymine (T), Guanine (G), and Cytosine (C)) exists at a specific genetic location and the rarer allele has a frequency of at least one percent in a population.[23, 24]  viii Xenobiotic: “A foreign chemical in the body. Any chemical, including drugs and foodstuffs, normally located outside the living organism, which exists in the organism (the opposite of an endogenous compound).”[29]  ix TagSNP: Particularly informative SNPs, which give information about the pattern of genetic variation in a chromosomal region that is generally in high linkage disequilibrium or inherited together. Essentially, the use of tagSNPs minimizes the number of SNPs that need to be genotyped to capture the underlying genetic variation of the region. [30, 31]  6 were not significant after controlling for multiple comparisons. These results suggested that organochlorines may increase NHL risk and that this may occur via the AHR pathway.  The objective of this thesis is to build on the previous analyses conducted in the parent study and assess further gene-environment analyses between the organochlorines and other genes selected a priori on the basis of their relevance to NHL and/or xenobiotic metabolism. This thesis is focused on genes belonging to the cytochrome P450 (CYP) superfamily. In humans, the CYP superfamily encodes 57 enzymes, which are involved in the metabolism of endogenous toxins, such as bilirubin, and diverse xenobiotics, such as some tricyclic antidepressant drugs.[28, 32] CYPs are also involved in the synthesis of cholesterol and steroids. Expression varies by specific CYP gene time, organ, tissue and cell.[29] The specific substratesx, inducersxi and inhibitors of CYPs also vary by the specific CYP gene. For example, CYP1A1 and CYP1B1 are known to be involved in the metabolism of polycyclic aromatic hydrocarbonsxii (PAHs), such as the procarcinogensxiii found in cigarette smoke, and polyhalogenated aromatic hydrocarbonsxiv (PHAHs), such as PCBs. These genes are regulated by the AHR pathway. Interestingly, CYPs are involved in both the detoxification and the activation of chemicals that can damage cellular DNA and proteins.[29] Such damage can result in tumour initiation, by impacting stress-signaling and/or checkpoint-signaling pathways. CYPs can also                                                   x Substrate: A molecule that is altered by an enzyme-catalyzed reaction.[33]   xi Inducer: An environmental agent that can initiate transcription (i.e. copying of DNA into Ribonucleic Acid (RNA). RNA can then be translated into protein (e.g. CYP enzyme)).[23]   xii Polycyclic aromatic hydrocarbons: “Chemicals that contain only C [carbon] and H [hydrogen] atoms arranged in three or more rings (cycles), and that have conjugating double bonds (π electrons), that is, not saturated.”[29]  xiii Procarcinogens: Chemicals that must be metabolically activated before they can cause cancer.[29]  xiv Polyhalogenated aromatic hydrocarbons: “Chemicals comprised of C [carbon], H [hydrogen], and Cl [chlorine], Br [bromine] or F [fluorine], and sometimes O atoms.”[29] Like PAHs, these are aromatic (i.e. have rings with unsaturated, conjugating double bonds (π electrons)). PCBs , DDT and HCB are PHAHs containing Cl.[34] The latter are also types of OCs. Not all OCs are PHAHs (e.g. aldrin, mirex, chlordane) because not all OCs are aromatic. Not all PHAHs are OCs because not all PHAHs are chlorinated. 7 affect the levels of other molecules and downstream signal transduction pathways. These alterations can result in tumour promotion and progression, by impacting the cell cycle, apoptosis and cell growth.  The association between CYPs and cancer was demonstrated in a recent four-centre U.S. population-based case-control study by De Roos et al. (1172 cases and 982 controls).[35] A borderline non-significant trend was noted between NHL and CYP1B1 V432L (rs1056836). This trend was significant in the B-cell and diffuse large B-cell lymphoma (DLBCL) subtypes of NHL. Significant associations were noted in NHL and B-cell subtypes for CYP2E1 -1054C>T (rs2031920). No significant associations were found for CYP1A1 I462V (rs1048943), CYP1A2 *1F (rs762551), CYP2C9 R144C (rs1799853), or CYP2E1 -332A>T (rs2070673).  Given the CYPs’ dual roles in carcinogenesis and organochlorine metabolism, the hypothesis of this thesis was that the positive association between many organochlorines and NHL could be mediated by a CYP-dependent pathway. Modification of the effect of organochlorines on NHL risk by CYP polymorphisms would appear to be biologically plausible and interactions may be detectable by statistical analysis. In order to test this hypothesis, this thesis’ objectives are as follows:  Primary objectives  1. To determine whether the organochlorines tested in the parent study are associated with increased or decreased risk of NHL in the subset of European participants  2. To determine whether SNPs of CYP genes are associated with increased or decreased risk of NHL in European participants  3. To determine whether significant associations between organochlorines and NHL are modified by CYP SNPs in European participants   8 Secondary objectives  4. To suggest biologically plausible mechanisms for carcinogenesis based on significant interactions or lack thereof  1.3 Outline   Chapter 2 gives some background on non-Hodgkin lymphoma, organochlorines and cytochrome P450 genes/ enzymes.   Chapter 3 provides details on the design of the study, the rationale for organochlorine and gene selection and the laboratory and quality control processes used to measure organochlorine levels and genotype CYP SNPs. The statistical methods used to analyze the organochlorine main effects, CYP main effects and gene-environment interactions are also described.   Chapter 4 presents the results of the organochlorine main effects, CYP main effects and gene-environment interaction analyses. These results are discussed in Chapter 5. This thesis concludes with Chapter 6, which summarizes the results, discusses the strengths and limitations of the study and touches on the implications of the results.      9 Chapter 2 Background  2.1 Background on NHL  2.1.1  Cancer  Cancer is an umbrella term describing more than 100 distinct diseases, which arise due to DNA changes that result in deregulated cellular proliferation.[36, 37] This allows for the invasion of surrounding tissue and metastasis or spread to distant areas of the body. Cancer development is often simplified into three stages: initiation, promotion and progression.[38, 39] During initiation, cellular DNA is non-lethally damaged, triggering DNA repair mechanisms. Promotion involves cellular proliferation and/or inhibition of apoptosisxv. Progression is the final irreversible stage during which DNA damage accumulates and benign cells become malignant.xvi  This deregulation of cellular proliferation occurs by virtue of somatically  acquired DNA mutationsxvii, such as nucleotide substitutions, rearrangements and epigenetic xviii changes.[36, 37] Mutations can accumulate during normal cell division or can be the result of environmental exposures and lifestyle factors that damage DNA.  It is important to note that some environmental exposures can impact cancer development without directly damaging DNA. A number of mechanisms have been                                                   xv Apoptosis: A form of programmed cell death.[33, 40]  xvi Malignant: Cancerous; capable of growth beyond normal boundaries and spread to other areas of the body.[41]  xvii Somatic mutations: DNA changes in non-germ cells (i.e. not cells that give rise to sperm and egg) that are acquired after conception.[41]  xviii Epigenetic: Changes in gene function that are not the result of changes in DNA base sequence.[23]  10 suggested by which OCs may affect carcinogenesis. These vary by cancer type and chemical. Because PCBs are not generally genotoxic, carcinogenesis is hypothesized to be driven by the induction of enzymes such as CYPs, tumour promotion and immune suppression.[42, 43] Not only has DDT been suggested to directly mutate through DNA adduct formation, but also to promote existing abnormal cell populations.[44] Chlordane has been shown to impact the immune system, through lymphocyte function alteration in vivo.[45]  Genetic susceptibilityxix or the increased risk of developing cancer due to an individual’s inherited genetics can impact carcinogen metabolism, somatic mutation rate and direct cancer growth.[37, 41] For example, the CYP genes can play a role in tumour initiation because the enzymes they encode are often involved in the metabolism of chemicals that can potentially damage cellular DNA and proteins.[29] CYPs can also impact tumour promotion and progression by affecting pathways involved in the cell cycle, apoptosis and cell growth.  It is estimated that five mutations in “driver” genesxx are required to produce cancer.[36, 37] This number can vary by type of cancer, with fewer mutations being required for hematopoietic neoplasms. The nature of mutations can vary by the type of cancer and exposure. For example, chromosomal translocations are typical of some non-Hodgkin lymphoma subtypes and tobacco carcinogens are known to induce C:G>A:T transversion mutations experimentally.[3, 37] Genetic and environmental risk factors can vary at each step of cancer development. Genes that are important for tumour development are not necessarily important for genetic susceptibility; ~10% of the genes that play a role in cancer development when somatic mutations occur in them also increase genetic susceptibility when germ line mutations occur in them.[36]                                                    xix Genetic susceptibility: An inherited increased risk of disease.[41] Constitutional mutations or abnormalities are present in the fertilized egg and therefore in every cell of an organism (e.g. mutations in the germ line).[46]  xx Driver genes: Genes that favour the expansion of cancer cells over normal cells .[37] 11 2.1.2 Non-Hodgkin lymphoma   2.1.2.1 Classification   Lymphomas are cancers that originate in lymphocytes, which are white blood cells that are part of the lymphatic system.[47] The lymphatic system plays a role in immunity and is comprised of the bone marrow, thymus, spleen, lymph nodes, and lymphatic vessels. Lymphocytes are produced in the bone marrow and are further classified into B-lymphocytes (B-cells) or T-lymphocytes (T-cells), depending on whether they mature in the bone marrow or the thymus, respectively.[48] Although unclear, it is thought that lymphocytes known as Natural killer (NK) cells mature in secondary lymphoid tissue, such as the lymph nodes and tonsils.[49] Lymphocytes then circulate in the lymph nodes, spleen and other lymphoid tissues via blood and lymphatic vessels, in order to carry out their role in the immune response.[50] The current classification of lymphoid neoplasms is the World Health Organization (WHO) classification, which is based on the cell lineage, maturity and stage of differentiation.[51] Morphology, immunophenotype, etiology, and clinical and genetic features of the neoplasms are taken into account.  In Canada, approximately 90% of lymphomas are non-Hodgkin lymphomasxxi.[1] NHLs differ from Hodgkin lymphomas in a number of ways, including the absence of Reed-Sternberg cells.[52] However, the overlap between NHL and classical Hodgkin lymphoma is increasingly appreciated.[51] NHL is an etiologically and clinically heterogeneous collection of cancers, comprised of several subtypes.[3] These subtypes originate at different stages along the development of a number of cell types, namely B-, T- and NK-cells.[51] NHL subtypes can differ in gene expression patterns and prognosis.[3] Chromosomal abnormalities and genetic mutations that can activate or inactivate cancer-relevant genes or alter proteins can also differ between NHL subtypes.                                                   xxi Excludes chronic lymphocytic leukemia  12 In the U.S., B-cell malignancies make up the majority of non-Hodgkin lymphomas.[3] The most common B-cell NHL subtypes are diffuse large B-cell lymphoma (DLBCL), follicular lymphoma (FL) and chronic lymphocytic leukemia/ small lymphocytic lymphoma (CLL/ SLL). CLL and SLL are grouped together under the WHO classification as they are recognized as different presentations of the same disease.[51, 53] However, CLL and SLL are also sometimes considered separately with leukemias and lymphomas, respectively. Less common B-cell subtypes include marginal zone B-cell lymphoma (MZL), mantle cell lymphoma (MCL) and lymphoplasmacytic lymphoma (LPL).[3] T- and NK-cell lymphomas, which are closely related and less common, include peripheral T-cell lymphomas (PTCL) and mycosis fungoides (MF)/ Sézary syndrome (SS).[3, 51]  2.1.2.2 Epidemiology of NHL  NHL is the fifthxxii most common form of cancer in Canada.[1] The estimated age-standardized NHL incidence rates for 2012 in Canadian males and females are 20 and 14 cases per 100,000, respectively. The estimated age-standardized NHL mortality rates in Canadian males and females are 7 and 5 deaths per 100,000 respectively. Worldwide, NHL incidence rates are highest in Western Europe and North America and lowest in Asia. [3] Incidence rates increased substantially between the 1970s and 1990s, after which point rates have begun to stabilize or decline in some, but not all, Western countries.[2-4]  According to the Surveillance, Epidemiology, and End Results registries in the U.S., the changes in incidence of NHL overall ranged from increases of 33% in white females to 77% in black males between 1978 and 1995 (1978–1983 to 1990-1995).[54] These patterns vary by NHL subtype. Increases were greatest in immunoblastic NHL, small  noncleaved NHL and lymphoblastic NHL. Small lymphocytic NHL, diffuse large NHL, peripheral T-cell and not otherwise specified (NOS) NHL also increased. FL was stable in black females, but increased in whites and black males. Only diffuse small NHL (diffuse                                                   xxii Tied with bladder cancer and excluding non-melanoma skin cancers 13 small cleaved cell malignant lymphoma, diffuse lymphocytic intermediate differentiation malignant lymphoma, centrocytic malignant lymphoma) declined.  In the U.S., the rates of lymphoid neoplasms, which include NHL, Hodgkin lymphoma, multiple myeloma, and acute and chronic lymphocytic leukemia, decreased by 1% per year in males and remained stable in females between 1992 and 2001.[55] Overall incidence rates of DLBCL, CLL/ SLL and unknown type lymphoid neoplasms declined during this period. However, rates increased for MZL, MCL, Burkitt lymphoma/ leukemia, and T-/ NK-cell neoplasms, particularly peripheral T-cell lymphomas. Rates were unchanged for FL and SS/ MF. It is important to note that some of the increases in MZL and MCL and decreases in unknown type lymphoid neoplasms can be attributed to changes in diagnostic practices. Other increases in MZL may be attributed to MZL of mucosa-associated lymphoid tissue (MALT) due to autoimmune diseases. The decrease in the unknown types can also be attributed to better classification of other subtypes, which leads to increases in the incidence of other subtypes.  In Canada, incidence rates stabilized between 1998 and 2012 for females and continued to slightly increase by 0.8% per year for males.[1] However, the proportion of NHL attributable to human Immunodeficiency virus (HIV) decreased due to the introduction of anti-retrovirals in the 1990s. Mortality rates have declined since 2000 by  2.6% and 2.7% per year in Canadian males and females, respectively, likely reflecting advances in treatments such as immunotherapy (e.g. In 2001, rituximab was recommended for treatment of advanced-stage DLBCL in BC).[1, 56] The increase and subsequent plateau/ decline in NHL incidence can only be partially explained by known risk factors, such as HIV and transplantation, and changes in diagnostic practisesxxiii.[57] It has been suggested that this trend can be partially attributed to the previous widespread use and subsequent heightened regulation of certain chemicals, such as organochlorines.[4] However, other factors may have also had                                                   xxiii Based on the increase in incidence in the U.S. from 1947-50 to 1984-88  14 an influence.[58] Changes in NHL classification over the years and varied incidence patterns between subtypes must be acknowledged as well.[55]   2.1.2.3 NHL risk factors   NHL risk is generally higher in males than in females and varies by ethnicity. [3] For example, incidence and mortality rates were highest in non-Hispanic Whites [European-Americans] and lowest in American Indians [Native Americans] in SEER areas between 1996 and 2000. Rates were intermediate for Blacks [African-Americans], Hispanics and Asian/ Pacific Islanders. The risk of NHL increases exponentially with age, regardless of sex or ethnicity. Prior history or immediate family history of NHL also increases NHL risk. Although a variety of genetic and environmental (non-genetic) risk factors for NHL are known, the causes of most lymphoma cases are unclear.   2.1.2.3.1 Environmental risk factors (non-genetic risk factors)  Immune suppression is a known risk factor for NHL and can be inherited or acquired.[3] Several microorganisms are known to increase the risk of NHL or a specific subtype of NHL, most likely by affecting or working in conjunction with immune system deregulation.[3, 59] The risk of a number of NHL subtypes is increased in those infected with HIV. The Epstein-Barr virus (EBV) and the human herpes virus-8 (HHV-8) are often associated with the B-cell subtypes, Burkitt lymphoma and Primary effusion lymphoma (PEL), respectively. Helicobacter pylori (H. pylori) bacterial infections are associated with increased risk of gastric MALT lymphomas. NHL risk can also increase following transplantation, depending on the type of transplantation and degree of immunosuppression. 15  Exposure to a number of chemicals, some of which have immunotoxicxxiv properties, can also increase the risk of NHL.[4] Potential chemical risk factors include organic solvents, PCBs and pesticides, such as organochlorine pesticides and phenoxyacetic acid herbicides. Refer to Section 2.2.2 for a review of the literature on the relationship between NHL and organochlorines. Exposure to these chemicals may explain the higher rates of lymphoma associated with a number of occupations, including farming and forestry.[3] However, results often vary between studies due to sample size, specific exposures and assessment methods. Interestingly, risk of NHL has been noted to be elevated in some fishermen. Diet has been suggested as a possible source of the increased risk. Although the evidence is not strong, several studies have noted excess lymphomas in those with diets high in fat and protein or low in fruits and vegetables.   2.1.2.3.2 Genetic risk factors  Immune deficiencies can also be the result of genetic susceptibility factors.[3] The risk of lymphoproliferative disease is increased in those with inherited immune deficiencies, such as the autosomal recessive disorder ataxia-telangiectasia (AT), in which the ATM gene that regulates cell division after DNA damage is mutated, X-linked lymphoproliferative disease (XLP), in which the SAP/SH2D1A gene is mutated, and the X-linked recessive disorder Wiskott-Aldrich syndrome (WAS), in which the WAS protein gene is mutated. Autoimmune disorders, such as rheumatoid arthritis, Sjogren syndrome, celiac disease, and systemic lupus erythematosus, are also associated with increased lymphoma risk. However, it is important to note that many of these conditions are not purely the result of inherited genetic mutations, but also involve environmental (non-genetic) components.  Variations in a number of genes have been found to be associated with NHL based on genome-wide association studies (GWAS). A meta-analysis of two GWAS in the U.S.,                                                   xxiv Immunotoxicity: An adverse effect on the immune system.[60] 16 Denmark and Sweden noted that variants in HLA (human leukocyte antigen), which play a role in immunity, are important in follicular lymphoma.[61-63] A Japanese GWAS in DLBCL implicated CDC42BPB (CDC42 binding protein kinase beta (DMPK-like); cellular structure, interaction with tumour-promoting agent) and LNX2 (ligand of numb-protein X 2; E3 ubiquitin ligase, signaling).[28, 64] A European GWAS in B-cell chronic lymphocytic leukemia (B-CLL) suggested roles for FARP2 (FERM, rhoGEF and pleckstrin domain protein 2; signaling), IRF8 (interferon regulatory factor 8; B-cell lineage specification), NEDD4 (neural precursor cell expressed, developmentally down-regulated 4, E3 ubiquitin protein ligase; B-cell development), CPEB1 (cytoplasmic polyadenylation element binding protein 1; cell division and differentiation), CXXC1 (CXXC finger protein 1; gene regulation), and MBD1 (methyl-CpG binding domain protein 1; gene regulation).[28, 65] A subset of this sample suggested SP140 (SP140 nuclear body protein; immunity), IRF4 (interferon regulatory factor 4; lymphocyte development and proliferation), PRKD2 (protein kinase D2; cellular proliferation), and GRAMD1B (GRAM domain containing 1B) associations.[66]  Furthermore, candidate gene studies have looked at variations in genes encoding enzymes that metabolize xenobiotic and endogenous chemicals. Among these are the cytochrome P450 genes, which encode enzymes that metabolize various substrates ranging from steroids to organochlorines, the aryl hydrocarbon receptor gene, which encodes transcription factors that regulate various xenobiotic-metabolizing genes, including CYP1A1, and the N-acetyltransferase (NAT) genes, which encode enzymes that metabolize aromatic and heterocyclic amines.[25, 28, 29, 35, 67] For example, Morton et al. found that the risk of NHL was significantly increased in individuals with the NAT1*10/*10 genotype compared to other NAT1 genotypes.[67] Refer to Section 2.3.2 for a review of the literature on the relationship between NHL and CYPs. Genes that play a role in hormonal modulation have also been suggested to increase NHL susceptibility. Among these genes are prolactin (PRL), which is involved in prolactin production, CYP17A1, which is involved in estrogen and testosterone biosynthesis, and catechol-O-transferase (COMT), which is involved in estrogen metabolism.[68] Skibola et al. noted 17 borderline increased risks of NHL in major allele homozygotes of PRL -1149G>T (rs1341239) and COMT 701A>G (rs737865).[68] The risk of DLCBL was significantly increased in minor allele homozygotes of CYP17A1 -34T>C (rs743572).   2.2 Background on organochlorines  2.2.1 Organochlorines  Organochlorines are chlorine-containing organic molecules for which large-scale production began after World War II.[8, 69, 70] Organochlorines have raised much public health and environmental concern, as many are persistent organic pollutants. POPs are toxic and highly stable carbon-based chemicals that are distributed throughout the environment and bioaccumulate.[6] The latter may be attributed to the lipophilicity of many OCs.[7] Biochemical transformation rates can vary by specific organochlorine, by isomer and by environmental context.[7] The biological half-lives of some OCs are given in Table B.1. Because OCs are generally lipid-soluble and water-insoluble, they are often stored in animal tissue and adsorb onto organic matter in soil and particulates in water.[7] Although many organochlorines have a low vapour pressure, large surface exposure and direct aerial application/ spray drift can result in spread in the air. Transport of these chemicals over long distances in the air has been noted.  Diet is the major source of OC exposure in the general human population.[7, 71] Segments of the population are also exposed via occupation, public health use (e.g. see DDT below) and accidents (e.g. rice oil contamination with PCB mixture Kanechlor 400 in Japan). Organochlorines can be classified according to whether or not they are used as pesticides. PCBs are synthetic non-pesticides. Details about specific organochlorine pesticides/ pesticide metabolites and PCBs are given below.   Polychlorinated biphenyls (PCBs) PCBs can be classified into 209 different congeners, according to the number and position of chlorine atoms on their two benzene rings. These congeners can also be classified 18 according to whether they are dioxin-like (i.e. their toxicity is mediated by AHR) or non-dioxin-like.[8] In 1867, dichlorodiphenyl, a PCB congener, was first made.[71] Commercial PCB production in the U.S. began in 1929. Aroclors are mixtures of commercially produced PCBs. PCB mixtures are also sometimes contaminated with polychlorinated dibenzofurans (PCDFs) and chlorinated naphthalenes. In the 1970s, a number of European countries, including Italy and France, were producing PCBs and a number of countries, including Canada and Finland, were importing them. More than one million tons of PCBs have been produced worldwide.[9]  PCBs were produced for a number of purposes, including use in electrical equipment, such as capacitors, condensers and transformers, as well as electrical wire/ cable insulation and fluorescent light ballasts.[71] Other uses have been in hydraulics/ lubricants, heat transfer systems, pesticide extenders, and insecticides/ bactericides. PCBs have also been included in adhesives, investment casting waxes, plasticizers, surface coatings, and inks. PCBs are sometimes used in chemical processes, such as polymerization and permeability conversion, and in laboratories to induce the activity of various enzymes, such as the CYPs and aryl hydrocarbon hydroxylase.  PCBs were banned in most countries, including Canada, during the 1970s. PCBs were never manufactured in Canada; they are prohibited and their use is restricted to existing closed electrical and hydraulic systems.[9, 72] PCBs are included on the List of Toxic Substances in Schedule 1 of the Canadian Environmental Protection Act, 1999 (CEPA 1999), which prohibits the manufacturing/ processing, sale and import of some toxic substances, with minor exceptions. It is scheduled for eventual elimination under the Stockholm convention.  Some existing PCBs have been destroyed, but others still exist in storage and old PCB-containing devices.[9] Environmental exposure to PCBs can therefore still occur through leakage. PCBs are quite stable, but are thought to degrade by photochemical reactions.[71]  PCBs ingested by animals accumulate in fatty tissue and concentrate up the food chain.[9] The nature of PCB metabolism, excretion and storage depends on the position and degree of chlorination.[73, 74] Generally, the stability of PCBs increases with the degree of chlorination. It has been noted that PCBs are hydroxylated during metabolism.  19 Although PCBs have been detected in soil, water, air, and occupational environments, the majority of people are exposed via diet.[71] Fish, eggs, cheese, and meat are dietary sources of PCB exposure. Between 1976 and 1978, the average dietary PCB intake in Canadians was reported to be 1 ng/kg body weight (bw)/day by the Total Diet study.[75] In 1994, levels were 7.73 ng/kg bw/day in Halifax and in 2002, they were 2.14 ng/kg bw/day in Vancouver.[76] PCB levels are higher in the Eastern Arctic than the rest of Canada and are of concern because of the traditional diet of the residents.[9] Of note, the levels in Yukon fish have declined between 1993 and 2008.   Aldrin  Aldrin was first synthesized in 1948.[7] Commercial production in the U.S. began in 1950. Technical grade aldrin can contain a number of components, including hexachlorohexahydrodimethanonaphthalene (HHDN), isodrin, hexachlorocyclopentadiene, hexachlorobutadiene, HHDN diadduct, bicycloheptadiene, and toluene. It was used as an insecticide for both agricultural and non-agricultural purposes, particularly against termites. In the 1970s, a number of restrictions on aldrin use were introduced in various countries. Aldrin has never been produced in Canada.[72] Registration of this pesticide was discontinued in Canada in 1990.   In the environment, plants and animals, aldrin is converted to dieldrin, which is very stable.[7] Aldrin and dieldrin have been detected in soil and foods, but levels have been declining over time.[77] Aldrin and dieldrin can enter the soil from direct pesticide treatment or leakage at disposal sites. They both bind tightly to soil, but dieldrin is more often detected. Runoff from contaminated soil can expose surface water to these compounds. Volatilization from soil and water into the air and dry and wet deposition back to the ground can also occur. In animals, aldrin, dieldrin and their metabolites are excreted in feces and urine or stored in adipose tissue. Dieldrin accumulates in the food chain.  Because of aldrin’s conversion to dieldrin in the environment, the general population is mostly exposed to dieldrin via diet.[77] In 1985, aldrin was not detected in any analyzed foods, whereas dieldrin was detected in all food composites in a Canadian food survey. The average 20 dieldrin intake in the Canadian Total Diet study was fairly stable at 2 ng/kg bw/day between 1976 and 1978 and 1.94 ng/kg bw/day between 1993 and 1996.[75, 76] Other sources of exposure can be via inhalation of contaminated air, particularly in pesticide-treated homes and waste sites.[77] In the past, occupational exposures may have been important in workers at chemical factories and in those applying pesticides.[7]    Hexachlorocyclohexane (HCCH) HCCH was first synthesized in 1825 and commercially produced in the U.S. in 1945.[78] Technical grade HCCH is comprised of various isomers (e.g. α, β, γ, δ, ε) and other components. Lindane is mainly comprised of γ-HCCH and has been used as a human therapeutic agent against scabies and lice.[78, 79] HCCH was used as an insecticide for agricultural and non-agricultural purposes.[78] γ-HCCH is responsible for the insecticidal properties of HCCH. Technical gradexxv HCCH and lindane are no longer produced in the U.S.[81] β-HCCH and lindane are not manufactured in Canada and their pesticide registrations were discontinued in the 1970s and 2005 respectively.[82]   HCCH isomers have been detected in soil, air, water, and food samples throughout the world.[78] Isomers can remain in soil for long periods, but are eventually degraded by bacteria, algae and fungi.[81] Isomers can also persist in air, but can also be degraded or enter rain water. HCCH in water also eventually degrades. β-HCCH is the most persistent HCCH isomer and accumulates in the food chain. In humans, the β-Isomer persists in fatty tissue. Greater proportions of the other isomers are excreted in the urine and feces. Isomers are metabolized to a number of compounds, including toxic chlorophenols.  General human exposure to HCCH occurs primarily through the consumption of contaminated food.[81] The average HCCH intake in Canadians in the Total Diet study was 10 ng/kg bw/day between 1976 and 1978.[75] Levels for β-HCCH and γ-HCCH were 0.39 and 1.32 ng/kg bw/day respectively between 1993 and 1996.[76] β-HCCH was detected at higher levels                                                   xxv Technical grade: suitable for industrial use.[80] 21 than γ-HCCH, but not as frequently, in all food composites from six Canadian cities between 1992 and 1996 (% Positive: 2.5%; Mean positive: 1.9 ng/g).[83] Exposure to contaminated air, soil and water, particularly at hazardous waste sites, is also possible.[81] Exposure to γ-HCCH can occur through pharmaceuticals. In the past, occupational exposures could have occurred through inhalation or dermal contact in those manufacturing or applying pesticides (e.g. agricultural workers, pesticide applicators, factory workers).  Hexachlorobenzene (HCB)  HCB was first prepared in 1893.[78] Commercial production began in the U.S. in 1933 and ended in 1978. A number of countries have set restrictions on its use. HCB has never been produced in Canada and the registration of this pesticide was discontinued in 1976. CEPA 1999 also set concentration limits above which this chemical is prohibited in certain products. It is scheduled for eventual elimination under the Stockholm convention.[72] HCB has been used as a fungicide/ pesticide on various crops.[78] HCB has also been used as an additive in military pyrotechnics, a porosity controller for electrodes and a wood preservative. Other applications include use as a polymer additive, a plasticizer for polyvinyl chloride and a peptizing agent for nitroso and styrene-type rubbers. Technical and commercial HCB formulations have been found to contain polychlorinated benzenes, hexachlorobutadiene, hepta- and octachlorodibenzofurans, and/or octachlorodibenzo-para-dioxin. HCB has been noted as a contaminant in other pesticides and as a byproduct in the production of other chemicals and in the petrochemical industry.  HCB is released into the environment via waste water and air emissions from chemical production facilities, gases and ash from municipal incinerators, air emissions from pyrotechnics, leaks in hazardous waste sites, and its past use in pesticides or pesticides contaminated with HCB.[84] HCB is quite stable.[78] It is water-insoluble and is not usually broken down by environmental physical or chemical processes. HCB has been detected in water, soil and sediments in North America and Europe. Monitoring of the Great Lakes region has occurred, as this was the site of much production of chlorobenzenes, like HCB.[84] HCB 22 accumulates in the food chain. In animal studies, HCB was excreted unchanged in feces or stored in fatty tissue. Excretion in urine also occurred, albeit to a lesser extent, as polychlorinated phenols. As HCB is not produced as an end-product or used as a pesticide in the U.S. anymore, dermal or airborne occupational exposure is limited to workers producing chemicals where HCB is a byproduct, farmers and sprayers using contaminated pesticides, those using pyrotechnics (e.g. military, firefighting), and those involved in incineration and waste disposal.[84] Other potentially higher exposure groups include those living near hazardous waste sites and factories where HCB is a byproduct and those consuming greater amounts of fish and game (e.g. fishermen, Aboriginal populations).   The main route of HCB exposure in the general human population is through consumption of contaminated foods, particularly fish, meat and poultry.[84] HCB residues have been detected in fish from Lake Superior and foods present in the average Canadian diet.[78] The average dietary HCB intake in Canadians was <1 ng/kg bw/day between 1976 and 1978 and 0.10 ng/kg bw/day between 1993 and 1996.[75, 76]   Mirex Mirex was first synthesized in the mid-1940s.[78] In the U.S., commercial production started in the 1950s and over 200 thousand kilograms were produced. Chlordecone, another insecticide, has been detected in some mirex formulations. One use of mirex was as an insecticide, particularly against fire ants. It was also an industrial chemical used as a fire-retardant additive with applications in paper, paint, rubber, electrical, adhesive, and textile.  Mirex is no longer produced commercially in the U.S.[85] In 1978, it was banned in the U.S., with certain exceptions. Mirex has never been produced or even registered in Canada, which means that this pesticide has never been legally imported, sold or used within Canada.[72] It is also included in the List of Toxic Substances in CEPA 1999.  Environmental exposure to mirex has occurred from its application to soil, leakage from waste disposal sites and runoff into surface water.[85] Mirex does not evaporate easily into air and is water-insoluble. It is quite stable and can remain in soil, water and sediment for 23 years.[85] Eventually, it degrades to chlordecone.[78] Mirex also reacts with sunlight to form the even more poisonous compound photomirex.[85] Mirex residues have been detected in animals living on mirex-treated land, as well as in cow milk and crops.[78] Mirex accumulates in the food chain and has been detected in aquatic animals.[85] It is not metabolized in the human body, but excreted unchanged in feces, and in urine to a lesser extent, within days or stored in fatty tissue for weeks or months.  In the general human population, mirex exposure most likely occurs via diet, particularly from eating fish from contaminated waters (e.g. Lake Ontario).[85] People living in contaminated areas, such as near waste sites, mirex-treated regions and manufacturing facilities, may also be exposed by touching and inadvertently ingesting contaminated soil (e.g. unwashed food). Because mirex is no longer produced, occupational exposure is now limited to workers involved in the disposal and cleaning of contaminated soil and sediment.  Chlordane/ chlordane-related compounds  Chlordane was first prepared in the 1940s.[78] Commercial production in the U.S. began in 1947.[86] Generally, “chlordane” refers to a mixture of chlordane isomers and other components. Technical-grade chlordane is composed of chlordane isomers, such as trans-chlordane (γ-isomer) and cis-chlordane (α-isomer) and additional components, such as nonachlor and the insecticide heptachlor. Heptachlor formulations also contain chlordane. Since the 1950s, chlordane has been used as an insecticide for agricultural and non-agricultural purposes, particularly for termite control. During the 1970s, many countries began restricting the use of chlordane. Chlordane has never been produced in Canada and registration of this pesticide was discontinued in 1998.[72]   Chlordane can persist for years in the air of buildings treated for termite control.[87] Residents of treated buildings would continue to be exposed years after treatment. Chlordane is also detectable in soil years after application. Because chlordane binds strongly to upper soil, it does not easily enter groundwater. Chlordane can evaporate from soil into the air. It can also react with light and chemicals in the atmosphere. Chlordane and its metabolites accumulate in the food chain. Chlordane, heptachlor and nonachlor have been detected in humans.[86] Once 24 absorbed, heptachlor is metabolized to heptachlor epoxide and trans- and cis-chlordane are metabolized to oxychlordane. Although trans-nonachlor is a minor component of technical chlordane, it is a major residue detected in human samples. These compounds are stored in adipose tissue or excreted in feces or urine.  Although the most common source of chlordane exposure is through the consumption of contaminated food, chlordane residues are not frequently detected in foods.[86, 87] trans-Nonachlor (% Positive: 3.2%; Mean positive: 0.5 ng/g), α-chlordane (% Positive: 2.7%; Mean positive: 0.5 ng/g), oxychlordane (% Positive: 1.5%; Mean positive: 1.5 ng/g), and γ-chlordane (% Positive: 0.9%; Mean positive: 1.0 ng/g) were detected in descending frequency in all food composites from six Canadian cities for the period of 1992 to 1996. The average chlordane (total) intake in Canadians was <1 ng/kg bw/day between 1976 and 1978 and 0.1 ng/kg bw/day between 1993 and 1996.[75, 76]  Before the restrictions on chlordane use were established, those involved in the manufacture or application of chlordane, such as factory workers, farmers and pest control workers would have had occupational exposure via dermal contact and inhalation.[87] Exposure can also occur in non-occupational settings through inhalation of air and dermal contact with soil at hazardous waste sites and contaminated areas.   1,1,1’-trichloro-2,2-bis(p-chlorophenyl)ethane/ 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene (DDT/ DDE) DDT was first synthesized in 1874 and was being commercially produced by 1943.[86] Although it was registered for use in Canada in 1946, it was never manufactured there.[72] Technical grade DDT generally contains p, p’-DDT/ DDE/ TDE (1,1-dichloro-2,2-bis-(p-chlorophenyl) ethane), the o, p’-isomers of these and other components.[86] DDT is used as an insecticide for agricultural and non-agricultural purposes, particularly public health initiatives against malaria. A number of countries placed restrictions on DDT use in the 1970s. The registration of DDT was discontinued in Canada in 1985 and is prohibited under CEPA 1999.[72] However, DDT continues to be used for vector control in some countries.  25 DDT and its metabolites, such as DDE and TDE, are not easily degraded.[86] p, p’-DDT is water-insoluble and fat soluble, and thus accumulates in the adipose tissue of animals.[44, 86] It has also been detected in air, soil and water.[7] DDT is also stored unchanged or as DDE or TDE, or excreted in the urine as 2,2-bis(p-chlorophenyl)acetic acid (DDA).[44] The occupations that may involve or may have involved exposure include agricultural and forestry workers, pesticide applicators and factory workers.[7]  Food is a major source of exposure for humans, although levels in food have been decreasing due to legislation in many countries.[44] p, p’-DDE was one of the most frequently detected pesticide residues in all food composites from six cities in the Canadian Market Basket Survey/ Canadian Total Diet Study between 1992 and 1996 (% Positive: 28.9%; Mean positive: 1.3 ng/g).[83] p, p’-DDT was detected less frequently (% Positive: 5.6%; Mean positive: 0.7 ng/g). The average DDT (total) intake in the Canadian Total Diet study decreased from 23 ng/kg bw/day between 1976 and 1978 to 2.44 ng/kg bw/day between 1993 and 1996.[75, 76]   2.2.2 Literature review of associations between OCs and NHL  The cancer risk for a number of OCs has been evaluated by the International Agency for Research on Cancer (IARC) Working Group. IARC classified PCBs as carcinogenic to humans (Group 1).[88] Chlordane and heptachlor were classified as possibly carcinogenic to humans (Group 2B), as evidence is inadequate in humans, but sufficient in experimental animals.[89] Hexachlorocyclohexanes were also classified as possible human carcinogens (Group 2B) because evidence in humans was inadequate and the evidence in animals was sufficient for the α-isomer and technical grade HCCH and limited for the β- and γ- isomers.[90] Similarly, mirex, chlordecone, DDT, and HCB are possible human carcinogens (Group 2B).[88] Aldrin and dieldrin were not classifiable (Group 3), as the evidence was inadequate for humans and limited for animals.[90]  Research assessing the relationship between organochlorine exposure and NHL has been reviewed elsewhere.[11-14] The evidence is somewhat inconsistent. This may be attributed to differences in the specific chemicals analyzed, as well as to methodological 26 limitations, such as insufficient power and/or imprecise exposure assessment. [12, 13] Much data has also been obtained from occupational groups. However, this evidence is often limited due to the small number of expected NHL cases and the use of mortality rather than incidence measures and surrogates for exposure, such as duration of employment.[12] In addition, discrepancies between findings in occupational groups and the general population may be attributed to the different composition of industrial-grade chemicals and chemicals in the environment. A subset of studies is summarized here.  Engel et al. noted exposure-response trends for NHLxxvi and lipid-corrected, prediagnostic blood levels of PCB-118, 138 and 153, using data collected between the 1970s and 1990 from three prospective cohorts: the CLUE 1 cohort from Maryland (74 cases and 147 controls), the American multistate NHS cohort (30 cases and 78 controls) and the community-based Janus cohort from Norway (190 cases and 190 controls).[15] These trends were statistically significant across cohorts and congeners, except for PCB-153 in NHS. Other PCB congeners common to the three cohorts were not the focus as they were detectable in less than 50% of samples and/or had high measurement error. Obvious trends were also not found for p, p’-DDE.  The CLUE 1 study also found no significant trends between NHL and prediagnostic serum levels of dieldrin, β-HCCH, HCB, trans-nonachlor, oxychlordane, heptachlor, heptachlor epoxide, summed chlordane/ heptachlor-related compounds (trans-nonachlor, oxychlordane. heptachlor, heptachlor epoxide), or summed DDT/ DDE (o, p’-DDT, p, p’-DDT, o, p’-DDE, p, p’-DDE).[16] A significant trend was found for summed PCBs (28 congeners), particularly in Epstein-Barr virus early antigen (EBV-EA) seropositive participants.[17] Lindane and mirex were not reported because fewer than ten percent of samples had detectable levels. Between 1969 and 1983, a nested case-control study, which used adipose samples from cadavers and surgical patients collected for the USEPA National Human Adipose                                                   xxvi NHL defined as ICD-8 codes 200 or 202 for Janus and CLUE I and 202 for NHS  27 Tissue Survey (175 casesxxvii and 481 controls), found significant trends for lipid-adjusted dieldrin, β-HCCH, oxychlordane, heptachlor epoxide, p, p’-DDT, and p, p’-DDE.[91] However, no significant associations were found for PCBs (Aroclor 1254 and 1260), HCB or trans-nonachlor. Because fewer than 20% of the samples had detectable levels for either cases or controls, aldrin, α-HCCH, δ-HCCH, γ-HCCH, mirex, and heptachlor were not analysed. The use of postdiagnostic samples, postmortem testing, different laboratories over many years, and cases with severe disease must be acknowledged.  In the 1990s, a Swedish case-control study by Hardell et al. examined the association between NHL and lipid-adjusted levels of HCB, cis-nonachlordane [cis-nonachlor], trans-nonachlordane [trans-nonachlor]. MC6 [nonachlor III], oxychlordane, p, p’-DDE, summed PCBs (36 congeners), summed immunotoxic PCBs (PCB-66, 74, 105, 110, 118, 128/ 167, 138, 156, 170/ 190), and summed chlordanes (trans-nonachlordane, cis-nonachlordane, MC6, oxychlordane), in blood and adipose tissue (82 cases and 83 controls).[92-94] Although the risk was increased for all analytes at levels above the median compared to below, the results were only statistically significant for HCB, trans-nonachlordane (borderline), MC6, and summed immunotoxic PCBs. However, the risks for the other analytes were significantly increased in participants with both high OC exposure and high EBV-EA titers, compared to low OC exposure/ low EBV-EA titer participants, except in the case of p, p’-DDE. In a subset of 100 cases and 100 controls from a four centre U.S. population-based case-control study conducted between 1998 and 2000, De Roos et al. did not find significant trends between NHLxxviii and the lipid-corrected plasma levels of the PCB congeners measured (PCB-74, 99, 118, 126, 138/ 158, 146, 153, 169, 170, 183, 187), nor with summed noncoplanar PCBs, dieldrin, β-HCCH, trans-nonachlor, oxychlordane, heptachlor epoxide, p, p’-DDT, or p, p’-DDE, .[18] Nevertheless, significant trends with                                                   xxvii NHL defined as ICD-8 codes 200 (lymphosarcoma and reticulum-cell sarcoma) or 202 (other neoplasms of lymphoid tissue); specific subtypes unclear  xxviii Includes CLL/ SLL  28 PCB-156, 180 and 194 were found. Because levels were not detectable in at least 30% of samples, the following OCs were not analyzed: some PCB congers (18, 28, 44, 49, 52, 66, 77, 81, 87, 101, 105, 110, 128, 149, 151, 157, 167, 172, 177, 178, 189, 195, 201, 206, 209, 196-203), aldrin, γ-HCCH, HCB, mirex, and o, p’-DDT. Significant trends between NHLxxix and PCB-180 and p, p’-DDE were found using carpet dust from 603 cases and 443 controls from the same parent study as the previous.[19] No significant trends were found for PCB-105, 138, 153, 170, α-chlordane, γ-chlordane, or p, p’-DDT. However, risk was significantly increased when total PCBs were categorized into detectable and undetectable groups. Because levels were not detected in at least 10% of samples, the following OCs were not analyzed: aldrin, dieldrin, lindane, and heptachlor.  Brauner et al. used a case-cohort design within a prospective Danish cohort (239 cases and 245 controls).[20] Participants provided prediagnostic adipose tissue at enrolment, which occurred between 1993 and 1997. Incidence rate ratios were examined for 18 lipid-adjusted OCs: PCB-99, 118, 138, 153, 156, 170, 180, 183, 187, 201, dieldrin, β-HCCH, HCB, cis-nonachlor, trans-nonachlor, oxychlordane, p, p´-DDT, p, p´-DDE, summed PCBs, summed immunotoxic PCBs (138, 153, 180), summed chlordanes (cis-nonachlor, trans-nonachlor, oxychlordane), and summed DDT/ DDE. A significant dose-response trend was only noted for p, p'-DDT.  2.2.2.1 Parent study (British Columbia non-Hodgkin lymphoma case-control study)  Between 2000 and 2004, data collection for the parent study was conducted at the BCCRC.[13, 14] This population-based case-control study was the largest study to date assessing associations between plasma OC levels and NHLxxx. Organochlorines were measured in the plasma of approximately half of the participants (422 cases and 460 controls). These OCs included 14 PCB congeners (dioxin-like PCBs: 105, 118, 156; non-dioxin-like PCBs: 28, 52, 99,                                                   xxix Excludes CLL  xxx Includes CLL/ SLL 29 101, 128, 138, 153, 170, 180, 183, 187) and 11 pesticides/ pesticide metabolites (aldrin, β-HCCH, HCB, mirex, α-chlordane, γ-chlordane, cis-nonachlor and trans-nonachlor, oxychlordane, p, p’-DDT, p, p’-DDE). Only 19 of these OCs were analyzed, as less than five percent of samples had detectable levels of PCB-52, PCB-101, PCB-128, aldrin, α-chlordane, or γ-chlordane.  Associations between these OCs and NHL were assessed in participants of all ethnicities. Significantly increased odds of NHL in the highest compared to the lowest OC categories and significant trends were found for total summed PCBs (OR=2.14 [1.38, 3.30]; p-trend=0.001), summed dioxin-like PCBs (OR=2.40 [1.53, 3.77]; p-trend <0.001) and summed non-dioxin-like PCBs (OR=2.18 [1.41, 3.38]; p-trend <0.001). These findings can be attributed to significantly increased odds of NHL in the highest compared to the lowest OC categories and significant trends for the dioxin-like congeners, PCB-118 and 156, and the non-dioxin-like congeners, PCB-153, 170, 180, and 187. Significance was maintained even after adjustment for multiple testing. PCB-99 and 138 showed significant trends, but not significantly increased odds in the highest compared to the lowest OC categories. The trend for PCB-138 also remained significant after adjustment for multiple testing.  Significant trends and significantly increased odds of NHL in the highest compared to the lowest OC categories were also found for the following pesticide analytes: β-HCCH, HCB, mirex, trans-nonachlor, and oxychlordane. p, p’-DDE showed only a significant trend. All associations, except for β-HCCH and mirex, remained significant after adjustment for multiple testing. No significant trends or significantly increased odds of NHL in the highest compared to the lowest OC categories were found for PCB-28, 105, 183, cis-nonachlor, or p, p’-DDT. However, it is important to note that aside from PCB-183, all of these organochlorines had over 60% of samples below the detection limit.  Associations were consistent across NHL subtypes (DLBCL, FL, other B-cell subtypes excluding DLBCL and FL, all T-cell subtypes). An exception to this was PCB-28 where OC levels were associated with increased risk of DLBCL and decreased risk of FL (borderline non-significant test for heterogeneity p=0.052).  In order to determine whether combinations of organochlorines give better risk estimates than individual OCs, a forward stepwise selection method was also implemented. 30 When all organochlorines were included in the stepwise procedure, only oxychlordane and summed non-dioxin-like PCBs remained in the final model (oxychlordane: ORhighest vs. lowest=2.28 [1.33, 3.91]; p-trend=0.002; summed non-dioxin-like PCBs: ORhighest vs. lowest=1.61 [0.94, 2.73]; p-trend=0.045). In separate stepwise models of PCBs only and pesticides only, PCB-153 and oxychlordane were the only OCs maintained in their respective models.  No significant multiplicative statistical interactions were found between organochlorines and age, sex, region, education, or body mass index (BMI). However, significant interactions were found between ethnicity (European vs. non-European) and both PCB-105 and p, p’-DDT. Those of non-European ethnic origin had increased and decreased risks of NHL with higher levels of PCB-105 and p, p’-DDT, respectively, whereas little effect was noted in Europeans. These interactions may be false positives, as a number of interactions were tested and non-Europeans are of diverse ethnic backgrounds, but were analyzed together because of small numbers.    2.3 Background on CYPs  2.3.1 Cytochrome P450 genes and protein products   The metabolism of chemicals by xenobiotic-metabolizing enzymesxxxi (XMEs) can be simplified into two phases: Phase I: functionalization and Phase II: conjugation.[29] Phase I generally involves adding a functional group to activate substrates, such as in the epoxidation of benzo[a]pyrene (BaP) to BaP trans-7,8-dihydrodiol-9,10-epoxide.  Subsequent conjugation by Phase II enzymes with other groups, such as glucuronide and mercapturic acid, results in hydrophilic products for excretion. Although metabolism can occur via this two-phase process, some compounds bypass Phase I and directly enter Phase II and others are further metabolized                                                   xxxi Xenobiotic-metabolizing enzymes: “Enzymes that can metabolize a foreign chemical. It has been suggested that no XME exists for the metabolism of xenobiotics alone, that is, all XMEs have one or more endogenous substrates or functions. Often regulated by ‘xenobiotic -sensing’ receptors.”[29]  31 by kidney and colon enzymes after Phase II, resulting in new mutagenic or carcinogenic intermediates.    Cytochrome P450 (CYP) enzymes play a major role in Phase I metabolism by catalyzing reactions, such as oxidation and dealkylation.[95] CYPs are membrane-bound heme-thiolate monooxygenases found in all domains of life.[29, 96, 97] At the C-terminus of all CYPs is the peptide motif, Phenylalanine–X(6–9)–Cysteine–Xxxxii–Glycine. The sulphurous thiol group of cysteine binds heme iron and is involved in transferring an oxygen atom to the substrate during catalysis. The CYP enzymes are encoded by their respective CYP genes. In humans, the CYP gene superfamily encodes 57 genes and more than 59 pseudogenes belonging to 18 families and 43 subfamilies.   Interestingly, CYPs are involved in both the detoxification and the activationxxxiii of chemicals that can damage cellular DNA and proteins.[29] It is estimated that three quarters of xenobiotics become carcinogenic after CYP-mediated metabolism and only one quarter are carcinogenic in their original, parent form. During CYP-mediated metabolism, unstable reactive oxygenated intermediates (ROMs) can be generated and these can potentially bind to DNA and proteins. This can mutate oncogenes and tumour suppressor genes and initiate tumours.[29, 97] CYPs can also affect the levels of other molecules and downstream signal transduction pathways. These alterations can result in tumour promotion and progression, by impacting the cell cycle, apoptosis and cell growth.   CYP expression varies by specific CYP, time, organ, tissue, and cell.[29] Differences in CYP expression and activity at the individual level can be attributed to variations in the genes encoding the CYPs and differences in environmental exposures.[99]  Not only are CYPs involved in the metabolism of diverse xenobiotics, such as some tricyclic antidepressant drugs, but they are also involved in the metabolism of endogenous toxins, such as bilirubin.[28, 32] The specific substrates, inducers and inhibitors of CYPs vary by the specific CYP gene.[29] In general, CYP1 to CYP4 metabolize a broad spectrum of xenobiotics and endogenous substrates, with redundancy                                                   xxxii “X” denotes any amino acid  xxxiii Bioactivation: Transformation of a compound to a more biochemically active state.[98]  32 in their substrates. Below, some functions of specific CYPs within select CYP families are highlighted.   CYP1 There is much overlap in the substrate specificity of CYP1 enzymes.[100] CYP1A1 and CYP1B1 are known to be involved in the metabolism of polycyclic aromatic hydrocarbons, such as the procarcinogens found in cigarette smoke, and polyhalogenated aromatic hydrocarbons, such as PCBs.[29] CYP1A2 is involved in the metabolism of various drugs, and endogenous compounds, such as steroids.[100] CYP1A1, CYP1A2 and CYP1B1 are under the regulation of the AHR pathway, which also controls the transcription of genes involved in growth, cell cycle and apoptosis.[29] Analogous regulatory loops most likely exist between other CYP families and their XME receptorsxxxiv.   CYP2 CYP2A6 and CYP2A13 are 94% identical in amino acidxxxv sequence, but are primarily expressed in the liver and respiratory systems, respectively.[101] CYP2A6 metabolizes nicotine, certain drugs and endogenous substrates, like retinoid acids and steroids.[102] CYP2A13 metabolizes various toxins and procarcinogens, including aflatoxin B1.[101] CYP2B6 and CYP2C9 also metabolize both xenobiotic and endogenous compounds and are important liver CYPs.[103, 104] CYP2C19 is involved in the biotransformation of various drugs, including some proton pump inhibitors and tricyclic antidepressants.[32] CYP2E1 activates many xenobiotics to toxic or carcinogenic products.[97] Among its substrates are benzene, ethanol and chlorinated solvents.                                                    xxxiv Xenobiotic-metabolizing enzyme receptors: These “xenobiotic-sensing” receptors respond to stimuli by regulating XME gene expression. AHR is an example of an XME receptor.[29]  xxxv Amino acid: Building block of protein.[33]  33 CYP3  The CYP3As metabolize a broad spectrum of substrates and are major CYPs in the liver and intestine.[105] CYP3As, which include CYP3A4, CYP3A5 and CYP3A7, are heavily involved in drug metabolism, although they also metabolize endogenous steroids.[99] Less is known about the more recently discovered CYP3A43; however, it has shown testosterone 6β-hydroxylase activity in an E. coli expression system.[106]   Other CYP families  CYP7A1 is involved in cholesterol metabolism and bile acid synthesis.[107] Similarly, CYP7B1 acts on oxysterols and steroids.[108] The CYP17s are involved in steroid biosynthesis.[109] CYP19A1, which is encoded by the aromatase gene, converts androgens into estrogens and is active in both gonadal and extra-gonadal tissues.[96] CYP24A1 functions as a vitamin D hydroxylase.[110]   2.3.2 Literature review of associations between CYPs and NHL   The associations between CYPs and NHL have been examined in several candidate gene studies. These are summarized here. Refer to Appendix A for details on the literature search strategy. Details on CYP variant nomenclature can be found on the Human Cytochrome P450 (CYP) Allele Nomenclature Database.[111] In a subset of 1172 cases and 982 controls from the four-centre U.S. population-based case-control study described in Section 2.2.2, De Roos et al. assessed associations between NHLxxxvi and the following CYP variants: CYP1A1 I462V (rs1048943), CYP1A2*1F (rs762551), CYP1B1 V432L (rs1056836), CYP2C9 R144C (rs1799853), and CYP2E1 -332A>T (rs2070673) and -1054C>T (rs2031920).[35] None of the trends for risk of NHL were statistically significant. However, a borderline non-significant trend was noted between NHL and CYP1B1 V432L (rs1056836). This trend was significant in B-cell lymphomas and DLBCL. Individuals heterozygous or homozygous TT for CYP2E1 (rs2031920), who were combined                                                   xxxvi Includes CLL/ SLL 34 due to small numbers, had a significantly decreased risk of NHL (ORCT/TT vs. CC= 0.59 [0.37, 0.93]) and B-cell lymphomas. The same pattern was noted in other NHL subtypes.  Al-Dayel et al. looked at associations between CYP1A1 variants and DLBCL in Saudi Arabia (182 cases and 513 controls).[112, 113] DLBCL patients had been diagnosed at the King Faisal Specialist Hospital and Research Centre (KFSHRC). DNA samples from healthy individuals recruited from the general Saudi population were randomly selected from the institutional Tissue Biorepository Bank at KFSHRC and used as controls. Statistically significant associations were found between DLBCL and CYP1A1 4887C>A (rs1799814, 2453C>A, 2452C>A, BsaI; ORAA vs. CC=2.34 [1.11, 4.95]) and CYP1A1*2C (3801T(Wild-type (WT))+4889G(mutation); OR2C allele= 6.63 [1.56, 28.1]). However, no statistically significant associations were found between DLBCL and CYP1A1 4889A>G (rs1048943, 2455A>G, 2454A>G, BsrDI), 3801T>C (rs4646903, 3798T>C, MspI), or the variant combinations: CYP1A1 WT (3801T(WT)+4889A(WT)), CYP1A1*2A (3801C(mutation)+4889A(WT)) or CYP1A1*2B (3801C(mutation)+4889G(mutation)). It is important to note that because archival samples were used, the biospecimen type was different between cases and controls (paraffin-embedded lymph node DLBCL tissues vs. blood). Although the assay was replicated, rs1048943 was also out of Hardy-Weinberg equilibrium in controls.   Barry et al. performed a genetic analysis on a population-based case-control study of Connecticut women conducted between 1996 and 2000 (518 cases and 597 controls).[114] No statistically significant associations were found between CYP2E1 (rs2031920) and NHL or its subtypes (i.e. FL, DLBCL, CLL/ SLL, MZL, T-cell lymphoma). In addition, no statistically significant interactions were found between this variant and ever being exposed to any organic solvent, any chlorinated solvent or individual organic solvents. Although no statistically significant associations were found between CYP2E1 (rs2070673) and NHL or its subtypes, statistically significant interactions were found between this variant and some of the chlorinated solvents.  Analyzing data from the same study of Connecticut women, Kilfoy et al. found no statistically significant associations between CYP1A1 (rs1048943), CYP1A2 (rs762551) or CYP1B1 (rs1056836) and NHL.[115] No significant interactions were found with smoking in 35 NHL, DLBCL, FL, CLL/ SLL or MZL. However, heterozygous or minor allele homozygous genotypes for CYP1A1 and CYP1B1 did confer borderline increases and decreases, respectively, in the risk of DLBCL compared to major allele homozygous genotypes in non-smokers. CYP1A2 was associated with a borderline increase in risk of NHL in smokers.   Also in this sample of Connecticut women, Li et al. found no statistically significant associations between NHL and the SNPs, CYP1A1 (rs1048943), CYP1A2 (rs762551), CYP1B1 (rs1056836), CYP2C9 (rs1799853), CYP2E1 (rs2070673), or CYP21A2 (rs6474), within categories of alcohol consumption.[116] Furthermore, there were no significant interactions in NHL. When analyses were limited to DLBCL, a significant association was only found with CYP1A1 in liquor-drinkers. The associations between DLBCL and CYP1A1 and CYP2C9 were significantly different between liquor-drinkers and non-drinkers. Although numbers were small, the association between CYP2E1 and T-cell histologic types was also significantly different between non-drinkers and wine-drinkers. No significant interactions or associations within alcohol consumption categories were found in marginal zone lymphomas.  An analysis of a subset of these Connecticut women who provided blood samples (461 cases and 535 controls) was conducted by Zhang et al. to assess modification of the risk of NHL conferred by hair-dye use by CYP1A1 Ex7+131A>G (rs1048943), CYP1A2 IVS1-154C>A (rs762551), CYP1B1 Ex3+251G>C (rs1056836), CYP2C9 Ex3-52C>T (rs1799853), CYP2E1 -332T>A (rs2070673), or CYP2E1 (rs2031920).[117] No significant associations were found between any of these variants and NHL, DLBCL, FL, or CLL/ SLL. A significant interaction was found between CYP2C9 and hair-dye use before 1980. Similar trends were found in FL.   In a population-based case-control study conducted in the San Francisco Bay area between 1988 and 1995 (308 cases and 684 controls), Skibola et al. found a significantly higher risk of NHLxxxvii in those homozygous for a variant of CYP17A1 75C>G (rs3740397; ORGG vs. CC =1.6 [1.1, 2.5]).[119] Similar trends were noted in men and women. Significantly or borderline increased risks of DLBCL were found in homozygotes of the CYP17A1                                                   xxxvii CLL not included.[118]  36 variants -34T>C (rs743572), 137G>A (rs6162), 195C>A (rs6163), 105A>C (rs743575), 75C>G, and 2930G>T (rs4919685). No associations were significant for CYP17A1 -270A>C (rs3781287), 35T>C (rs1004467), or 114A>T (rs10883783) or in FL. In a British population-based case-control study conducted between 1998 and 2001 (620 cases and 762 controls), Skibola et al. reported statistically significant increased risks of NHL (ORCC vs. TT=1.44 [1.02, 2.03]) and DLBCL for CYP17A1 -34T>C (rs743572).[120, 121] Similar patterns were noted in males and females. Non-significantly increased risks of NHL, DLBCL and FL were found for minor allele homozygotes of CYP17A1 IVS2 105A>C (rs743575). However, when stratified by sex, a statistically significant increased risk of FL was seen in males.   Kim et al. looked at the association between NHL and CYP1A1, and whether smoking modifies this relationship, in a Korean population-based case-control study conducted between 1997 and 2008 (713 cases and 1700 controls).[122] CYP1A1 4889A>G (rs1048943) was significantly associated with NHL (ORAG/GG vs. AA=1.26 [1.06, 1.51]) and DLBCL. Similar but non-significant trends were noted in T-cell lymphoma. Although no statistically significant associations were found for CYP1A1 6235T>C (rs4646903) in NHL or DLBCL, a borderline association was found in T-cell lymphoma. No interactions with smoking status were found.   Gra et al. assessed the risk of B-cell chronic lymphocytic leukemia and T-cell NHL conferred by variants of CYP1A1, CYP2C9, CYP2C19, and CYP2D6 in a Russian case-control study (83 B-CLL, 76 T-cell and 177 controls).[123, 124] No statistically significant associations were found for CYP2C19 (*1, *2 (rs4244285, 681G>A)) or CYP2D6 (*1; *3 (rs35742686, 2637delA, 2549delA), *4 (rs3892097, 1934G>A, 1846G>A)). However, a statistically significant effect was found for CYP2C9*2/*3 (*2 (rs1799853, 430C>T), *3 (rs1057910, 1075A>C); OR=9.48 [1.03, 86.9]) in B-CLL. This effect was also significant in males, as was CYP2C9*2. Combinations of CYP1A1 polymorphisms (4887C>A (rs1799814, 2452C>A); 4889A>G (rs1048943, 2454A>G); 6235T>C (rs4646903, 3798T>C)) resulted in a borderline increased risk of B-CLL. This was significant in males.   An Australian case-control study by Kerridge et al. did not find any significant 37 associations between the CYP1A1 m1 mutation (rs4646903, MspI) and NHL, follicle centre cell [FL], diffuse large cell [DLBCL] or “other” NHL subtypes (169 cases and 205 controls).[125, 126] It is important to note that in this study lymph node or bone marrow specimen were used for genotyping. Similarly, Lemos et al. did not find any significant differences in the distribution of CYP2D6*4 (rs3892097, 1934G>A, 1846G>A) genotypes or alleles between 128 controls and 71 NHL or 13 CLL cases.[127, 128] This Portuguese case control study used a convenience sample of hospital and research staff and students as controls.   Finally, a case-control study conducted in the Czech Republic (143 cases and 455 controls) by Sarmonova et al. found no significant associations between NHL or its subtypes (follicular or diffuse) and the CYP1A1 3'-flanking region (*2A: rs4646903, 3801T>C, 3798T>C, MspI), CYP2E1 5'-flanking region (*5B: rs2031920, -1053C>T, RsaI; rs3813867, -1293G>C, PstI) or CYP2E1-intron 6 (*6: 7632T>A, DraI).[129] It is important to note that control recruitment was conducted among public health and medical staff and residents of elderly housing. Similar findings were found in a follow-up publication by Soucek et al.[130]  To summarize, very little evidence was found of an association between NHL and variation in CYP1A1, CYP1A2, CYP1B1, CYP2C9, CYP2C19, CYP2D6, CYP2E1, or CYP21A2. There is limited evidence of an association between NHL and CYP17A1.    [] 2.4 Chapter synopsis  Lymphomas are cancers that originate in lymphocytes, which are white blood cells involved in the immune response. In Canada, approximately 90% of lymphomas are non-Hodgkin lymphomas. NHL is the fifth most common cancer in Canada. Although a variet y of genetic and environmental (non-genetic) risk factors for NHL are known, the causes of most lymphoma cases are unclear. Exposure to a number of chemicals, some of which have immunotoxic properties, can increase the risk of NHL. Potential chemical risk factors 38 include exposure to organochlorines, which are ubiquitous chlorine-containing organic molecules. These include PCBs and organochlorine pesticides. Variations in a number of genes encoding enzymes that metabolize xenobiotic and endogenous chemicals have also been suspected to increase the risk of NHL. Among these are the cytochrome P450 genes, which encode enzymes that metabolize substrates ranging from steroids to some organochlorines.    39 Chapter 3 Study design and methods  3.1 Study design    Study design details for the parent study have been described elsewhere.[13, 14, 25] A brief description is given here, followed by specific details of the Europeans-only subset where applicable.  3.1.1 Study population   For the parent study, individuals between the ages of 20 and 79 years residing in the Greater Vancouver regional district (GVRD) and greater Victoria (Capital regional district; CRD) were recruited for study participation. The GVRD and CRD comprise ~60% of the British Columbian population.[131] Data collection occurred between March 2000 and February 2004. Cases and controls were first contacted to participate in the parent study by letter and followed up by telephone, if they did not respond after two weeks. Individuals were excluded if they were unable to give informed consent or were unable to complete the questionnaire due to language limitations. Study material was available in English, Chinese, Punjabi, and Tagalog as these are the most commonly spoken languages in the catchment area.   The study questionnaire was conducted by self-report and a computer-assisted telephone interview, administered by trained interviewers using a standardized protocol. The interview collected data on demography, medical history and various exposures of interest. Participants’ ethnicities were derived from the self-reported ethnicities of their four grandparents. Participants were also asked to provide a non-fasting blood sample, or in the case of refusal, a mouthwash or saliva sample. All biospecimen types were used for genotyping assays. However, organochlorine levels were only measured in blood.   40 3.1.2 Cases   For the parent study, all newly diagnosed cases of NHL during the study period were ascertained from the BC Cancer Registry. The BC Cancer Agency oversees this population-based registry of cancer diagnoses in BC residents.[132] Cases were excluded from the parent study if their NHL diagnoses could not be reviewed by expert lymphoma pathologists and classified to a subtype under the WHO classification.[133] Cases were also excluded if they were HIV-positive or had a post-transplantation-related lymphoproliferative disorder, as immunosuppression is a rare and important NHL risk factor.[3, 134, 135]  Refer to Figure 3.1 for a flowchart of sample selection. In the parent study, 1263 cases were potentially eligible for participation. At the time of the all ethnicities analyses, 828 cases had consented and had at least partially completed the study questionnaire. Since the original study analyses, a subsequent review has resulted in the removal of an additional seven cases who were found to have been initially diagnosed outside the study period and an additional two cases who were found to be HIV-positive.  Plasma organochlorine levels, adjusted for lipids, were measured in a subset of participants. More specifically, the organochlorine analyses previously conducted in all ethnicities were restricted to the 422 cases who not only had sufficient plasma for lipid-adjustment, but had also provided blood prior to chemotherapy and had not experienced >10% weight loss within the year preceding blood collection. Weight loss and chemotherapy are known to alter blood organochlorine levels.[136, 137] Weight loss due to diet and exercise is expected to occur in both cases and controls, however >10% weight loss is also a general B symptom of cancer.[138] Weight loss was ascertained through chart reviews, followed by contacting family physicians and then the participants. Here, a re-evaluation of the organochlorine analyses previously conducted in all ethnicities was restricted to the 331 cases who were of European ethnic origin and passed organochlorine quality control. Refer to Section 4.1.1.  DNA (blood and/or mouthwash and/or saliva) was available for 797 cases for both the Illumina and Taqman genotyping assays.[139] Refer to Figure 3.2, Figure 3.3 and Section 4.2.1. 41 Six hundred-three European cases were available for CYP genetic analyses (Illumina or Taqman) after quality control. For the identification of gene-environment interactions between CYP SNPs and organochlorines, 329 European cases were available for analysis.   42    Figure 3.1  Participant selection for parent NHL study and subsequent organochlorine analyses  Details on sample selection up to the 422 cases and 460 controls included in the parent OC analyses ca be found in [13, 14].    1263 potential cases 2373 potential controls133 deceased 13 deceased62 not contacted 504 not contacted1068 contacted cases 1856 contacted controls147 refusals 856 refusals73 poor health 103 poor health8 language barrier 49 language barrier840 consented cases12 not classified/prior transplant828 cases in parent study 848 consented controls in parent study59 no blood 169 no blood314 post-chemotherapy blood455 cases with pre-chemotherapy blood 679 controls with blood24 >10% weight loss8 unknown weight loss 463 frequency-matched controls with blood1 no lipid-adjustment 3 no lipid-adjustment422 cases in parent OC analysis 460 controls in parent OC analysis1 diagnosis outside study period 5 sample switches1 HIV-positive 1 control to case conversion7 sample switches1 sibling pair412 cases in updated OC study 454 controls in updated OC study81 non-European 91 non-European331 European cases in updated OC analysis 363 European controls in updated OC analysis43  Figure 3.2  Participant selection for Illumina genetic analyses     828 cases in parent study 848 consented controls in parent study797 cases with DNA (blood, mouthwash, saliva) 790 controls with DNA (blood, mouthwash, saliva)15 sample switches 9 sample switches782 cases 781 controls61 call rate <0.98 85 call rate <0.981 excess heterozygosity 1 excess heterozygosity1 control to case conversion720 cases2 sibling pairs1 ethnicity discrepancy717 cases7 diagnoses outside study period1 HIV-positive709 eligible cases for Illumina study 694 eligible controls for Illumina study145 non-European 147 non-European564 European cases for Illumina analysis 547 European controls for Illumina analysis44  Figure 3.3  Participant selection for Taqman genetic analyses   828 cases in parent study 848 consented controls in parent study797 cases with DNA (blood, mouthwash, saliva) 791 controls with DNA (blood, mouthwash, saliva)7 diagnoses outside study period2 HIV-positive 1 missing sample788 eligible cases for genotyping 1 sample 0 DNA1 missing sample787 cases 789 controls14 sample switches 8 sample switches2 sibling pairs 1 control to case conversion771 cases 780 controls18 missing >1 genotype 29 missing >1 genotype753 eligible cases for Taqman study 751 eligible controls for Taqman study155 non-European 169 non-Europeans598 European cases for Taqman analysis 582 European controls for Taqman analysis45  3.1.3 Controls   For the parent study, random population controls were ascertained from the BC Ministry of Health Client Registry, which includes subscribers to the provincial medical services plan (MSP). All eligible BC residents are required to enroll in MSP.[140] MSP captures Canadian citizens or permanent residents who reside in BC for at least half the year, but not visitors to BC. Persons in BC for fewer than three months are generally excluded. Also excluded are members of the Canadian Armed Forces, whose healthcare falls under federal jurisdiction, and individuals receiving services from First Nations under self-government agreements.[141] Controls were frequency-matched by age (5-year age groups), sex (Male or Female) and region (GVRD or CRD) in a ~1:1 ratio to cases.  Refer to Figure 3.1 for a flowchart of sample selection. In the parent study, 2373 controls were potentially eligible for participation. Eight hundred forty-eight controls consented and at least partially completed the study questionnaire. A random subset of 463 controls frequency-matched to cases was initially selected for the organochlorine analysis of all ethnicities. This number was reduced to the 460 controls who had sufficient plasma for lipid measurement. Of these, 363 were included in the Europeans-only re-evaluation conducted here after organochlorine quality control. Refer to Section 4.1.1.  DNA (blood and/or mouthwash and/or saliva) was available for 790 controls for the Illumina genotyping assay and 791 for the Taqman genotyping assay.[139] Refer to Figure 3.2, Figure 3.3 and Section 4.2.1. Five hundred ninety-four European controls were available for CYP genetic analyses (Illumina or Taqman) after quality control. For identification of gene-environment interactions between CYP SNPs and organochlorines, 360 European controls were available for analysis.   46 3.2 Organochlorine measurement methods  3.2.1 Organochlorine measurement  Organochlorine measurement protocols have been described elsewhere.[13, 14] Briefly, the following 14 PCB congers were quantified in plasma samples by gas chromatography: 28, 52, 99, 101, 105, 118, 128, 138, 153, 156, 170, 180, 183, and 187. The following 11 pesticides/ pesticide metabolites were also quantified: aldrin, β-HCCH, HCB, mirex, α-chlordane, γ-chlordane, cis-nonachlor, trans-nonachlor, oxychlordane, p, p’-DDT, and p, p’-DDE. These organochlorines belonged to a pre-specified panel from the Toxicology Centre at the Québec National Public Health Institute. Organochlorine detection limits were as follows: 0.025 µg/L for PCB-52, β-HCCH, p, p’-DDT, and p, p’-DDE, and 0.015 µg/L for all other organochlorines of interest.   3.2.2 Organochlorine quality control  Participants were removed if they had been identified as problematic during the Illumina QC process described in Section 3.3.6.1 (e.g. possible sample swaps, control to case conversion, relatedness between participants). Ethnicity updates made during Illumina QC were also replicated here.   3.2.3 Organochlorine lipid-adjustment  Lipid adjustment methods have been described elsewhere.[13, 14] Briefly, whole-weight organochlorine measurements were divided by total lipid concentrationsxxxviii, in order to calculate lipid-adjusted organochlorine concentrations. Undetectable                                                   xxxviii The Patterson et al. formula was used: Total lipid = 1.677 (TC-FC) + FC + TG + PL; FC=free cholesterol; TC=total cholesterol; TG=triglycerides; PL=phospholipids .[142]  47 organochlorines were set as the detection limit divided by √2 prior to lipid -adjustment.[143] Median lipid-adjusted organochlorine detection limits were calculated by dividing the detection limits by the median total lipid concentration in all samples.  Summed PCB measures were also created and adjusted for total lipid  concentration. Lipid-adjusted total summed PCBs was calculated by summing the whole-weight PCB measurements for each sample and dividing by the total lipid concentration. Undetectable PCB values were set as the detection limit divided by √2 prior to summation. An analogous method was used to calculate lipid-adjusted summed dioxin-like PCBs and lipid-adjusted summed non-dioxin-like PCBs. Dioxin-like PCBs include PCB-105, 118 and 156, and non-dioxin-like PCBs include all other PCB congeners analysed. These lipid-adjusted organochlorine concentrations, with undetectables set as the detection limit divided by √2, were used in subsequent analyses.   3.3 Genetic measurement methods  3.3.1 Gene selection  Candidate xenobiotic-metabolizing genes, including the CYPs, were selected for study, on the basis of their purported involvement in the metabolism of xenobiotics  and/or role in cancer, by reviewing the literature and mining BioMartxxxix (ensembl.org/biomart/martview/).                                                      xxxix BioMart (ensembl.org/biomart/martview/) is a search-engine used to mine the Ensembl project, which brings together genomic data for a number of chordates, including humans. [144, 145]   48 3.3.2 TagSNP selection   TagSNPs corresponding to these genes were selected from publicly available data for the CEPH population (Utah residents with ancestry from northern and western Europe; hereafter referred to as “CEU”) in HapMapxl.[24] Eligible SNPs had a Minor Allele Frequency (MAF)xli ≥ 0.05, in order to focus on relatively common variants, and were selected using Tagger Pairwisexlii. The collection of tagSNPs was chosen to ensure good proxies for the untyped SNPs in the region. This was done by setting the minimal coefficient of determination, r2, to 0.8. r2 is a measure of linkage disequilibrium or “the nonrandom association of alleles at different loci”.[146, 147] Values range from 0 to 1, with 1 indicating that the markers are perfect proxies.  3.3.3 DNA extraction and whole genome amplification  Procedures for genomic DNA extraction from biospecimen and DNA quantification have been described elsewhere.[14, 139] Samples with insufficient DNA were whole genome amplified (WGA). Protocols have been described elsewhere.[14, 25, 139]                                                     xl The International HapMap Project (http://www.hapmap.org) seeks to map common genetic variation in humans in order to better understand disease.[31]  xli Minor Allele Frequency: The frequency of the less abundant allele of a SNP in a population.[24]  xlii “Pairwise Tagging: Pairwise tagging means that all tag SNPs will act as direct proxies to all other untyped SNPs because they are highly correlated with one another. In pairwise tagging mode, Tagger should behave similarly to ldSelect developed by Carlson et al. Multimarker Tagging: Aggressively searches for multi-marker predictors to capture all alleles of interest (SNPs and/or haplotypes). Increases the efficiency of tagging.”[24] 49 3.3.4 Sequencing  3.3.4.1 Taqman  CYP1A1 tagSNPs were supplemented with variants found in 31 randomly selected cases through bidirectional sequencing. Sequencing was conducted by staff at the Brooks-Wilson lab in the Genome Sciences Centre at the BC Cancer Agency (BCCA). The sequencing protocol is similar to that described by Ng et al. for AHR.[14, 25]  3.3.5 Genotyping  3.3.5.1 Illumina  Genotyping of the majority of CYP SNPs (119) was performed using a custom Illumina GoldenGate Genotyping Assay (Illumina, San Diego, CA), which included SNPs from other hypotheses as well.[148] The genotyping protocol has been described elsewhere.[139] Of note, prior to genotyping, tagSNPs that were predicted to fail assay design were replaced with equivalent tagSNPs, wherever possible. Reasons for failing assay design include SNPs positioned too close together. Fifty-three replicate pairs of non-WGA samples, as well as 52 WGA samples with matched pre-whole genome amplification counterparts (52 WGA/ non-WGA pairs) were included in the assay, as quality control measures.  In order to detect population stratification, which can confound the disease-gene relationship, the assay also included 51 ancestry informative markers (AIMs) selected from Halder et al.[149, 150] AIMs are SNPs for which the ancestral populations of interest have very different allele frequencies.[150] The presence of population stratification was assessed using genomic control methods, by calculating the genomic inflation factor, lambda (λ).[151, 152] A λ value far from 1 is suggestive of population stratification.   50 3.3.5.2 Taqman  In addition to genotyping using Illumina, seven CYP1A1 SNPs derived from sequencing results and three additional CYP1A1 tagSNPs selected from HapMap were genotyped using TaqMan allelic discrimination assay optimized with the Assays-by-DesignSM service (Applied Biosystems, Foster City, CA, USA).[153] The genotyping protocol was similar to that described by Ng et al. for AHR.[14, 25] The sequenced CYP1A1 SNPs were selected for inclusion in the Taqman assay based on the following criteria: MAF ≥ 0.05 and/or the variant resulted in an amino acid residue change that would not be tolerated according to SIFT xliii or was probably damaging according to PolyPhenxliv. Please note that rs4646421 was also genotyped by Illumina.  3.3.6 Genetic quality control  3.3.6.1 Illumina  The quality control process for the Illumina assay has been described elsewhere.[139] Briefly, the rationale behind the QC process was to remove low quality SNPs followed by low quality samples, in order to preserve sample size, by applying progressively stringent quality criteria. Samples were removed from the analysis if sample call rates were <0.98. Samples were also removed if they demonstrated excess heterozygosity in autosomal SNPs, as determined by more observed heterozygote calls than would be expected in European non-WGA samples. Excess sample heterozygosity may indicate cross-contamination between samples.[157] Other reasons for sample removal included participants who changed from control                                                   xliii SIFT is an algorithm used to predict the effects of coding non-synonymous variants on protein function.[154]  xliv PolyPhen automates the functional annotation of coding non-synonymous SNPs.[155, 156]  51 to case status (and were therefore only included as cases) or participants who were related. These types of sample problems were identified using identity-by-state (IBS) values. Essentially, IBS assesses the average number of alleles that are identical between participants across all assayed SNPs to infer the degree of relatedness.[158] Possible relationships between participants were then confirmed by further review of study records.  Samples were also removed if possible sample swapping could have occurred during sample processing. One of the indicators of sample swapping was discrepancies between self-reported sex and sex determined through genetic verification. Another indicator of possible sample swapping was discrepancies between self-reported ethnic origin and ethnicity predicted using multidimensional scaling (MDS) plots with HapMap populations set as the references.[159] In these instances, the samples were removed or their ethnicity changed to “other”, depending on the outcome of a manual review.  SNPs were excluded if any of the following criteria were met: GenTrain xlv scores <0.4, GenTrain scores greater than or equal to 0.4 and less than 0.7 in combination with poor clustering (i.e. non-discrete or abnormal numbers of clusters), possible copy number variants (CNVs) as indicated by more than three clusters, possible problematic genotyping as indicated by monomorphism (MAF=0), or SNP call rates <0.95 overall. SNPs with call rates <0.95 in a specific sample type (whole blood, lymphocytesxlvi, mouthwash or saliva) were only removed from that sample type due to differences in their chemistries.  Furthermore, SNPs were removed if there was discordance between any of the 53                                                   xlv GenTrain scores are assigned based on a clustering algorithm to assess the shape and relative distances of the three possible genotype clusters for each SNP.[160] Based on Illumina recommendations, the GenCall threshold was also set at 0.25 for all genotypes. [161] GenCall scores provide a method of filtering out failed genotypes based on the GenTrain score and a genotype-calling algorithm.[160]  xlvi Until 2003, genomic DNA was extracted from lymphocytes isolated using the Ficoll gradient technique. However, this was replaced by extraction from whole blood due to low DNA yields  in the original protocol.[14]  52 pairs of replicate non-WGA samples included in the assay.xlvii Additionally, SNPs out of Hardy-Weinberg Equilibriumxlviii (HWE; p<0.001) in European controls were removed from the analysis, as this could indicate problematic genotyping.  The preceding QC steps were conducted separately on non-WGA and WGA samples due to possible differences in the intensity of their genotype signals. [164] Any SNPs that were removed from the non-WGA samples were also removed from the WGA samples, but not vice-versa again due to the purported higher quality of non-WGA samples. If all the above criteria were met but there was discordance between any of the 52 WGA/ non-WGA pairs, the discordant SNPs were excluded from WGA samples only.xlix  3.3.6.2 Taqman   A similar QC process as was implemented for the Illumina assay was implemented for the Taqman assay used to genotype CYP1A1 SNPs. Samples identified as problematic and excluded in the Illumina QC process (e.g. related samples, controls who became cases, sample switches) were automatically excluded from the Taqman dataset. The exceptions to this automatic exclusion were samples that were excluded due to excess heterozygosity, as contamination issues were likely specific to the Illumina plates, and samples that were inadvertently not genotyped using the Illumina assay, but were not problematic per se.  Samples missing more than one genotype (i.e. sample call rates of <8/9 or 0.89,                                                   xlvii N.B. At the start of the QC process, there were 53 non-WGA replicates, but only 49/50 remained by the time this criterion was applied. SNPs in which the genotype was called in one replicate, but not in the other, were not considered discordant. These SNPs were still included with the called genotype being kept.   xlviii Hardy-Weinberg equilibrium is the theoretical allele frequencies and genotypes that would be present in a stable, randomly mating population free of mutations, gene flow (e.g. migration) and selection (e.g. natural selection).[162, 163]   xlix N.B. At the start of the QC process, there were 52 WGA/ non-WGA pairs, but only 28 remained by the time this criterion was applied. SNPs in which the genotype was called in one member of the WGA/ non-WGA pair, but not in the other, were also not considered discordant. 53 rather than <0.98 as in Illumina) were excluded. Similarly, samples were removed if they were found to have low quality extracted DNA, as assessed by a DNA concentration of 0 and failed amplification (using the Polymerase Chain Reaction (PCR))  or if they were misplaced. Sample contamination and assay quality were assessed by the laboratory using negative controls and checks for Mendelian l errors in positive controls from multigenerational CEPH families.[166]  As in the Illumina assay, QC processes were carried out separately in WGA and non-WGA samples. If a SNP was removed from non-WGA samples, it was also removed from WGA samples, but not vice-versa. If SNP call rates were <0.95 within a specific sample type, they were removed from that sample type. Because so few SNPs were being assessed, they were removed if they departed from HWE in European controls at a threshold of 0.05.  Furthermore, 33 replicate pairs were used to assess genotyping inconsistencies in the Taqman assay. These replicates were comprised of non-WGA samples from two individuals who were genotyped twice and 31 individuals who were both genotyped and sequenced. The latter 31 individuals were the randomly chosen cases discussed previously, in whom CYP1A1 variants had been discovered through sequencing (i.e. the non-tagSNPs). SNPs were removed if there was discordance between any of these 33 pairs of replicate samples.li With respect to WGA/ non-WGA sample pairs, 269 non-WGA samples were also whole genome amplified. If there were discrepancies between the genotypes of these WGA/ non-WGA pairs, the discordant SNPs were excluded from the WGA samples only.lii                                                   l Mendelian errors: Genotyping errors which are detected if they result in non-Mendelian inheritance patterns.[165]  li N.B. At the start of the QC process, there were 31 individuals with sequencing replicates, but only 30 remained by the time this criterion was applied. SNPs in which the genotype was called in one replicate, but not in the other, were not considered discordant. These SNPs were still included with the called genotype being kept.   lii N.B. At the start of the QC process, there were 269 WGA/ non-WGA pairs, but only 196 remained by the time this criterion was applied. SNPs in which the genotype was called in one 54 Finally, 130 individuals in the Taqman assay were requested to donate additional samples, in order to obtain higher DNA concentrations than their earlier original samples. These additional samples are hereafter referred to as “rebleeds”. If there was discordance in called genotypes between earlier original samples and these rebleed samples, the rebleed genotype was kept.liii    3.4 Statistical analyses  Unless otherwise indicated, analyses were conducted using PASW 18/ IBM SPSS 19/ 20 (IBM, Armonk, NY). Statistical significance was set at p ≤ 0.05.   3.4.1 Organochlorine analyses   3.4.1.1 Descriptive statistics   The distribution of covariates by case/ control status was assessed in European participants with organochlorine measurements that had passed quality control, as this ethnicity was the main analytic group used in genetic analyses. The number of samples above the detection limits of each organochlorine, as well as the detection limits of these organochlorines adjusted for the median lipid concentration, were calculated. The Spearman rank coefficients (rho) for PCBs, summed PCB measures and pesticides/                                                                                                                                                               member of the WGA/ non-WGA pair, but not in the other, were not considered discordant. These SNPs were still included with the called genotype being kept.   liii N.B. With respect to rebleed/ original pairs, discordance between 2 called genotypes did not result in removal of the SNP. The genotype of the rebleed was chosen over the original. In the case of 1 undetermined/ 1 called genotype, the called genotype was used, even if the call was from the original sample. The exception to this was when there were additional discordances between 2 called genotypes in the rebleed/ original pairs. As this may indicate a sample switch, the rebleed sample was chosen over the original sample for all genotypes in that individual. For all other rebleed/ original sample pairs without differences, the rebleed sample was  selected over the original (although this did not change the final genotype). 55 pesticide metabolites were also examined in controls to assess the degree of correlation between analytes in these OC groupings.    3.4.1.2 Organochlorine categorization  Data coding has been described elsewhere.[13, 14] Briefly, OCs were excluded from this and subsequent analyses if fewer than five percent of samples were detectable. This was the case for PCB-52, 101, 128, aldrin, α-chlordane, and γ-chlordane. The remaining OCs were categorized into two, three or four categories, depending on the percentage of detectable values in controls. If 50% or fewer samples were detectable, OCs were categorized into two categories: undetectable and detectable. This was the case for PCB-28, 105, mirex, cis-nonachlor, and p, p’-DDT. If more than 50%, but fewer than 75% of samples, were detectable, then three categories were made: an undetectable category and two categories whereby the detectable samples were divided above and below the median level in the detectable controls. This was the case for PCB-183. If at least 75% of samples were detectable, OCs were categorized into quartiles based on OC levels in controls. All undetectable samples were in the lowest quartile. This was the case for PCB-99, 118, 138, 153, 156, 170, 180, 187, β-HCCH, HCB, trans-nonachlor, oxychlordane, and p, p’-DDE. Summed lipid-adjusted PCB measures were also categorized into quartiles based on the distribution in controls.  3.4.1.3 Descriptive organochlorine analysis  Organochlorine concentrations by case/ control status, age, sex, region, education level, BMI one year prior to study participation, and having a family history of NHL were assessed in participants with OC measurements that had passed QC. liv Briefly, differences                                                   liv The previous analyses of all ethnicities had also reported bivariate statistics.[14] In addition to these covariates, ethnicity and self-reported having ever lived/ worked on a farm were also presented. These were not presented here because ethnicity is non-informative in these 56 in lipid-adjusted organochlorine concentrations, with undetectable levels set as the detection limit divided by √2, between these covariates were assessed using the Wilcoxon rank-sum test. For simplicity, the covariates were dichotomized (i.e. age: <60 years vs. ≥60 years; education: less than post-secondary graduate vs. post-secondary graduate; BMI: <25 kg/m2 vs. ≥25 kg/m2) and unknown values were excluded. Education was used as a surrogate for socioeconomic status in this and subsequent analyses, rather than income or per capita income, which had high levels of missing data. Other relevant metrics were also computed (e.g. minimum, maximum, mean, standard deviation, median, 25 th and 75th percentiles).  3.4.1.4 Occupational analyses  3.4.1.4.1 Occupational coding  Study participants were asked by questionnaire what is, or was if they were retired, the usual occupation and usual industry in which they work(ed). A number of separate coders standardized these self-reported usual occupations, using different classification methods. These codes were then consolidated. Briefly, occupational codes were assigned based on the 1980 version of Statistics Canada’s Standard Occupational Classification (SOC), by a student trained by experienced occupational coders at BCCRC.[167] Assignment of occupational codes was based primarily on participants’ self -reported usual occupation, with additional specificity being gleaned from self-reported usual industry. A subset of the assigned codes were checked and verified. Occupational codes were also assigned to the self-reported usual occupations by ‘t Mannetje et al., using the International Labour Office’s 1968 revision of the International Standard Classification of Occupation (ISCO).[168, 169]                                                                                                                                                               Europeans-only analyses. Having ever lived/ worked on a farm was combined with other occupations hypothesized to have possible OC exposure in Section 3.4.1.4.2.  57 Unblinded to both the SOC and ISCO codes that had been already assigned, a second set of SOC codes for usual occupation was assigned to each participant by the author. To increase the specificity of farming-related occupational codes, additional details on farming history provided by participants who had self-reported having ever lived/ worked on a farm in other sections of the questionnaire were utilized. If participants reported multiple occupations as the usual occupation, only the first occupation was coded. The exception to this was if the first occupation was unclear/ non-specific (i.e. classified to SOC Major Group 00 – Persons not classifiable by occupation or Major Group 99 – Occupations not elsewhere classified).  Finally, the occupational codes assigned by the author, another student and ‘t Mannetje et al. were reviewed and a final set of SOC codes was assigned. This final version of SOC codes attempted to consolidate differences between coders.  When assigning the final occupational codes, a code that reflected the common usage of a job title was selected. When it was difficult to assign an occupation to a more specific level, the occupation was classified in the closest residual group (i.e. the “not elsewhere classified” (n.e.c.) group). These residual categories are heterogeneous. Final coding decisions were made by the author, not by consensus.   3.4.1.4.2 Occupations with hypothesized organochlorine exposure  In order to assess whether occupation is associated with OC exposure, hypotheses were made as to which occupations may have higher exposure to OCs. These hypotheses were based on reviewing IARC monographs specific to each OC analyzed. This review focused on the uses of the OC that are likely to occur while working, as well as direct or indirect mention of exposed occupations. These areas were often highlighted in sections dedicated to OC use and human epidemiological studies conducted in specific groups. Additional sources were consulted for clarification of OC uses.[8, 10, 85, 170-173] The SOC codes corresponding to these possibly exposed occupational groups were then determined. Refer to Table B.2 for a complete list of possibly exposed occupational codes. For 58 example, one of the uses of mirex is as an insecticide, particularly against fire ants for pastures and croplands.[7] Possibly exposed occupations would be mirex-manufacturing occupations, farming-related occupations and pesticide-application. Therefore, the following SOC codes were flagged as corresponding to possibly exposed occupations: 816/817-Chemicals, Petroleum, Rubber, Plastic and Related Materials Processing Occupations, 71-Farming, Horticultural and Animal Husbandry Occupations and 6199-Other Service Occupations, n.e.c. (includes Extermination/ Pesticide Application).  As noted above, some SOC codes can be broad in scope and the same SOC code can refer to more than one possible exposure occupation. Therefore, the full-text self-reported usual occupations corresponding to these SOC codes of interest were consulted before assigning a participant to a possible OC exposure occupational group. In situations where possible exposure occupations were specific to a unit group, rather than to an entire major or minor group, the full-text usual occupations for participants in the supervisor/ foreman/ forewoman, inspector/ tester/ grader/ sampler, labourer, and residual categories of the same minor group were also inspected for potentially exposed participants. These generally corresponded to unit codes ending in 0, 6, 8, and 9, respectively.  Because information on whether participants had ever lived/ worked on a farm was also available from the questionnaire and this could be a source of OC exposure, additional farming-related exposure groups were also created. Individuals who self-reported their first usual occupation as farming (i.e. the 1st reported usual occupation was classified in SOC Major group 71–Farming, Horticultural and Animal Husbandry Occupations) were placed in one group. A group was also created for individuals who self-reported a farming-related usual occupation, regardless of whether it was reported 1st or coded to SOC Major group 71. This group included those reporting a usual occupation directly involving farming (i.e. Major group 71), even if this was not their 1st reported occupation, as well as those reporting usual occupations that were farming-related, but did not code to Major group 71. This means that the “Farming-related usual occupation” group included everyone found in the “Farming 1st usual occupation” group, but not vice-versa. Another group was created for individuals who had self-reported ever living/ working on a farm elsewhere in the questionnaire. These three farming-related groups were 59 also collapsed into an inclusive category that included individuals who had either self-reported ever living/ working on farm or who had reported a farming-related usual occupation, regardless of whether it was the first usual occupation reported or whether it coded to SOC Major group 71.  Participants not identified as being possibly exposed to OCs (i.e. no farming-related usual occupation, not reporting ever living/ working on a farm and no other possible exposure 1st usual occupation) were combined into a single unlikely occupational OC exposure reference group. It is important to note that participants were placed in an Unknown category if no information suggestive of OC exposure was found, but exposure could not be ruled out. In other words, participants were excluded if (1) No information on whether they had ever lived/ worked on a farm was available (i.e. missing questionnaire information) and their usual occupation was not considered possible exposure or (2) None of their usual occupations were clear (i.e. SOC codes: 0011-Workers Reporting Unidentifiable or Inadequately Described Occupations (Not Codeable), 0021-Workers Not Reporting Any Occupation or 99-Occupations Not Elsewhere Classified (if tasks unclear)) and either no information on whether they had ever lived/ worked on a farm was available (i.e. missing questionnaire information) or they reported not having ever lived/ worked on a farm (i.e. they responded “No”). Because of small numbers, all exposed groups were combined into one possible exposure category. The distribution of the possible occupational OC exposure group compared to the unlikely exposure group between cases and controls with OC measurements was assessed. Differences in lipid-adjusted organochlorine concentrations, with undetectable values set as the detection limit divided by √2, between this possible occupational OC exposure group and the unlikely exposure group, were assessed using the Wilcoxon rank-sum test. Unknown exposure values were excluded. In order to further explore the relationship between occupation and plasma organochlorine levels, multivariate linear regression models were also run. Natural log-transformed lipid-adjusted organochlorine concentrations, with undetectable values set as the detection limit divided by √2, were set as the dependent variables. The occupational OC exposure group was set as the independent variable. Models were adjusted for age, sex, region, 60 BMI, education and case/ control status.   3.4.1.5 Organochlorine main effects  3.4.1.5.1 Organochlorine models in NHL   The effects of organochlorine levels on the risk of NHL were estimated. Odds ratios and 95% CIs were obtained from unconditional logistic regression models. NHL case/  control status was set as the binary outcome. Each lipid-adjusted OC and summed PCB measure was set as a categorical independent variable in separate models, due to high correlation between OCs. The reference group was set as the undetectable OC category or the lowest OC quartile. Models were adjusted for confounders based on the change-in-estimate criterion (CIE).[7, 174] Refer to Section 3.4.1.5.2.  The significance of the Wald statistic of the categorical OC variable was assessed for the two-category OCs. The significance of OCs with three or four categories was examined using tests for trend (p-trend). These were estimated by running unconditional logistic regression models in which the categorized lipid-adjusted OC and summed PCB variables were transformed into continuous variables, by assigning the median value of the controls to each category. The significance of the OCs was based on the Wald statistic for this continuous variable.   3.4.1.5.2 Confounder selection   Potential confounders in the organochlorine logistic regression models were the categorical study design variables: age (20–49 years, 50–59 years, 60–69 years, ≥70 years), sex (male, female) and region (GVRD, CRD), as well as categorized education level (<high school, high school graduate (includes those with some university education), post -secondary graduate (includes trade, vocational, community college, or university graduates)), BMI one year prior to study participation (≤24.9 kg/m2, 25-27.4 kg/m2, 27.5-61 29.9 kg/m2, ≥30 kg/m2lv), family history of NHL (yes, no), and occupations with possible OC exposurelvi (Possible OC Exposure, Unlikely OC Exposure). Occupation was included as a potential confounder to account for non-OC occupational exposures. As five cases and four controls were missing education level, six cases and five controls were missing BMI, 14 cases and one control were missing family history of NHL, and 22 cases and nine controls were missing occupational information, these individuals were excluded from confounder selection and the subsequent analyses that included these variables.  Confounders were selected from these potential confounders for inclusion in the final logistic regression models for each OC as follows. First, all seven potential confounders were included in the logistic regression model containing the categorized OC in question (i.e. the full model). ORs were estimated for each OC category relative to the reference category. All seven potential confounders were then removed from the model and the ORs of the resulting completely unadjusted model were compared to the ORs in the full model. If none of the ORs of the OC categories changed by more than 5% ([ORunadjusted model-ORfull model]/ORfull model), then no confounders were included in the final model. If any of the ORs changed by more than 5%, the change-in-estimate criterion was used to select confounders.  In CIE, all seven potential confounders were added to build the full model and then each potential confounder was removed one at a time from the full model. The ORs of the OC categories in the reduced models were compared to the ORs in the full model containing all seven potential confounders. For OCs with more than two categories, the absolute values of the percentage changes of each indicator variable were summed to find the overall change. The confounder that resulted in the smallest overall change in O R, where none of the indicators changed by more than 5%, was removed from the model. The remaining potential confounders were once again removed one at a time from the resulting reduced model. The process was repeated until none of the potential                                                   lv Unlikely BMI values (i.e. <16 kg/m2 or >56 kg/m2) set to unknown  lvi Occupations with possible OC exposure includes individuals who self-reported having ever lived/ worked on a farm 62 confounders could be removed without at least one of indicator variables changing by more than 5%. These remaining confounders were included in the final model for the OC in question.   3.4.1.5.3 Multiple comparisons   The false discovery rate, the expected proportion of falsely rejected hypotheses, was controlled to account for testing multiple individual OCs and summed PCB measures. The Benjamini-Hochberg method was applied to calculate a new false discovery rate threshold for the p-trend of each OC measure.[175] The p-value of the Wald statistic was used for OCs with only 2 categories.   3.4.1.5.4 Interactions with covariates   Multiplicative statistical interactions between the OCs and the potential confounders, age, sex, region, education level, BMI, family history of NHL, and occupations with possible OC exposure, were assessed by introducing product terms to the final logistic regression models. All covariates were dichotomized to allow this (i.e. age: <60 years vs. ≥60 years; education: less than post -secondary graduate vs. post-secondary graduate; BMI: <25 kg/m2 vs. ≥25 kg/m2). The significance of adding the interaction product term to the final models was based on the likelihood ratio test. For significant interactions, logistic regression models were fit within the categories of the covariate.  3.4.1.5.5 NHL subtypes    Sub-analyses were conducted by NHL subtype: FL, DLBCL, other B-cell, and all T-cell. The category “other B-cell subtypes” includes all B-cell NHL excluding FL and DLBCL. 63 Tests for heterogeneity between NHL subtypes (FL vs. DLBCL vs. other B-cell vs. all T-cell) were also performed using polytomous regression. The same confounders selected for the final OC models with NHL as the binary outcome were included in these subtype models. Participants with multiple primary NHL subtype diagnoses were consolidated, according to an adaptation of the InterLymph Consortium harmonization, whereby the less aggressive subtype was used. For example, FL/ DLBCL diagnoses were treated as FL, as this was selected as the etiologically relevant subtype. Second non-NHL hematological malignancies were ignored (i.e. participants were only included in the relevant NHL subtype). When multiple diagnoses belonged to different phenotype groups (e.g. B-cell and T-cell), these were excluded from subtype analyses.   3.4.2 Genetic analyses  3.4.2.1 Descriptive statistics   The distribution of covariates by case/ control status in the subset of Europeans with CYP genotyping measurements that passed quality control for the Illumina or Taqman assays was assessed.   3.4.2.2 Genetic main effects  3.4.2.2.1 Genetic models in NHL   Genetic main effect analyses in NHL were conducted using PLINK (v.1.07).[159, 176] The effects of CYP genetic variants on the risk of NHL were estimated using odds ratios for individual SNPs. The phenotypic expression of each SNP was fitted using a codominant genetic model.[177] Refer to Table 3.1. The codominant model is the most 64 flexible model in that the major allele lvii homozygotes, minor allele lviii homozygotes and heterozygoteslix are each permitted to have different effects on NHL risk, through the coding of each SNP as two indicator variables. The major allele homozygous genotype was set as the reference group. If fewer than ten individuals (cases and controls combined) were present in the minor allele homozygous group, a dominant genetic model was analyzed instead (see below).  The significance of each SNP was examined using a test for trend (p-trend). This was calculated by constructing an additive genetic model in which the log-odds of NHL changes by a factor equal to the number of copies of the minor allele. Under this model, each SNP was coded as one continuous variable with three levels representing the number of copies of the minor allele (i.e. 0, 1 or 2).  The significance of the SNPs was based on the Wald statistic for this continuous variable. ORs and 95% CIs were calculated for these models, using binary unconditional logistic regression with NHL case/ control status as the outcome. Models were adjusted for the study design matching variables (i.e. age, sex, region).  For SNPs that were selected for gene-environment interaction analyses, recessive and dominant genetic models were also fitted. These additional models, as well as re-evaluation of the codominant and additive models, were run using PASW 18/ IBM SPSS 19/ 20. In the dominant model, the SNP was coded as one indicator variable. The minor allele was set to confer a dominant phenotype (i.e. either one or two copies of the minor allele result in the phenotypic expression of the minor allele) and was compared to the major allele homozygous genotype, which was once again set as the reference group. In the recessive model, the minor allele was set to confer a recessive phenotype (i.e. two copies of the minor allele are required for phenotypic expression of the allele). The major allele homozygous and heterozygous genotypes were combined and set as the reference                                                   lvii Major allele: Most abundant allele in the entire sample  lviii Minor allele: Least abundant allele in the entire sample  lix Heterozygotes: Two different alleles at the same DNA position.[46] 65 group. These models were adjusted for the same covariates included in the previous codominant models. Heterogeneity between the codominant model and the recessive, dominant and additive models was determined using the likelihood ratio test (LRT). The model with the least significant likelihood ratio test p-value, among the models that were not statistically different from the codominant model, was considered the most comparable to the codominant model.    66 Table 3.1  Genetic models Genetic model Genotype Code x1 x2 Codominant G G (Ref) 0 0 A G 1 0 A A 0 1 Additive G G (Ref) 0    A G 1    A A 2   Dominant G G (Ref) A G or A A 0 1     Recessive A G or G G (Ref) A A 0 1     Abbreviations  Ref: Reference group;  N.B. Example using rs16974799 (CYP2B6), where A is the minor allele;    67 3.4.2.2.2 Multiple comparisons  The issue of multiple comparisons was addressed using methods similar to those employed by Morton et al.[178] Essentially, permutation-based re-sampling (10 000 permutations) was used to account for testing multiple SNPs within each CYP gene and the false discovery rate was controlled to account for testing multiple CYP genes. First, the statistical significance of the smallest p-trendlx, min P, among all the SNPs within each gene was determined by calculating p-perm. P-perm was taken to represent the entire gene and takes into account the number of SNPs genotyped for each gene and underlying linkage disequilibrium.[159, 176, 179, 180] A new false discovery rate threshold for the p-perm of each gene was then calculated. (The observed p-value rather than p-perm was used for genes for which only one SNP was analyzed.)   3.4.2.2.3 Interactions with covariates  Multiplicative statistical interactions between the SNPs and each of the study design matching variables (i.e. age, sex, region) were assessed by introducing product interaction terms into the final logistic regression models under an additive genetic model. In the interaction term, age was dichotomized into <60 years and ≥60 years. Significance was based on the Wald statistic of the product interaction term. For significant interactions, logistic regression models were fit within the categories of the covariate.   3.4.2.2.4 NHL subtypes  Sub-analyses were conducted by subtype, as described in Section 3.4.1.5.5.                                                    lx N.B. p-trend or the p-value from the additive genetic model was used for all SNPs, regardless of whether the dominant or codominant model had been used to estimate ORs. 68  3.4.3 Gene-environment interaction analyses  3.4.3.1 Gene-environment models  In order to minimize the number of potential gene-environment interactions tested, an a priori decision was made to only assess GxE interactions between SNPs and organochlorines/ summed PCB measures that had significant main effect associations with NHL, before adjusting for multiple comparisons. These potential tests were further reduced by focusing on SNPs and OCs that were not highly correlated with each other. For OCs, this meant that if two significant PCBs/ summed PCB measures had a Spearman rank ρ greater than 0.800, then only the most significant PCB/ summed PCB measure would be included in GxE tests. Similarly, the most significant pesticide/ pesticide metabolite was included, if two significant pesticides/ pesticide metabolites were found to be highly correlated. The same reasoning was applied to SNPs, using an r2 greater than 0.80 as the threshold. (A posteriori, one SNP of borderline non-significance was also included in GxE tests because it was in the same gene cluster as one of the significant SNPs that were found. Another borderline non-significant SNP was also in the same cluster as one of the significant SNPs, but was not included in GxE tests as it was in high linkage disequilibrium with the significant SNP).  The interactions between the remaining SNPs and organochlorines were assessed by adding a product term to a logistic regression model containing both these variables. The continuous OC variable was used to represent the OC. The binary genetic model most comparable to the codominant model was used to represent the SNP for ease of interpretation. Confounders from the main effects models, which assessed the effect of the SNPs alone and the organochlorines alone on NHL risk, were also included. The significance of adding the interaction term to these models was based on the likelihood ratio test. For significant gene-environment interactions, logistic regression models of the associations between the OC of interest and NHL were re-run within genotypes of the SNP of interest. The same binary genetic model was used. 69  3.4.3.2 Multiple comparisons  The false discovery rate was controlled to account for multiple GxE comparisons, using the p-value distribution of the interaction likelihood ratio test.   3.5 Power calculations  Power calculations were carried out using QUANTO v.1.2.[181] Minimum detectable odds ratios for detecting main OC effects, main SNP effects and GxE interactions are shown in Table 3.2 to Table 3.4.     70 Table 3.2  Minimum detectable odds ratios for main effect of OCs and risk of NHL Population prevalence/ proportion of controls above vs. below the median organochlorine exposure Minimum detectable OR Power ~ 80% Power ~ 90% 0.50 1.59 1.71 Abbreviations  OC: Organochlorine; OR: Odds Ratio; Based on sample size of 300 cases and 300 controls; Unmatched case-control (1:1); α=0.05, 2-sided; Kplxi=0.0001; Binary environment model;       Table 3.3  Minimum detectable odds ratios for main effect of SNPs and risk of NHL  Allele frequency in controls Minimum detectable OR Power ~ 80% Power ~ 90% 0.05 1.61 1.72 0.10 1.43 1.50 0.15 1.36 1.42 0.20 1.32 1.37 0.25 1.30 1.35 0.30 1.28 1.33 0.35 1.27 1.32 0.40 1.26 1.31 0.45 1.26 1.31 0.50 1.26 1.31 Abbreviations  OR: Odds Ratio; Based on sample size of 600 cases and 600 controls; Unmatched case-control (1:1); α=0.05, 2-sided; Kp=0.0001; Log-additive modellxii;                                                       lxi Kp: Overall NHL risk in the general population  lxii Log-additive model: Equivalent to additive genetic model 71 Table 3.4  Minimum detectable odds ratios for gene-environment interactions             Abbreviations  OR: Odds Ratio; Based on sample size of 300 cases and 300 controls; Unmatched case-control (1:1); α=0.05, 2-sided; Kp=0.0001; Pe lxiv=0.5; Rğlxv=1.3; Rělxvi=1.3; Log-additive Model; Binary environmental factor; Power=80%;                                                      lxiii Interaction effect: the odds of NHL for organochlorine exposure above the median level in the non-reference genotype compared to organochlorine exposure below the median in the reference genotype, divided by the product of the OR for organochlorine exposure in those with the  reference genotype and the OR for genotype in those with organochlorine exposure below the median. For AHR IVS + 4640 and oxychlordane, assuming highest quartile is like the above the median group: [2.64]/[3.16 x 2.06]=0.41 (reciprocal = 2.47)  lxiv Pe: Population prevalence/ proportion of controls above vs. below the median organochlorine exposure in controls is 0.5 by definition  lxv Rğ: Marginal genetic OR/ Population-average OR for NHL for SNP.  lxvi Rĕ: Marginal environmental OR/ Population-average OR for NHL for the above vs. below median organochlorine exposure categories Allele frequency in controls Minimum detectable ORlxiii 0.05 4.95 0.10 2.99 0.15 2.48 0.20 2.28 0.25 2.24 0.30 2.11 0.35 1.98 0.40 1.95 0.45 1.93 0.50 1.93 72  3.6 Ethical considerations  Ethics approval for the study was obtained from the University of British Columbia - British Columbia Cancer Agency (UBC BCCA) Research Ethics Board. All study participants provided written informed consent prior to participation.[13, 14]   3.7 Chapter synopsis  Between 2000 and 2004, 828 cases and 848 controls were recruited for a study of NHL risk factors in British Columbia, Canada. All newly diagnosed cases of NHL during the study period were ascertained from the BC Cancer Registry. Cases were excluded if they were HIV-positive or had prior transplantation. Random population controls, who were frequency-matched by age, sex and region in a ~1:1 ratio to cases, were ascertained from the BC Ministry of Health Client Registry.  Organochlorines were measured in cases and controls with sufficient plasma for lipid adjustment. Only measurements from cases who had provided blood prior to chemotherapy and had not experienced >10% weight loss within one year of blood collection were included. The following 14 PCB congeners were quantified in the blood samples: 28, 52, 99, 101, 105, 118, 128, 138, 153, 156, 170, 180, 183, and 187. The following 11 pesticides/ pesticide metabolites were also quantified: aldrin, β-HCCH, HCB, mirex, α-chlordane, γ-chlordane, trans-nonachlor, cis-nonachlor, oxychlordane, p, p’-DDT, and p, p’-DDE.  One hundred twenty-nine genetic variations known as single nucleotide polymorphisms in 18 CYP genes were selected for genotyping, using the Taqman and Illumina assays, in participants with sufficient blood, mouthwash or saliva.  Prior to statistical analysis, quality control procedures removed low quality SNPs and samples, including potential sample swaps and related samples. Analyses were restricted to participants of European ethnic origin. The relative risks of NHL conferred by lipid -adjusted organochlorines and CYP SNPs (Odds ratios and 95% CIs) were estimated from 73  unconditional logistic regression models. Subanalyses were conducted by subtype: FL, DLBCL, other B-cell subtypes, and all T-cell subtypes. Tests for heterogeneity between NHL subtypes were performed using polytomous regression. The interactions between these organochlorines and SNPs were assessed by adding a product term to a logistic regression model containing both these variables.    74  Chapter 4 Results   4.1 Organochlorine analyses  4.1.1 Organochlorine quality control   Refer to Figure 3.1 for a flowchart of the quality control process. In light of the sample switches identified during the Illumina quality control process, 12 participants (7 cases; 5 controls) were removed. Similarly, one case was removed as the participant was part of a sibling-pair and one control who was later diagnosed with NHL and also participated as a case was removed. One control’s ethnicity was changed from Asian to “other”. This left 866 participants (412 cases; 454 controls) of all ethnicities for organochlorine analyse s. This subset was further restricted to 694 Europeans (331 cases; 363 controls) for this analysis.  4.1.2 Descriptive statistics  The frequencies of various demographic variables and other covariates by case/  control status for the 694 Europeans with organochlorine measurements that passed quality control are presented in Table 4.1. As noted in previous analyses of all ethnicities, DLBCL cases are under-represented in the organochlorine analyses compared to the parent study and the genetic analyses.[13, 14] This is because cases were excluded from organochlorine analyses if their blood samples were taken after the start of chemotherapy. As DLBCL is an aggressive NHL subtype, chemotherapy is often started quickly after diagnosis, thereby making it more difficult to obtain pre-chemotherapy samples from DLBCL cases.     75  Table 4.1  Characteristics of European cases and controls with organochlorine measurements included in analyses [frequency (percentage)]     Controls (n=363) Cases (n=331) Histology path4gp consolidated for multiple diagnoses   Multiple phenotypes (B-cell and T-cell)   1 (0.30) DLBCL    49 (14.8) FL    110 (33.2) Other B-cell    138 (41.7)  All T-cell   33 (9.97) Age at selection (years)  Agegp   20 - 49  61 (16.8) 50 (15.1) 50 - 59  87 (24.0) 79 (23.9) 60 - 69  88 (24.2) 85 (25.7) ≥70  127 (35.0) 117 (35.3) Sex SEX   Male  199 (54.8) 178 (53.8) Female  164 (45.2) 153 (46.2) Region Reg   GVRD  304 (83.7) 275 (83.1) CRD  59 (16.3) 56 (16.9) Family history of NHL fam hist – nhl   No  350 (96.4) 303 (91.5) Yes  12 (3.31) 14 (4.23) Unknown*  1 (0.28) 14 (4.23) Highest level of education  educgp3New   <High school  48 (13.2) 62 (18.7) High school graduate  118 (32.5) 116 (35.0) Post-secondary graduate  193 (53.2) 148 (44.7) Unknown**  4 (1.10) 5 (1.51) Income (dollars)  IncngpNew   <25 000  56 (15.4) 63 (19.0) 25 000 - 49 999  107 (29.5) 80 (24.2) 50 000 - 74 999  72 (19.8) 68 (20.5) 75 000 - 99 999  42 (11.6) 42 (12.7) ≥100 000  41 (11.3) 38 (11.5) Unknown**  45 (12.4) 40 (12.1) Per capita income   PcigpNew   <15 000  67 (18.5) 69 (20.8) 15 000 - 29 999  102 (28.1) 91 (27.5) 30 000 - 44 999  108 (29.8) 92 (27.8) ≥45 000  40 (11.0) 37 (11.2) Unknown**  46 (12.7) 42 (12.7)   76      Controls (n=363) Cases (n=331) BMI (kg/m2) BmigpNew   <25  150 (41.3) 138 (41.7) 25-27.4  102 (28.1) 90 (27.2) 27.5-29.9  56 (15.4) 39 (11.8) ≥30  50 (13.8) 58 (17.5) Unknown***  5 (1.38) 6 (1.81) Abbreviations  BMI: Body mass index; CRD: Capital regional district; DLBCL: Diffuse large B-cell lymphoma; FL: Follicular lymphoma; GVRD: Greater Vancouver regional district;  Histology: 1 case excluded from subtype analyses because individual had 2 NHL diagnoses with different phenotypes (still included in overall NHL analyses);  Highest level of education: "High-school graduate" includes those with some university education. "Post-secondary graduate" includes trade, vocational, community college, or university graduates. Note that this categorization differs from that published in [13];  Income: Average annual household income before taxes in the year prior to interview;  Per capita income: Midpoint of average annual household income ranges divided by the number of family members supported by this income (if 0 members reported, set to unknown);  BMI: BMI one year prior to study participation (values <16 kg/m2 or >56 kg/m2 set to "unknown"); * "Unknown" includes missing data; ** "Unknown" includes refusal to respond/ don't know responses/ missing data; *** "Unknown" includes don't know responses/ missing data; N.B. Columns do not add up to 100% due to rounding;      77  4.1.3 Descriptive organochlorine analysis   The number of samples above the detection limits of each organochlorine, as well as the detection limits of these organochlorines adjusted for the median lipid concentration, are shown in Table 4.2. The Spearman rank rho coefficients for PCBs/ summed PCB measures and for pesticides/ pesticide metabolites are shown in Table 4.3 and Table 4.4.  The distribution of organochlorines by case/ control status and various covariates are shown in Table 4.5 to Table 4.11. In general, OC levels were higher in cases than controls and in those ≥60 years compared to those <60 years. Those in the ≥25 kg/m 2 BMI group generally had significantly higher median OC levels than those in the <25 kg/m 2 group (e.g. PCB-99, 118, β-HCCH, HCB, trans-nonachlor, oxychlordane, and p, p’-DDE). The opposite was true for mirex, which also had a moderate number of samples below the detection limit.  Levels varied between males and females as well. Statistically higher levels were found in females compared to males for PCB-105, 118, β-HCCH, HCB, and p, p'-DDE. The opposite was true for mirex, trans-nonachlor and cis-nonachlor. For a number of OCs, significantly higher median OC levels were found in those with less education compared to those with more education (e.g. PCB-99, 170, 180, 183, 187, HCB, mirex, trans-nonachlor, oxychlordane, and p, p’-DDE). Median OC levels did not tend to vary by region (except for a small difference in PCB-180). Median OC levels did not tend to vary by family history of NHL, except for differences in β-HCCH, p, p’-DDT and p, p’-DDE.  78  Table 4.2  Organochlorine measurements above the detection limit in blood plasma Organochlorine DL/Median lipid (µg/kg lipid) Samples above DL  n* % PCB-28 2.20 135 19.5 PCB-52 3.67 8 1.15 PCB-99 2.20 576 83.0 PCB-101 2.20 21 3.03 PCB-105 2.20 219 31.6 PCB-118 2.20 652 93.9 PCB-128 2.20 2 0.29 PCB-138 2.20 688 99.1 PCB-153 2.20 692 99.7 PCB-156 2.20 643 92.7 PCB-170 2.20 679 97.8 PCB-180 2.20 693 99.9 PCB-183 2.20 447 64.4 PCB-187 2.20 675 97.3 Aldrin 2.20 0 0.00 β-HCCH 3.67 609 87.8 HCB 2.20 693 99.9 Mirex 2.20 265 38.2 α-Chlordane  2.20 0 0.00 γ-Chlordane 2.20 0 0.00 trans-Nonachlor 2.20 685 98.7 cis-Nonachlor 2.20 179 25.8 Oxychlordane 2.20 673 97.0 p, p’-DDT 3.67 179 25.8 p, p’-DDE 3.67 694 100 Abbreviations  DL: Detection limit;  * Total of 694 European samples passed organochlorine quality control;    79  Table 4.3  Spearman rho for PCBs/ summed PCB measures in European controls  PCB-28 PCB-99 PCB-105 PCB-118 PCB-138 PCB-153 PCB-156 PCB-170 PCB-180 PCB-183 PCB-187 Total Dioxin Non-dioxin PCB-28 1.00 .233** .530** .266** .140** .082 .053 .006 .003 .171** .083 .124* .239** .100 PCB-99 .233** 1.00 .524** .760** .833** .709** .569** .453** .419** .703** .568** .691** .761** .667** PCB-105 .530** .524** 1.00 .624** .447** .339** .329** .192** .166** .383** .275** .383** .612** .335** PCB-118 .266** .760** .624** 1.00 .752** .673** .611** .467** .434** .574** .553** .683** .937** .632** PCB-138 .140** .833** .447** .752** 1.00 .907** .814** .711** .660** .859** .768** .879** .850** .863** PCB-153 .082 .709** .339** .673** .907** 1.00 .859** .863** .844** .880** .883** .963** .807** .963** PCB-156 .053 .569** .329** .611** .814** .859** 1.00 .869** .818** .688** .795** .892** .824** .881** PCB-170 .006 .453** .192** .467** .711** .863** .869** 1.00 .983** .739** .923** .928** .669** .942** PCB-180 .003 .419** .166** .434** .660** .844** .818** .983** 1.00 .719** .927** .910** .626** .929** PCB-183 .171** .703** .383** .574** .859** .880** .688** .739** .719** 1.00 .824** .858** .681** .863** PCB-187 .083 .568** .275** .553** .768** .883** .795** .923** .927** .824** 1.00 .936** .704** .947** Total .124* .691** .383** .683** .879** .963** .892** .928** .910** .858** .936** 1.00 .837** .996** Dioxin .239** .761** .612** .937** .850** .807** .824** .669** .626** .681** .704** .837** 1.00 .794** Non-dioxin .100 .667** .335** .632** .863** .963** .881** .942** .929** .863** .947** .996** .794** 1.00 Abbreviations  Dioxin: Summed dioxin-like PCBs; Non-dioxin: Summed non-dioxin-like PCBs; Total: Total summed PCBs;  ** Correlation is significant at the 0.01 level (2-tailed); * Correlation is significant at the 0.05 level (2-tailed); 80  Table 4.4  Spearman rho for pesticides/ pesticide metabolites in European controls    β-HCCH HCB Mirex trans-Nonachlor cis-Nonachlor Oxychlordane p, p’-DDT p, p’-DDE β-HCCH 1.00 .622** -.001 .471** .232** .485** .133* .415** HCB .622** 1.00 -.020 .527** .176** .633** .066 .481** Mirex -.001 -.020 1.00 .184** .343** .133* .238** .019 trans-Nonachlor .471** .527** .184** 1.00 .475** .875** .115* .448** cis-Nonachlor .232** .176** .343** .475** 1.00 .317** .476** .171** Oxychlordane .485** .633** .133* .875** .317** 1.00 .038 .448** p, p’-DDT .133* .066 .238** .115* .476** .038 1.00 .257** p, p’-DDE .415** .481** .019 .448** .171** .448** .257** 1.00 ** Correlation is significant at the 0.01 level (2-tailed); * Correlation is significant at the 0.05 level (2-tailed); 81  Table 4.5  European lipid-adjusted organochlorine concentrations (μg/kg) by case/ control group and p-values for the Wilcoxon rank-sum tests Organochlorine Case/ Control group* Mean** SD Min. Max. 25p Median 75p p*** PCB-28 Control 2.19 3.35 0.53 57.5 1.41 1.62 1.97 0.005 PCB-28 Case 2.20 1.60 0.82 17.9 1.45 1.76 2.23  PCB-99 Control 6.27 5.24 0.53 61.3 3.29 4.97 7.81 0.286 PCB-99 Case 6.96 6.35 0.82 53.1 2.95 5.65 8.69  PCB-105 Control 2.55 2.86 0.53 33.5 1.47 1.72 2.54 0.018 PCB-105 Case 2.80 3.46 0.82 37.6 1.52 1.87 2.60  PCB-118 Control 10.8 12.8 1.28 202 4.70 8.05 13.4 0.088 PCB-118 Case 12.3 13.2 0.82 130 5.12 8.84 15.3  PCB-138 Control 24.2 17.7 1.41 156 12.7 20.0 30.6 0.294 PCB-138 Case 26.6 23.4 1.44 289 12.5 21.6 33.3  PCB-153 Control 50.2 52.8 1.61 736 27.2 40.7 60.0 0.010 PCB-153 Case 58.8 56.0 1.44 659 30.1 46.1 69.5  PCB-156 Control 7.31 7.44 1.11 113 4.03 5.97 8.63 0.079 PCB-156 Case 7.85 5.91 0.82 56.7 4.27 6.39 9.63  PCB-170 Control 19.0 51.7 1.41 902 7.90 11.8 18.3 0.034 PCB-170 Case 21.6 33.2 0.82 276 8.50 12.7 20.0  PCB-180 Control 65.4 216 1.61 3790 24.3 37.2 58.3 0.012 PCB-180 Case 73.6 127 2.72 1140 27.2 42.0 70.1  PCB-183 Control 4.18 6.51 0.98 84.9 1.75 2.84 4.68 0.077 PCB-183 Case 4.66 5.99 0.82 54.0 1.91 3.07 5.10  PCB-187 Control 17.7 52.0 1.21 833 6.49 10.1 15.5 0.024 PCB-187 Case 20.7 39.3 0.82 319 7.23 11.2 18.6  Total summed PCBs Control 216 389 23.6 6570 108 159 227 0.008 Total summed PCBs Case 244 267 33.4 2130 118 176 270  Summed dioxin-like PCBs Control 20.6 19.2 3.88 281 10.7 16.5 25.4 0.024 Dioxin-like summed PCBs Case 23.0 18.9 2.47 170 12.3 18.3 28.2  Summed non-dioxin-like PCBs Control 195 381 18.8 6450 95.8 141 204 0.010 Non-dioxin-like summed PCBs Case 221 259 28.1 2070 106 153 243  82  Organochlorine Case/ Control group* Mean** SD Min. Max. 25p Median 75p p*** β-HCCH Control 19.2 30.7 1.87 321 8.89 13.3 19.6 0.004 β-HCCH Case 24.7 53.3 1.37 839 9.60 15.8 23.2  HCB Control 19.2 13.7 1.61 175 11.7 16.4 22.4 0.001 HCB Case 27.8 63.8 5.81 1050 12.6 18.7 28.6  Mirex Control 2.64 2.98 0.92 37.5 1.48 1.74 2.66 <0.001 Mirex Case 3.41 5.22 0.82 60.5 1.58 2.03 3.45  trans-Nonachlor Control 15.8 10.4 1.41 90.2 9.32 14.0 19.7 0.007 trans-Nonachlor Case 18.3 12.7 1.44 106 9.99 15.8 23.5  cis-Nonachlor Control 1.95 0.99 0.79 10.8 1.46 1.67 2.09 <0.001 cis-Nonachlor Case 2.22 1.22 0.82 11.3 1.49 1.84 2.37  Oxychlordane Control 10.8 5.99 1.25 36.6 6.42 9.89 14.0 <0.001 Oxychlordane Case 13.2 8.23 0.82 58.2 7.76 11.7 16.9  p, p’-DDT Control 4.57 6.56 1.42 92.1 2.36 2.74 3.84 0.001 p’, p’-DDT Case 5.13 6.43 1.37 82.8 2.50 3.05 5.46  p, p’-DDE Control 334 332 14.9 2880 124 237 433 0.068 p’, p’-DDE Case 433 558 9.28 5820 129 270 520  Abbreviations  25p: 25th percentile; 75p: 75th percentile; Max.: Maximum lipid-adjusted organochlorine concentration (μg/kg); Min.: Minimum lipid-adjusted organochlorine concentration (μg/kg); SD: Standard deviation (μg/kg);  * Sample sizes for Case/ control group: Control (n=363); Case (n=331);  ** Mean and median values are based on lipid-adjusted organochlorine concentrations where undetectable values are replaced by the detection limit divided by √2;  *** Wilcoxon rank-sum asymptotic 2-sided p-value excludes unknown data;   83  Table 4.6  European lipid-adjusted organochlorine concentrations (μg/kg) by age group and p -values for the Wilcoxon rank-sum tests Organochlorine Age group (years)* Mean** SD Min. Max. 25p Median 75p p*** PCB-28 <60 1.98 1.60 0.53 21.9 1.39 1.66 1.99 0.041 PCB-28 ≥60 2.33 3.17 0.94 57.5 1.46 1.67 2.15  PCB-99 <60 4.85 4.44 0.53 53.1 2.28 3.87 5.95 <0.001 PCB-99 ≥60 7.76 6.29 0.94 61.3 3.98 6.32 9.67  PCB-105 <60 2.22 2.83 0.53 37.6 1.43 1.71 2.04 <0.001 PCB-105 ≥60 2.97 3.33 1.00 33.5 1.53 1.95 3.22  PCB-118 <60 7.64 8.42 0.82 90.4 3.40 5.54 9.06 <0.001 PCB-118 ≥60 14.1 14.8 1.32 202 6.92 10.8 17.9  PCB-138 <60 18.4 14.2 1.61 91.6 8.86 14.6 22.6 <0.001 PCB-138 ≥60 29.9 22.8 1.41 289 17.3 25.4 36.0  PCB-153 <60 37.2 28.0 1.61 201 20.2 30.7 46.1 <0.001 PCB-153 ≥60 65.7 64.1 1.44 736 36.4 52.0 72.5  PCB-156 <60 5.51 3.87 0.82 24.7 2.82 4.39 7.01 <0.001 PCB-156 ≥60 8.94 7.84 1.21 113 5.27 7.09 10.1  PCB-170 <60 12.5 16.4 0.82 200 5.45 8.77 13.8 <0.001 PCB-170 ≥60 25.3 54.4 1.41 902 10.2 14.9 20.9  PCB-180 <60 41.4 63.6 1.61 808 16.6 27.9 45.1 <0.001 PCB-180 ≥60 87.9 223 8.01 3790 33.2 48.5 70.2  PCB-183 <60 2.97 2.36 0.82 19.9 1.70 2.16 3.20 <0.001 PCB-183 ≥60 5.36 7.71 1.22 84.9 2.30 3.65 5.65  PCB-187 <60 10.3 12.9 0.82 165 4.49 7.25 11.4 <0.001 PCB-187 ≥60 25.0 58.2 1.41 833 9.12 12.8 19.6  Total summed PCBs <60 151 131 23.6 1380 80.4 116 179 <0.001 Total summed PCBs ≥60 281 413 31.8 6570 146 201 280  Summed dioxin-like PCBs <60 15.4 12.6 2.47 120 8.46 11.9 18.8 <0.001 Dioxin-like summed PCBs ≥60 26.0 21.3 5.09 281 14.7 20.9 31.9  Summed non-dioxin-like PCBs <60 135 125 18.8 1370 69.8 106 159 <0.001 Non-dioxin-like summed PCBs ≥60 255 404 25.9 6450 129 176 249  84  Organochlorine Age group (years)* Mean** SD Min. Max. 25p Median 75p p*** β-HCCH <60 17.4 29.0 1.37 315 7.44 11.7 18.4 <0.001 β-HCCH ≥60 24.8 50.1 2.18 839 10.5 16.3 23.1  HCB <60 16.1 18.6 1.61 299 10.1 13.5 18.7 <0.001 HCB ≥60 28.1 56.0 3.57 1050 14.8 19.9 29.9  Mirex <60 2.53 2.12 0.82 18.0 1.48 1.79 2.66 0.035 Mirex ≥60 3.32 5.13 0.92 60.5 1.53 1.92 3.28  trans-Nonachlor <60 12.1 10.3 1.61 106 6.69 9.96 14.3 <0.001 trans-Nonachlor ≥60 20.3 11.2 1.41 90.2 13.5 17.4 24.9  cis-Nonachlor <60 1.82 0.81 0.79 8.33 1.41 1.67 1.97 <0.001 cis-Nonachlor ≥60 2.25 1.25 0.92 11.3 1.50 1.81 2.54  Oxychlordane <60 8.32 5.78 0.82 58.2 4.75 7.35 10.4 <0.001 Oxychlordane ≥60 14.4 7.11 1.25 53.9 9.50 13.1 17.7  p, p’-DDT <60 4.50 6.58 1.37 82.8 2.38 2.89 3.79 0.115 p’, p’-DDT ≥60 5.06 6.44 1.49 92.1 2.45 2.92 5.45  p, p’-DDE <60 258 400 23.8 5820 95.0 166 300 <0.001 p’, p’-DDE ≥60 463 473 9.28 3810 162 322 566  Abbreviations  25p: 25th percentile; 75p: 75th percentile; Max.: Maximum lipid-adjusted organochlorine concentration (μg/kg); Min.: Minimum lipid-adjusted organochlorine concentration (μg/kg); SD: Standard deviation (μg/kg);  * Sample sizes for Age group: <60 (n=277); ≥60 (n=417);  ** Mean and median values are based on lipid-adjusted organochlorine concentrations where undetectable values are replaced by the detection limit divided by √2; *** Wilcoxon rank-sum asymptotic 2-sided p-value excludes unknown data;   85  Table 4.7  European lipid-adjusted organochlorine concentrations (μg/kg) by sex and p-values for the Wilcoxon rank-sum tests Organochlorine Sex* Mean** SD Min. Max. 25p Median 75p p*** PCB-28 Male 2.15 3.20 0.53 57.5 1.42 1.66 2.01 0.159 PCB-28 Female 2.25 1.81 0.85 21.9 1.45 1.69 2.19  PCB-99 Male 6.45 6.13 0.53 53.1 2.90 4.82 7.79 0.054 PCB-99 Female 6.78 5.39 0.97 61.3 3.29 5.58 8.74  PCB-105 Male 2.46 2.63 0.53 28.1 1.48 1.75 2.30 0.010 PCB-105 Female 2.92 3.68 0.94 37.6 1.52 1.84 3.08  PCB-118 Male 10.1 11.4 0.82 130 4.20 7.22 11.9 <0.001 PCB-118 Female 13.2 14.5 1.41 202 6.47 10.3 17.0  PCB-138 Male 24.7 22.0 1.48 289 12.3 19.9 31.5 0.110 PCB-138 Female 26.1 18.9 1.41 156 13.2 21.7 33.4  PCB-153 Male 55.6 63.6 1.61 736 27.9 42.2 65.4 0.788 PCB-153 Female 52.7 41.1 1.44 318 28.3 43.8 63.4  PCB-156 Male 7.73 7.83 0.82 113 4.12 6.06 9.11 0.919 PCB-156 Female 7.39 5.20 1.21 45.1 4.10 6.21 9.09  PCB-170 Male 22.0 54.8 0.82 902 8.30 12.6 19.3 0.240 PCB-170 Female 18.1 25.0 1.41 216 7.81 11.9 18.4  PCB-180 Male 78.2 229 1.61 3790 26.9 40.3 62.1 0.165 PCB-180 Female 58.8 85.5 5.00 794 24.1 37.6 60.9  PCB-183 Male 4.47 7.15 0.82 84.9 1.85 2.87 4.71 0.607 PCB-183 Female 4.32 5.03 0.97 47.8 1.81 3.04 5.00  PCB-187 Male 20.4 56.4 0.82 833 6.62 10.5 17.2 0.740 PCB-187 Female 17.6 30.4 1.41 297 7.16 10.6 16.8  Total summed PCBs Male 241 421 23.6 6570 109 163 246 0.530 Total summed PCBs Female 216 194 31.8 1450 116 171 247  Summed dioxin-like PCBs Male 20.3 17.8 2.47 170 10.4 15.4 24.6 <0.001 Dioxin-like summed PCBs Female 23.5 20.4 4.28 281 13.2 19.7 28.3  Summed non-dioxin-like PCBs Male 220 411 18.8 6450 99.4 146 220 0.989 Non-dioxin-like summed PCBs Female 192 186 25.9 1420 99.7 147 215  86  Organochlorine Sex* Mean** SD Min. Max. 25p Median 75p p*** β-HCCH Male 18.9 30.6 1.37 321 8.13 12.4 18.9 <0.001 β-HCCH Female 25.3 54.1 1.98 839 10.8 16.9 23.4  HCB Male 17.8 16.8 1.61 277 10.8 15.0 20.0 <0.001 HCB Female 29.8 64.0 5.79 1050 14.8 20.8 30.5  Mirex Male 3.50 4.99 0.82 60.5 1.61 2.14 3.45 <0.001 Mirex Female 2.42 2.93 0.92 45.1 1.44 1.71 2.29  trans-Nonachlor Male 17.4 10.7 1.61 83.6 9.75 15.6 23.1 0.050 trans-Nonachlor Female 16.5 12.5 1.41 106 9.69 14.2 19.4  cis-Nonachlor Male 2.15 1.11 0.79 11.3 1.49 1.79 2.36 0.005 cis-Nonachlor Female 2.00 1.10 0.92 10.8 1.44 1.68 2.08  Oxychlordane Male 11.5 6.85 0.82 47.4 6.45 10.2 14.9 0.081 Oxychlordane Female 12.4 7.66 1.25 58.2 7.71 11.1 16.0  p, p’-DDT Male 4.65 6.23 1.37 92.1 2.47 2.93 4.19 0.257 p’, p’-DDT Female 5.06 6.81 1.42 82.8 2.38 2.85 4.84  p, p’-DDE Male 319 318 16.7 2480 114 209 397 <0.001 p’, p’-DDE Female 454 571 9.28 5820 136 297 532  Abbreviations  25p: 25th percentile; 75p: 75th percentile; Max.: Maximum lipid-adjusted organochlorine concentration (μg/kg); M in.: Minimum lipid-adjusted organochlorine concentration (μg/kg); SD: Standard deviation (μg/kg);  * Sample sizes for Sex: Male (n=377); Female (n=317); ** Mean and median values are based on lipid-adjusted organochlorine concentrations where undetectable values are replaced by the detection limit divided by √2;  *** Wilcoxon rank-sum asymptotic 2-sided p-value excludes unknown data;   87  Table 4.8  European lipid-adjusted organochlorine concentrations (μg/kg) by region and p-values for the Wilcoxon rank-sum tests Organochlorine Regional group* Mean** SD Min. Max. 25p Median 75p p*** PCB-28 GVRD 2.21 2.85 0.82 57.5 1.42 1.67 2.06 0.363 PCB-28 CRD 2.13 1.35 0.53 9.71 1.47 1.70 2.12  PCB-99 GVRD 6.58 5.40 0.82 53.1 3.11 5.35 8.24 0.509 PCB-99 CRD 6.71 7.53 0.53 61.3 3.13 4.97 7.90  PCB-105 GVRD 2.61 2.91 0.82 37.6 1.48 1.78 2.56 0.303 PCB-105 CRD 2.99 4.19 0.53 33.5 1.56 1.78 2.78  PCB-118 GVRD 11.1 9.79 0.82 86.9 4.97 8.31 13.9 0.958 PCB-118 CRD 13.4 23.2 1.29 202 4.72 8.15 16.0  PCB-138 GVRD 25.3 20.1 1.44 289 12.8 21.5 32.0 0.419 PCB-138 CRD 25.6 23.0 1.41 156 12.5 19.1 31.1  PCB-153 GVRD 54.4 55.2 1.44 736 28.1 43.5 64.9 0.363 PCB-153 CRD 54.1 51.3 1.61 300 28.4 39.6 56.6  PCB-156 GVRD 7.65 6.91 0.82 113 4.16 6.18 9.21 0.232 PCB-156 CRD 7.19 5.94 1.21 45.1 3.96 5.70 8.75  PCB-170 GVRD 20.6 46.1 0.82 902 8.10 12.5 19.2 0.070 PCB-170 CRD 18.4 30.1 1.41 185 7.88 11.4 16.2  PCB-180 GVRD 70.8 189 2.72 3790 26.2 40.4 62.2 0.029 PCB-180 CRD 62.2 116 1.61 748 24.1 34.7 50.3  PCB-183 GVRD 4.36 5.99 0.82 84.9 1.85 3.04 4.82 0.143 PCB-183 CRD 4.61 7.56 1.24 48.0 1.74 2.77 4.17  PCB-187 GVRD 19.3 47.7 0.82 833 6.72 10.9 17.4 0.052 PCB-187 CRD 18.1 38.9 1.41 257 6.99 9.56 13.3  Total summed PCBs GVRD 231 350 33.4 6570 113 172 250 0.111 Total summed PCBs CRD 222 264 23.6 1670 107 150 222  Summed dioxin-like PCBs GVRD 21.4 15.5 2.47 127 11.6 17.6 27.0 0.573 Dioxin-like summed PCBs CRD 23.5 31.4 3.88 281 11.3 16.3 26.4  Summed non-dioxin-like PCBs GVRD 209 341 28.1 6450 100 151 221 0.082 Non-dioxin-like summed PCBs CRD 198 253 18.8 1630 94.5 128 189  88  Organochlorine Regional group* Mean** SD Min. Max. 25p Median 75p p*** β-HCCH GVRD 21.9 46.2 1.37 839 9.09 14.4 20.9 0.134 β-HCCH CRD 21.3 21.1 2.15 147 9.63 15.6 23.5  HCB GVRD 23.1 47.5 3.57 1050 12.2 17.2 25.5 0.860 HCB CRD 24.4 32.8 1.61 299 12.0 17.3 22.9  Mirex GVRD 3.03 3.87 0.82 45.1 1.52 1.87 3.06 0.251 Mirex CRD 2.90 5.64 1.09 60.5 1.47 1.75 2.80  trans-Nonachlor GVRD 17.1 11.6 1.44 106 9.66 14.6 21.3 0.867 trans-Nonachlor CRD 16.8 11.6 1.41 83.6 10.4 14.5 19.3  cis-Nonachlor GVRD 2.09 1.14 0.82 11.3 1.47 1.75 2.23 0.892 cis-Nonachlor CRD 2.00 0.93 0.79 7.49 1.47 1.73 2.18  Oxychlordane GVRD 12.1 7.29 0.82 58.2 7.21 10.8 15.6 0.240 Oxychlordane CRD 11.2 6.97 1.61 51.9 6.53 9.99 14.0  p, p’-DDT GVRD 4.72 5.98 1.37 92.1 2.42 2.91 4.30 0.607 p’, p’-DDT CRD 5.44 8.65 1.81 82.8 2.44 2.88 4.16  p, p’-DDE GVRD 379 405 9.28 3810 124 255 480 0.360 p’, p’-DDE CRD 391 657 14.9 5820 131 213 417  Abbreviations  25p: 25th percentile; 75p: 75th percentile; CRD: Capital regional district (greater Victoria); GVRD: Greater Vancouver regional district; Max.: Maximum lipid-adjusted organochlorine concentration (μg/kg); Min.: Minimum lipid -adjusted organochlorine concentration (μg/kg); SD: Standard deviation (μg/kg);  * Sample sizes for Regional group: GVRD (n=579); CRD (n=115);  ** Mean and median values are based on lipid-adjusted organochlorine concentrations where undetectable values are replaced by the detection limit divided by √2;  *** Wilcoxon rank-sum asymptotic 2-sided p-value excludes unknown data;   89  Table 4.9  European lipid-adjusted organochlorine concentrations (μg/kg) by body mass index group and p-values for the Wilcoxon rank-sum tests Organochlorine BMI group (kg/m2)* Mean** SD Min. Max. 25p Median 75p p*** PCB-28 <25 2.23 3.63 0.85 57.5 1.47 1.70 2.04 0.681 PCB-28 ≥25 2.19 1.68 0.53 17.9 1.41 1.66 2.17  PCB-99 <25 6.24 5.96 1.19 61.3 2.65 4.72 7.81 0.007 PCB-99 ≥25 6.92 5.71 0.53 53.1 3.47 5.72 8.69  PCB-105 <25 2.56 3.05 0.98 33.5 1.55 1.80 2.43 0.696 PCB-105 ≥25 2.77 3.28 0.53 37.6 1.47 1.77 2.96  PCB-118 <25 11.4 16.6 1.48 202 4.57 7.93 13.3 0.014 PCB-118 ≥25 11.8 9.83 0.82 86.9 5.37 8.93 15.6  PCB-138 <25 24.2 18.3 1.48 156 11.0 19.7 31.8 0.052 PCB-138 ≥25 26.4 22.3 1.41 289 13.7 22.0 32.0  PCB-153 <25 51.9 38.0 4.86 318 26.9 43.9 66.3 0.745 PCB-153 ≥25 56.5 64.3 1.44 736 29.9 42.4 63.2  PCB-156 <25 7.80 5.32 1.31 45.1 4.39 6.57 9.86 0.052 PCB-156 ≥25 7.46 7.69 0.82 113 4.05 5.97 8.55  PCB-170 <25 18.5 23.6 1.74 216 8.07 13.0 20.0 0.160 PCB-170 ≥25 21.5 54.4 0.82 902 8.08 11.8 18.4  PCB-180 <25 60.4 80.0 4.12 794 26.1 43.2 64.4 0.110 PCB-180 ≥25 76.1 227 2.72 3790 25.4 37.5 59.0  PCB-183 <25 3.88 3.64 0.98 34.5 1.78 2.68 4.67 0.143 PCB-183 ≥25 4.82 7.68 0.82 84.9 1.90 3.09 4.80  PCB-187 <25 16.6 26.7 1.72 297 6.43 10.9 17.6 0.826 PCB-187 ≥25 21.0 57.0 0.82 833 7.19 10.3 16.6  Total summed PCBs <25 212 179 33.9 1450 108 169 255 0.975 Total summed PCBs ≥25 243 418 31.8 6570 116 166 241  Summed dioxin-like PCBs <25 21.7 22.7 5.02 281 10.8 17.1 26.0 0.273 Dioxin-like summed PCBs ≥25 22.0 16.1 2.47 127 11.9 17.9 27.6  Summed non-dioxin-like PCBs <25 190 169 28.7 1420 94.5 151 228 0.849 Non-dioxin-like summed PCBs ≥25 221 409 25.9 6450 106 145 214  90  Organochlorine BMI group (kg/m2)* Mean** SD Min. Max. 25p Median 75p p*** β-HCCH <25 17.3 19.4 1.87 156 8.08 13.0 19.1 <0.001 β-HCCH ≥25 24.9 53.8 1.37 839 10.0 15.4 22.4  HCB <25 18.8 10.6 3.57 72.4 11.5 16.0 23.6 0.002 HCB ≥25 26.8 59.2 5.03 1050 13.0 18.2 26.2  Mirex <25 3.15 4.24 0.98 60.5 1.59 2.01 3.42 0.003 Mirex ≥25 2.93 4.25 0.82 45.1 1.48 1.78 2.80  trans-Nonachlor <25 16.0 11.4 1.66 95.6 9.00 13.9 19.4 0.003 trans-Nonachlor ≥25 18.0 11.6 1.41 106 10.4 15.7 22.9  cis-Nonachlor <25 2.02 0.96 0.98 9.14 1.49 1.74 2.14 0.835 cis-Nonachlor ≥25 2.13 1.22 0.79 11.3 1.46 1.74 2.32  Oxychlordane <25 11.2 7.14 1.46 58.2 6.06 9.66 14.8 0.002 Oxychlordane ≥25 12.6 7.23 0.82 53.9 7.81 11.1 15.7  p, p’-DDT <25 4.37 4.14 1.42 34.5 2.46 2.91 3.88 0.845 p’, p’-DDT ≥25 5.18 7.82 1.37 92.1 2.40 2.88 5.08  p, p’-DDE <25 342 418 30.1 3390 105 199 410 <0.001 p’, p’-DDE ≥25 410 484 9.28 5820 148 275 507  Abbreviations  25p: 25th percentile; 75p: 75th percentile; BMI: Body mass index; Max.: Maximum lipid-adjusted organochlorine concentration (μg/kg); Min.: Minimum lipid-adjusted organochlorine concentration (μg/kg); SD: Standard deviation (μg/kg);  * Sample sizes for BMI group: <25 (n=288); ≥25 (n=395);  ** Mean and median values are based on lipid-adjusted organochlorine concentrations where undetectable values are replaced by the detection limit divided by √2; *** Wilcoxon rank-sum asymptotic 2-sided p-value excludes unknown data;  91  Table 4.10  European lipid-adjusted organochlorine concentrations (μg/kg) by education group and p-values for the Wilcoxon rank-sum tests Organochlorine Education group* Mean** SD Min. Max. 25p Median 75p p*** PCB-28 Less than post-secondary graduate 2.06 1.47 0.53 17.9 1.42 1.64 2.04 0.241 PCB-28 Post-secondary graduate 2.34 3.49 0.82 57.5 1.47 1.70 2.09  PCB-99 Less than post-secondary graduate 7.00 6.06 0.53 61.3 3.20 5.76 8.93 0.033 PCB-99 Post-secondary graduate 6.22 5.56 0.82 53.1 2.95 4.76 7.79  PCB-105 Less than post-secondary graduate 2.68 3.05 0.53 33.5 1.49 1.80 2.68 0.465 PCB-105 Post-secondary graduate 2.67 3.30 0.82 37.6 1.48 1.77 2.48  PCB-118 Less than post-secondary graduate 12.1 14.9 1.32 202 5.13 8.42 14.7 0.374 PCB-118 Post-secondary graduate 10.9 10.9 0.82 130 4.64 8.12 13.3  PCB-138 Less than post-secondary graduate 26.0 18.8 1.41 156 13.4 21.9 33.5 0.073 PCB-138 Post-secondary graduate 24.7 22.5 1.61 289 12.0 19.5 30.9  PCB-153 Less than post-secondary graduate 57.5 60.8 1.44 736 29.1 45.3 67.4 0.075 PCB-153 Post-secondary graduate 51.2 47.9 1.61 659 27.2 40.4 61.1  PCB-156 Less than post-secondary graduate 7.87 7.81 1.11 113 4.27 6.49 9.19 0.140 PCB-156 Post-secondary graduate 7.27 5.55 0.82 56.7 3.90 5.97 8.94  PCB-170 Less than post-secondary graduate 22.6 56.9 1.41 902 8.51 13.1 19.2 0.013 PCB-170 Post-secondary graduate 17.8 25.1 0.82 200 7.38 11.6 18.4  PCB-180 Less than post-secondary graduate 78.3 235 2.72 3790 27.3 40.7 63.1 0.012 PCB-180 Post-secondary graduate 60.1 95.6 1.61 808 22.9 36.5 58.8  PCB-183 Less than post-secondary graduate 4.78 7.49 1.11 84.9 1.88 3.21 5.08 0.041 PCB-183 Post-secondary graduate 4.05 4.81 0.82 54.0 1.78 2.77 4.60  PCB-187 Less than post-secondary graduate 22.4 60.9 1.21 833 7.45 11.5 17.7 0.007 PCB-187 Post-secondary graduate 15.8 24.9 0.82 208 6.54 9.71 15.5  Total summed PCBs Less than post-secondary graduate 249 429 31.8 6570 119 174 255 0.029 Total summed PCBs Post-secondary graduate 209 211 23.6 2030 105 158 231  Summed dioxin-like PCBs Less than post-secondary graduate 22.6 21.5 4.25 281 11.9 17.7 27.5 0.231 Dioxin-like summed PCBs Post-secondary graduate 20.9 16.4 2.47 170 10.9 17.0 26.4  Summed non-dioxin-like PCBs Less than post-secondary graduate 227 420 25.9 6450 107 152 226 0.022 Non-dioxin-like summed PCBs Post-secondary graduate 188 202 18.8 1910 93.8 138 209  92  Organochlorine Education group* Mean** SD Min. Max. 25p Median 75p p*** β-HCCH Less than post-secondary graduate 21.7 48.6 2.30 839 9.73 15.1 21.6 0.201 β-HCCH Post-secondary graduate 22.1 37.4 1.37 321 8.26 14.1 21.6  HCB Less than post-secondary graduate 26.7 61.4 3.57 1050 12.6 18.1 27.8 0.001 HCB Post-secondary graduate 19.8 19.1 1.61 299 11.5 16.1 22.3  Mirex Less than post-secondary graduate 3.36 5.14 0.92 60.5 1.55 1.95 3.24 0.030 Mirex Post-secondary graduate 2.63 2.98 0.82 42.4 1.49 1.77 2.75  trans-Nonachlor Less than post-secondary graduate 18.8 13.1 1.41 106 10.7 16.2 23.2 <0.001 trans-Nonachlor Post-secondary graduate 15.3 9.58 1.61 83.6 8.84 13.6 19.0  cis-Nonachlor Less than post-secondary graduate 2.17 1.32 0.79 11.3 1.47 1.75 2.30 0.336 cis-Nonachlor Post-secondary graduate 1.99 0.85 0.82 5.98 1.48 1.73 2.15  Oxychlordane Less than post-secondary graduate 13.3 7.57 1.68 58.2 7.98 12.1 17.2 <0.001 Oxychlordane Post-secondary graduate 10.6 6.64 0.82 53.9 5.98 9.38 13.5  p, p’-DDT Less than post-secondary graduate 4.70 4.70 1.42 40.5 2.39 2.90 4.38 0.936 p’, p’-DDT Post-secondary graduate 5.02 7.97 1.37 92.1 2.45 2.91 4.22  p, p’-DDE Less than post-secondary graduate 412 467 9.28 3810 131 270 506 0.040 p’, p’-DDE Post-secondary graduate 352 448 23.8 5820 124 215 420  Abbreviations  25p: 25th percentile; 75p: 75th percentile; Max.: Maximum lipid-adjusted organochlorine concentration (μg/kg); M in.: Minimum lipid-adjusted organochlorine concentration (μg/kg); SD: Standard deviation (μg/kg);  * Sample sizes for Education group: Less than post-secondary graduate (n=344); Post-secondary graduate (n=341); ** Mean and median values are based on lipid-adjusted organochlorine concentrations where undetectable values are replaced by the detection limit divided by √2; *** Wilcoxon rank-sum asymptotic 2-sided p-value excludes unknown data; 93  Table 4.11  European lipid-adjusted organochlorine concentrations (μg/kg) by family history of NHL and p-values for the Wilcoxon rank-sum tests Organochlorine Family history of NHL* Mean** SD Min. Max. 25p Median 75p p*** PCB-28 No 2.17 2.70 0.53 57.5 1.43 1.66 2.05 0.097 PCB-28 Yes 2.68 1.99 1.32 9.84 1.47 1.83 3.22  PCB-99 No 6.62 5.89 0.53 61.3 3.12 5.23 8.20 0.678 PCB-99 Yes 6.67 4.42 1.43 16.7 3.33 5.79 9.10  PCB-105 No 2.68 3.24 0.53 37.6 1.48 1.77 2.56 0.289 PCB-105 Yes 2.60 1.48 1.26 6.37 1.54 1.93 3.48  PCB-118 No 11.5 13.3 0.82 202 4.85 8.17 13.9 0.227 PCB-118 Yes 12.4 8.54 1.32 32.6 6.08 9.54 19.2  PCB-138 No 25.4 21.0 1.41 289 12.6 21.0 32.1 0.539 PCB-138 Yes 24.8 13.1 5.43 58.6 18.2 20.8 31.6  PCB-153 No 54.7 55.9 1.44 736 27.9 42.8 64.4 0.979 PCB-153 Yes 46.6 22.1 13.1 98.5 32.0 42.0 62.2  PCB-156 No 7.64 6.93 0.82 113 4.07 6.14 9.12 0.893 PCB-156 Yes 6.47 2.76 1.49 12.5 5.02 5.92 8.62  PCB-170 No 20.6 45.1 0.82 902 8.04 12.2 19.1 0.396 PCB-170 Yes 12.0 4.73 2.91 20.9 8.09 10.8 15.2  PCB-180 No 70.8 184 1.61 3790 25.4 39.1 62.0 0.301 PCB-180 Yes 38.1 15.9 8.73 70.3 25.6 36.7 49.0  PCB-183 No 4.47 6.44 0.82 84.9 1.81 2.94 4.79 0.989 PCB-183 Yes 3.49 1.98 1.39 9.83 1.95 3.01 4.24  PCB-187 No 19.6 47.7 0.82 833 6.73 10.5 17.4 0.708 PCB-187 Yes 11.2 5.37 3.02 25.4 7.36 11.0 14.0  Total summed PCBs No 232 346 23.6 6570 110 165 249 0.742 Total summed PCBs Yes 173 71.3 56.0 340 127 166 209  Summed dioxin-like PCBs No 21.8 19.5 2.47 281 11.4 17.4 26.9 0.588 Dioxin-like summed PCBs Yes 21.5 11.9 7.73 48.4 13.0 18.1 31.1  Summed non-dioxin-like PCBs No 210 338 18.8 6450 98.7 145 220 0.680 Non-dioxin-like summed PCBs Yes 151 63.0 48.3 292 107 149 173  94  Organochlorine Family history of NHL* Mean** SD Min. Max. 25p Median 75p p*** β-HCCH No 21.2 43.3 1.37 839 9.04 14.2 21.0 0.012 β-HCCH Yes 30.9 36.2 8.89 173 13.7 17.2 34.6  HCB No 22.4 43.9 1.61 1050 12.0 17.1 24.8 0.226 HCB Yes 23.9 15.1 7.38 62.0 13.5 19.4 28.5  Mirex No 3.02 4.30 0.82 60.5 1.52 1.85 3.02 0.810 Mirex Yes 3.09 2.74 1.14 11.8 1.47 1.89 2.91  trans-Nonachlor No 16.9 11.7 1.41 106 9.67 14.4 21.0 0.060 trans-Nonachlor Yes 18.5 7.44 5.49 41.0 13.6 18.0 21.1  cis-Nonachlor No 2.06 1.12 0.79 11.3 1.47 1.73 2.20 0.227 cis-Nonachlor Yes 2.20 1.08 1.32 6.58 1.43 1.93 2.52  Oxychlordane No 11.9 7.29 0.82 58.2 6.95 10.5 15.2 0.325 Oxychlordane Yes 12.3 4.82 4.88 25.0 8.69 11.9 15.9  p, p’-DDT No 4.71 6.43 1.37 92.1 2.42 2.88 4.05 0.025 p, p’-DDT Yes 7.53 8.38 1.92 29.7 2.76 3.39 7.16  p, p’-DDE No 373 442 9.28 5820 121 242 455 0.037 p, p’-DDE Yes 525 659 105 3390 198 294 570  Abbreviations  25p: 25th percentile; 75p: 75th percentile; Max.: Maximum lipid-adjusted organochlorine concentration (μg/kg); Min.: Minimum lipid-adjusted organochlorine concentration (μg/kg); SD: Standard deviation (μg/kg);  * Sample sizes for Family history of NHL: Yes (n=653); No (n=26);  ** Mean and median values are based on lipid-adjusted organochlorine concentrations where undetectable values are replaced by the detection limit divided by √2;  *** Wilcoxon rank-sum asymptotic 2-sided p-value excludes unknown data;95  4.1.4 Occupational analyses  The frequencies of 1st usual occupations hypothesized to have possible OC exposure, as well as farming-related exposure groups, are presented by case/ control status in Table 4.12 for 694 European participants. Because the number of participants in individual possible exposure categories was small, individuals who self-reported ever living/ working on farm, reported a farming-related usual occupation or reported a 1st usual occupation with possible OC exposure were also combined into a single possible occupational OC exposure group. A similar proportion of cases and controls were in this possible OC exposure group.  In order to explore the relationship between occupation/ farming-related exposure and plasma organochlorine concentrations, the associations between OC levels and the occupational OC exposure group were examined. As shown in Table 4.13, no statistically significant differences were observed. Similarly, no significant differences were observed when associations were adjusted for age, sex, region, BMI, education, and case/ control status.     96  Table 4.12  Farming history and 1st usual occupations of European cases and controls with organochlorine measurements included in analyses [frequency (percentage)]    Controls (n=363) Cases (n=331) Farming 1st usual occupation  3 (0.83) 2 (0.60) Farming-related usual occupation  5 (1.38) 3 (0.91) Self-reported ever living/ working on farm  11 (3.03) 13 (3.93) Self-reported ever living/ working on farm OR farming-related usual occupation 16 (4.41) 16 (4.83) Electrical/ Electronics Manufacturing/ Repair 6 (1.65) 1 (0.30) Electrical Power 3 (0.83) 3 (0.91) Fishing  1 (0.28) 0 (0.00) Forestry 1 (0.28) 2 (0.60) Metal Machining 3 (0.83) 2 (0.60) Industrial Machinery Fabrication/ Assemblage/ Repair 3 (0.83) 2 (0.60) Painting 2 (0.55) 1 (0.30) Construction (Caulking and Coating) 3 (0.83) 1 (0.30) Cleaning  3 (0.83) 2 (0.60) Possible OC Exposure 38 (10.5) 29 (8.76) Unlikely OC Exposure (Reference group) 316 (87.1) 280 (84.6) N.B. Columns do not add up to 100% for the following reasons: 1. Rounding;  2. All categories are mutually exclusive except the farming occupations. “Farming 1st usual occupation” only includes those reporting their 1st usual occupation as directly involving farming (i.e. SOC Major Group 71-Farming, Horticultural and Animal Husbandry Occupations). “Farming-related usual occupation” includes those reporting a usual occupation directly involving farming (i.e. SOC Major group 71), even if this is not their 1 st reported occupation, as well as those reporting usual occupations that are farming-related, but do not code to Major group 71. This means that “Farming-related usual occupation” includes everyone found in “Farming 1st usual occupation”, but not vice-versa. “Self-reported ever living/ working on farm OR farming-related usual occupation” includes both those who self-reported ever living/ working on a farm and those with a farming-related usual occupation. 97  3. Also note that 3 controls and 1 case self-reported ever living/ working on a farm in addition to having a 1st usual occupation hypothesized to have possible OC exposure. They are therefore included in multiple categories. 4. 9 controls and 22 cases were marked as unknown because information suggestive of possible exposure was not found, but possible exposure could not be ruled out. 5. “Possible OC Exposure” combines all individuals who self-reported ever living/ working on farm, reported a farming-related usual occupation or reported a 1st usual occupation with possible OC exposure due to small numbers in individual possible OC exposure occupations.  98  Table 4.13  European lipid-adjusted organochlorine concentrations (μg/kg) by occupational OC exposure group and p -values for the Wilcoxon rank-sum tests Organochlorine Occupational OC Exposure group* Mean** SD Min. Max. 25p Median 75p p*** PCB-28 Unlikely 2.10 1.57 0.53 21.9 1.44 1.67 2.04 0.505 PCB-28 Possible 2.18 2.26 0.97 17.9 1.39 1.63 2.13  PCB-99 Unlikely 6.59 5.77 0.53 61.3 3.18 5.22 8.26 0.529 PCB-99 Possible 7.17 6.70 1.30 46.3 3.64 5.69 8.03  PCB-105 Unlikely 2.62 2.99 0.53 37.6 1.49 1.77 2.67 0.628 PCB-105 Possible 2.64 3.14 1.10 25.2 1.42 1.76 2.52  PCB-118 Unlikely 11.6 12.7 0.82 202 5.17 8.31 14.5 0.125 PCB-118 Possible 11.2 16.5 1.32 130 4.00 7.66 11.7  PCB-138 Unlikely 25.6 21.2 1.41 289 12.9 21.4 32.6 0.856 PCB-138 Possible 25.5 18.2 1.48 108 13.5 20.3 31.5  PCB-153 Unlikely 55.5 57.2 1.61 736 28.3 43.2 64.9 0.966 PCB-153 Possible 51.4 36.7 4.86 244 30.5 43.9 62.9  PCB-156 Unlikely 7.72 7.11 0.82 113 4.10 6.21 9.12 0.955 PCB-156 Possible 7.20 4.23 1.68 20.3 4.29 6.01 9.18  PCB-170 Unlikely 21.1 47.0 0.82 902 8.06 12.3 19.3 0.887 PCB-170 Possible 15.7 12.0 2.41 79.1 9.07 12.2 17.4  PCB-180 Unlikely 72.9 192 1.61 3790 25.4 39.3 62.1 0.739 PCB-180 Possible 50.5 36.7 7.59 232 28.4 39.7 60.0  PCB-183 Unlikely 4.54 6.66 0.82 84.9 1.83 2.98 4.84 0.935 PCB-183 Possible 3.90 3.16 1.28 21.5 1.95 3.01 4.17  PCB-187 Unlikely 20.2 49.8 0.82 833 6.86 10.7 17.3 0.933 PCB-187 Possible 13.9 12.1 2.30 84.9 7.48 10.4 18.8  Total summed PCBs Unlikely 236 360 23.6 6570 111 168 250 0.986 Total summed PCBs Possible 197 125 48.0 818 119 165 242  Summed dioxin-like PCBs Unlikely 21.9 18.9 2.47 281 11.8 17.6 27.2 0.274 Dioxin-like summed PCBs Possible 21.1 21.5 5.02 170 10.5 16.4 25.0  99  Organochlorine Occupational OC Exposure group* Mean** SD Min. Max. 25p Median 75p p*** Summed non-dioxin-like PCBs Unlikely 214 351 18.8 6450 99.5 146 220 0.883 Non-dioxin-like summed PCBs Possible 176 115 40.8 786 107 148 215  β-HCCH Unlikely 22.3 45.3 1.37 839 9.30 14.9 21.8 0.056 β-HCCH Possible 16.1 19.9 2.46 147 8.64 13.7 16.5  HCB Unlikely 22.9 45.8 1.61 1050 12.2 17.4 25.5 0.293 HCB Possible 19.4 11.7 6.56 57.3 11.8 16.1 21.0  Mirex Unlikely 2.84 3.44 0.82 45.1 1.50 1.83 2.92 0.428 Mirex Possible 4.48 8.68 0.97 60.5 1.55 2.00 3.19  trans-Nonachlor Unlikely 17.0 11.6 1.41 106 9.74 14.4 20.9 0.196 trans-Nonachlor Possible 18.2 12.0 1.68 83.6 10.8 16.3 24.3  cis-Nonachlor Unlikely 2.08 1.15 0.79 11.3 1.47 1.73 2.20 0.830 cis-Nonachlor Possible 2.04 0.94 1.10 5.70 1.41 1.69 2.35  Oxychlordane Unlikely 11.9 7.13 0.82 58.2 7.22 10.5 15.2 0.336 Oxychlordane Possible 12.8 8.13 1.68 53.9 7.76 12.1 15.6  p, p’-DDT Unlikely 4.82 6.67 1.37 92.1 2.43 2.90 4.17 0.640 p, p’-DDT Possible 4.05 2.70 1.52 14.2 2.26 2.82 4.30  p, p’-DDE Unlikely 380 446 14.9 5820 130 250 469 0.648 p, p’-DDE Possible 340 308 30.1 1600 122 224 495  Abbreviations  25p: 25th percentile; 75p: 75th percentile; Max.: Maximum lipid-adjusted organochlorine concentration (μg/kg); Min.: Minimum lipid-adjusted organochlorine concentration (μg/kg); SD: Standard deviation (μg/kg);  * Sample sizes for Occupational OC Exposure group: Unlikely (n=596); Possible (n=67);  ** Mean and median values are based on lipid-adjusted organochlorine concentrations where undetectable values are replaced by the detection limit divided by √2; *** Wilcoxon rank-sum asymptotic 2-sided p-value excludes unknown data; 100  4.1.5 Organochlorine main effects  4.1.5.1 Confounder selection   The confounders selected by the change-in-estimate criterion for inclusion in the final logistic regression models of the organochlorine-NHL relationship in Europeans are shown in Table 4.14. Some of these final confounders differ from those selected in previous analyses of all ethnicities, due to the restriction to Europeans and occupational OC exposure replacing having ever lived/ worked on a farm as a potential confounder.[13, 14]     101  Table 4.14  Confounders included in final logistic regression models of the associations between organochlorines and NHL in Europeans based on the change-in-estimate criterion  Organochlorine Confounders* PCB-28 None  PCB-99 Occupation  PCB-105 BMI  PCB-118 Age, Sex, BMI, Occupation  PCB-138 None  PCB-153 Age  PCB-156 Occupation  PCB-170 None  PCB-180 Age  PCB-183 Occupation  PCB-187 Family history of NHL, Age, BMI  Total summed PCBs Age, Occupation  Summed dioxin-like PCBs Age, BMI, Occupation  Summed non-dioxin-like PCBs None  β-HCCH None  HCB Age, Sex, BMI  Mirex Occupation  trans-Nonachlor Age, Occupation  cis-Nonachlor None  Oxychlordane Family history of NHL, Age, BMI  p, p’-DDT None  p, p’-DDE BMI  * Occupation: Possible Occupational OC Exposure group versus Unlikely Occupational OC Exposure group as defined in Section 3.4.1.4.2;  102  4.1.5.2 Final organochlorine models in NHL   The associations between organochlorines and NHL are shown in Table 4.15. For PCB congeners, significant trends were found for PCB-153, 180 and 187, with odd ratios for the highest OC exposure relative to the lowest OC exposure ranging from 1.58 [0.98, 2.54] for PCB-153 to 1.82 [1.10, 3.01] for PCB-187. Significant trends were also noted for the summed PCB measures. For the pesticides/ pesticide metabolites, significant trends were found for β-HCCH, HCB, mirex, trans-nonachlor, and oxychlordane, with ORs ranging from 1.44 [1.05, 1.97] for mirex to 2.62 [1.54, 4.45] for oxychlordane. After adjusting for multiple comparisons, PCB-187, total summed PCBs, β-HCCH, HCB, trans-nonachlor, and oxychlordane remained statistically significant.   103  Table 4.15  Lipid-adjusted organochlorine associations with non-Hodgkin lymphoma   Organochlorine Categories (μg/kg) Controls (n) Cases (n) OR L95 U95 p-trend PCB-28 Not detected 291 268 1   0.790 qc281c(1) ≥1.38 to 57.5 72 63 0.95 0.65 1.39  PCB-99 ≤3.29 87 87 1   0.261 qc991c(1) >3.29 to 4.97 89 56 0.63 0.40 0.98  qc991c(2) >4.97 to 7.81 89 68 0.77 0.50 1.19  qc991c(3) >7.81 to 61.3 89 98 1.10 0.73 1.67  PCB-105 Not detected 243 223 1   0.818 qc1051c(1) ≥1.32 to 37.6 115 102 0.96 0.70 1.33  PCB-118 ≤4.70 84 61 1   0.086 qc1181c(1) >4.70 to 8.05 89 75 1.19 0.74 1.91  qc1181c(2) >8.05 to 13.4 90 77 1.24 0.76 2.03  qc1181c(3) >13.4 to 202 87 92 1.56 0.94 2.60  PCB-138 ≤12.7 91 84 1   0.225 qc1381c(1) >12.7 to 20.0 91 69 0.82 0.53 1.26  qc1381c(2) >20.0 to 30.6 91 79 0.94 0.62 1.44  qc1381c(3) >30.6 to 289 90 99 1.19 0.79 1.80  PCB-153 ≤27.2 91 71 1   0.030 qc1531c(1) >27.2 to 40.7 91 74 1.05 0.66 1.67  qc1531c(2) >40.7 to 60.0 91 78 1.13 0.69 1.83  qc1531c(3) >60.0 to 736 90 108 1.58 0.98 2.54  PCB-156 ≤4.03 86 67 1   0.059 qc1561c(1) >4.03 to 5.97 90 69 0.99 0.63 1.55  qc1561c(2) >5.97 to 8.63 90 75 1.07 0.69 1.67  qc1561c(3) >8.63 to 113 88 98 1.44 0.93 2.21  104  Organochlorine Categories (μg/kg) Controls (n) Cases (n) OR L95 U95 p-trend PCB-170 ≤7.90 91 73 1   0.061 qc1701c(1) >7.90 to 11.8 91 72 0.99 0.64 1.53  qc1701c(2) >11.8 to 18.3 91 84 1.15 0.75 1.77  qc1701c(3) >18.3 to 902 90 102 1.41 0.93 2.15  PCB-180 ≤24.3 91 66 1   0.023 qc1801c(1) >24.3 to 37.2 91 76 1.21 0.75 1.94  qc1801c(2) >37.2 to 58.3 91 84 1.35 0.83 2.21  qc1801c(3) >58.3 to 3790 90 105 1.73 1.05 2.83  PCB-183 Not detected 122 107 1   0.527 qc1831c(1) ≥1.87 to 3.94 115 90 0.90 0.61 1.31  qc1831c(2) >3.94 to 84.9 117 112 1.09 0.75 1.57  PCB-187 ≤6.49 89 66 1   0.012** qc1871c(1) >6.49 to 10.1 89 75 1.22 0.75 2.00  qc1871c(2) >10.1 to 15.5 91 67 1.09 0.65 1.82  qc1871c(3) >15.5 to 833 89 105 1.82 1.10 3.01  Total summed PCBs ≤108 86 63 1   0.007** qtotpcblc(1) >108 to 159 90 71 1.13 0.69 1.87  qtotpcblc(2) >159 to 227 90 66 1.07 0.63 1.80  qtotpcblc(3) >227 to 6570 88 109 1.83 1.10 3.06  Summed dioxin-like PCBs ≤10.7 83 51 1   0.028 qcpcb_lc(1) >10.7 to 16.5 91 79 1.54 0.93 2.54  qcpcb_lc(2) >16.5 to 25.4 89 81 1.64 0.98 2.76  qcpcb_lc(3) >25.4 to 281 87 94 2.01 1.18 3.42  Summed non-dioxin-like PCBs <95.8 90 72 1   0.020 qnpcb_lc(1) ≥95.8 to 141 92 70 0.95 0.61 1.48  qnpcb_lc(2) >141 to 204 90 78 1.08 0.70 1.67  qnpcb_lc(3) ≥204 to 6450 91 111 1.53 1.01 2.31    105  Organochlorine Categories (μg/kg) Controls (n) Cases (n) OR L95 U95 p-trend β-HCCH ≤8.89 91 70 1   0.003** qbhc10c(1) >8.89 to 13.3 91 58 0.83 0.53 1.30  qbhc10c(2) >13.3 to 19.6 91 88 1.26 0.82 1.93  qbhc10c(3) >19.6 to 839 90 115 1.66 1.10 2.52  HCB ≤11.7 88 65 1   0.001** qhexa4c(1) >11.7 to 16.4 91 71 1.13 0.71 1.80  qhexa4c(2) >16.4 to 22.4 89 72 1.20 0.74 1.96  qhexa4c(3) >22.4 to 1050 90 117 2.10 1.26 3.48  Mirex Not detected 234 178 1   0.024 qmire4c(1) ≥1.43 to 60.5 120 131 1.44 1.05 1.97  trans-Nonachlor ≤9.32 87 60 1   0.012** qtran4c(1) >9.32 to 14.0 90 67 1.16 0.71 1.89  qtran4c(2) >14.0 to 19.7 88 83 1.53 0.92 2.55  qtran4c(3) >19.7 to 106 89 99 1.84 1.10 3.09  cis-Nonachlor Not detected 277 238 1   0.186 qcis4c(1) ≥0.79 to 11.3 86 93 1.26 0.90 1.77  Oxychlordane ≤6.42 88 56 1   <0.001** qoxy4c(1) >6.42 to 9.89 90 68 1.34 0.82 2.20  qoxy4c(2) >9.89 to 14.0 90 76 1.58 0.94 2.67  qoxy4c(3) >14.0 to 58.2 90 113 2.62 1.54 4.45  p, p’-DDT Not detected 275 240 1   0.329 qpddt10c(1) ≥3.13 to 92.1 88 91 1.19 0.84 1.67  p, p’-DDE ≤124 90 77 1   0.114 qpdde10c(1) >124 to 237 91 74 0.96 0.62 1.48  qpdde10c(2) >237 to 433 88 76 1.00 0.65 1.55  qpdde10c(3) >433 to 5820 89 98 1.32 0.86 2.03  Abbreviations  L95: Lower limit of 95% confidence interval; OR: Odds ratio; U95: Upper limit of 95% confidence interval; N.B. OR for reference group, comprised of the lowest organochlorine concentration group, is equal to 1 by d efinition.  Statistical significance after controlling false discovery rate is indicated by **; 106  4.1.5.3 Interactions with covariates  No statistically significant interactions were found between OCs and potential confounders.  4.1.5.4 NHL subtypes  The odds ratios were relatively consistent across NHL subtypes. However, significant heterogeneity between subtypes was observed for PCB-99 (p-het=0.035), as shown in Table 4.16. Although there was no overall association between NHL and PCB-99, there was a significant trend in other B-cell subtypes (highest OC quartile compared to the lowest OR=1.78 [1.02, 3.09]).  107  Table 4.16  Associations between PCB-99 and NHL subtypes Subtype Quartiles  (μg/kg) Controls       (n) Cases          (n) OR L95 U95 p-trend DLBCL ≤3.29 87 20 1   0.200 qc991c(1) >3.29 to 4.97 89 7 0.34 0.14 0.85  qc991c(2) >4.97 to 7.81 89 7 0.34 0.14 0.86  qc991c(3) >7.81 to 61.3 89 11 0.54 0.24 1.19  FL ≤3.29 87 29 1   0.471 qc991c(1) >3.29 to 4.97 89 17 0.56 0.29 1.10  qc991c(2) >4.97 to 7.81 89 26 0.89 0.49 1.64  qc991c(3) >7.81 to 61.3 89 31 1.05 0.59 1.89  Other B-cell ≤3.29 87 27 1   0.013 qc991c(1) >3.29 to 4.97 89 25 0.91 0.49 1.69  qc991c(2) >4.97 to 7.81 89 29 1.05 0.57 1.91  qc991c(3) >7.81 to 61.3 89 49 1.78 1.02 3.09  All T-cell ≤3.29 87 11 1   0.376 qc991c(1) >3.29 to 4.97 89 7 0.62 0.23 1.68  qc991c(2) >4.97 to 7.81 89 5 0.45 0.15 1.34  qc991c(3) >7.81 to 61.3 89 7 0.62 0.23 1.68  Abbreviations  DLBCL: Diffuse large B-cell lymphoma; FL: Follicular lymphoma; L95: Lower limit of 95% confidence interval; OR: Odds ratio; U95: Upper limit of 95% confidence interval;  N.B. OR for reference group, comprised of the lowest organochlorine exposure group, is eq ual to 1 by definition;   108  4.2 Genetic analyses  4.2.1 Genetic quality control  4.2.1.1 Illumina   One hundred-nineteen tagSNPs belonging to 18 CYP genes were selected for genotyping using the Illumina assay. Ninety-eight SNPs belonging to 15 CYP genes remained after QC for the Illumina assay. Nine CYP SNPs failed the Illumina assay (CYP1B1: rs162549; CYP2A6: rs2316213; CYP2A13: rs1645690; CYP2B6: rs2279342; CYP3A7: rs2687079; CYP7B1: rs16931360; CYP19A1: rs10519296, rs12439137; CYP24A1: rs6127118). One SNP had a GenTrain score less than 0.4 (CYP2C19: rs3758581). Five SNPs had GenTrain scores between 0.4 and 0.7 in addition to poor clustering (CYP1A1: rs4646421; CYP2E1: rs2070674; CYP7A1: rs8192871; CYP24A1: rs2245153, rs2762939). Three SNPs were in regions containing possible CNVs (CYP2E1: rs2480259, rs2515641; CYP19A1: rs1062033). One SNP was removed for having one inconsistency between replicates and imperfect clustering (CYP2B6: rs1042389) and another was removed for having a call rate <0.95 (CYP2C9: rs9332238).   rs2302989 (CYP2A6) had a call rate <0.95 in non-WGA mouthwash samples and a GenTrain score between 0.4 and 0.7 in addition to bad clustering in WGA samples. It was first only removed from these sample types, but was eventually removed from all remaining samples for being out of HWE. rs2153628 (CYP2C9), rs4917623 (CYP2C19), rs915908 (CYP2E1), rs4258041 (CYP7B1), rs743572 (CYP17A1), rs3751592 (CYP19A1), rs7172156 (CYP19A1), rs13038432 (CYP24A1), rs2244719 (CYP24A1), and rs2585428 (CYP24A1) were only problematic in WGA samples and were therefore only removed from WGA samples. With respect to the AIMs, six out of the original 51 failed QC. Although population stratification cannot be ruled out, the resulting genomic inflation factor (based on the median chi-squared statistic) in Europeans was equal to λ=1, which provides no evidence of stratification.  Details on the samples/ individuals removed during the QC process for the Illumina 109  assay have been described elsewhere.[139] After QC, 1403 individuals remained for analysis (709 cases and 694 controls). Of these, 1111 were Europeans (564 cases and 547 controls).   4.2.1.2 Taqman   Refer to Figure 3.3 for a flowchart of the QC process. During QC, two samples (1 case; 1 control) were found to be missing and were therefore used as negative controls. One control was removed because the extracted DNA was determined to have a concentration of 0 and also failed amplification.  Twenty-five individuals whose samples were excluded during QC of the Illumina assay were also removed (14 cases and 8 controls with possible sample switches during processing; 2 cases belonging to sibling pairs; 1 sample belonged to a control who became a case). The ethnicities of two middle-Eastern controls who clustered with the CEU population in Illumina were also changed to an ethnicity of “other” from “Asian”. Forty-seven individuals (18 cases and 29 controls) were removed because their samples were missing more than one genotype. After QC, 1504 individuals (753 cases and 751 controls) remained for analysis. Of these, 1180 were Europeans (598 cases and 582 controls). Ten CYP1A1 SNPs were selected for genotyping using the Taqman assay. CYP1A1_7_3UTR5v11 failed at the assay design stage and was therefore not genotyped. Nine CYP1A1 SNPs were genotyped using Taqman, but only eight SNPs passed further QC procedures. CYP73UT1_1-v8 (rs1799814) had an inconsistent genotype with non-WGA genotyping replicates and three inconsistencies with sequencing replicates. It was excluded. CYP1A1_83t2 (rs7495708) genotypes were only removed from non-WGA saliva samples because the call rate was <0.95 in this sample type. Several SNPs failed only in WGA samples and were therefore only removed from this sample type. CYP1A1_109t3 (rs1456432) had a call rate of <0.95 in WGA lymphocytes, WGA blood and WGA mouthwash. This SNP was removed from all WGA samples, including 110  WGA saliva, as few samples of this type were present. CYP5UTRNC2v2 (rs4646421) had a call rate <0.95 in WGA mouthwash and was first only removed from this sample type before being removed from all WGA samples for being out of HWE. CYP_73UTR4v10 (rs4646903), CYP73UT1_2v9 (rs1048943) and CYP1A1_83t2 (rs7495708) were removed from WGA samples as they were out of HWE. Discordance between WGA/ non-WGA pairs resulted in the removal of CYP1A1_14t1 (rs2606345) from WGA samples.   4.2.2 Descriptive statistics    The genotyped samples for 1529 participants passed quality control for the Illumina or the Taqman assays. Of these, 1197 participants were European, 159 were Asian, 61 were South Asian, 69 were other/ mixed, and 43 refused to respond or were of unknown ethnic origin. The frequencies of various demographic variables and other covariates by case/ control status for the 1197 Europeans with genotyping measurements that passed quality control are presented in Table 4.17.    111  Table 4.17  Characteristics of European cases and controls with genotyping measurements included in analyses [frequency (percentage)]     Controls (n=594) Cases (n=603) Histology path4gp consolidated for multiple diagnoses   Multiple phenotypes (B-cell and T-cell)    1 (0.17) DLBCL    155 (25.7) FL    174 (28.9) Other B-cell    225 (37.3) All T-cell    48 (7.96) Age at selection (years) Agegp   20 – 49  131 (22.1) 96 (15.9) 50 – 59  132 (22.2) 146 (24.2) 60 – 69  165 (27.8) 172 (28.5) ≥70  166 (27.9) 189 (31.3) Sex SEX   Male  315 (53.0) 346 (57.4) Female  279 (47.0) 257 (42.6) Region Reg   GVRD  428 (72.1) 477 (79.1) CRD  166 (27.9) 126 (20.9) Family history of NHL . fam hist – nhl   No  575 (96.8) 560 (92.9) Yes  17 (2.86) 23 (3.81) Unknown*  2 (0.34) 20 (3.32) Highest level of education  educgp3New   <High school  69 (11.6) 109 (18.1) High school graduate  189 (31.8) 205 (34.0) Post-secondary graduate  330 (55.6) 280 (46.4) Unknown**  6 (1.01) 9 (1.49) Income (dollars)  incngpNew  <25 000 90 (15.2) 111 (18.4) 25 000 - 49 999 166 (27.9) 152 (25.2) 50 000 - 74 999 132 (22.2) 120 (19.9) 75 000 - 99 999 74 (12.5) 81 (13.4) ≥100 000 67 (11.3) 72 (11.9) Unknown** 65 (10.9) 67 (11.1)   112      Controls (n=594) Cases (n=603) Per capita income (dollars/family member)  pcigpNew   <15 000  122 (20.5) 132 (21.9) 15 000 - 29 999  164 (27.6) 155 (25.7) 30 000 - 44 999  174 (29.3) 178 (29.5) ≥45 000  68 (11.4) 67 (11.1) Unknown**  66 (11.1) 71 (11.8) Body mass index (kg/m2)  bmigpNew   <25  255 (42.9) 236 (39.1) 25 - 27.4  155 (26.1) 177 (29.4) 27.5 - 29.9  81 (13.6) 76 (12.6) ≥30  93 (15.7) 106 (17.6) Unknown***  10 (1.68) 8 (1.33) Abbreviations  BMI: Body mass index; CRD: Capital regional district; DLBCL: Diffuse large B-cell lymphoma; FL: Follicular lymphoma; GVRD: Greater Vancouver regional district;  Histology: 1 case excluded from subtype analyses because individual had 2 NHL diagnoses with different phenotypes (still included in overall NHL analyses);  Highest level of education: "High-school graduate" includes those with some university education. "Post-secondary graduate" includes trade, vocational, community college, or university graduates; Note that this categorization differs from that published in [13]; Income: Average annual household income before taxes in the year prior to interview;  Per capita income: Midpoint of average annual household income ranges divided by the number of family members supported by this income (if 0 members reported, set to unknown);  Body mass Index: BMI one year prior to study participation (values <16 kg/m2 or >56 kg/m2 set to "unknown") * "Unknown" includes missing data; ** "Unknown" includes refusal to respond/ don't know responses/ missing data; *** "Unknown" includes don't know responses/ missing data;    113  4.2.3 Genetic main effects  4.2.3.1 Genetic models in NHL   Because CYP1A1_2_1v4 (rs149846182) and CYP1A1_5_6v6 (rs191792412) had fewer than 10 individuals in the heterozygous/ minor allele homozygous group, they were not analysed. The associations between 104 CYP SNPs and NHL are shown in Table C.1. Supplementary tables for the codominant, additive, dominant, and recessive models for each SNP are in Table C.2 to Table C.5.  Only rs743572 (CYP17A1) and rs1322179 (CYP2C19) have statistically significant p-trends (rs743572: ORAG vs. AA=1.16 [0.88, 1.55], ORGG vs. AA=1.75 [1.18, 2.61], p-trend=0.009; rs1322179: ORAG vs. GG=1.12 [0.84, 1.48], ORAA vs. GG=4.45 [1.45, 13.7], p-trend=0.048). However, significance was not maintained after controlling for multiple testing.  In addition, there were two SNPs, rs10509679 (CYP2C9) and rs9332197 (CYP2C9), for which p-trends approached significance (rs10509679: ORAG vs. GG=1.10 [0.84, 1.44], ORAA vs. GG=3.24 [1.33, 7.90], p-trend=0.051; rs9332197: ORAG vs. AA=1.50 [1.02, 2.20], ORGG vs. AA=1.10 [0.31, 3.86], p-trend=0.066). As the two significant SNPs, these were also found on chromosome 10. As shown in an LD plot of chromosome 10, generated by HaploView 4.2 using European controls (Figure 4.1), rs743572 and rs9332197 are not in high LD with each other, rs1322179 or rs10509679. However, rs1322179 is in high LD with rs10509679, with an r2 value of 0.81. This is not surprising as the tagSNP selection process from HapMap would have prevented redundant SNPs within the same gene, but not between different genes.   114   Figure 4.1  Linkage disequilibrium plot of pairwise r2 values displayed as percentages between CYP SNPs analyzed on chromosome 10.  Plot generated in HaploView 4.2 using genotyping measurements from European controls. SNPs of particular interest are indicated by *. Colour scheme: r2 = 0 (white); 0 < r2 < 1 (shades of grey); r2 = 1 (black);  CYP2C19 CYP2C9 CYP17A1 CYP2E1 * * * * 115  4.2.3.1.1 Additional genetic models in NHL   Comparisons between the codominant model and the additive, recessive and dominant models for the two SNPs with statistically significant p-trends, rs743572 and rs1322179, and the low LD borderline significant SNP, rs9332197 are shown in Table 4.18. As the additive and recessive models were of similar comparability to the codominant for rs743572, the recessive model was selected for GxE analyses as it was the most comparable binary model. Recessive and dominant models were selected for GxE analyses for rs1322179 and rs9332197, respectively.  116   Table 4.18  Various genetic models for CYP SNPs selected for gene-environment interaction analyses SNP Model Genotype OR L95 U95 LRT p-value** rs743572      Codominant A A 1     A G 1.16 0.88 1.55   G G 1.75 1.18 2.61   Additive nG 1.29 1.06 1.55 0.353 Dominant A G or G G 1.28 0.98 1.68 0.035 Recessive G G 1.60 1.12 2.31 0.296 rs1322179      Codominant G G 1     A G 1.12 0.84 1.48   A A 4.45 1.45 13.7   Additive nA 1.28 1.00 1.65 0.030 Dominant A G or A A 1.21 0.92 1.59 0.009 Recessive A A 4.33 1.41 13.3 0.443 rs9332197      Codominant A A 1     A G 1.50 1.02 2.20   G G 1.10 0.31 3.86   Additive nG 1.37 0.98 1.91 0.333 Dominant A G or G G 1.46 1.01 2.12 0.637 Recessive G G 1.05 0.30 3.70 0.037 Abbreviations  L95: Lower limit of 95% confidence interval; LRT: Likelihood ratio test; OR: Odds ratio; SNP: Single nucleotide polymorphism; U95: Upper limit of 95% confidence interval;  N.B. OR for reference group, comprised of major allele homozygotes (as well as heterozygotes under the recessive model), is equal to 1 by definition;  Slight differences between results shown in this table and genetic models in other tables are due to analyses being conducted using different software (PASW/ SPSS and PLINK);  ** Model selected for use in gene-environment interaction analyses shown in bold (i.e. binary model most similar to the codominant model);  117  4.2.3.2 Interactions with covariates    We examined multiplicative statistical interactions of the CYP SNPs with sex, age and region. Four SNPs were shown to have statistically different effects between males and females. They were rs730154 (ORmales=1.26 [0.91, 1.75], ORfemales=0.65 [0.45, 0.92], pinteraction=0.009), rs2470155 (ORmales=1.47 [1.00, 2.15], ORfemales=0.77 [0.52, 1.14], pinteraction=0.027), CYP73UT1_2v9 (rs1048943) (ORmales=1.93 [0.93, 4.01], ORfemales=0.57 [0.23, 1.41], p interaction=0.046), and rs9332242 (ORmales=1.14 [0.80, 1.63], ORfemales=0.66 [0.44, 0.99], p interaction=0.049).  The effects of two SNPs were significantly different between those less than 60 years of age and 60 years of age or older. They were rs8192712 (OR<60=1.54 [0.96, 2.46], OR≥60=0.73 [0.51, 1.05], p interaction=0.012) and rs915908 (OR<60=1.50 [1.00, 2.24], OR≥60=0.80 [0.56, 1.14], p interaction=0.028).  Three SNPs showed statistically significant differences between regions. They were rs2472304 (ORCRD=0.65, [0.45, 0.95], ORGVRD=1.16 [0.95, 1.41], p interaction=0.016), rs762551 (ORCRD=0.65 [0.44, 0.97], ORGVRD=1.15 [0.93, 1.42], p interaction=0.024) and rs679320 (ORCRD=2.16 [0.96, 4.89], ORGVRD=0.81 [0.55, 1.18], p interaction=0.046).  4.2.3.3 NHL subtypes   The adjusted odds ratios were relatively consistent across NHL subtypes. However, statistically significant heterogeneity was found for several SNPs. For rs1007219, rs2470152, rs743535, and rs915906, the B-cell subtypes and the all T-cell subtypes had differing effects. For rs12911554, rs4774584 and rs7172156, DLBCL had an effect in the opposite direction from the other subtypes. ORs and 95% CIs under an additive genetic model are given for these SNPs for each subtype in Table 4.19. ORs and 95% CIs under codominant or dominant genetic models with additive p-trends are given in Table C.6 to Table C.12.    118  Table 4.19  Additive associations for CYP SNPs with heterogeneity between NHL subtypes  NHL subtype OR [95% CI]  SNP DLBCL FL Other B-cell All T-cell p-het rs743535  1.37 [0.87, 2.17] 1.48 [0.96, 2.28] 1.12 [0.75, 1.68] 0.24 [0.06, 1.02] 0.025 rs915906  0.80 [0.54, 1.18] 1.16 [0.83, 1.63] 1.07 [0.77, 1.47] 0.37 [0.16, 0.88] 0.020 rs1007219  1.29 [0.98, 1.69] 1.23 [0.95, 1.59] 0.92 [0.71, 1.19] 2.02 [1.30, 3.14] 0.010 rs2470152  1.28 [0.99, 1.67] 0.91 [0.71, 1.16] 1.03 [0.82, 1.29] 0.65 [0.42, 1.01] 0.018 rs4774584  0.77 [0.58, 1.01] 1.19 [0.93, 1.53] 1.08 [0.85, 1.36] 1.30 [0.84, 2.01] 0.048 rs7172156  1.41 [1.07, 1.86] 0.79 [0.60, 1.06] 0.89 [0.69, 1.15] 0.77 [0.47, 1.26] 0.002 rs12911554  1.42 [1.08, 1.87] 0.87 [0.67, 1.13] 0.98 [0.78, 1.24] 0.68 [0.43, 1.09] 0.015 Abbreviations  95% CI: 95% Confidence interval; DLBCL: Diffuse large B-cell lymphoma; FL: Follicular lymphoma;  OR: Odds ratio; p-het: Statistical significance of heterogeneity test; SNP: Single nucleotide polymorphism;   119  4.3 Gene-environment interaction analyses   Eleven OCs/ summed PCB measures had statistically significant trends with NHL in Europeans: PCB-153, 180, 187, total summed PCBs, summed dioxin-like PCBs, summed non-dioxin-like PCBs, β-HCCH, HCB, mirex, trans-nonachlor, and oxychlordane. Because of the high correlations between the individual PCB congeners and summed PCB measures, only total summed PCBs was tested for GxE interactions. Similarly, oxychlordane was tested in lieu of trans-nonachlor. Two SNPs had statistically significant trends with NHL: rs743572 and rs1322179. Two SNPs were of borderline significance: rs9332197 and rs10509679. rs10509679 was excluded because it was correlated with rs1322179. This resulted in five OCs (total summed PCBs, β-HCCH, HCB, mirex, and oxychlordane) crossed with three SNPs (rs743572, rs1322179 and rs9332197) to give 15 GxE comparisons.   The significance (p-values) of the interactions are shown in Table 4.20. One significant interaction was seen between rs743572 and mirex. Significance was maintained after controlling for multiple comparisons.  The effect estimates for OC exposure categories within SNP genotypes are shown for this GxE interaction in Table 4.21. The risk of NHL conferred by higher mirex levels is reversed depending on the rs743572 genotype (Genotype AA or A G: ORhighest OC exposure vs. lowest OC exposure=1.69 [1.14, 2.52]; Genotype GG: ORhighest OC exposure vs. lowest OC exposure=0.21 [0.07, 0.68]).     120  Table 4.20  Likelihood ratio test p-values for CYP SNPs and organochlorines tested for gene-environment interactions SNP Organochlorine LRT p-value rs743572 Total summed PCBs 0.750 rs743572 β-HCCH 0.618 rs743572_G HCB 0.443 rs743572_G Mirex 0.002** rs743572_G Oxychlordane 0.436 rs1322179 Total summed PCBs 0.904 rs1322179 β-HCCH 0.210 rs1322179_A HCB 0.926 rs1322179_A Mirex 0.097 rs1322179_A Oxychlordane 0.558 rs9332197 Total summed PCBs 0.112 rs9332197 β-HCCH 0.086 rs9332197_G HCB 0.168 rs9332197_G Mirex 0.903 rs9332197_G Oxychlordane 0.136 Abbreviations  LRT: Likelihood ratio test; SNP: Single nucleotide polymorphism;  Statistical significance after controlling false discovery rate is indicated by **;   121  Table 4.21  Odds ratio estimates for interaction between rs743572 (CYP17A1) and mirex in Europeans Organochlorine rs743572 Mirex Genotype = A G or A A Genotype = G G  categories  (μg/kg lipid) Controls (n) Cases (n) OR L95 U95 p Controls (n) Cases (n) OR L95 U95 p Not detected 176 116 1    16 32 1    ≥1.43 to 60.5 78 86 1.69 1.14 2.52 0.010 19 14 0.21 0.07 0.68 0.009 Abbreviations  L95: Lower limit of 95% confidence interval; OR: Odds ratio; U95: Upper limit of 95% confidence interval; N.B. p value shown is equivalent to p-trend because there are only 2 exposure categories for mirex;  Models are adjusted for age, sex, region, and possible occupational OC exposure; 122  4.4 Chapter synopsis  Significant trends were noted between NHL and lipid-adjusted levels of PCB 153, 180 and 187, as well as summed PCB measures. With respect to the pesticides/  pesticide metabolites, significant trends were found for β-HCCH, HCB, mirex, trans-nonachlor, and oxychlordane. Significance was maintained after controlling the false discovery rate for PCB-187, total summed PCBs, β-HCCH, HCB, trans-nonachlor, and oxychlordane. The final models in were consistent across NHL subtypes, except for PCB-99.  The risk of NHL was significantly increased for each additional copy of the minor allele of two CYP SNPs, rs743572 (CYP17A1) and rs1322179 (CYP2C19). Significance was not maintained after controlling for testing multiple SNPs within each gene and testing multiple genes. Two other SNPs, rs9332197 (CYP2C9) and rs10509679 (CYP2C9), were in the same gene cluster as one of these significant SNPs and were of borderline non-significance. The latter was also in high linkage disequilibrium with rs1322179. Models were consistent across NHL subtypes, except for seven SNPs, none of which showed significant overall associations.  One significant GxE interaction was found between rs743572 and mirex, even after controlling for multiple testing. The risk of NHL conferred by higher mirex levels was decreased for minor allele homozygotes of rs743572.    123  Chapter 5 Discussion   This population-based case-control study conducted in British Columbia, Canada examined the risk of non-Hodgkin lymphoma associated with plasma organochlorine levels and genetic variations in the cytochrome P450 gene superfamily, as well as whether these environmental and genetic factors are modified by each other.   In this analysis, 25 organochlorines were measured in plasma. However, PCB-52, 101, 128, aldrin, α-chlordane, and γ-chlordane had high numbers of samples below the detection limit and were excluded from further analyses. Given that in humans aldrin is rapidly metabolized to dieldrin, which was not measured, aldrin’s low detection rate is expected.[182] The biological half-life of chlordane is also relatively short.[182, 183] The metabolite, oxychlordane, and trans-nonachlor are more often detected than chlordanes in fat.[87] cis-nonachlor tends to be more readily metabolized than trans-nonachlor. As well, fewer detectable samples in the less chlorinated PCBs, is consistent with the general rule that lower chlorination decreases lipophilicity and persistence.[73, 74] The more detectable, more persistent, highly chlorinated PCBs tended to be more correlated to each other than the less detectable, less persistent, less chlorinated PCBs.  The two chlordane-related compounds with high detection rates, trans-nonachlor and oxychlordane, were highly correlated with each other and less correlated with the less detectable chlordane-related compound, cis-nonachlor. Similarly, the less detectable p, p’-DDT was not highly correlated with its highly detectable metabolite , p, p’-DDE. The lower detectability of the former is expected given its shorter half -life compared to the latter.[8]  In general, OC levels were higher in cases than controls, with the highest median PCB levels being for the more detectable, highly chlorinated PCBs and the highest median pesticide/ pesticide metabolite levels being for the more detectable, long-lived metabolite, p, p’-DDE. Median organochlorine levels also tended to be higher in those ≥60 years compared to those <60 years. Higher levels in older age groups is consistent with 124  the storage of organochlorines in tissue over time and decreasing levels in the environment in recent years.[4, 7]  Some levels varied between males and females as well. Sex differences may reflect both external differences (e.g. different behaviour) and internal differences (e.g. body physiology).[184] Those in the ≥25 kg/m2 BMI group had some significantly higher median OC levels than those in the <25 kg/m2 group. This can be attributed to the lipophilicity of many OCs and the relationship between BMI and body fat.[185] However, the relationship between OCs and BMI is complex.[186]  For a number of OCs, significantly higher median OC levels were found in those with less education compared to those with more education. Education is a surrogate for socioeconomic status and may reflect a number of differences, including diet. It is important to note that although statistically significant differences were noted above, some of these differences were small. Median OC levels did not tend to vary by region or family history of NHL.  These results and the interrelationships between OCs are comparable to those reported in previous analyses of all ethnicities.[13, 14] Some differences can be attributed to borderline statistical significance and the large number of comparisons made. As well, as noted in previous analyses, non-Europeans had significantly higher levels of PCB-183, β-HCCH, cis-nonachlor, p, p’-DDT, and p, p’-DDE, and lower levels of PCB-138, 153, 156, 170, 180, 187, HCB (borderline), and oxychlordane, compared to Europeans.[14] Differences between ethnicities may be attributed to genetic differences (e.g. metabolic differences) and/or exposure differences (e.g. dietary differences). For example, DDT is still used for anti-malaria initiatives in the developing world even though it was banned in North America in the 1970s.[72]  The levels in our controls were compared to the studies reported by Smith et al. in a systematic review of biological markers of exposure to environmental contaminants.[187] Smith et al. assessed published and grey literature from January 1990 to January 2007 on Canadian populations with measurements from any biospecimen type 125  (urine, feces, blood (serum, plasma), breast milk, fat (adipose) or other tissue, saliva, semen, nails, hair).  It is important to note that most of the studies reviewed involved populations suspected of being highly exposed (e.g. Northern communities reliant on marine mammals and fish), rather than random samples recruited from the general population. It is also difficult to compare OC levels across studies because of the different age and sex distributions of participants, the different biospecimen analyzed and the different concentration units reported. Nevertheless, it is important to situate our sample among these other Canadian samples. To facilitate comparisons, both the lipid-adjusted and non-lipid-adjusted OC levels of our cases and controls stratified by sex are reported in Table B.3 to Table B.6. Overall, the levels of a number of OCs in our study controls (PCB-118, 138, 153, β-HCCH, HCB, oxychlordane, p, p’-DDT, p, p’-DDE) were comparable to those reported in the blood (plasma/ serum, non-cord) of other urban Canadian adults. However, our controls did at times fall in the range of the “higher” exposure groups. For example, our controls had PCB-180 levels similar to those reported by Muckle et al. in pregnant Inuit women (postnatal) from Northern Quebec and Cole et al. in Great Lakes fish eaters.[188, 189] It should be noted that there is much overlap in the levels between hypothesized high exposure and low exposure groups in the literature and results varied depending on whether or not lipid-adjustment was conducted.  Because our study used current plasma organochlorine levels as a surrogate measure for past organochlorine exposure, a number of assumptions are being made. If organochlorine exposure was not consistent over time (i.e. current levels are not reflective of the relevant past levels), current levels must be interpreted with greater caution. For example, OC exposure due to occupation may not be consistent over time. Such a subset of individuals inconsistently exposed to organochlorines may complicate the comparison of their current OC levels to other individuals’.  Although diet is most likely the major source of OC exposure, this study also tested whether OC levels varied by occupation.[7, 71] No significant differences were found in 126  the median levels of any OC, regardless of OC half-life, between the occupations with possible and unlikely OC exposure. There was therefore no evidence that occupation is an important source of OC exposure in this sample.  Although, no evidence was found to support occupation as an important predictor of OC exposure in our study, small numbers limited our ability to detect an association.  In order to examine the association between having ever worked in a specific occupation and risk of NHL, DLBCL and FL, ‘t Mannetje et al. conducted a preliminary meta-analysis of data from ten international case-control studies, including this study, from the InterLymph Consortium.[168] Both occupations hypothesized a priori to be associated with NHL based on the literature (i.e. either the occupation has been associated with NHL or the occupation involves exposures (e.g. pesticides, solvents) that have been associated with NHL) and hypothesis-free occupations were assessed. Of particular interest, farmers were hypothesized to be exposed to pesticides and infectious agents. ‘t Mannetje et al. found that for animal farmers, the risk was reduced significantly for NHL and reduced non-significantly for FL and DLBCL. Risks were non-significantly increased for NHL, FL and DLBCL for crop farmers. In our analysis, the risk of NHL was approximately 1.5 to 2.5 times greater in individuals with the highest OC exposures compared to the lowest OC exposures , where trends were significant. With respect to PCB exposure, significant trends were noted for PCB-153, 180, 187, and the summed PCB measures. With respect to the pesticides/ pesticide metabolites, significant trends were found for β-HCCH, HCB, mirex, trans-nonachlor, and oxychlordane. Significance was maintained after controlling the false discovery rate for PCB-187, total summed PCBs, β-HCCH, HCB, trans-nonachlor, and oxychlordane. No significant multiplicative statistical interactions were found between OCs and potential confounders. The final models were relatively consistent across NHL subtypes. However, in PCB-99, a statistically significant trend was only detected in other B-cell subtypes, but not FL, DLBCL or all T-cell subtypes.  The main effect associations between organochlorines and NHL were similar in Europeans-only and all ethnicities.[13, 14] However, unlike in all ethnicities, significant 127  trends were not found for PCB-99, 118, 138, 156, 170, or p, p’-DDE in Europeans-only. Nevertheless, the odds ratios were relatively comparable, although they were somewhat attenuated in Europeans-only.  The associations between some of these organochlorines and NHL have been replicated in other studies. As in our analysis, Engel et al. noted significant trends for PCB-153 in the CLUE 1 cohort and the Janus cohort.[15] Engel et al. also noted significant trends with PCB-118 and 138 in these and the Nurses’ Health Study cohort, which we did not. However, as in our study, a significant trend was found for summed PCBs (28 congeners) in CLUE 1.[17] In a Danish nested case-control study, Brauner et al. did not note significant trends for PCB-153, 180 or 187, unlike in our study.[20] As in our study, significant trends were not found for PCB-99, 118, 138, 156, 170, or 183.  In an American population-based case-control study, De Roos et al. did not find significant trends for a number of PCB congeners that were significant in our study (PCB-153, 187, summed noncoplanar PCBs), as well as for a number that were also not significant in our study (PCB-99, 118, 138/ 158, 170, 183).[18] De Roos et al. also noted significant trends with PCB-156, unlike our study, and PCB-180, like our study. The De Roos et al. findings were replicated in carpet dust from the same sample for PCB-138, 153, 170, and 180.[19] Like our study, no significant trends were found for PCB-105 either. Unlike our study, Hardell et al. did not note a significantly increased risk of NHL in those with high compared to low levels of summed PCBs (36 congeners).[92] Similarly, no significant associations were found for PCBs (Aroclor 1254 and 1260) in a USEPA study.[91] With respect to pesticides/ pesticide metabolites, the CLUE 1 cohort and Brauner et al. did not find significant trends with β-HCCH, HCB, trans-nonachlor, or oxychlordane, unlike our study.[16, 20] As in our study, obvious trends were not found for p, p’-DDE or cis-nonachlor by Brauner et al. or p, p’-DDE by Engel et al.[15] Brauner et al. also noted a significant trend for p, p’-DDT, which was not replicated in our study.  128  In a USEPA study, significant trends were found for β-HCCH and oxychlordane, similar to our study, and for p, p’-DDT and p, p’-DDE, dissimilar to our study.[91] In contrast to our study, no significant associations were found for HCB or trans-nonachlor. Unlike our study, De Roos et al. did not find significant trends with β-HCCH, trans-nonachlor or oxychlordane.[18] Like our study, De Roos et al. did not note significant trends for p, p’-DDT or p, p’-DDE. This was replicated in carpet dust from the same sample for p, p’-DDT, but not p, p’-DDE.[19] No significant trends were found for α-chlordane and γ-chlordane, which were not analyzed in our study because of many undetectable values.  As in our study, Hardell et al. noted significantly increased NHL risks in those with higher compared to lower levels of HCB and trans-nonachlor (borderline).[92, 93] Similar to our study, risks were not significant for p, p’-DDE or cis-nonachlor. Unlike our study, risk was not significant for oxychlordane.   One hundred nineteen tagSNPs, belonging to 18 CYP genes, were selected for genotyping using the Illumina assay. Ninety-eight of these SNPs, belonging to 15 CYP genes, remained after QC. In addition, ten CYP1A1 SNPs were selected for genotyping using the Taqman assay. Eight of these passed QC. Statistically increased risks of NHL were noted for each additional copy of the minor allele of two CYP SNPs, namely rs743572 (CYP17A1) and rs1322179 (CYP2C19). For rs743572, the risk of NHL was 75% higher in minor allele homozygotes compared to the major allele homozygotes. For rs1322179, the risk of NHL was 4.45 times greater in the minor allele homozygotes compared to the major allele homozygotes. Significance was not maintained after controlling for multiple testing. Two SNPs, rs9332197 and rs10509679, located in CYP2C9 in the same gene cluster as CYP2C19, were of borderline non-significance. SNP rs10509679 was in high LD with rs1322179.   Trends were relatively consistent across NHL subtypes. However, statistically significant heterogeneity between subtypes was found for seven SNPs: rs1007219, rs2470152, rs743535, rs915906, rs12911554, rs4774584, and rs7172156.  Although some of these differences may be attributed to chance, subtype differences can be biologically 129  plausible. Not only do the different NHL subtypes arise in different cell types or along difference points of cell development, but they can also vary in incidence.[55]  Significant multiplicative statistical interactions were found between a few SNPs under an additive genetic model and sex, age or region. With respect to sex, statistical interactions were found for rs730154, rs2470155, CYP73UT1_2v9 (rs1048943), and rs9332242. The effects of two SNPs, rs8192712 and rs915908, were significantly different between participants less than 60 years of age and participants 60 years of age or older. Statistically significant differences between regions were found for rs2472304, rs762551 and rs679320. Differences in effect by age and sex may be attributed to differences in metabolism, environmental exposures or random chance. Differences by region are likely due to chance, as exposures and demographics are likely similar between regions.  The top four most significant SNPs are found in CYP17A1, CYP2C19 and CYP2C9 on chromosome 10. CYP17A1 is involved in steroid biosynthesis.[109] It encodes steroid 17-alpha-hydroxylase/ 17,20-lyase, which is involved in the synthesis of mineralocorticoids (e.g. aldosterone), glucocorticoids (e.g. cortisol) and sex hormones (e.g. androgens and estrogens) by acting on progesterone and pregnenolone.[109, 190]  CYP2C9 and CYP2C19 are located in a 500 kb cluster along with CYP2C8 and CYP2C18.[104] These four enzymes have >82% amino acid identity with some substrate overlap.[191] CYP2s are known to metabolize eicosanoidslxvii, xenobiotics and drugs.[29] CYP2C19 is involved in the biotransformation of some proton pump inhibitors and tricyclic antidepressants.[32] CYP2C19 has also been shown to metabolize endogenous compounds, such as the 21-hydroxylation of progesterone and the catalysis of testosterone to androstendione in human liver microsomes.[192] CYP2C9 is an important liver enzyme that metabolizes a number of xenobiotic compounds, including BaP, the OC methoxychlor and a broad spectrum of drugs.[104, 191] CYP2C9 also metabolize progesterone and testosterone in liver microsomes.[192]                                                   lxvii Eicosanoids: Arachidonic acid derivatives that are physiologically active.[29] 130  These top SNPs are located in non-coding regions of their respective CYP genes. Although the coding exonic regions encode the actual gene product (i.e. CYP enzyme/ protein), non-coding regions can also play a number of important roles.[23, 193] Regulatory regions take activation cues from other areas of the genome or the environment and regulate the location and timing of transcription. Transcription termination is signalled by other sequences. In humans, the intronic regions that interrupt exons and are spliced out of the transcript are not included in the final gene product. Accurate transcript production and splicing require specific intronic sequences. The stability and localization of transcripts, and by consequence translation regulation, can be affected by upstream and downstream regions of the transcript that are not translated into protein, such as the 5’ and 3’ untranslated regions (5’/ 3’-UTR).  The Ensembl project (http://www.ensembl.org; Release 72 – June 2013) was used to explore any known regulatory effects of these variants (i.e. “Genes and regulation” link).[144] rs743572 was noted to be located in the 5' untranslated region of the gene encoding the CYP17A1 protein. According to the literature, variation in rs743572 is hypothesized to enhance transcription and enzymatic activity, by creating an additional promoter site 34 bp upstream of the translation initiation site.[120] Intuitively, increasing the activity of enzymes that activate chemicals to potentially more carcinogenic forms would increase the risk of cancer.  Ensembl also indicated that rs1322179 is located in an intron of the gene encoding the CYP2C19 protein. rs10509679 and rs9332197 are located in introns of the gene encoding the CYP2C9 protein. No Ensembl Regulatory features were reported for these three SNPs.   However, as all the CYP SNPs genotyped using Illumina were tagSNPs, it is likely that these SNPs are in high LD with causal SNPs rather than being the source of risk themselves. Therefore, the HapMap database (http://www.hapmap.org/; Release 27 PhaseII+III, Feb09, on NCBI B36 assembly, dbSNP b126) was searched for other  SNPs within 200 kbp that may be in high LD (r2 ≥ 0.80) with these SNPs in the CEU population.[31, 194] A number of SNPs were found to be in high LD with rs743572, 131  rs10509679 and rs1322179. Several SNPs, including rs3740397 (intron – Ensembl Release 75 February 2014), rs6162 (synonymous variant; no change in protein), rs6163 (synonymous variant), and rs3781287 (intron), in high LD with rs743572, were investigated by Skibola et al. (see below) in CYP17A1.[119, 144] One of the SNPs in high LD with rs1322179, rs4244285, was investigated by Gra et al. (see below) in CYP2C19 and appears to result in defective splicing.[111, 123] No SNPs were found to be in high LD with rs9332197. This may indicate that there are other SNPs in high LD yet to be discovered, that another type of causal genetic variant is in high LD with rs9332197, that this is a chance finding, or that rs9332197 really is the causal SNP.   Some of these genetic results have been replicated in other studies. In a British population-based case-control study by Skibola et al., rs743572 was found to be associated with significantly increased risks of NHL and DLBCL, but not FL.[120, 121] In a San Francisco Bay area population-based case-control study by Skibola et al., this variant significantly increased the risk of DLBCL, but not NHL or FL. As in our study, effects did not appear to vary substantially between the sexes. This suggests that factors shared between the sexes in the CYP17A1 pathway, such as cortisol, may be involved in NHL.   SNPs in high LD with rs743572 were also examined in the San Francisco Bay study. Significantly and borderline increased risks of DLBCL were found in homozygotes of the CYP17A1 variants 137G>A (rs6162) and 195C>A (rs6163), respectively. Associations were not significant in NHL or FL. No associations were significant for CYP17A1 -270A>C (rs3781287) in NHL, DLBCL or FL. Although our variants in CYP2C9 were not tested in a sample of Connecticut women, Zhang et al. assessed whether another variant in the gene, Ex3-52C>T (rs1799853), modified the risk of NHL conferred by hair-dye use.[117] This was hypothesized because some hair dyes contain polycyclic aromatic hydrocarbons and heterocyclic aromatic amines that CYPs may activate to more toxic compounds. No significant associations were found between the variant and NHL, DLBCL, FL, or CLL/ SLL. A significant interaction was found between CYP2C9 and hair-dye use before 1980. 132   In the same study, Li et al. tested whether rs1799853 modified the risk conferred by alcohol consumption, as alcohol may induce the activity of carcinogen-metabolizing enzymes and alcohol metabolites can be genotoxic.[116] No significant associations within categories of alcohol consumption or interactions were found in NHL, T-cell subtypes or MZL. However, the association in DLBCL was significantly different between liquor-drinkers and non-drinkers, although neither effect was significant. De Roos et al. looked at associations between NHL and rs1799853 in a four-centre U.S. population-based case-control study.[35] No significant associations or trends were found. Gra et al. assessed the risk of B-CLL and T-cell NHL conferred by variants of CYP2C9 and CYP2C19 in Russian adults.[123, 124] No statistically significant associations were found for CYP2C19 (*1, *2 (rs4244285, 681G>A)). However, a statistically significant effect was found for CYP2C9*2/*3 in B-CLL (*2 (rs1799853, 430C>T), *3 (rs1057910, 1075A>C)). This effect was also significant in males, as was CYP2C9*2. It is important to note that although CYP17A1, and to some extent CYP2C9 and CYP2C19, play a role in steroid metabolism, other CYPs with related roles were not found to have significant effects. For example, CYP19A1 is involved downstream of CYP17A1, converting testosterone and androstenedione into estradiol and estrone.[96] Although CYP1s are known to be important in the metabolism of eicosanoids and xenobiotics, estradiol is also a substrate for CYP1A1, CYP1A2 and CYP1B1.[29, 100, 195] Similarly, CYP7B1 also acts on pregnenolone in steroid hormone metabolism, albeit 7α-hydroxylation in the brain.[108] Cholesterol, which is a precursor for pregnenolone, is converted into hydroxysterols and bile acids via a CYP7A1-mediated pathway.[107] Vitamin D is derived from cholesterol and CYP24A1 is a vitamin D hydroxylase.[110, 190] CYP3s are known to be involved in the metabolism of eicosanoids, xenobiotics and drugs.[29] However, both CYP3A4 and CYP3A5 can also form the cholesterol metabolite, 4β-hydroxycholesterol.[196] CYP3A4 is involved in the metabolism of estradiol, estrone, testosterone, androstenedione, and progesterone as well.[100] CYP3A43 has shown low levels of testosterone 6β-hydroxylase activity in an E. coli expression system.[106]  133  The lack of significant associations between NHL and these other SNPs with hormonal/ steroid-related functions may indicate that hormonal modulation is irrelevant to NHL and that the significant and borderline findings in CYP17A1, CYP2C9 and CYP2C19 were false positives. Conversely, this may indicate that the effects of some CYPs are masked by redundant functions and differences in their relative importance. As well, their effects may only be expressed in the context of specific environmental exposures. It is also possible that the analyzed variants did not capture the relevant variation. It is important to note that although these pathways are related and interconnected, not all branches may be relevant to NHL. For example, the effect of CYP17A1 on NHL may be influenced more by cortisol rather than vitamin D or estrogen pathways.   In order to examine whether the risks conferred by organochlorines were modified by CYP genetic variants, a number of gene-environment multiplicative statistical interactions were tested. One significant GxE interaction was found between rs743572 in CYP17A1 and mirex, even after controlling for multiple comparisons. The risk of NHL conferred by higher mirex levels is decreased in minor allele homozygotes of rs743572. The fact that the effect of mirex is modified by variations in this CYP gene lends support to the hypothesis that the risk of NHL conferred by mirex may proceed by a CYP17A1-related pathway.  A major use of mirex is as an insecticide.[78] Although a gene-environment interaction was noted between mirex and rs743572 in this study, the Comparative Toxicogenomics Database did not note any examples of mirex binding or altering the expression or activity of CYP17A1 protein or RNA.[197] However, mirex was noted to affect other CYPs, specifically CYP2B1, CYP2B2, CYP2B10, and CYP2E1. Only CYP2E1 was included in this study, but it was not assessed for GxE interactions as no significant trends were found in the main genetic analyses. The database also noted that mirex affects androgen receptor (AR), estrogen receptor 1 (ESR1), nuclear receptor subfamily 3, group C, member 1 (glucocorticoid receptor) (NR3C1), ornithine decarboxylase 1 (ODC1), and tyrosine aminotransferase (TAT). The studies that noted these effects were largely 134  laboratory-based in vitro and in vivo studies, using both animal models and animal and human cell lines.  Although no record of mirex affecting CYP17A1 was found in the database, it is interesting that mirex was found to affect steroid hormone receptors (i.e. AR and ESR1), as CYP17A1 is involved in steroid biosynthesis.[109] CYP17A1 has also been shown to be capable of xenobiotic metabolism. For example, human CYP17A1 expressed in E. coli demonstrated N-demethylation of aminopyrine. CYP17A1 has been noted to oxidize aflatoxin B1 to a genotoxic epoxide as well. CYP17A1’s activity is most likely regulated by phosphorylation, which affects enzyme stability, and retention of intermediates in the enzyme’s active site. Other regulators are other proteins, namely NAPH-cytochrome P450 reductase and cytochrome b5, which may play a role in electron transfer.  Xenobiotics and genetic mutations may also affect CYP17A1 activity. Associations have been noted between the genetic variant rs743572 and sex hormone and cortisol levels in the blood, human physical performance and risk of breast and prostate cancer.[109] Some azole pesticides have been shown to bind to CYP17A1 with high affinity. Perhaps exposure to mirex affects CYP17A1 activity and this is one mechanism mediating mirex’s carcinogenicity. As mirex does not appear to be metabolized in the human body, it is unlikely that it is a xenobiotic substrate for CYP17A1.[85] However, in rats, mirex was found to induce cytochromes P450b and P450e (CYP2B1 and CYP2B2).[198] Alternatively, CYP17A1 and mirex may be involved independently in NHL through a common steroid-related pathway.[119, 120]  Skibola et al. suggested that lymphoma risk may involve hormonal modulation because of crosstalk between the endocrine and immune systems.[119, 120] Estrogen plays a role in antibody production and immune cell activation, growth and apoptosis in both sexes. Estrogen receptors are expressed on lymphocytes and lymphoma cell lines. Cortisol also has both immune stimulatory and inhibitory properties. As noted above, CYP17A1 is involved with sex hormone production and is associated with cortisol levels. Mirex has been noted to increase levels of corticosterone, a hormone related to corti sol, in rats.[190, 199] Mirex is also degraded to chlordecone in the environment and 135  chlordecone is found in mirex preparations.[78] Not only is chlordecone structurally related to mirex, it also displays estrogenic properties.[200] Chlordecone has been noted to induce liver cancer in female rodents by interacting with estrogen. The interaction between CYP17A1 (rs743572) and mirex lends support to the involvement of hormonal modulation in NHL risk.   136  Chapter 6 Conclusion  6.1 Summary of results   This is the largest study to date assessing the association between plasma organochlorine levels and NHL.[13] Significant trends were found between several organochlorines and risk of NHL in previous analyses of all ethnicities. Many of these were replicated here in Europeans only. This analysis also found significant trends b etween several CYP SNPs and risk of NHL. To our knowledge, this is the first study to find a significant gene-environment interaction between CYP17A1 and mirex in NHL. This suggests that the risk of NHL conferred by CYP17A1 and mirex may occur by hormonal modulation.   6.2 Strengths and limitations  This study has a number of strengths and limitations that have been described elsewhere.[13, 14] As previously stated, the parent study is a population-based case-control study. It therefore shares many of the typical strengths and weaknesses of this study design. As cases and controls were drawn from virtually complete registries of cancer diagnoses in BC residents and recipients of health care coverage in BC, respectively, case and control sampling frames were representative of the target BC population and were virtually complete.[132, 201] While selection bias was thereby minimized, participation rates varied between cases and controls, with 78.7% of contacted cases and 45.7% of contacted controls consenting to participate in the parent study. In general, participation rates tend to be higher among those with the condition of interest (i.e. cases) and those of higher socioeconomic status.[202] Although differential participation rates between cases and controls may lead to non-response bias, it is unlikely that participants knew their genotypes or plasma organochlorine levels and that this knowledge influenced their participation. Other factors such as education, which 137  influence participation rates and may be associated with organochlorine exposure, were adjusted for in the analysis.  NHL subtypes were reviewed by one of two pathologists, minimizing outcome misclassification.[13] The use of newly diagnosed cases minimized survival bias. Unlike classic case-control studies, which often suffer from recall bias and error in the measurement of exposure due to their retrospective nature, our study benefitted from the use of biospecimen to determine organochlorine levels and genetic variation. However, it must be acknowledged that organochlorine levels were measured in plasma, as it is a less invasive, less costly and less analytically laborious option than adipose tissue.[20, 203] As adipose tissue contains most of the body burden of organochlorines and better represents long-term exposure and storage, it is preferred for measuring organochlorines.[204] Nevertheless, a strong correlation between some fasting organochlorine levels in serum and adipose tissue has been previously noted.  The reliability of the organochlorine measurements was also high. As previously reported in all ethnicities, sample replicates of ~10% of all samples had high intraclass correlations (ICCslxviii; Median: 0.998) and low average coefficients of variations (CVs; Median: 0.04).[13, 14] Bias was minimized by excluding cases with factors that are known to alter plasma organochlorine levels, namely chemotherapy and rapid weight loss.[136, 137] There was an almost two-fold difference in some OC levels between included and excluded cases of all ethnicities.[14] Removal of individuals who had started chemotherapy or lost weight gives us more confidence in attributing differences between cases and controls to the exposure rather than to disease progression. However, while chemotherapy was not an issue in controls, the inability to exclude weight loss in controls may have biased the effect toward the null. It is also important to note that weight loss may alter lipid levels.[206] Cases with rapid weight loss were excluded in previous analyses of all ethnicities, but no differences in lipid levels were observed between included and excluded cases.[14] To control for OC differences solely due to lipid differences between included cases and controls, lipid levels were adjusted for in OC analyses. However, the intraindividual variability of lipid levels may be large, resulting                                                   lxviii The Fleiss et al. ICC formula was used: ICC=(BMS-WMS)/(BMS + (ko-1)WMS); BMS=mean square between samples; WMS=mean square within samples (i.e. within a QC pair); ko=2.[205]  138  in non-differential measurement error in the exposure from lipid-adjustment.[207] If the differences in lipid levels are related to changes in the rate of OC release into the bloodstream, lipid-adjustment may also be inappropriate.[18] Nevertheless, the differences in lipid levels between cases and controls of all ethnicities were small (~5%).[14] Another limitation of using current plasma organochlorine levels as a measure of exposure is that we are unable to establish the temporality of the exposure-outcome relationship. Although it is unlikely that this non-occupational sample was exposed to substantial organochlorine levels following NHL diagnosis, due to the short time interval between diagnosis and first blood collection (Median in 315 European cases: 48 days), this does not necessarily mean that current organochlorine levels are representative of levels before NHL onset during the health-relevant “time-window”.[208] Engel et al. noted that prediagnostic serum PCB levels in two cohorts were most strongly associated with NHL risk when collected 0 to 12 years or 2 to 16 years before diagnosis versus >12 or >16 years before diagnosis, respectively.[15] Therefore, blood collected closer to the diagnosis, as was the case in our study, may be the most relevant OC measure for the evaluation of NHL risk. Nevertheless, the possibility of biological changes due to disease progression rather than cancer causation by OCs must be acknowledged.  Current OC levels reflect the absorbed dose from all routes of exposure over time, the distribution of the exposure to various tissue compartments, endogenous metabolic and elimination processes, and the timing of measurement relative to exposure.[209] Measurements of current exposure levels best serve as a surrogate for past exposure during the health-relevant “time-window” when the exposures have long biological half-lives and levels are unaltered by the outcome. The stability of many organochlorines, by virtue of their long half-lives, makes current blood levels reasonable estimates of cumulative exposure. Axmon et al. also noted that although past organochlorine blood concentrations estimated using complex decay models were better proxies for past concentrations than were current concentrations, the rank order of exposure levels in sampled individuals, and therefore outcome risks, were comparable between categorized current exposure levels and modeled past estimates.  139  The use of summed PCB measures may also provide better estimates of body burden and risk than individual PCB congeners alone because of the correlation between congeners and shared sources of exposure (e.g. diet).[14] It should be noted that attributing risk to specific OCs is limited by potential confounding by other unmeasured lipophilic compounds and by the high correlation between congeners.[15] Correction for multiple testing for the main effects analyses, both organochlorine and genetics, as well as for the gene-environment analyses minimized the probability of chance findings.  A number of quality control checks were undertaken to ensure the quality of samples and SNPs for the genetic analyses. As well, no evidence of population stratification was found, although the number of markers used was relatively small.[210] In this analysis of Europeans from the parent NHL study, we had a relatively large sample. However, a limitation of interaction analyses is that large samples, approximately fourfold larger than those required to detect main effects, are required.[211] These analyses were likely underpowered.  With respect to the occupational analyses, a number of limitations must be acknowledged. Sensitive, but non-specific, definitions were used to classify occupations as having possible OC exposure. Some currently obsolete uses of organochlorines were used to hypothesize exposed occupations and occupations where either exposure would be minimal or only a fraction of those in the occupation would be exposed were sometimes included. The result is likely misclassification of OC exposed occupations.  Furthermore, this analysis was limited by the lack of complete occupational histories and the reliance on self-reported “usual” occupations. The difficulty in classifying occupations was evident in the lack of overlap between those who reported a farming-related usual occupation and those who self-reported having ever lived/ worked on a farm. Misclassification of occupation is therefore likely.  The time period (e.g. date and duration) in which this “usual” occupation was held was also unknown, making its relationship with organochlorine levels unclear. Other sources of exposure (e.g. hobbies where pesticide exposure could have occurred) and 140  whether protective equipment was worn in potentially exposed occupations were not examined either.  The misclassification noted above may have diluted differences between the possible occupational OC exposure and unlikely occupational OC exposure groups. As well, given the small numbers in individual exposed occupations, these individual occupations had to be combined together for analysis, precluding conclusions about specific occupational groups. Nevertheless, given the moderate number of potentially occupationally exposed individuals (n=67), it is unlikely that their inclusion in the study greatly influenced the final conclusions that were drawn about the OC-NHL relationship.   6.3 Implications  Although the use of many organochlorines has been restricted in Canada and other developed countries for decades, exposure to these compounds is ongoing due to their accumulation in fatty tissue and their persistence in the environment and food chain.[8, 9, 170] As well, some organochlorines, such as DDT, continue to be used in other parts of the world. Even regions where certain organochlorines have never been used have been exposed, due to long-range atmospheric transport and animal migration.  These results have the potential to inform new policies and strategies for the prevention and early detection of NHL. Intake of foods with high levels of organochlorines may need to be reconsidered. For example, limiting intake of fish from contaminated areas may be useful. These findings may motivate sustained monitoring of organochlorines and increased vigilance with newer chemicals. This analysis also provides proof of concept in evaluating the health risks of chemicals. The evaluation of new chemicals can thereby be conducted in a similar manner.  In general, gene-environment interaction analyses may guide cancer control efforts in targeting groups that are at higher risk for disease due to greater environmental exposure and/or the presence of certain genotypes. Although this study did not provide much information in this effort, the formulation of biologically plausible mechanisms by which 141  organochlorines and CYPs may increase the risk of NHL may shed light on potential new molecular targets for screening.   6.4 Future directions  Because the statistical interaction between mirex and CYP17A1 (rs743572) can be explained by a number of different biological mechanisms, future basic laboratory studies may be useful.[174] How do mirex and CYP17A1 affect each other and what is their combined effect on hormones, such as cortisol, and carcinogenesis pathways? Rothman et al. has also suggested that biologic interactions could be assessed by departures from additivity on a risk difference scale, rather than risk ratio scale as was done here. Consideration of this definition of interaction may be warranted.[174] Replication of our results by other studies is also necessary. Data pooling with the InterLymph Consortium, which has resulted in the active dissemination of research and an extensive body of publications, would be useful to gain additional power for subtype-specific analyses and to replicate findings.[212]       142  Bibliography  1. Canadian Cancer Society's Steering Committee on Cancer Statistics. Canadian cancer statistics 2012. Toronto: Canadian Cancer Society; 2012. 2. Canadian Cancer Society, National Cancer Institute of Canada. Canadian cancer statistics 2008. Toronto; 2008. 3. Hartge P, Wang SS, Bracci PM, Devesa SS, Holly EA. Non-Hodgkin lymphoma. In: Schottenfeld D, Fraumeni JF,Jr, editors. Cancer epidemiology and prevention. 3rd ed. Oxford; New York: Oxford University Press; 2006. p. 898-918. 4. Hardell L, Eriksson M. Is the decline of the increasing incidence of non-Hodgkin lymphoma in Sweden and other countries a result of cancer preventive measures?. Environ Health Perspect 2003; Nov;111(14):1704-6. 5. Bioaccumulation. Available at: http://toxics.usgs.gov/definitions/bioaccumulation.html. Accessed 06/19, 2014. 6. What are POPs? Available at: http://chm.pops.int/TheConvention/ThePOPs/tabid/673/Default.aspx. Accessed 02/17, 2014. 7. International Agency for Research on Cancer. IARC monographs on the evaluation of the carcinogenic risk of chemicals to man: some organochlorine pesticides. Lyon: International Agency for Research on Cancer; 1974. 143  8. Longnecker MP, Rogan WJ, Lucier G. The human health effects of DDT (dichlorodiphenyltrichloroethane) and PCBS (polychlorinated biphenyls) and an overview of organochlorines in public health. Annu Rev Public Health 1997;18:211-44. 9. PCBs: polychlorinated biphenyls. Available at: http://www.ainc-inac.gc.ca/. Accessed 01/02, 2011. 10. Fact sheet: sources of polychlorinated biphenyls. Available at: http://www.deq.state.or.us/lq/cu/nwr/PortlandHarbor/docs/SourcePCBs.pdf. Accessed 02/14, 2013. 11. Dreiher J, Kordysh E. Non-Hodgkin lymphoma and pesticide exposure: 25 years of research. Acta Haematol 2006;116(3):153-64. 12. Engel LS, Lan Q, Rothman N. Polychlorinated biphenyls and non-Hodgkin lymphoma. Cancer Epidemiol Biomarkers Prev 2007; Mar;16(3):373-6. 13. Spinelli JJ, Ng CH, Weber JP, Connors JM, Gascoyne RD, Lai AS, et al. Organochlorines and risk of non-Hodgkin lymphoma. Int J Cancer 2007; Dec 15;121(12):2767-75. 14. Ng CH. Plasma organochlorines, interaction between the aryl hydrocarbon receptor gene and organochlorines, and risk of Non-Hodgkin lymphoma 2007; 05. 15. Engel LS, Laden F, Andersen A, Strickland PT, Blair A, Needham LL, et al. Polychlorinated biphenyl levels in peripheral blood and non-Hodgkin's lymphoma: a report from three cohorts. Cancer Res 2007; Jun 1;67(11):5545-52. 144  16. Cantor KP, Strickland PT, Brock JW, Bush D, Helzlsouer K, Needham LL, et al. Risk of non-Hodgkin's lymphoma and prediagnostic serum organochlorines: beta-hexachlorocyclohexane, chlordane/heptachlor-related compounds, dieldrin, and hexachlorobenzene. Environ Health Perspect 2003; Feb;111(2):179-83. 17. Rothman N, Cantor KP, Blair A, Bush D, Brock JW, Helzlsouer K, et al. A nested case-control study of non-Hodgkin lymphoma and serum organochlorine residues. Lancet 1997; Jul 26;350(9073):240-4. 18. De Roos AJ, Hartge P, Lubin JH, Colt JS, Davis S, Cerhan JR, et al. Persistent organochlorine chemicals in plasma and risk of non-Hodgkin's lymphoma. Cancer Res 2005; Dec 1;65(23):11214-26. 19. Colt JS, Severson RK, Lubin J, Rothman N, Camann D, Davis S, et al. Organochlorines in carpet dust and non-Hodgkin lymphoma. Epidemiology 2005; Jul;16(4):516-25. 20. Brauner EV, Sorensen M, Gaudreau E, LeBlanc A, Eriksen KT, Tjonneland A, et al. A prospective study of organochlorines in adipose tissue and risk of non-Hodgkin lymphoma. Environ Health Perspect 2012; Jan;120(1):105-11. 21. Ottman R. Gene-environment interaction: definitions and study designs. Prev Med 1996; Nov-Dec;25(6):764-70. 22. Hunter DJ. Gene-environment interactions in human diseases. Nat Rev Genet 2005; Apr;6(4):287-98. 145  23. Griffiths AJF, Gelbart WM, Miller JH, Lewontin RC. Modern genetic analysis. New York: W. H. Freeman; 1999. 24. International HapMap project. Available at: http://hapmap.ncbi.nlm.nih.gov/.  25. Ng CH, Janoo-Gilani R, Sipahimalani P, Gallagher RP, Gascoyne RD, Connors JM, et al. Interaction between organochlorines and the AHR gene, and risk of non-Hodgkin lymphoma. Cancer Causes Control 2010; Jan;21(1):11-22. 26. Bock KW, Kohle C. Ah receptor: dioxin-mediated toxic responses as hints to deregulated physiologic functions. Biochem Pharmacol 2006; Aug 14;72(4):393-404. 27. Denison MS, Nagy SR. Activation of the aryl hydrocarbon receptor by structurally diverse exogenous and endogenous chemicals. Annu Rev Pharmacol Toxicol 2003;43:309-34. 28. GeneCards version 3. Available at: www.genecards.org.  29. Nebert DW, Dalton TP. The role of cytochrome P450 enzymes in endogenous signalling pathways and environmental carcinogenesis. Nat Rev Cancer 2006; Dec;6(12):947-60. 30. Johnson GC, Esposito L, Barratt BJ, Smith AN, Heward J, Di Genova G, et al. Haplotype tagging for the identification of common disease genes. Nat Genet 2001; Oct;29(2):233-7. 31. Thorisson GA, Smith AV, Krishnan L, Stein LD. The International HapMap project web site. Genome Res 2005; Nov;15(11):1592-3. 146  32. Tomalik-Scharte D, Lazar A, Fuhr U, Kirchheiner J. The clinical role of genetic polymorphisms in drug-metabolizing enzymes. Pharmacogenomics J 2008; Feb;8(1):4-15. 33. Alberts B, Johnson A, Lewis J, Raff M, Roberts K, Walter P. Molecular biology of the cell. 4th ed. New York: Garland Science; 2002. 34. Fried KW, Rozman KK. Persistent polyhalogenated aromatic hydrocarbons. In: Greim H, Snyder R, editors. Toxicology and risk assessment: a comprehensive introduction. John Wiley & Sons; 2008. p. 698. 35. De Roos AJ, Gold LS, Wang S, Hartge P, Cerhan JR, Cozen W, et al. Metabolic gene variants and risk of non-Hodgkin's lymphoma. Cancer Epidemiol Biomarkers Prev 2006; Sep;15(9):1647-53. 36. Stratton MR, Campbell PJ, Futreal PA. The cancer genome. Nature 2009; Apr 9;458(7239):719-24. 37. Stratton MR. Exploring the genomes of cancer cells: progress and promise. Science 2011; Mar 25;331(6024):1553-8. 38. Valko M, Rhodes CJ, Moncol J, Izakovic M, Mazur M. Free radicals, metals and antioxidants in oxidative stress-induced cancer. Chem Biol Interact 2006; Mar 10;160(1):1-40. 39. Klaunig JE, Kamendulis LM. The role of oxidative stress in carcinogenesis. Annu Rev Pharmacol Toxicol 2004;44:239-67. 147  40. Green DR, Victor B. The pantheon of the fallen: why are there so many forms of cell death?. Trends Cell Biol 2012; Nov;22(11):555-6. 41. NCI dictionary of cancer terms. Available at: http://www.cancer.gov/dictionary. 2014. 42. Carpenter DO. Polychlorinated biphenyls (PCBs): routes of exposure and effects on human health. Rev Environ Health 2006; Jan-Mar;21(1):1-23. 43. Silberhorn EM, Glauert HP, Robertson LW. Carcinogenicity of polyhalogenated biphenyls: PCBs and PBBs. Crit Rev Toxicol 1990;20(6):440-96. 44. Agency for Toxic Substances and Disease Registry. Toxicological profile for DDT, DDE, and DDD. U.S. Department of Health and Human Services; Public Health Service; 2002 09. 45. McConnachie PR, Zahalsky AC. Immune alterations in humans exposed to the termiticide technical chlordane. Arch Environ Health 1992; Jul-Aug;47(4):295-301. 46. Strachan T, Read AP. Human molecular genetics. 2nd ed. New York: Wiley-Liss; 1999. 47. Lymphomas. Available at: http://www.bccancer.bc.ca/PPI/TypesofCancer/Lymphomas.htm. Accessed 05/02, 2014. 48. What is non-Hodgkin lymphoma? Available at: http://www.cancer.ca/en/cancer-information/cancer-type/non-hodgkin-lymphoma/overview/?region=on. Accessed 05/02, 2014. 49. Caligiuri MA. Human natural killer cells. Blood 2008; Aug 1;112(3):461-9. 148  50. Janeway CA, Travers P, Walport M. Immunobiology: the immune system in health and disease. 5th ed. New York: Garland Science; 2001. 51. Swerdlow SH, Campo E, Harris NL, Jaffe ES, Pileri SA, Stein H, et al. WHO classification of tumours of haematopoietic and lymphoid tissues 2008. 52. What you need to know about non-Hodgkin lymphoma. Available at: http://www.cancer.gov/cancertopics/wyntk/non-hodgkin-lymphoma.pdf. Accessed 05/03, 2014. 53. SEER training modules. Available at: http://training.seer.cancer.gov/leukemia/abstract-code-stage/keys.html. Accessed 04/03, 2014. 54. Groves FD, Linet MS, Travis LB, Devesa SS. Cancer surveillance series: non-Hodgkin's lymphoma incidence by histologic subtype in the United States from 1978 through 1995. J Natl Cancer Inst 2000; Aug 2;92(15):1240-51. 55. Morton LM, Wang SS, Devesa SS, Hartge P, Weisenburger DD, Linet MS. Lymphoma incidence patterns by WHO subtype in the United States, 1992-2001. Blood 2006; Jan 1;107(1):265-76. 56. Sehn LH, Donaldson J, Chhanabhai M, Fitzgerald C, Gill K, Klasa R, et al. Introduction of combined CHOP plus rituximab therapy dramatically improved outcome of diffuse large B-cell lymphoma in British Columbia. J Clin Oncol 2005; Aug 1;23(22):5027-33. 149  57. Hartge P, Devesa SS. Quantification of the impact of known risk factors on time trends in non-Hodgkin's lymphoma incidence. Cancer Res 1992; Oct 1;52(19 Suppl):5566s-9s. 58. Spinelli J. Pesticides, soft tissue sarcoma and non-Hodgkin lymphoma. Acta Oncol 2008;47(3):335-6. 59. Non-Hodgkin lymphoma. Available at: http://www.cancer.org/acs/groups/cid/documents/webcontent/003126-pdf.pdf. Accessed 05/05, 2014. 60. U.S. Department Of Health And Human Services, Food and Drug Administration, Center for Devices and Radiological Health. Immunotoxicity testing guidance. 1999 05/06. 61. Skibola CF, Conde L, Foo JN, Riby J, Humphreys K, Sille FC, et al. A meta-analysis of genome-wide association studies of follicular lymphoma. BMC Genomics 2012; Oct 1;13:516,2164-13-516. 62. Smedby KE, Foo JN, Skibola CF, Darabi H, Conde L, Hjalgrim H, et al. GWAS of follicular lymphoma reveals allelic heterogeneity at 6p21.32 and suggests shared genetic susceptibility with diffuse large B-cell lymphoma. PLoS Genet 2011; Apr;7(4):e1001378. 63. Conde L, Halperin E, Akers NK, Brown KM, Smedby KE, Rothman N, et al. Genome-wide association study of follicular lymphoma identifies a risk locus at 6p21.32. Nat Genet 2010; Aug;42(8):661-4. 150  64. Kumar V, Matsuo K, Takahashi A, Hosono N, Tsunoda T, Kamatani N, et al. Common variants on 14q32 and 13q12 are associated with DLBCL susceptibility. J Hum Genet 2011; Jun;56(6):436-9. 65. Crowther-Swanepoel D, Broderick P, Di Bernardo MC, Dobbins SE, Torres M, Mansouri M, et al. Common variants at 2q37.3, 8q24.21, 15q21.3 and 16q24.1 influence chronic lymphocytic leukemia risk. Nat Genet 2010; Feb;42(2):132-6. 66. Di Bernardo MC, Crowther-Swanepoel D, Broderick P, Webb E, Sellick G, Wild R, et al. A genome-wide association study identifies six susceptibility loci for chronic lymphocytic leukemia. Nat Genet 2008; Oct;40(10):1204-10. 67. Morton LM, Schenk M, Hein DW, Davis S, Zahm SH, Cozen W, et al. Genetic variation in N-acetyltransferase 1 (NAT1) and 2 (NAT2) and risk of non-Hodgkin lymphoma. Pharmacogenet Genomics 2006; Aug;16(8):537-45. 68. Skibola CF, Bracci PM, Paynter RA, Forrest MS, Agana L, Woodage T, et al. Polymorphisms and haplotypes in the cytochrome P450 17A1, prolactin, and catechol-O-methyltransferase genes and non-Hodgkin lymphoma risk. Cancer Epidemiol Biomarkers Prev 2005;14(10):2391-401. 69. Lee CC. Environmental engineering dictionary. Government Institutes; 2005. 70. The Organochlorines programme. Available at: http://www.mfe.govt.nz/. Accessed 01/06, 2011. 151  71. International Agency for Research on Cancer. IARC monographs on the evaluation of the carcinogenic risk of chemicals to human: polychlorinated biphenyls and polybrominated biphenyls. Lyon: International Agency for Research on Cancer; 1978. 72. Government of Canada. Canada’s national implementation plan under the Stockholm convention on persistent organic pollutants. Ottawa: Environment Canada; 2006 05. Report No.: En84-39/2006E. 73. United States Environmental Protection Agency. Technical factsheet on: polychlorinated biphenyls (PCBs).  74. World Health Organization Regional Office for Europe. Chapter 5.10 polychlorinated biphenyls (PCBs). Air quality guidelines. 2nd ed. Copenhagen; 2000. 75. McLeod HA, Smith DC, Bluman N. Pesticide residues in the total diet in Canada, v: 1976 to 1978. J Food Saf 1980; 07/01;2(3):141-64. 76. Dietary intakes of contaminants & other chemicals for different age-sex groups of Canadians. Available at: http://www.hc-sc.gc.ca/fn-an/surveill/total-diet/intake-apport/index-eng.php. Accessed 12/09, 2013. 77. Agency for Toxic Substances and Disease Registry. Toxicological profile for aldrin/dieldrin. U.S. Department of Health and Human Services; Public Health Service; 2002 09. 152  78. International Agency for Research on Cancer. IARC monographs on the evaluation of the carcinogenic risks of chemicals to humans: some halogenated hydrocarbons. Lyon: International Agency for Research on Cancer : World Health Organization distributor; 1979. 79. Canadian Paediatric Society, Infectious Diseases And Immunization Committee. Head lice infestations: a clinical update. Paediatrics & Child Health 2008; October;13(8):692-6. 80. Grades of chemicals. Available at: http://www.reagents.com/products/reagents/grades.html. Accessed 06/19, 2014. 81. Agency for Toxic Substances and Disease Registry. Toxicological profile for alpha-, beta-, gamma-, and delta-hexachlorocyclohexane. U.S. Department of Health and Human Services; Public Health Service; 2005 08. 82. Update to Canada's national implementation plan under the Stockholm convention on persistent organic pollutants. Available at: http://www.ec.gc.ca/lcpe-cepa/default.asp?lang=En&n=E0F02793-1&printfullpage=true. Accessed 02/19, 2014. 83. Newsome WH, Doucet J, Davies D, Sun WF. Pesticide residues in the Canadian Market Basket Survey--1992 to 1996. Food Addit Contam 2000; Oct;17(10):847-54. 84. Agency for Toxic Substances and Disease Registry. Toxicological profile for hexachlorobenzene. U.S. Department of Health and Human Services; Public Health Service; 2002 09. 153  85. Agency for Toxic Substances and Disease Registry. Toxicological profile for mirex and chlordecone. U.S. Department of Health and Human Services; Public Health Service; 1995 08. 86. International Agency for Research on Cancer. IARC monographs on the evaluation of carcinogenic risks to humans: occupational exposures in insecticide application, and some pesticides. Lyon: International Agency for Research on Cancer : World Health Organization distributor; 1991. 87. Agency for Toxic Substances and Disease Registry. Toxicological profile for chlordane. U.S. Department of Health and Human Services; Public Health Service; 1994 05. 88. Agents classified by the IARC monographs, volumes 1–109. Available at: http://monographs.iarc.fr/ENG/Classification/ClassificationsAlphaOrder.pdf. Accessed 02/19, 2014. 89. International Agency for Research on Cancer. IARC monographs on the evaluation of carcinogenic risks to humans: some thyrotropic agents. Lyon: International Agency for Research on Cancer : World Health Organization distributor; 2001. 90. International Agency for Research on Cancer. IARC monographs on the evaluation of carcinogenic risks to humans. Overall evaluations of carcinogenicity: an updating of IARC monographs, volumes 1 to 42. Lyon: International Agency for Research on Cancer; 1987. 154  91. Quintana PJ, Delfino RJ, Korrick S, Ziogas A, Kutz FW, Jones EL, et al. Adipose tissue levels of organochlorine pesticides and polychlorinated biphenyls and risk of non-Hodgkin's lymphoma. Environ Health Perspect 2004; Jun;112(8):854-61. 92. Hardell E, Eriksson M, Lindstrom G, Van Bavel B, Linde A, Carlberg M, et al. Case-control study on concentrations of organohalogen compounds and titers of antibodies to Epstein-Barr virus antigens in the etiology of non-Hodgkin lymphoma. Leuk Lymphoma 2001; Aug;42(4):619-29. 93. Hardell L, Liljegren G, Lindstrom G, Vanbavel B, Broman K, Fredrikson M, et al. Increased concentrations of chlordane in adipose tissue from non-Hodgkin's lymphoma patients compared with controls without a malignant disease. Int J Oncol 1996; Dec;9(6):1139-42. 94. Product synonyms. Available at: http://www.chemicalbook.com. Accessed 04/26, 2014. 95. Androutsopoulos VP, Tsatsakis AM, Spandidos DA. Cytochrome P450 CYP1A1: wider roles in cancer progression and prevention. BMC Cancer 2009; Jun 16;9:187. 96. Czajka-Oraniec I, Simpson ER. Aromatase research and its clinical significance. Endokrynol Pol 2010; Jan-Feb;61(1):126-34. 97. Trafalis DT, Panteli ES, Grivas A, Tsigris C, Karamanakos PN. CYP2E1 and risk of chemically mediated cancers. Expert Opin Drug Metab Toxicol 2010; Mar;6(3):307-19. 98. Schwab M. Encyclopedia of cancer. Springer; 2011. 155  99. Wojnowski L, Kamdem LK. Clinical implications of CYP3A polymorphisms. Expert Opin Drug Metab Toxicol 2006; Apr;2(2):171-82. 100. Zhou SF, Wang B, Yang LP, Liu JP. Structure, function, regulation and polymorphism and the clinical significance of human cytochrome P450 1A2. Drug Metab Rev 2010; May;42(2):268-354. 101. DeVore NM, Smith BD, Wang JL, Lushington GH, Scott EE. Key residues controlling binding of diverse ligands to human cytochrome P450 2A enzymes. Drug Metab Dispos 2009; Jun;37(6):1319-27. 102. Di YM, Chow VD, Yang LP, Zhou SF. Structure, function, regulation and polymorphism of human cytochrome P450 2A6. Curr Drug Metab 2009; Sep;10(7):754-80. 103. Mo SL, Liu YH, Duan W, Wei MQ, Kanwar JR, Zhou SF. Substrate specificity, regulation, and polymorphism of human cytochrome P450 2B6. Curr Drug Metab 2009; Sep;10(7):730-53. 104. Van Booven D, Marsh S, McLeod H, Carrillo MW, Sangkuhl K, Klein TE, et al. Cytochrome P450 2C9-CYP2C9. Pharmacogenet Genomics 2010; Apr;20(4):277-81. 105. Burk O, Wojnowski L. Cytochrome P450 3A and their regulation. Naunyn Schmiedebergs Arch Pharmacol 2004; Jan;369(1):105-24. 106. Daly AK. Significance of the minor cytochrome P450 3A isoforms. Clin Pharmacokinet 2006;45(1):13-31. 156  107. Luoma PV. Cytochrome P450 and gene activation--from pharmacology to cholesterol elimination and regression of atherosclerosis. Eur J Clin Pharmacol 2008; Sep;64(9):841-50. 108. Stiles AR, McDonald JG, Bauman DR, Russell DW. CYP7B1: one cytochrome P450, two human genetic diseases, and multiple physiological functions. J Biol Chem 2009; Oct 16;284(42):28485-9. 109. Gilep AA, Sushko TA, Usanov SA. At the crossroads of steroid hormone biosynthesis: the role, substrate specificity and evolutionary development of CYP17. Biochim Biophys Acta 2011; Jan;1814(1):200-9. 110. Masuda S, Prosser DE, Guo YD, Kaufmann M, Jones G. Generation of a homology model for the human cytochrome P450, CYP24A1, and the testing of putative substrate binding residues by site-directed mutagenesis and enzyme activity studies. Arch Biochem Biophys 2007; Apr 15;460(2):177-91. 111. The Human cytochrome P450 (CYP) allele nomenclature database . Available at: http://www.cypalleles.ki.se/index.htm. 2014. 112. Al-Dayel F, Al-Rasheed M, Ibrahim M, Bu R, Bavi P, Abubaker J, et al. Polymorphisms of drug-metabolizing enzymes CYP1A1, GSTT and GSTP contribute to the development of diffuse large B-cell lymphoma risk in the Saudi Arabian population. Leuk Lymphoma 2008; Jan;49(1):122-9. 157  113. Bu R, Gutierrez MI, Al-Rasheed M, Belgaumi A, Bhatia K. Variable drug metabolism genes in Arab population. Pharmacogenomics J 2004;4(4):260-6. 114. Genetic variation in metabolic genes, occupational solvent exposure, and risk of non-hodgkin lymphoma. Accessed 4, 173. 115. Kilfoy BA, Zheng T, Lan Q, Han X, Qin Q, Rothman N, et al. Genetic polymorphisms in glutathione S-transferases and cytochrome P450s, tobacco smoking, and risk of non-Hodgkin lymphoma. Am J Hematol 2009; May;84(5):279-82. 116. Genetic polymorphisms in cytochrome P450s, GSTs, NATs, alcohol consumption and risk of non-Hodgkin lymphoma. Accessed 3, 85. 117. Genetic variations in xenobiotic metabolic pathway genes, personal hair dye use, and risk of non-Hodgkin lymphoma. Accessed 10, 170. 118. Holly EA, Lele C. Non-Hodgkin's lymphoma in HIV-positive and HIV-negative homosexual men in the San Francisco Bay Area: allergies, prior medication use, and sexual practices. J Acquir Immune Defic Syndr Hum Retrovirol 1997; Jul 1;15(3):211-22. 119. Skibola CF, Bracci PM, Paynter RA, Forrest MS, Agana L, Woodage T, et al. Polymorphisms and haplotypes in the cytochrome P450 17A1, prolactin, and catechol-O-methyltransferase genes and non-Hodgkin lymphoma risk. Cancer Epidemiol Biomarkers Prev 2005; Oct;14(10):2391-401. 158  120. Skibola CF, Lightfoot T, Agana L, Smith A, Rollinson S, Kao A, et al. Polymorphisms in cytochrome P450 17A1 and risk of non-Hodgkin lymphoma. Br J Haematol 2005; Jun;129(5):618-21. 121. Willett EV, Smith AG, Dovey GJ, Morgan GJ, Parker J, Roman E. Tobacco and alcohol consumption and the risk of non-Hodgkin lymphoma. Cancer Causes Control 2004; Oct;15(8):771-80. 122. Kim HN, Kim NY, Yu L, Kim YK, Lee IK, Yang DH, et al. Polymorphisms of drug-metabolizing genes and risk of non-Hodgkin lymphoma. Am J Hematol 2009; Dec;84(12):821-5. 123. Gra OA, Glotov AS, Nikitin EA, Glotov OS, Kuznetsova VE, Chudinov AV, et al. Polymorphisms in xenobiotic-metabolizing genes and the risk of chronic lymphocytic leukemia and non-Hodgkin's lymphoma in adult Russian patients. Am J Hematol 2008; Apr;83(4):279-87. 124. Glotov AS, Nasedkina TV, Ivaschenko TE, Urasov RA, Surzhikov SA, Pan'kov SV, et al. Development of a biochip for analyzing polymorphism of the biotransformation genes. Mol Biol (N Y ) 2005; 05/01;39(3):357-65. 125. Kerridge I, Lincz L, Scorgie F, Hickey D, Granter N, Spencer A. Association between xenobiotic gene polymorphisms and non-Hodgkin's lymphoma risk. Br J Haematol 2002; Aug;118(2):477-81. 159  126. Cascorbi I, Brockmoller J, Roots I. A C4887A polymorphism in exon 7 of human CYP1A1: population frequency, mutation linkages, and impact on lung cancer susceptibility. Cancer Res 1996; Nov 1;56(21):4965-9. 127. Genetic polymorphism of CYP2D6, GSTM1 and NAT2 and susceptibility to haematological neoplasias. Accessed 7, 20. 128. Gough AC, Miles JS, Spurr NK, Moss JE, Gaedigk A, Eichelbaum M, et al. Identification of the primary gene defect at the cytochrome P450 CYP2D locus. Nature 1990; Oct 25;347(6295):773-6. 129. Sarmanova J, Benesova K, Gut I, Nedelcheva-Kristensen V, Tynkova L, Soucek P. Genetic polymorphisms of biotransformation enzymes in patients with Hodgkin's and non-Hodgkin's lymphomas. Hum Mol Genet 2001;10(12):1265-73. 130. Soucek P, Sarmanova J, Kristensen VN, Apltauerova M, Gut I. Genetic polymorphisms of biotransformation enzymes in patients with Hodgkin's and non-Hodgkin's lymphomas. Int Arch Occup Environ Health 2002;75 Suppl:S86-92. 131. Population estimates. Available at: http://www.bcstats.gov.bc.ca/StatisticsBySubject/Demography/PopulationEstimates.aspx. Accessed 04/29, 2014. 132. BC cancer statistics. Available at: http://www.bccancer.bc.ca/HPI/CancerStatistics/default.htm. Accessed 09/01, 2010. 160  133. Jaffe ES, Harris NL, Stein H, Vardiman JW. World Health Organization classification of tumours: pathology and genetics of tumours of haematopoietic and lymphoid tissues 2001. 134. British Columbia Centre for Disease Control, Society STD/AIDS Control. HIV / AIDS update: year end 2000   . 2000. 135. Stats & FAQ's. Available at: http://www.transplant.bc.ca/stats_faqs_main.htm. Accessed 04/29, 2014. 136. Baris D, Kwak LW, Rothman N, Wilson W, Manns A, Tarone RE, et al. Blood levels of organochlorines before and after chemotherapy among non-Hodgkin's lymphoma patients. Cancer Epidemiol Biomarkers Prev 2000; Feb;9(2):193-7. 137. Chevrier J, Dewailly E, Ayotte P, Mauriege P, Despres JP, Tremblay A. Body weight loss increases plasma and adipose tissue concentrations of potentially toxic pollutants in obese individuals. Int J Obes Relat Metab Disord 2000; Oct;24(10):1272-8. 138. Signs and symptoms of non-Hodgkin lymphoma. Available at: http://www.cancer.ca/en/cancer-information/cancer-type/non-hodgkin-lymphoma/signs-and-symptoms/?region=on. Accessed 04/29, 2014. 139. Schuetz JM, Daley D, Graham J, Berry BR, Gallagher RP, Connors JM, et al. Genetic variation in cell death genes and risk of non-Hodgkin lymphoma. PLoS One 2012;7(2):e31560. 140. Medical services plan. Available at: http://www.health.gov.bc.ca/msp/infoben/pdf/msp-brochure.pdf. Accessed 11/06, 2012. 161  141. Eligibility and enrolment. Available at: http://www.health.gov.bc.ca.ezproxy.library.ubc.ca/msp/infoben/eligible.html. Accessed 04/29, 2014. 142. Patterson DG,Jr, Isaacs SG, Alexander LR, Turner WE, Hampton L, Bernert JT, et al. Determination of specific polychlorinated dibenzo-p-dioxins and dibenzofurans in blood and adipose tissue by isotope dilution-high-resolution mass spectrometry. IARC Sci Publ 1991;11(108):299-342. 143. Hornung RW, Reed LD. Estimation of average concentration in the presence of nondetectable values. Appl Occup Environ Hyg 1990;5(1):46-51. 144. Flicek P, Ahmed I, Amode MR, Barrell D, Beal K, Brent S, et al. Ensembl 2013. Nucleic Acids Res 2013; Jan;41(Database issue):D48-55. 145. Data mining in Ensembl with BioMart. Available at: http://uswest.ensembl.org/info/website/tutorials/mining_with_biomart.pdf. Accessed 09/18, 2013. 146. Slatkin M. Linkage disequilibrium--understanding the evolutionary past and mapping the medical future. Nat Rev Genet 2008; Jun;9(6):477-85. 147. Grady BJ, Torstenson ES, Ritchie MD. The effects of linkage disequilibrium in large scale SNP datasets for MDR. BioData Min 2011; May 5;4:11,0381-4-11. 162  148. GoldenGate genotyping assay. Available at: http://www.illumina.com/technology/goldengate_genotyping_assay.ilmn. Accessed 09/06, 2010. 149. Halder I, Shriver M, Thomas M, Fernandez JR, Frudakis T. A panel of ancestry informative markers for estimating individual biogeographical ancestry and admixture from four continents: utility and applications. Hum Mutat 2008; May;29(5):648-58. 150. Knowles JW, Assimes TL, Li J, Quertermous T, Cooke JP. Genetic susceptibility to peripheral arterial disease: a dark corner in vascular biology. Arterioscler Thromb Vasc Biol 2007; Oct;27(10):2068-78. 151. Devlin B, Roeder K. Genomic control for association studies. Biometrics 1999; Dec;55(4):997-1004. 152. Genetic epidemiology - how to quantify, localize and identify genetic influences on human traits. Available at: http://dare.ubvu.vu.nl/bitstream/handle/1871/16312/chapter_2.pdf?sequence=8. Accessed 05/01, 2014. 153. Livak KJ. Allelic discrimination using fluorogenic probes and the 5′ nuclease assay. Genet Anal : Biomol Eng 1999; 2;14(5–6):143-9. 154. Kumar P, Henikoff S, Ng PC. Predicting the effects of coding non-synonymous variants on protein function using the SIFT algorithm. Nat Protoc 2009;4(7):1073-81. 163  155. Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, Bork P, et al. A method and server for predicting damaging missense mutations. Nat Methods 2010; Apr;7(4):248-9. 156. Ramensky V, Bork P, Sunyaev S. Human non-synonymous SNPs: server and survey. Nucleic Acids Res 2002; Sep 1;30(17):3894-900. 157. Jorgensen TJ, Ruczinski I, Kessing B, Smith MW, Shugart YY, Alberg AJ. Hypothesis-driven candidate gene association studies: practical design and analytical considerations. Am J Epidemiol 2009; Oct 15;170(8):986-93. 158. Stevens EL, Heckenberg G, Roberson ED, Baugher JD, Downey TJ, Pevsner J. Inference of relationships in population data using identity-by-descent and identity-by-state. PLoS Genet 2011; Sep;7(9):e1002287. 159. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007; Sep;81(3):559-75. 160. Illumina GenCall data analysis software. Available at: http://res.illumina.com/documents/products/technotes/technote_gencall_data_analysis_software.pdf. Accessed 06/19, 2014. 161. Analyzing GoldenGate genotyping data. Available at: http://res.illumina.com/documents/products/datasheets/datasheet_gggt_dataanalysis_tn.pdf. Accessed 06/19, 2014. 164  162. Andrews CA. The Hardy-Weinberg principle. Nat Educ Knowl 2010;3(10):65. 163. McCarthy MI, Abecasis GR, Cardon LR, Goldstein DB, Little J, Ioannidis JP, et al. Genome-wide association studies for complex traits: consensus, uncertainty and challenges. Nat Rev Genet 2008; May;9(5):356-69. 164. Mead S, Poulter M, Beck J, Uphill J, Jones C, Ang CE, et al. Successful amplification of degraded DNA for use with high-throughput SNP genotyping platforms. Hum Mutat 2008; Dec;29(12):1452-8. 165. Saunders IW, Brohede J, Hannan GN. Estimating genotyping error rates from Mendelian errors in SNP array genotypes and their impact on inference. Genomics 2007; Sep;90(3):291-6. 166. CEPH families reference panel   . Available at: www.cephb.fr/en/familles_CEPH.php. Accessed 06/15, 2014. 167. Statistics Canada. Standard occupational classification, 1980. Ottawa: Statistics Canada, Standards Division; 1981. 168. ‘t Mannetje A. Interlymph occupation 2011. 169. International Labour Office. International standard classification of occupations, revised edition 1968. Geneva; 1969. 170. DDT: dichlorodiphenyltrichloroethane. Available at: http://www.ainc-inac.gc.ca/. Accessed 01/02, 2011. 165  171. United States Environmental Protection Agency. PCBs in the United States industrial use and environmental distribution. Washington, D.C.: United States Environmental Protection Agency Office of Toxic Substances; 1976 02/25. Report No.: Task I. 172. Proceedings of the regional workshop on the management of persistent organic pollutants (POPs). 03/16-19/1999; Hanoi, Vietnam. United Nations Environment Programme Chemicals. 173. Inter-organization Programme for the Sound Management of Chemicals. Guidelines for the identification of PCBs and materials containing PCBs. United Nations Environment Programme Chemicals; 1999 08/1999. 174. Rothman KJ, Greenland S, Lash TL. Modern Epidemiology. 3rd ed.Wolters Kluwer Health/Lippincott Williams & Wilkins; 2008. 175. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society.Series B (Methodological) 1995;57(1):pp. 289-300. 176. PLINK v1.07. Available at: http://pngu.mgh.harvard.edu/purcell/plink/.  177. Elston RC. Introduction and overview. Statistical methods in genetic epidemiology. Stat Methods Med Res 2000; Dec;9(6):527-41.   166  178. Morton LM, Purdue MP, Zheng T, Wang SS, Armstrong B, Zhang Y, et al. Risk of non-Hodgkin lymphoma associated with germline variation in genes that regulate the cell cycle, apoptosis, and lymphocyte development. Cancer Epidemiol Biomarkers Prev 2009; Apr;18(4):1259-70. 179. Westfall PH, Zaykin DV, Young SS. Multiple tests for genetic effects in association studies. Methods Mol Biol 2002;184:143-68. 180. Chen BE, Sakoda LC, Hsing AW, Rosenberg PS. Resampling-based multiple hypothesis testing procedures for genetic case-control association studies. Genet Epidemiol 2006; Sep;30(6):495-507. 181. Gauderman WJ, Morrison JM. QUANTO 1.1: a computer program for power and sample size calculations for genetic-epidemiology studies 2006. 182. International Programme on Chemical Safety. Aldrin and dieldrin. Geneva: World Health Organization; 1989. Report No.: Environmental health criteria 91. 183. International Programme on Chemical Safety. Chlordane: health and safety guide. Geneva: World Health Organization; 1988. Report No.: 13. 184. Salihovic S, Lampa E, Lindstrom G, Lind L, Lind PM, van Bavel B. Circulating levels of persistent organic pollutants (POPs) among elderly men and women from Sweden: results from the Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS). Environ Int 2012; Sep;44:59-67. 167  185. Calculate your body mass index. Available at: http://www.nhlbi.nih.gov/guidelines/obesity/BMI/bmicalc.htm. Accessed 06/09, 2014. 186. Wolff MS, Britton JA, Teitelbaum SL, Eng S, Deych E, Ireland K, et al. Improving organochlorine biomarker models for cancer research. Cancer Epidemiol Biomarkers Prev 2005; Sep;14(9):2224-36. 187. Smith LF, Do MT. Systematic review of human biomonitoring studies of environmental contaminants in Canada January 1990 - January 2007. National Collaborating Centre for Environmental Health; 2008 01/31. Report No.: EXS00005. 188. Muckle G, Ayotte P, Dewailly EE, Jacobson SW, Jacobson JL. Prenatal exposure of the northern Quebec Inuit infants to environmental contaminants. Environ Health Perspect 2001; Dec;109(12):1291-9. 189. Cole DC, Sheeshka J, Murkin EJ, Kearney J, Scott F, Ferron LA, et al. Dietary intakes and plasma organochlorine contaminant levels among Great Lakes fish eaters. Arch Environ Health 2002; Sep-Oct;57(5):496-509. 190. Nussey S, Whitehead S. Endocrinology: an integrated spproach. Oxford: BIOS Scientific Publishers; 2001. 191. Mo SL, Zhou ZW, Yang LP, Wei MQ, Zhou SF. New insights into the structural features and functional relevance of human cytochrome P450 2C9. Part I. Curr Drug Metab 2009; Dec;10(10):1075-126. 168  192. Yamazaki H, Shimada T. Progesterone and testosterone hydroxylation by cytochromes P450 2C19, 2C9, and 3A4 in human liver microsomes. Arch Biochem Biophys 1997; Oct 1;346(1):161-9. 193. Ward LD, Kellis M. Interpreting non-coding variation in complex disease genetics. Nat Biotechnol 2012; Nov;30(11):1095-106. 194. International HapMap Consortium. The International HapMap project. Nature 2003; Dec 18;426(6968):789-96. 195. Kristensen VN, Kure EH, Erikstein B, Harada N, Borresen-Dale A. Genetic susceptibility and environmental estrogen-like compounds. Mutat Res 2001; Oct 1;482(1-2):77-82. 196. Diczfalusy U, Nylen H, Elander P, Bertilsson L. 4beta-Hydroxycholesterol, an endogenous marker of CYP3A4/5 activity in humans. Br J Clin Pharmacol 2011; Feb;71(2):183-9. 197. Comparative toxicogenomics database. Available at: http://CTDbase.org. Accessed 08, 2013. 198. Kocarek TA, Schuetz EG, Guzelian PS. Selective induction of cytochrome P450e by kepone (chlordecone) in primary cultures of adult rat hepatocytes. Mol Pharmacol 1991; Aug;40(2):203-10. 199. Brown LD, Wilson DE, Yarbrough JD. Alterations in the hepatic glucocorticoid response to mirex treatment. Toxicol Appl Pharmacol 1988; Feb;92(2):203-13. 169  200. Faroon O, Kueberuwa S, Smith L, DeRosa C. ATSDR evaluation of health effects of chemicals. II. Mirex and chlordecone: health effects, toxicokinetics, human exposure, and environmental fate. Toxicol Ind Health 1995;11(6):1-203. 201. Krenten-Boaretto B, Buxton JA, Dore K, Fyfe M, Middleton D, McEwen S. Using provincial client registries for selection of control subjects: lessons learned. Can Commun Dis Rep 2003; Oct 15;29(20):173-9. 202. Galea S, Tracy M. Participation rates in epidemiologic studies. Ann Epidemiol 2007; Sep;17(9):643-53. 203. Ryan PB, Burke TA, Cohen Hubal EA, Cura JJ, McKone TE. Using biomarkers to inform cumulative risk assessment. Environ Health Perspect 2007; May;115(5):833-40. 204. Archibeque-Engle SL, Tessari JD, Winn DT, Keefe TJ, Nett TM, Zheng T. Comparison of organochlorine pesticide and polychlorinated biphenyl residues in human breast adipose tissue and serum. J Toxicol Environ Health 1997; Nov;52(4):285-93. 205. Fleiss JL. Reliability of measurement. John Wiley & Sons, Inc; 1999. p. 1-32. 206. Hallak MH, Nomani MZ. Body weight loss and changes in blood lipid levels in normal men on hypocaloric diets during Ramadan fasting. Am J Clin Nutr 1988; Nov;48(5):1197-210. 207. Brenner H, Heiss G. The intraindividual variability of fasting triglyceride--a challenge for further standardization. Eur Heart J 1990; Dec;11(12):1054-8. 170  208. Heederik D, Teschke K. Exposure assessment for pesticides in epidemiological studies. In: Franklin C, Worgan J, editors. Occupational and residential exposure assessment for pesticides Hoboken, NJ, USA: John Wiley & Sons, Incorporated; 2005. 209. Centers for Disease Control and Prevention. Fourth national report on human exposure to environmental chemicals. Atlanta: Department of Health and Human Services; Centers for Disease Control and Prevention; 2009. Report No.: Fourth report. 210. Barnholtz-Sloan JS, McEvoy B, Shriver MD, Rebbeck TR. Ancestry estimation and correction for population stratification in molecular epidemiologic association studies. Cancer Epidemiology Biomarkers & Prevention 2008; March 01;17(3):471-7. 211. Smith PG, Day NE. The design of case-control studies: the influence of confounding and interaction effects. Int J Epidemiol 1984; Sep;13(3):356-65. 212. International Lymphoma Epidemiology Consortium. Available at: http://epi.grants.cancer.gov/InterLymph/. Accessed 09/01, 2010. 213. Luotamo M, Jarvisalo J, Aitio A. Assessment of exposure to polychlorinated biphenyls: analysis of selected isomers in blood and adipose tissue. Environ Res 1991; Apr;54(2):121-34. 214. Brown Jr. JF, Lawton RW, Ross MR, Feingold J, Wagner RE, Hamilton SB. Persistence of PCB congeners in capacitor workers and yusho patients. Chemosphere 1989;19(1-6):829-34. 171  215. Yakushiji T, Watanabe I, Kuwabara K, Tanaka R, Kashimoto T, Kunita N, et al. Rate of decrease and half-life of polychlorinated biphenyls (PCBs) in the blood of mothers and their children occupationally exposed to PCBs. Arch Environ Contam Toxicol 1984; May;13(3):341-5. 216. Wolff MS, Fischbein A, Selikoff IJ. Changes in PCB serum concentrations among capacitor manufacturing workers. Environ Res 1992; Oct;59(1):202-16. 217. Ritter R, Scheringer M, Macleod M, Moeckel C, Jones KC, Hungerbuhler K. Intrinsic human elimination half-lives of polychlorinated biphenyls derived from the temporal evolution of cross-sectional biomonitoring data from the United kingdom. Environ Health Perspect 2011; Feb;119(2):225-31. 218. Chen PH, Luo ML, Wong CK, Chen CJ. Comparative rates of elimination of some individual polychlorinated biphenyls from the blood of PCB-poisoned patients in Taiwan. Food Chem Toxicol 1982; Aug;20(4):417-25. 219. Ogura I. Half-life of each dioxin and PCB congener in the human body. Organohalogen Compounds 2004;66:3376-80. 220. Bühler F, Schmid P, Schlatter C. Kinetics of PCB elimination in man. Chemosphere 1988;17(9):1717-26.   172  221. Ryan JJ, Levesque D, Panopio LG, Sun WF, Masuda Y, Kuroki H. Elimination of polychlorinated dibenzofurans (PCDFs) and polychlorinated biphenyls (PCBs) from human blood in the Yusho and Yu-Cheng rice oil poisonings. Arch Environ Contam Toxicol 1993; May;24(4):504-12. 222. Masuda Y. Fate of PCDF/PCB congeners and change of clinical symptoms in patients with Yusho PCB poisoning for 30 years. Chemosphere 2001; May-Jun;43(4-7):925-30. 223. Dearth MA, Hites RA. Depuration rates of chlordane compounds from rat fat. Environ Sci Technol 1991; 06/01; 2014/06;25(6):1125-8. 224. Kang JH, Chang YS. Organochlorine pesticides in human serum. In: Stoytcheva M, editor. Pesticides - strategies for pesticides analysis. InTech; 2011. 225. Nordstrom M, Hardell L, Lindstrom G, Wingfors H, Hardell K, Linde A. Concentrations of organochlorines related to titers to Epstein-Barr virus early antigen IgG as risk factors for hairy cell leukemia. Environ Health Perspect 2000; May;108(5):441-5. 226. International Programme on Chemical Safety. Mirex: health and safety guide. Geneva: World Health Organization; 1990. Report No.: 39. 227. Toxnet: toxicology data network. Available at: http://toxnet.nlm.nih.gov/. Accessed 06/18, 2014. 228. Hunter DJ, Hankinson SE, Laden F, Colditz GA, Manson JE, Willett WC, et al. Plasma organochlorine levels and the risk of breast cancer. N Engl J Med 1997; Oct 30;337(18):1253-8. 173  229. Shirai JH, Kissel JC. Uncertainty in estimated half-lives of PCBS in humans: impact on exposure assessment. Sci Total Environ 1996; Sep 9;187(3):199-210.   174  Appendices   Appendix A Search strategy for literature review of associations between CYPs and NHL  Database: "Ovid MEDLINE(R) In-Process & Other Non-Indexed Citations and Ovid MEDLINE(R) 1948 to Present"   Search name: NHL-CYP-general ADhalla 5 pm 06SEP2011  1. exp Lymphoma, Non-Hodgkin/ 2. non hodgkin: lymphoma:.tw. 3. non-hodgkin: lymphoma:.tw. 4. nonhodgkin: lymphoma:.tw. 5. nhl.tw. 6. or/1-5 7. exp Cytochrome P-450 Enzyme System/ 8. Cytochrome: P-450:.tw. 9. Cytochrome: P 450:.tw. 10. Cytochrome: P450:.tw. 11. CYP:.tw. 12. or/7-11 13. 6 and 12   Ovid search features were used to limit to English papers for practical reasons. Also, although only MEDLINE was searched, two duplicate citations were found.  In the 1st round of screening, abstracts and titles were assessed for Human studies with data on NHL and CYPs. Human studies, rather than other species or molecular studies using human cell lines, were accepted as these were most comparable to our study design. Studies were also marked as “Human” if several species, including humans, were assessed. Data on NHL susceptibility (e.g. risk), rather than other NHL-related papers (e.g. survival and treatment), were accepted. CYP genotyping data, rather than other CYP-related papers (e.g. CYP expression), were accepted. 175  All these fields were attempted to be filled. If there was no abstract, the keywords in the Ovid MEDLINE citation could be used to clarify (e.g. if NHL or CYP-related keywords were present, the full text could be obtained rather than omitting the paper altogether). If a field was unclear and none of the other fields were unacceptable (i.e. “No”), or there was no abstract and the keywords and title were not helpful, full-texts were obtained and a 2nd screening of these fields was conducted. A note was made as to whether the citations were primary or not in the 1st round, but non-primaries were not excluded until the 2nd screening. At a later time, non-primary sources can be reviewed for new, possibly relevant primary sources. These primary sources can then be obtained. This has yet to be done. Primary, English papers, in Humans, with data on NHL susceptibility and CYP genotypes were abstracted.   176        Ovid MEDLINE n=78 Unique, English citations n=71 Non-English  n=5 Duplicates  n=2 Non-Human (Animal n=8;  Molecular n=14) No NHL data (n=23) No CYP data (n=7) English, Unclear citations n=3 English, Human, NHL, and CYP data  n=16 1st round screening (abstracts, titles) English, Human, NHL, and CYP data n=16   2nd round screening (full text) No NHL data (n=2) No CYP data (n=1) 177                        English, Human, NHL, and CYP data  n=16 Non-primary (n=2) Primary, English, Human, NHL, and CYP data n=14  Data Abstraction (full-text) Unique samples n=10 178  Appendix B Supplementary organochlorine tables  Table B.1  Organochlorine half-lives Organochlorine Half-life (years) Reference PCB-28 0.05 1.4 3.0 4.8 5.5 Luotamo et al.lxix [213]  Brown et al.lxix [214] Yakushiji et al.lxx [215] Wolff et al.lxix [216] Ritter et al.lxxii [217] PCB-52 2.6 5.5 Ritter et al.lxxii [217] Wolff et al.lxix,[216] PCB-99 3.3 5.7 Brown et al.lxix [214] Wolff et al.lxix  [216] PCB-101 0.02 5.7 Luotamo et al.lxix [213] Wolff et al.lxix [216] PCB-105 0.51, 0.58 2.4 3.9 4 ∞ Chen et al.lxxi [218] Ogura et al.lxxii [219] Brown et al.lxix [214] Ritter et al.lxxii [217] Wolff et al.lxix [216] PCB-118 0.3-0.8 0.77, 0.83 1.1 1.6, 17.6 3.8 5.8 9.3 9.6 Bühler et al.lxxi [220] Chen et al.lxxi [218] Ryan et al.lxix [221] Masuda et al.lxxii [222] Ogura et al.lxxii [219] Brown et al.lxix [214] Ritter et al.lxxii [217] Wolff et al.lxix [216] PCB-128 5.2, 5.4 7.9 Chen et al.lxxi [218] Wolff et al.lxix [216]                                                     lxix Serum  lxx Not reported  lxxi Whole blood 179  Organochlorine Half-life (years) Reference PCB-138 0.88 3.4 4.5, 12.8 6-7 10.8 16.3 16.7 20, 32 Bühler et al.lxxi [220] Ryan et al.lxix [221] Masuda et al.lxxii [222] Brown et al.lxix [214] Ritter et al.lxxii [217] Yakushiji et al.lxx [215] Wolff et al.lxix [216] Chen et al.lxxi [218] PCB-153 0.93 3.9 4.2, 9.1 12.4 14.4 26, 47 27.5 ∞ Bühler et al.lxxi [220] Ryan et al.lxix [221] Masuda et al.lxxii [222] Brown et al.lxix [214] Ritter et al.lxxii [217] Chen et al.lxxi [218] Yakushiji et al.lxx [215] Wolff et al.lxix [216] PCB-156 1.62 5.3, 13.2 ∞ Ryan et al.lxix [221] Masuda et al.lxxii [222] Chen et al.lxxi [218] PCB-170 4.5 5.9, 18.4 15.5 47, 71 Ryan et al.lxix [221] Masuda et al.lxxii [222] Ritter et al.lxxii [217] Chen et al.lxxi [218] PCB-180 0.34 4.8 6.0, 16.7 9.9 11.5 ∞ Bühler et al.lxxi [220] Ryan et al.lxix [221] Masuda et al.lxxii [222] Wolff et al.lxix [216] Ritter et al.lxxii [217] Chen et al.lxxi [218] PCB-183 7.9 Wolff et al.lxix [216] PCB-187 10.5 Ritter et al.lxxii [217] Aldrin Rapidly converted to dieldrin (little aldrin found in human tissue) International Programme on Chemical Safety [182]                                                     lxxii Blood  180  Organochlorine Half-life (years) Reference Chlordane 0.06, 0.09, 0.24 Agency for Toxic Substances and Disease Registry lxix,lxxiii [87] α-Chlordane 0.02 Dearth et al.lxxiv,lxxv [223] γ-Chlordane 0.02 Dearth et al.lxxiv,lxxv [223] trans-Nonachlor 0.04 Dearth et al.lxxiv,lxxv [223] cis-Nonachlor 0.03 Dearth et al.lxxiv,lxxv [223] Oxychlordane  0.07 Dearth et al.lxxiv,lxxv [223] β-HCCH 7 Kang et al.lxxii [224] HCB Not documented in humans 0.59 Nordstrom et al. [225] Agency for Toxic Substances and Disease Registrylxxiii [84] Mirex Several months International Programme on Chemical Safety [226] p, p’-DDT 1 7 Toxnet [227] Longnecker et al. [8] p, p’-DDE >7 8 10 Longnecker et al. [8] Toxnet [227] Hunter et al.lxxvi [228] Note the substantial variability in estimates both between and within organochlorines. Among the reasons for these discrepancies are different laboratory assays, sample size, short sampling intervals, different exposure concentrations, sampling before the organochlorine has reached equilibrium, and not accounting for ongoing exposures.[229] Extreme values should be regarded with caution. Multiple values for the same study reflect different methods used in the study. Secondary sources were sometimes used. For example, values for the studies by Brown et al. [214], Bühler et al. [220], Chen et al. (re-calculated) [218], Luotamo et al. [213], Ryan et al. [221], Wolff et al. [216], and Yakushiji et al. [215] were taken directly from a review by Shirai et al. [229]. Wherever possible, values from studies using blood (whole, serum, plasma) from human adults are given.                                                      lxxiii Children  lxxiv Rats  lxxv Adipose  lxxvi Plasma  181  Table B.2  Possible organochlorine exposed occupational groups with corresponding Standard Occupational Classification codes (1980) and number of European participants with organochlorine measurements included in analyses reporting as 1st usual occupation Occupational group Possible OC exposures SOC codes n  Chemical Processing All OCs  816/817 Chemicals, Petroleum, Rubber, Plastic a 0 Farming Pesticides [aldrin, β-HCCH, HCB, mirex, chlordanes, DDT; to lesser extent PCBs] 71 Farming, Horticultural and Animal Husbandry Occupations 5 Extermination/ Pesticide Application Pesticides [aldrin, β-HCCH, HCB, mirex, chlordanes, DDT; to lesser extent PCBs] 6199 Other Service  0 Electrical/ Electronics Manufacturing/ Repair PCBs, HCB, mirex 853 Fabricating, Assembling, Installing and Repairing  E 8599 Other Product Fabricating, 7 Electrical Power PCBs, HCB, mirex 873  6 Fishing PCBs 7311 Captains and Other Officers, Fishing Vessels 7313 Net, Trap and Line Fishing Occupations 7319 Fishing, Trapping and Related Occupations, n.e.c. 1 Forestry DDT 75 Forestry and Logging Occupations 3 Wood Preservation HCB 8235 Wood treating occupations 0 Paper (Finishing) PCBs, mirex  8253 Papermaking and Finishing Occupations 0 Tile Manufacturing (Finishing) PCBs 8159 Clay, Glass and Stone Processing, Forming and R 0 Textiles (Finishing) PCBs, mirex, DDT 8275 Textile Finishing and Calendering Occupations 8278 Occupations in Labouring and Other Elemental Work: Textile  0 Investment Casting (Ceramic Step) PCBs  8155  Clay, Glass and Stone  8158 Occupations in Labouring and Other Elemental  0 Metal Processing PCBs  8130 Foremen/women: Metal Processing and Related Occupation 8131 Metal Smelting, Converting and Refining Occupations 8137 Moulding, Coremaking and Metal Casting Occupations 8141 Metal Extruding and Drawing Occupations 8143 Plating, Metal Spraying and Related Occupations 8146 Inspecting, Testing, Grading and Sampling Occupations: Me 8148 Occupations in Labouring and Other Elemental Work: Metal  8149 Metal Processing and Related Occupations, n.e.c. 0 Metal Machining PCBs  831 Metal Machining Occupations 5 Industrial Machinery Fabrication/ Assemblage/ Repair PCBs  8511 Engine and Related Equipment Fabricating and Assem 8523 Industrial, Farm, Construction and Other Mechan 8584 Industrial, Farm and Construction Machinery Mechanics and  5 Pyrotechnics HCB 8599 Other Product Fabricating, Assembling and Repairing Occup 0 Painting PCBs, mirex  8595 Painting and Decorating Occupations, n.e.c. 8785 Painters, Paperhangers and Related Occupations 3   182  Occupational group Possible OC exposures SOC codes n  Construction (Caulking and Coating) PCBs 8782 Brick and Stone Masons and Tile Setters 8783 Concrete Finishing and Related Occupations 8784 Plasterers and Related Occupations 8786 Insulating Occupations, Construction 8787 Roofing, Waterproofing and Related Occupations 8791 Pipefitting, Plumbing and Related Occupations 8795 Glaziers 8799 Other Construction Trades Occupations, n.e.c. 4 Printing Press PCBs 9511 Typesetting and Composing Occupations 9512 Printing Press Occupations 0 Cleaning PCBs 6133 Lodging Cleaners, Except Private Household 6142 Housekeepers, Servants and Related Occupations 6191 Janitors, Charworkers and Cleaners 6198 Occupations in Labouring and Other Elemental Work: Other  5 Scrap yard/ Collection All OCs 8148 Occupations in Labouring and Other Elemental Work: Metal  8149 Metal Processing and Related Occupations, n.e.c. 9318 Occupations in Labouring and Other Elemental Work: Mater 9918 Occupations in Labouring and Other Elemental Work, n.e.c.  9919 Other Occupations, n.e.c.  0 Abbreviations  SOC: Standard Occupational Classification;   183  Table B.3  Organochlorine concentrations (μg/L plasma) by sex in European controls  with organochlorine measurements included in analyses  Organochlorine Sex* Mean** SD GM Min. Max. Median PCB-28 Female 0.02 0.01 0.01 0.01 0.14 0.01 PCB-28 Male 0.01 0.03 0.01 0.01 0.37 0.01 PCB-99 Female 0.05 0.04 0.04 0.01 0.30 0.04 PCB-99 Male 0.04 0.03 0.03 0.01 0.24 0.03 PCB-105 Female 0.02 0.03 0.02 0.01 0.25 0.01 PCB-105 Male 0.02 0.01 0.01 0.01 0.10 0.01 PCB-118 Female 0.09 0.10 0.07 0.01 0.99 0.07 PCB-118 Male 0.06 0.06 0.05 0.01 0.34 0.05 PCB-138 Female 0.19 0.13 0.15 0.01 0.76 0.17 PCB-138 Male 0.16 0.11 0.13 0.01 0.80 0.14 PCB-153 Female 0.36 0.25 0.29 0.03 1.38 0.33 PCB-153 Male 0.36 0.50 0.27 0.01 6.28 0.28 PCB-156 Female 0.05 0.04 0.04 0.01 0.23 0.05 PCB-156 Male 0.05 0.07 0.04 0.01 0.97 0.04 PCB-170 Female 0.12 0.13 0.08 0.01 0.92 0.09 PCB-170 Male 0.15 0.56 0.08 0.01 7.70 0.09 PCB-180 Female 0.39 0.46 0.27 0.03 3.27 0.27 PCB-180 Male 0.55 2.35 0.27 0.01 32.3 0.28 PCB-183 Female 0.03 0.03 0.02 0.01 0.27 0.02 PCB-183 Male 0.03 0.06 0.02 0.01 0.72 0.02 PCB-187 Female 0.11 0.15 0.08 0.01 1.14 0.08 PCB-187 Male 0.14 0.53 0.07 0.01 7.12 0.07 Total summed PCBs Female 1.47 1.14 1.18 0.22 8.09 1.22 Total summed PCBs Male 1.63 4.11 1.09 0.16 56.1 1.09 Summed dioxin-like PCBs Female 0.17 0.14 0.14 0.03 1.37 0.13 Dioxin-like summed PCBs Male 0.13 0.11 0.10 0.03 1.06 0.10   184  Organochlorine Sex* Mean** SD GM Min. Max. Median Summed non-dioxin-like PCBs Female 1.30 1.06 1.03 0.19 7.79 1.07 Non-dioxin-like summed PCBs Male 1.50 4.04 0.97 0.12 55.1 0.98 β-HCCH Female 0.15 0.18 0.11 0.02 1.74 0.11 β-HCCH Male 0.13 0.28 0.08 0.02 2.80 0.08 HCB Female 0.17 0.13 0.15 0.03 1.11 0.15 HCB Male 0.11 0.06 0.10 0.01 0.41 0.10 Mirex Female 0.01 0.01 0.01 0.01 0.08 0.01 Mirex Male 0.02 0.03 0.02 0.01 0.29 0.01 trans-Nonachlor Female 0.12 0.10 0.09 0.01 0.74 0.09 trans-Nonachlor Male 0.11 0.07 0.09 0.01 0.53 0.10 cis-Nonachlor Female 0.01 0.01 0.01 0.01 0.09 0.01 cis-Nonachlor Male 0.01 0.01 0.01 0.01 0.06 0.01 Oxychlordane Female 0.09 0.06 0.07 0.01 0.33 0.07 Oxychlordane Male 0.07 0.04 0.06 0.01 0.32 0.07 p, p’-DDT Female 0.03 0.03 0.02 0.02 0.27 0.02 p, p’-DDT Male 0.03 0.06 0.02 0.02 0.69 0.02 p, p’-DDE Female 2.88 3.23 1.91 0.11 29.2 1.86 p, p’-DDE Male 2.03 1.92 1.43 0.16 11.2 1.47 Abbreviations  GM: Geometric mean; Max.: Maximum organochlorine concentration (μg/L plasma);  Min.: Minimum organochlorine concentration (μg/L plasma); SD: Standard deviation (μg/L plasma);  * Sample sizes for European controls: Female (n=164); Male (n=199);  ** N.B. Mean and median values are based on organochlorine concentrations where undetectable values are replaced by the detection limit divided by √2;  185  Table B.4  Organochlorine concentrations (μg/L plasma) by sex in European cases  with organochlorine measurements included in analyses Organochlorine Sex* Mean** SD GM Min. Max. Median PCB-28 Female 0.02 0.01 0.01 0.01 0.08 0.01 PCB-28 Male 0.01 0.01 0.01 0.01 0.11 0.01 PCB-99 Female 0.05 0.04 0.04 0.01 0.17 0.04 PCB-99 Male 0.05 0.05 0.03 0.01 0.43 0.03 PCB-105 Female 0.02 0.03 0.02 0.01 0.27 0.01 PCB-105 Male 0.02 0.02 0.01 0.01 0.23 0.01 PCB-118 Female 0.09 0.08 0.07 0.01 0.56 0.08 PCB-118 Male 0.07 0.09 0.05 0.01 0.66 0.05 PCB-138 Female 0.18 0.14 0.14 0.01 0.81 0.16 PCB-138 Male 0.17 0.19 0.13 0.01 2.02 0.13 PCB-153 Female 0.40 0.37 0.30 0.01 2.64 0.30 PCB-153 Male 0.39 0.42 0.29 0.03 4.61 0.30 PCB-156 Female 0.05 0.04 0.04 0.01 0.21 0.04 PCB-156 Male 0.05 0.05 0.04 0.01 0.40 0.04 PCB-170 Female 0.14 0.23 0.09 0.01 1.70 0.09 PCB-170 Male 0.15 0.23 0.09 0.01 1.89 0.09 PCB-180 Female 0.47 0.77 0.29 0.03 5.54 0.27 PCB-180 Male 0.51 0.95 0.30 0.02 7.80 0.28 PCB-183 Female 0.03 0.05 0.02 0.01 0.42 0.02 PCB-183 Male 0.03 0.04 0.02 0.01 0.38 0.02 PCB-187 Female 0.15 0.30 0.08 0.01 2.34 0.08 PCB-187 Male 0.14 0.27 0.08 0.01 2.18 0.07 Total summed PCBs Female 1.65 1.72 1.23 0.20 12.3 1.21 Total summed PCBs Male 1.62 1.99 1.17 0.19 14.6 1.14 Summed dioxin-like PCBs Female 0.17 0.11 0.14 0.03 0.75 0.14 Dioxin-like summed PCBs Male 0.14 0.14 0.11 0.03 0.88 0.10   186  Organochlorine Sex* Mean** SD GM Min. Max. Median Summed non-dioxin-like PCBs Female 1.48 1.69 1.08 0.17 12.1 1.05 Non-dioxin-like summed PCBs Male 1.48 1.92 1.05 0.15 14.2 1.03 β-HCCH Female 0.21 0.55 0.12 0.02 6.41 0.12 β-HCCH Male 0.13 0.17 0.08 0.02 1.18 0.08 HCB Female 0.24 0.53 0.16 0.04 5.86 0.15 HCB Male 0.13 0.14 0.11 0.03 1.57 0.10 Mirex Female 0.02 0.03 0.01 0.01 0.34 0.01 Mirex Male 0.02 0.03 0.02 0.01 0.25 0.01 trans-Nonachlor Female 0.12 0.10 0.10 0.01 0.78 0.10 trans-Nonachlor Male 0.12 0.08 0.10 0.01 0.45 0.11 cis-Nonachlor Female 0.01 0.01 0.01 0.01 0.06 0.01 cis-Nonachlor Male 0.01 0.01 0.01 0.01 0.06 0.01 Oxychlordane Female 0.09 0.06 0.08 0.01 0.47 0.08 Oxychlordane Male 0.08 0.06 0.07 0.01 0.39 0.07 p, p’-DDT Female 0.04 0.05 0.03 0.02 0.41 0.02 p, p’-DDT Male 0.03 0.02 0.02 0.02 0.16 0.02 p, p’-DDE Female 3.64 4.67 2.10 0.07 31.4 2.35 p, p’-DDE Male 2.22 2.32 1.49 0.09 17.3 1.32 Abbreviations  GM: Geometric mean; Max.: Maximum organochlorine concentration (μg/L plasma);  Min.: Minimum organochlorine concentration (μg/L plasma); SD: Standard deviation (μg/L plasma);  * Sample sizes for European cases: Female (n=153); Male (n=178);  ** N.B. Mean and median values are based on organochlorine concentrations where undetectable values are replaced by the detection limit divided by √2 ; 187  Table B.5  Lipid-adjusted organochlorine concentrations (μg/kg) by sex in European controls  with organochlorine measurements included in analyses Organochlorine Sex* Mean** SD GM Min. Max. Median PCB-28 Female 2.16 1.98 1.87 0.85 21.9 1.66 PCB-28 Male 2.22 4.15 1.77 0.53 57.5 1.60 PCB-99 Female 6.61 5.83 5.26 1.22 61.3 5.34 PCB-99 Male 6.00 4.69 4.63 0.53 30.0 4.69 PCB-105 Female 2.88 3.66 2.23 0.94 33.5 1.79 PCB-105 Male 2.28 1.93 1.93 0.53 15.8 1.66 PCB-118 Female 12.7 16.7 9.68 1.41 202 9.42 PCB-118 Male 9.11 7.95 6.60 1.28 45.3 6.49 PCB-138 Female 25.7 19.2 20.8 1.41 156 21.0 PCB-138 Male 22.9 16.3 18.1 1.48 97.5 19.2 PCB-153 Female 48.6 34.0 40.5 4.28 241 41.5 PCB-153 Male 51.6 64.4 38.7 1.61 736 40.4 PCB-156 Female 7.41 5.74 6.00 1.38 45.1 6.15 PCB-156 Male 7.23 8.61 5.72 1.11 113 5.96 PCB-170 Female 16.2 18.8 11.7 1.41 146 11.3 PCB-170 Male 21.3 67.7 12.2 1.61 902 12.2 PCB-180 Female 52.6 67.1 36.9 5.00 549 35.5 PCB-180 Male 76.0 285 38.9 1.61 3790 39.3 PCB-183 Female 3.99 4.43 3.11 0.98 45.4 2.90 PCB-183 Male 4.33 7.83 3.03 1.03 84.9 2.81 PCB-187 Female 15.2 21.5 10.5 1.41 190 10.3 PCB-187 Male 19.7 67.5 10.2 1.21 833 9.95 Total summed PCBs Female 200 165 164 31.8 1360 164 Total summed PCBs Male 229 504 157 23.6 6570 157 Summed dioxin-like PCBs Female 23.0 23.6 18.9 5.15 281 18.6 Dioxin-like summed PCBs Male 18.6 14.2 15.1 3.88 125 14.3  188  Organochlorine Sex* Mean** SD GM Min. Max. Median Summed non-dioxin-like PCBs Female 177 153 143 25.9 1310 142 Non-dioxin-like summed PCBs Male 210 495 140 18.8 6450 139 β-HCCH Female 20.4 24.9 14.8 1.98 236 15.8 β-HCCH Male 18.1 34.8 11.3 1.87 321 11.7 HCB Female 23.5 17.3 20.2 5.79 175 19.7 HCB Male 15.6 8.26 13.9 1.61 57.3 14.2 Mirex Female 2.01 1.25 1.81 0.92 11.3 1.64 Mirex Male 3.16 3.79 2.39 1.03 37.5 1.93 trans-Nonachlor Female 15.4 10.5 13.0 1.41 90.2 13.4 trans-Nonachlor Male 16.2 10.2 13.4 1.61 83.6 15.0 cis-Nonachlor Female 1.91 1.06 1.76 0.92 10.8 1.62 cis-Nonachlor Male 1.99 0.93 1.85 0.79 9.14 1.70 Oxychlordane Female 11.4 6.23 9.73 1.25 36.6 10.6 Oxychlordane Male 10.3 5.75 8.76 1.61 32.0 9.38 p, p’-DDT Female 4.43 4.52 3.40 1.42 30.7 2.69 p, p’-DDT Male 4.69 7.86 3.45 1.52 92.1 2.77 p, p’-DDE Female 383 384 264 14.9 2880 268 p, p’-DDE Male 293 276 206 23.8 1700 203 Abbreviations  GM: Geometric mean; Max.: Maximum lipid-adjusted organochlorine concentration (μg/kg); Min.: Minimum lipid-adjusted organochlorine concentration (μg/kg); SD: Standard deviation (μg/kg);  * Sample sizes for European controls: Female (n=164); Male (n=199);  ** N.B. Values are based on lipid-adjusted organochlorine concentrations where undetectable values are replaced by the detection limit divided by √2;  189  Table B.6  Lipid-adjusted organochlorine concentrations (μg/kg) by sex in European cases  with organochlorine measurements included in analyses Organochlorine Sex* Mean** SD GM Min. Max. Median PCB-28 Female 2.34 1.62 2.02 0.97 11.0 1.75 PCB-28 Male 2.07 1.58 1.86 0.82 17.9 1.76 PCB-99 Female 6.96 4.89 5.39 0.97 27.7 6.00 PCB-99 Male 6.95 7.39 5.06 0.82 53.1 5.43 PCB-105 Female 2.95 3.72 2.31 0.97 37.6 1.88 PCB-105 Male 2.66 3.23 2.12 0.82 28.1 1.87 PCB-118 Female 13.6 11.8 10.4 1.43 90.4 10.7 PCB-118 Male 11.2 14.3 7.69 0.82 130 7.65 PCB-138 Female 26.5 18.7 20.5 1.44 108 22.2 PCB-138 Male 26.6 26.9 20.2 1.96 289 21.3 PCB-153 Female 57.1 47.3 44.3 1.44 318 46.8 PCB-153 Male 60.2 62.6 45.8 6.20 659 44.3 PCB-156 Female 7.36 4.57 6.19 1.21 24.6 6.33 PCB-156 Male 8.28 6.84 6.52 0.82 56.7 6.41 PCB-170 Female 20.1 30.2 13.1 1.92 216 12.5 PCB-170 Male 22.9 35.5 14.3 0.82 276 12.7 PCB-180 Female 65.5 101 42.3 5.32 794 41.9 PCB-180 Male 80.6 145 47.2 2.72 1140 42.3 PCB-183 Female 4.68 5.59 3.40 0.97 47.8 3.20 PCB-183 Male 4.64 6.33 3.34 0.82 54.0 2.96 PCB-187 Female 20.1 37.6 11.9 1.44 297 11.2 PCB-187 Male 21.2 40.7 12.0 0.82 319 11.0 Total summed PCBs Female 233 221 181 34.1 1450 176 Total summed PCBs Male 254 302 185 33.4 2130 174 Summed dioxin-like PCBs Female 23.9 16.3 20.1 4.28 120 20.9 Dioxin-like summed PCBs Male 22.2 20.9 17.3 2.47 170 16.8  190  Organochlorine Sex* Mean** SD GM Min. Max. Median Summed non-dioxin-like PCBs Female 209 216 158 28.2 1420 153 Non-dioxin-like summed PCBs Male 232 291 166 28.1 2070 155 β-HCCH Female 30.6 73.3 17.9 2.38 839 18.0 β-HCCH Male 19.7 25.0 13.0 1.37 169 14.0 HCB Female 36.4 90.1 23.7 6.04 1050 22.5 HCB Male 20.3 22.5 16.6 5.81 277 15.8 Mirex Female 2.86 3.97 2.20 0.97 45.1 1.77 Mirex Male 3.88 6.06 2.76 0.82 60.5 2.30 trans-Nonachlor Female 17.7 14.3 14.2 1.44 106 14.7 trans-Nonachlor Male 18.9 11.1 15.7 1.76 60.0 16.4 cis-Nonachlor Female 2.09 1.14 1.91 0.97 8.33 1.77 cis-Nonachlor Male 2.33 1.27 2.11 0.82 11.3 1.96 Oxychlordane Female 13.6 8.83 11.3 1.66 58.2 11.9 Oxychlordane Male 12.9 7.69 10.7 0.82 47.4 11.1 p, p’-DDT Female 5.74 8.57 3.93 1.62 82.8 2.96 p, p’-DDT Male 4.60 3.65 3.80 1.37 29.7 3.13 p, p’-DDE Female 530 713 309 9.28 5820 354 p, p’-DDE Male 349 358 236 16.7 2480 219 Abbreviations  GM: Geometric mean; Max.: Maximum lipid-adjusted organochlorine concentration (μg/kg); Min.: Minimum lipid-adjusted organochlorine concentration (μg/kg); SD: Standard deviation (μg/kg);  * Sample sizes for European cases: Female (n=153); Male (n=178);  ** N.B. Values are based on lipid-adjusted organochlorine concentrations where undetectable values are replaced by the detection limit divided by √2;  191  Appendix C Supplementary genetics tables  Table C.1  Codominant associations between CYP SNPs and NHL with additive p-trends (dominant associations shown when minor allele homozygotes <10)  Gene SNP Genotype Controls (n) Cases (n) OR L95 U95 p-trend CYP1A1 CYP_73UTR4v10 T T 407 413 1   0.648 CYP1A1 CYP_73UTR4v10 C T or C C 78 89 1.08 0.77 1.51  CYP1A1 CYP1A1_109t3x A A 366 362 1   0.528 CYP1A1 CYP1A1_109t3x A G 104 125 1.18 0.87 1.59  CYP1A1 CYP1A1_109t3x G G 9 7 0.76 0.28 2.07  CYP1A1 CYP1A1_14t1x T T 233 241 1   0.726 CYP1A1 CYP1A1_14t1x G T 209 212 0.95 0.73 1.24  CYP1A1 CYP1A1_14t1x G G 44 45 0.95 0.60 1.51  CYP1A1 CYP1A1_83t2 T T 346 357 1   0.628 CYP1A1 CYP1A1_83t2 C T 103 130 1.18 0.87 1.60  CYP1A1 CYP1A1_83t2 C C 11 8 0.67 0.27 1.71  CYP1A1 CYP5UTRNC2v2 C C 407 411 1   0.717 CYP1A1 CYP5UTRNC2v2 C T or T T 78 88 1.07 0.77 1.51  CYP1A1 CYP73UT1_2v9 A A 454 465 1   0.479 CYP1A1 CYP73UT1_2v9 A G or G G 23 31 1.25 0.71 2.18  CYP1A2 rs2472304 A A 235 237 1   0.807 CYP1A2 rs2472304 A G 240 249 1.00 0.78 1.30  CYP1A2 rs2472304 G G 72 78 1.06 0.73 1.54  CYP1A2 rs762551 A A 275 287 1   0.888 CYP1A2 rs762551 A C 225 219 0.91 0.71 1.17  CYP1A2 rs762551 C C 47 56 1.18 0.77 1.81     192  Gene SNP Genotype Controls (n) Cases (n) OR L95 U95 p-trend CYP1B1 rs1056836 G G 163 183 1   0.974 CYP1B1 rs1056836 C G 271 253 0.85 0.64 1.12  CYP1B1 rs1056836 C C 112 127 1.04 0.74 1.45  CYP1B1 rs10916 A A 335 340 1   0.603 CYP1B1 rs10916 A C 179 186 1.03 0.80 1.33  CYP1B1 rs10916 C C 33 38 1.16 0.70 1.91  CYP1B1 rs1800440 A A 360 388 1   0.380 CYP1B1 rs1800440 A G 165 153 0.86 0.66 1.12  CYP1B1 rs1800440 G G 21 23 0.97 0.53 1.80  CYP2A13 rs16974961 C C 455 481 1   0.443 CYP2A13 rs16974961 A C or A A 89 81 0.85 0.61 1.19  CYP2B6 rs11673270 A A 318 310 1   0.270 CYP2B6 rs11673270 A C 197 216 1.10 0.86 1.42  CYP2B6 rs11673270 C C 31 38 1.27 0.77 2.11  CYP2B6 rs16974799 G G 307 304 1   0.274 CYP2B6 rs16974799 A G 210 217 1.02 0.80 1.31  CYP2B6 rs16974799 A A 30 43 1.47 0.89 2.41  CYP2B6 rs2099361 A A 216 228 1   0.644 CYP2B6 rs2099361 A C 247 255 0.94 0.73 1.22  CYP2B6 rs2099361 C C 84 80 0.93 0.65 1.34  CYP2B6 rs7260329 G G 257 294 1   0.125 CYP2B6 rs7260329 A G 236 221 0.82 0.64 1.06  CYP2B6 rs7260329 A A 54 49 0.80 0.52 1.23  CYP2B6 rs8100458 A A 254 278 1   0.379 CYP2B6 rs8100458 A G 233 229 0.89 0.70 1.15  CYP2B6 rs8100458 G G 60 57 0.88 0.59 1.32     193  Gene SNP Genotype Controls (n) Cases (n) OR L95 U95 p-trend CYP2B6 rs8192712 A A 446 463 1   0.891 CYP2B6 rs8192712 A G 97 93 0.91 0.66 1.25  CYP2B6 rs8192712 G G <5 7 1.74 0.50 6.01  CYP2C9 rs10509679 G G 398 390 1   0.051 CYP2C9 rs10509679 A G 142 154 1.10 0.84 1.44  CYP2C9 rs10509679 A A 7 19 3.24 1.33 7.90  CYP2C9 rs1934967 G G 340 323 1   0.154 CYP2C9 rs1934967 A G 177 211 1.28 0.99 1.65  CYP2C9 rs1934967 A A 30 30 1.07 0.62 1.82  CYP2C9 rs1934968 G G 434 450 1   0.885 CYP2C9 rs1934968 A G 106 104 0.95 0.70 1.28  CYP2C9 rs1934968 A A 7 9 1.18 0.43 3.22  CYP2C9 rs2153628 A A 291 283 1   0.407 CYP2C9 rs2153628 A G 149 169 1.14 0.86 1.51  CYP2C9 rs2153628 G G 25 26 1.10 0.61 1.96  CYP2C9 rs2253635 A A 214 204 1   0.662 CYP2C9 rs2253635 A G 243 270 1.17 0.90 1.52  CYP2C9 rs2253635 G G 90 88 1.02 0.71 1.45  CYP2C9 rs9325473 G G 472 495 1   0.354 CYP2C9 rs9325473 A G or A A 75 69 0.88 0.62 1.26  CYP2C9 rs9332172 A A 347 382 1   0.164 CYP2C9 rs9332172 A G 178 165 0.86 0.67 1.12  CYP2C9 rs9332172 G G 21 17 0.72 0.37 1.40  CYP2C9 rs9332197 A A 490 486 1   0.066 CYP2C9 rs9332197 A G 52 73 1.50 1.02 2.20  CYP2C9 rs9332197 G G 5 5 1.10 0.31 3.86     194  Gene SNP Genotype Controls (n) Cases (n) OR L95 U95 p-trend CYP2C9 rs9332242 G G 415 445 1   0.402 CYP2C9 rs9332242 C G 125 110 0.84 0.63 1.13  CYP2C9 rs9332242 C C 6 8 1.26 0.43 3.70  CYP2C19 rs10786172 A A 234 234 1   0.341 CYP2C19 rs10786172 A G 240 247 1.04 0.81 1.35  CYP2C19 rs10786172 G G 68 80 1.22 0.84 1.78  CYP2C19 rs11597626 C C 335 339 1   0.811 CYP2C19 rs11597626 C G 182 195 1.03 0.80 1.33  CYP2C19 rs11597626 G G 29 30 1.03 0.60 1.77  CYP2C19 rs1322179 G G 417 410 1   0.048 CYP2C19 rs1322179 A G 126 139 1.12 0.84 1.48  CYP2C19 rs1322179 A A <5 15 4.45 1.45 13.7  CYP2C19 rs1555474 C C 184 180 1   0.385 CYP2C19 rs1555474 C G 270 277 1.04 0.80 1.37  CYP2C19 rs1555474 G G 93 107 1.18 0.83 1.67  CYP2C19 rs4388808 A A 359 391 1   0.350 CYP2C19 rs4388808 A G 174 151 0.78 0.60 1.02  CYP2C19 rs4388808 G G 14 21 1.35 0.67 2.70  CYP2C19 rs4917623 G G 134 141 1   0.478 CYP2C19 rs4917623 A G 224 234 0.98 0.73 1.33  CYP2C19 rs4917623 A A 109 103 0.87 0.61 1.25  CYP2C19 rs7088784 A A 471 498 1   0.213 CYP2C19 rs7088784 A G or G G 76 66 0.83 0.58 1.19  CYP2E1 rs2070675 G G 371 376 1   0.655 CYP2E1 rs2070675 A G 160 171 1.07 0.82 1.39  CYP2E1 rs2070675 A A 16 17 1.05 0.52 2.12    195  Gene SNP Genotype Controls (n) Cases (n) OR L95 U95 p-trend CYP2E1 rs6413421 A A 494 511 1   0.775 CYP2E1 rs6413421 A G or G G 53 53 0.94 0.63 1.41  CYP2E1 rs743535 G G 458 458 1   0.196 CYP2E1 rs743535 A G or A A 89 105 1.19 0.87 1.62  CYP2E1 rs8192772 A A 468 480 1   0.750 CYP2E1 rs8192772 A G or G G 79 84 1.03 0.74 1.44  CYP2E1 rs915906 A A 385 403 1   0.745 CYP2E1 rs915906 A G 149 149 0.96 0.73 1.26  CYP2E1 rs915906 G G 10 11 0.93 0.39 2.23  CYP2E1 rs915908 G G 353 349 1   0.689 CYP2E1 rs915908 A G 102 123 1.23 0.91 1.67  CYP2E1 rs915908 A A 13 6 0.50 0.18 1.34  CYP3A4 rs2242480 G G 446 469 1   0.343 CYP3A4 rs2242480 A G 93 91 0.90 0.65 1.24  CYP3A4 rs2242480 A A 7 <5 0.58 0.16 2.01  CYP3A4 rs4646437 G G 438 458 1   0.443 CYP3A4 rs4646437 A G 100 100 0.93 0.68 1.27  CYP3A4 rs4646437 A A 9 6 0.67 0.23 1.91  CYP3A5 rs15524 A A 465 486 1   0.627 CYP3A5 rs15524 A G or G G 82 76 0.86 0.62 1.22  CYP3A5 rs4646450 G G 391 402 1   0.866 CYP3A5 rs4646450 A G 139 146 1.00 0.76 1.32  CYP3A5 rs4646450 A A 17 16 0.89 0.44 1.81  CYP3A43 rs11981167 A A 494 508 1   0.999 CYP3A43 rs11981167 A T or T T 52 55 1.04 0.70 1.56    196  Gene SNP Genotype Controls (n) Cases (n) OR L95 U95 p-trend CYP3A43 rs472660 G G 427 425 1   0.470 CYP3A43 rs472660 A G 113 133 1.17 0.88 1.56  CYP3A43 rs472660 A A 6 <5 0.64 0.18 2.32  CYP3A43 rs533486 A A 167 172 1   0.778 CYP3A43 rs533486 A G 272 288 1.03 0.79 1.36  CYP3A43 rs533486 G G 107 103 0.94 0.66 1.33  CYP3A43 rs679320 A A 479 492 1   0.847 CYP3A43 rs679320 A G or G G 68 72 1.04 0.72 1.49  CYP7A1 rs11786580 G G 341 339 1   0.845 CYP7A1 rs11786580 A G 176 200 1.13 0.87 1.45  CYP7A1 rs11786580 A A 30 25 0.82 0.47 1.43  CYP7A1 rs6997473 A A 495 507 1   0.725 CYP7A1 rs6997473 A G or G G 52 57 1.04 0.70 1.55  CYP7A1 rs8192877 A A 378 404 1   0.241 CYP7A1 rs8192877 A G 152 143 0.86 0.66 1.13  CYP7A1 rs8192877 G G 16 14 0.80 0.38 1.69  CYP7B1 rs1007219 G G 275 257 1   0.073 CYP7B1 rs1007219 A G 219 241 1.17 0.91 1.51  CYP7B1 rs1007219 A A 52 66 1.39 0.93 2.09  CYP7B1 rs10112206 C C 141 138 1   0.452 CYP7B1 rs10112206 A C 274 285 1.06 0.80 1.42  CYP7B1 rs10112206 A A 131 141 1.14 0.81 1.60  CYP7B1 rs13272719 A A 345 354 1   0.878 CYP7B1 rs13272719 A G 176 184 1.02 0.78 1.31  CYP7B1 rs13272719 G G 26 26 1.04 0.59 1.83  CYP7B1 rs13276608 G G 443 457 1   0.790 CYP7B1 rs13276608 A G or A A 103 106 1.01 0.75 1.37    197  Gene SNP Genotype Controls (n) Cases (n) OR L95 U95 p-trend CYP7B1 rs2884074 C C 319 337 1   0.570 CYP7B1 rs2884074 A C 194 196 0.96 0.74 1.23  CYP7B1 rs2884074 A A 34 30 0.87 0.52 1.46  CYP7B1 rs3779874 A A 294 293 1   0.425 CYP7B1 rs3779874 A G 214 223 1.05 0.82 1.35  CYP7B1 rs3779874 G G 39 48 1.22 0.77 1.92  CYP7B1 rs4236934 A A 228 218 1   0.190 CYP7B1 rs4236934 A G 246 255 1.09 0.84 1.41  CYP7B1 rs4236934 G G 73 89 1.28 0.89 1.85  CYP7B1 rs4258041 C C 248 268 1   0.450 CYP7B1 rs4258041 C G 188 179 0.89 0.68 1.16  CYP7B1 rs4258041 G G 31 30 0.92 0.54 1.57  CYP7B1 rs7002672 A A 306 284 1   0.210 CYP7B1 rs7002672 A G 203 250 1.33 1.04 1.70  CYP7B1 rs7002672 G G 37 29 0.91 0.54 1.52  CYP7B1 rs7834854 A A 333 320 1   0.288 CYP7B1 rs7834854 A G 187 219 1.22 0.95 1.57  CYP7B1 rs7834854 G G 27 25 0.98 0.55 1.73  CYP17A1 rs743572 A A 182 161 1   0.009 CYP17A1 rs743572 A G 227 229 1.16 0.88 1.55  CYP17A1 rs743572 G G 59 88 1.75 1.18 2.61  CYP19A1 rs1004984 G G 204 218 1   0.689 CYP19A1 rs1004984 A G 269 269 0.96 0.74 1.24  CYP19A1 rs1004984 A A 74 76 0.94 0.64 1.37  CYP19A1 rs10459592 C C 167 176 1   0.862 CYP19A1 rs10459592 A C 282 275 0.90 0.69 1.19  CYP19A1 rs10459592 A A 97 110 1.07 0.75 1.51    198  Gene SNP Genotype Controls (n) Cases (n) OR L95 U95 p-trend CYP19A1 rs10519295 A A 444 468 1   0.390 CYP19A1 rs10519295 A G 97 92 0.91 0.67 1.25  CYP19A1 rs10519295 G G 6 <5 0.57 0.16 2.06  CYP19A1 rs12911554 A A 155 178 1   0.875 CYP19A1 rs12911554 A G 288 257 0.76 0.58 1.01  CYP19A1 rs12911554 G G 101 127 1.09 0.77 1.53  CYP19A1 rs17523880 C C 422 408 1   0.077 CYP19A1 rs17523880 A C 116 142 1.29 0.97 1.72  CYP19A1 rs17523880 A A 9 12 1.30 0.54 3.14  CYP19A1 rs17523922 C C 453 457 1   0.595 CYP19A1 rs17523922 C G 87 99 1.13 0.82 1.56  CYP19A1 rs17523922 G G 7 6 0.85 0.28 2.57  CYP19A1 rs17647707 C C 508 537 1   0.133 CYP19A1 rs17647707 A C or A A 38 25 0.64 0.38 1.08  CYP19A1 rs1870049 A A 418 445 1   0.486 CYP19A1 rs1870049 A G 123 108 0.81 0.61 1.09  CYP19A1 rs1870049 G G 5 11 1.80 0.61 5.28  CYP19A1 rs1902584 T T 458 477 1   0.749 CYP19A1 rs1902584 A T or A A 89 87 0.92 0.66 1.27  CYP19A1 rs2414099 A A 381 403 1   0.562 CYP19A1 rs2414099 A G 155 146 0.88 0.68 1.16  CYP19A1 rs2414099 G G 11 15 1.13 0.51 2.50  CYP19A1 rs2445762 A A 284 312 1   0.351 CYP19A1 rs2445762 A G 218 209 0.88 0.68 1.13  CYP19A1 rs2445762 G G 42 42 0.89 0.56 1.42    199  Gene SNP Genotype Controls (n) Cases (n) OR L95 U95 p-trend CYP19A1 rs2470152 G G 155 156 1   0.844 CYP19A1 rs2470152 A G 261 265 0.99 0.74 1.31  CYP19A1 rs2470152 A A 131 143 1.04 0.75 1.44  CYP19A1 rs2470155 C C 430 442 1   0.635 CYP19A1 rs2470155 A C 109 113 1.03 0.76 1.39  CYP19A1 rs2470155 A A 6 8 1.44 0.49 4.25  CYP19A1 rs2470169 A A 477 504 1   0.161 CYP19A1 rs2470169 A C or C C 70 60 0.80 0.56 1.17  CYP19A1 rs2899472 C C 308 318 1   0.572 CYP19A1 rs2899472 A C 207 202 0.97 0.75 1.25  CYP19A1 rs2899472 A A 32 43 1.31 0.80 2.13  CYP19A1 rs3751591 A A 366 394 1   0.362 CYP19A1 rs3751591 A G 162 157 0.93 0.72 1.22  CYP19A1 rs3751591 G G 19 13 0.70 0.34 1.46  CYP19A1 rs3751592 A A 224 211 1   0.094 CYP19A1 rs3751592 A G 201 210 1.16 0.88 1.52  CYP19A1 rs3751592 G G 43 57 1.43 0.92 2.22  CYP19A1 rs4646 C C 290 299 1   0.683 CYP19A1 rs4646 A C 227 226 0.98 0.76 1.25  CYP19A1 rs4646 A A 30 39 1.24 0.75 2.06  CYP19A1 rs4774584 G G 167 175 1   0.654 CYP19A1 rs4774584 A G 277 274 0.96 0.73 1.26  CYP19A1 rs4774584 A A 102 114 1.10 0.78 1.56  CYP19A1 rs6493494 G G 175 182 1   0.381 CYP19A1 rs6493494 A G 283 273 0.95 0.72 1.24  CYP19A1 rs6493494 A A 88 108 1.23 0.86 1.75    200  Gene SNP Genotype Controls (n) Cases (n) OR L95 U95 p-trend CYP19A1 rs6493495 A A 417 443 1   0.362 CYP19A1 rs6493495 A G 123 110 0.83 0.62 1.11  CYP19A1 rs6493495 G G 7 10 1.17 0.43 3.13  CYP19A1 rs7172156 G G 189 189 1   0.821 CYP19A1 rs7172156 A G 204 216 1.03 0.78 1.36  CYP19A1 rs7172156 A A 75 73 0.93 0.63 1.37  CYP19A1 rs7174997 C C 385 365 1   0.075 CYP19A1 rs7174997 A C 147 181 1.29 0.99 1.68  CYP19A1 rs7174997 A A 15 18 1.21 0.60 2.44  CYP19A1 rs730154 A A 377 404 1   0.484 CYP19A1 rs730154 A G 157 149 0.91 0.69 1.19  CYP19A1 rs730154 G G 12 11 0.91 0.39 2.10  CYP19A1 rs8023263 A A 147 165 1   0.729 CYP19A1 rs8023263 A C 293 267 0.81 0.61 1.07  CYP19A1 rs8023263 C C 107 131 1.10 0.78 1.55  CYP19A1 rs9944225 C C 459 476 1   0.737 CYP19A1 rs9944225 A C or A A 87 88 0.99 0.72 1.38  CYP24A1 rs13038432 A A 397 419 1   0.379 CYP24A1 rs13038432 A G or G G 71 59 0.78 0.53 1.13  CYP24A1 rs1570669 A A 225 235 1   0.815 CYP24A1 rs1570669 A G 258 256 0.95 0.74 1.23  CYP24A1 rs1570669 G G 64 73 1.11 0.75 1.63  CYP24A1 rs2181874 G G 309 334 1   0.258 CYP24A1 rs2181874 A G 206 204 0.93 0.72 1.20  CYP24A1 rs2181874 A A 32 25 0.71 0.41 1.24    201  Gene SNP Genotype Controls (n) Cases (n) OR L95 U95 p-trend CYP24A1 rs2244719 A A 142 119 1   0.109 CYP24A1 rs2244719 A G 224 242 1.25 0.92 1.71  CYP24A1 rs2244719 G G 102 117 1.34 0.93 1.93  CYP24A1 rs2585428 G G 121 133 1   0.350 CYP24A1 rs2585428 A G 236 246 0.95 0.70 1.29  CYP24A1 rs2585428 A A 111 99 0.84 0.58 1.21  CYP24A1 rs2762934 G G 356 370 1   0.633 CYP24A1 rs2762934 A G 169 177 1.01 0.78 1.31  CYP24A1 rs2762934 A A 22 16 0.73 0.37 1.42  CYP24A1 rs3787554 G G 444 472 1   0.136 CYP24A1 rs3787554 A G or A A 101 85 0.80 0.58 1.10  CYP24A1 rs3787557 A A 402 412 1   0.818 CYP24A1 rs3787557 A G 130 139 1.07 0.81 1.42  CYP24A1 rs3787557 G G 15 13 0.89 0.41 1.91  CYP24A1 rs4809958 A A 378 389 1   0.805 CYP24A1 rs4809958 A C 157 159 0.99 0.76 1.29  CYP24A1 rs4809958 C C 12 15 1.26 0.58 2.76  CYP24A1 rs4809959 G G 161 154 1   0.504 CYP24A1 rs4809959 A G 276 286 1.08 0.82 1.43  CYP24A1 rs4809959 A A 110 124 1.12 0.80 1.58  CYP24A1 rs4809960 A A 299 305 1   0.541 CYP24A1 rs4809960 A G 212 218 1.05 0.82 1.35  CYP24A1 rs4809960 G G 34 39 1.16 0.71 1.89  CYP24A1 rs6022999 A A 319 325 1   0.653 CYP24A1 rs6022999 A G 199 209 1.05 0.82 1.35  CYP24A1 rs6022999 G G 28 29 1.09 0.63 1.88    202  Gene SNP Genotype Controls (n) Cases (n) OR L95 U95 p-trend CYP24A1 rs6068816 G G 440 454 1   0.965 CYP24A1 rs6068816 A G or A A 107 109 0.97 0.72 1.32  CYP24A1 rs6097809 A A 498 515 1   0.687 CYP24A1 rs6097809 A G or G G 49 49 0.97 0.64 1.48  CYP24A1 rs912505 A A 331 348 1   0.396 CYP24A1 rs912505 A G 190 197 0.98 0.76 1.26  CYP24A1 rs912505 G G 26 18 0.66 0.35 1.24  CYP24A1 rs927650 G G 164 151 1   0.898 CYP24A1 rs927650 A G 272 310 1.21 0.92 1.60  CYP24A1 rs927650 A A 111 102 0.98 0.69 1.40  Abbreviations  L95: Lower limit of 95% confidence interval; OR: Odds ratio; SNP: Single nucleotide polymorphism;  U95: Upper limit of 95% confidence interval;  N.B. OR for reference group, comprised of major allele homozygotes, is equal to 1 by definition ;  No SNPs are significant after controlling for multiple comparisons;    203  Table C.2  Codominant associations between CYP SNPs and NHL  Gene SNP Genotype OR L95 U95 p CYP1A1 CYP_73UTR4v10 T T 1   0.899 CYP1A1 CYP_73UTR4v10 C T 1.07 0.76 1.51 0.699 CYP1A1 CYP_73UTR4v10 C C 1.20 0.32 4.54 0.789 CYP1A1 CYP1A1_109t3x A A 1   0.477 CYP1A1 CYP1A1_109t3x A G 1.18 0.87 1.59 0.295 CYP1A1 CYP1A1_109t3x G G 0.76 0.28 2.07 0.586 CYP1A1 CYP1A1_14t1x T T 1   0.925 CYP1A1 CYP1A1_14t1x G T 0.95 0.73 1.24 0.703 CYP1A1 CYP1A1_14t1x G G 0.95 0.60 1.51 0.838 CYP1A1 CYP1A1_83t2 T T 1   0.362 CYP1A1 CYP1A1_83t2 C T 1.18 0.87 1.60 0.276 CYP1A1 CYP1A1_83t2 C C 0.67 0.27 1.71 0.406 CYP1A1 CYP5UTRNC2v2 C C 1   0.906 CYP1A1 CYP5UTRNC2v2 C T 1.08 0.76 1.53 0.663 CYP1A1 CYP5UTRNC2v2 T T 0.95 0.23 3.87 0.946 CYP1A1 CYP73UT1_2v9 A A 1   0.723 CYP1A1 CYP73UT1_2v9 A G 1.26 0.71 2.24 0.424 CYP1A1 CYP73UT1_2v9 G G 0.88 0.05 14.4 0.926 CYP1A2 rs2472304 A A 1   0.952 CYP1A2 rs2472304 A G 1.00 0.78 1.30 0.984 CYP1A2 rs2472304 G G 1.06 0.73 1.54 0.765 CYP1A2 rs762551 A A 1   0.464 CYP1A2 rs762551 A C 0.91 0.71 1.17 0.453 CYP1A2 rs762551 C C 1.18 0.77 1.81 0.449   204  Gene SNP Genotype OR L95 U95 p CYP1B1 rs1056836 G G 1   0.322 CYP1B1 rs1056836 C G 0.85 0.64 1.12 0.237 CYP1B1 rs1056836 C C 1.04 0.74 1.45 0.822 CYP1B1 rs10916 A A 1   0.843 CYP1B1 rs10916 A C 1.03 0.80 1.33 0.829 CYP1B1 rs10916 C C 1.16 0.70 1.91 0.566 CYP1B1 rs1800440 A A 1   0.521 CYP1B1 rs1800440 A G 0.86 0.66 1.12 0.255 CYP1B1 rs1800440 G G 0.97 0.53 1.80 0.932 CYP2A13 rs16974961 C C 1   0.560 CYP2A13 rs16974961 A C 0.83 0.60 1.16 0.281 CYP2A13 rs16974961 A A inf 0 Inf 0.999 CYP2B6 rs11673270 A A 1   0.540 CYP2B6 rs11673270 A C 1.10 0.86 1.42 0.446 CYP2B6 rs11673270 C C 1.27 0.77 2.11 0.347 CYP2B6 rs16974799 G G 1   0.317 CYP2B6 rs16974799 A G 1.02 0.80 1.31 0.868 CYP2B6 rs16974799 A A 1.47 0.89 2.41 0.132 CYP2B6 rs2099361 A A 1   0.882 CYP2B6 rs2099361 A C 0.94 0.73 1.22 0.653 CYP2B6 rs2099361 C C 0.93 0.65 1.34 0.710 CYP2B6 rs7260329 G G 1   0.257 CYP2B6 rs7260329 A G 0.82 0.64 1.06 0.126 CYP2B6 rs7260329 A A 0.80 0.52 1.23 0.315 CYP2B6 rs8100458 A A 1   0.639 CYP2B6 rs8100458 A G 0.89 0.70 1.15 0.384 CYP2B6 rs8100458 G G 0.88 0.59 1.32 0.543   205  Gene SNP Genotype OR L95 U95 p CYP2B6 rs8192712 A A 1   0.560 CYP2B6 rs8192712 A G 0.91 0.66 1.25 0.553 CYP2B6 rs8192712 G G 1.74 0.50 6.01 0.383 CYP2C9 rs10509679 G G 1   0.031 CYP2C9 rs10509679 A G 1.10 0.84 1.44 0.488 CYP2C9 rs10509679 A A 3.24 1.33 7.90 0.010 CYP2C9 rs1934967 G G 1   0.166 CYP2C9 rs1934967 A G 1.28 0.99 1.65 0.058 CYP2C9 rs1934967 A A 1.07 0.62 1.82 0.817 CYP2C9 rs1934968 G G 1   0.887 CYP2C9 rs1934968 A G 0.95 0.70 1.28 0.724 CYP2C9 rs1934968 A A 1.18 0.43 3.22 0.749 CYP2C9 rs2153628 A A 1   0.636 CYP2C9 rs2153628 A G 1.14 0.86 1.51 0.348 CYP2C9 rs2153628 G G 1.10 0.61 1.96 0.755 CYP2C9 rs2253635 A A 1   0.464 CYP2C9 rs2253635 A G 1.17 0.90 1.52 0.244 CYP2C9 rs2253635 G G 1.02 0.71 1.45 0.930 CYP2C9 rs9325473 G G 1   0.403 CYP2C9 rs9325473 A G 0.93 0.65 1.33 0.682 CYP2C9 rs9325473 A A 0.34 0.06 1.75 0.195 CYP2C9 rs9332172 A A 1   0.379 CYP2C9 rs9332172 A G 0.86 0.67 1.12 0.262 CYP2C9 rs9332172 G G 0.72 0.37 1.40 0.338 CYP2C9 rs9332197 A A 1   0.116 CYP2C9 rs9332197 A G 1.50 1.02 2.20 0.038 CYP2C9 rs9332197 G G 1.10 0.31 3.86 0.888   206  Gene SNP Genotype OR L95 U95 p CYP2C9 rs9332242 G G 1   0.447 CYP2C9 rs9332242 C G 0.84 0.63 1.13 0.242 CYP2C9 rs9332242 C C 1.26 0.43 3.70 0.675 CYP2C19 rs10786172 A A 1   0.575 CYP2C19 rs10786172 A G 1.04 0.81 1.35 0.751 CYP2C19 rs10786172 G G 1.22 0.84 1.78 0.294 CYP2C19 rs11597626 C C 1   0.968 CYP2C19 rs11597626 C G 1.03 0.80 1.33 0.808 CYP2C19 rs11597626 G G 1.03 0.60 1.77 0.906 CYP2C19 rs1322179 G G 1   0.028 CYP2C19 rs1322179 A G 1.12 0.84 1.48 0.443 CYP2C19 rs1322179 A A 4.45 1.45 13.7 0.009 CYP2C19 rs1555474 C C 1   0.654 CYP2C19 rs1555474 C G 1.04 0.80 1.37 0.753 CYP2C19 rs1555474 G G 1.18 0.83 1.67 0.361 CYP2C19 rs4388808 A A 1   0.111 CYP2C19 rs4388808 A G 0.78 0.60 1.02 0.070 CYP2C19 rs4388808 G G 1.35 0.67 2.70 0.403 CYP2C19 rs4917623 G G 1   0.718 CYP2C19 rs4917623 A G 0.98 0.73 1.33 0.917 CYP2C19 rs4917623 A A 0.87 0.61 1.25 0.458 CYP2C19 rs7088784 A A 1   0.329 CYP2C19 rs7088784 A G 0.87 0.60 1.25 0.449 CYP2C19 rs7088784 G G 0.33 0.06 1.74 0.192 CYP2E1 rs2070675 G G 1   0.889 CYP2E1 rs2070675 A G 1.07 0.82 1.39 0.631 CYP2E1 rs2070675 A A 1.05 0.52 2.12 0.897   207  Gene SNP Genotype OR L95 U95 p CYP2E1 rs6413421 A A 1   NA CYP2E1 rs6413421 A G NA NA NA NA CYP2E1 rs6413421 G G NA NA NA NA CYP2E1 rs743535 G G 1   0.268 CYP2E1 rs743535 A G 1.14 0.83 1.57 0.407 CYP2E1 rs743535 A A 4.78 0.55 41.7 0.157 CYP2E1 rs8192772 A A 1   0.998 CYP2E1 rs8192772 A G 1.01 0.72 1.42 0.953 CYP2E1 rs8192772 G G inf 0 Inf 0.999 CYP2E1 rs915906 A A 1   0.948 CYP2E1 rs915906 A G 0.96 0.73 1.26 0.769 CYP2E1 rs915906 G G 0.93 0.39 2.23 0.870 CYP2E1 rs915908 G G 1   0.137 CYP2E1 rs915908 A G 1.23 0.91 1.67 0.181 CYP2E1 rs915908 A A 0.50 0.18 1.34 0.168 CYP3A4 rs2242480 G G 1   0.572 CYP3A4 rs2242480 A G 0.90 0.65 1.24 0.523 CYP3A4 rs2242480 A A 0.58 0.16 2.01 0.386 CYP3A4 rs4646437 G G 1   0.687 CYP3A4 rs4646437 A G 0.93 0.68 1.27 0.636 CYP3A4 rs4646437 A A 0.67 0.23 1.91 0.453 CYP3A5 rs15524 A A 1   0.246 CYP3A5 rs15524 A G 0.82 0.58 1.16 0.260 CYP3A5 rs15524 G G 2.71 0.54 13.7 0.228 CYP3A5 rs4646450 G G 1   0.951 CYP3A5 rs4646450 A G 1.00 0.76 1.32 0.990 CYP3A5 rs4646450 A A 0.89 0.44 1.81 0.755   208  Gene SNP Genotype OR L95 U95 p CYP3A43 rs11981167 A A 1   0.927 CYP3A43 rs11981167 A T 1.08 0.72 1.63 0.698 CYP3A43 rs11981167 T T 0 0 Inf 0.999 CYP3A43 rs472660 G G 1   0.429 CYP3A43 rs472660 A G 1.17 0.88 1.56 0.281 CYP3A43 rs472660 A A 0.64 0.18 2.32 0.501 CYP3A43 rs533486 A A 1   0.840 CYP3A43 rs533486 A G 1.03 0.79 1.36 0.822 CYP3A43 rs533486 G G 0.94 0.66 1.33 0.715 CYP3A43 rs679320 A A 1   0.855 CYP3A43 rs679320 A G 1.11 0.77 1.60 0.576 CYP3A43 rs679320 G G 0 0 Inf 0.999 CYP7A1 rs11786580 G G 1   0.458 CYP7A1 rs11786580 A G 1.13 0.87 1.45 0.366 CYP7A1 rs11786580 A A 0.82 0.47 1.43 0.481 CYP7A1 rs6997473 A A 1   1.00 CYP7A1 rs6997473 A G 1.00 0.67 1.50 0.987 CYP7A1 rs6997473 G G inf 0 Inf 0.999 CYP7A1 rs8192877 A A 1   0.494 CYP7A1 rs8192877 A G 0.86 0.66 1.13 0.275 CYP7A1 rs8192877 G G 0.80 0.38 1.69 0.564 CYP7B1 rs1007219 G G 1   0.201 CYP7B1 rs1007219 A G 1.17 0.91 1.51 0.214 CYP7B1 rs1007219 A A 1.39 0.93 2.09 0.111 CYP7B1 rs10112206 C C 1   0.753 CYP7B1 rs10112206 A C 1.06 0.80 1.42 0.674 CYP7B1 rs10112206 A A 1.14 0.81 1.60 0.452   209  Gene SNP Genotype OR L95 U95 p CYP7B1 rs13272719 A A 1   0.988 CYP7B1 rs13272719 A G 1.02 0.78 1.31 0.910 CYP7B1 rs13272719 G G 1.04 0.59 1.83 0.904 CYP7B1 rs13276608 G G 1   0.584 CYP7B1 rs13276608 A G 0.99 0.73 1.34 0.932 CYP7B1 rs13276608 A A 2.47 0.44 13.7 0.303 CYP7B1 rs2884074 C C 1   0.842 CYP7B1 rs2884074 A C 0.96 0.74 1.23 0.727 CYP7B1 rs2884074 A A 0.87 0.52 1.46 0.598 CYP7B1 rs3779874 A A 1   0.686 CYP7B1 rs3779874 A G 1.05 0.82 1.35 0.715 CYP7B1 rs3779874 G G 1.22 0.77 1.92 0.394 CYP7B1 rs4236934 A A 1   0.404 CYP7B1 rs4236934 A G 1.09 0.84 1.41 0.525 CYP7B1 rs4236934 G G 1.28 0.89 1.85 0.180 CYP7B1 rs4258041 C C 1   0.677 CYP7B1 rs4258041 C G 0.89 0.68 1.16 0.382 CYP7B1 rs4258041 G G 0.92 0.54 1.57 0.760 CYP7B1 rs7002672 A A 1   0.056 CYP7B1 rs7002672 A G 1.33 1.04 1.70 0.025 CYP7B1 rs7002672 G G 0.91 0.54 1.52 0.706 CYP7B1 rs7834854 A A 1   0.283 CYP7B1 rs7834854 A G 1.22 0.95 1.57 0.121 CYP7B1 rs7834854 G G 0.98 0.55 1.73 0.944 CYP17A1 rs743572 A A 1   0.022 CYP17A1 rs743572 A G 1.16 0.88 1.55 0.296 CYP17A1 rs743572 G G 1.75 1.18 2.61 0.006   210  Gene SNP Genotype OR L95 U95 p CYP19A1 rs1004984 G G 1   0.918 CYP19A1 rs1004984 A G 0.96 0.74 1.24 0.727 CYP19A1 rs1004984 A A 0.94 0.64 1.37 0.736 CYP19A1 rs10459592 C C 1   0.543 CYP19A1 rs10459592 A C 0.90 0.69 1.19 0.461 CYP19A1 rs10459592 A A 1.07 0.75 1.51 0.715 CYP19A1 rs10519295 A A 1   0.606 CYP19A1 rs10519295 A G 0.91 0.67 1.25 0.577 CYP19A1 rs10519295 G G 0.57 0.16 2.06 0.394 CYP19A1 rs12911554 A A 1   0.040 CYP19A1 rs12911554 A G 0.76 0.58 1.01 0.056 CYP19A1 rs12911554 G G 1.09 0.77 1.53 0.622 CYP19A1 rs17523880 C C 1   0.186 CYP19A1 rs17523880 A C 1.29 0.97 1.72 0.075 CYP19A1 rs17523880 A A 1.30 0.54 3.14 0.558 CYP19A1 rs17523922 C C 1   0.705 CYP19A1 rs17523922 C G 1.13 0.82 1.56 0.440 CYP19A1 rs17523922 G G 0.85 0.28 2.57 0.777 CYP19A1 rs17647707 C C 1   0.195 CYP19A1 rs17647707 A C 0.61 0.36 1.04 0.071 CYP19A1 rs17647707 A A Inf 0 Inf 0.999 CYP19A1 rs1870049 A A 1   0.197 CYP19A1 rs1870049 A G 0.81 0.61 1.09 0.165 CYP19A1 rs1870049 G G 1.80 0.61 5.28 0.288 CYP19A1 rs1902584 T T 1   0.586 CYP19A1 rs1902584 A T 0.89 0.64 1.24 0.482 CYP19A1 rs1902584 A A 1.69 0.41 6.89 0.466   211  Gene SNP Genotype OR L95 U95 p CYP19A1 rs2414099 A A 1   0.624 CYP19A1 rs2414099 A G 0.88 0.68 1.16 0.372 CYP19A1 rs2414099 G G 1.13 0.51 2.50 0.770 CYP19A1 rs2445762 A A 1   0.575 CYP19A1 rs2445762 A G 0.88 0.68 1.13 0.307 CYP19A1 rs2445762 G G 0.89 0.56 1.42 0.631 CYP19A1 rs2470152 G G 1   0.954 CYP19A1 rs2470152 A G 0.99 0.74 1.31 0.937 CYP19A1 rs2470152 A A 1.04 0.75 1.44 0.836 CYP19A1 rs2470155 C C 1   0.793 CYP19A1 rs2470155 A C 1.03 0.76 1.39 0.839 CYP19A1 rs2470155 A A 1.44 0.49 4.25 0.508 CYP19A1 rs2470169 A A 1   0.256 CYP19A1 rs2470169 A C 0.85 0.58 1.24 0.391 CYP19A1 rs2470169 C C 0.21 0.02 1.80 0.154 CYP19A1 rs2899472 C C 1   0.499 CYP19A1 rs2899472 A C 0.97 0.75 1.25 0.807 CYP19A1 rs2899472 A A 1.31 0.80 2.13 0.281 CYP19A1 rs3751591 A A 1   0.587 CYP19A1 rs3751591 A G 0.93 0.72 1.22 0.610 CYP19A1 rs3751591 G G 0.70 0.34 1.46 0.343 CYP19A1 rs3751592 A A 1   0.241 CYP19A1 rs3751592 A G 1.16 0.88 1.52 0.299 CYP19A1 rs3751592 G G 1.43 0.92 2.22 0.114 CYP19A1 rs4646 C C 1   0.660 CYP19A1 rs4646 A C 0.98 0.76 1.25 0.853 CYP19A1 rs4646 A A 1.24 0.75 2.06 0.405   212  Gene SNP Genotype OR L95 U95 p CYP19A1 rs4774584 G G 1   0.703 CYP19A1 rs4774584 A G 0.96 0.73 1.26 0.782 CYP19A1 rs4774584 A A 1.10 0.78 1.56 0.578 CYP19A1 rs6493494 G G 1   0.309 CYP19A1 rs6493494 A G 0.95 0.72 1.24 0.685 CYP19A1 rs6493494 A A 1.23 0.86 1.75 0.262 CYP19A1 rs6493495 A A 1   0.430 CYP19A1 rs6493495 A G 0.83 0.62 1.11 0.215 CYP19A1 rs6493495 G G 1.17 0.43 3.13 0.761 CYP19A1 rs7172156 G G 1   0.877 CYP19A1 rs7172156 A G 1.03 0.78 1.36 0.839 CYP19A1 rs7172156 A A 0.93 0.63 1.37 0.722 CYP19A1 rs7174997 C C 1   0.152 CYP19A1 rs7174997 A C 1.29 0.99 1.68 0.055 CYP19A1 rs7174997 A A 1.21 0.60 2.44 0.605 CYP19A1 rs730154 A A 1   0.768 CYP19A1 rs730154 A G 0.91 0.69 1.19 0.480 CYP19A1 rs730154 G G 0.91 0.39 2.10 0.818 CYP19A1 rs8023263 A A 1   0.090 CYP19A1 rs8023263 A C 0.81 0.61 1.07 0.130 CYP19A1 rs8023263 C C 1.10 0.78 1.55 0.578 CYP19A1 rs9944225 C C 1   0.283 CYP19A1 rs9944225 A C 0.94 0.67 1.31 0.694 CYP19A1 rs9944225 A A 3.44 0.70 16.8 0.127 CYP24A1 rs13038432 A A 1   0.250 CYP24A1 rs13038432 A G 0.72 0.49 1.06 0.096 CYP24A1 rs13038432 G G Inf 0 Inf 0.999   213  Gene SNP Genotype OR L95 U95 p CYP24A1 rs1570669 A A 1   0.724 CYP24A1 rs1570669 A G 0.95 0.74 1.23 0.697 CYP24A1 rs1570669 G G 1.11 0.75 1.63 0.600 CYP24A1 rs2181874 G G 1   0.453 CYP24A1 rs2181874 A G 0.93 0.72 1.20 0.570 CYP24A1 rs2181874 A A 0.71 0.41 1.24 0.228 CYP24A1 rs2244719 A A 1   0.230 CYP24A1 rs2244719 A G 1.25 0.92 1.71 0.149 CYP24A1 rs2244719 G G 1.34 0.93 1.93 0.118 CYP24A1 rs2585428 G G 1   0.618 CYP24A1 rs2585428 A G 0.95 0.70 1.29 0.751 CYP24A1 rs2585428 A A 0.84 0.58 1.21 0.341 CYP24A1 rs2762934 G G 1   0.634 CYP24A1 rs2762934 A G 1.01 0.78 1.31 0.952 CYP24A1 rs2762934 A A 0.73 0.37 1.42 0.349 CYP24A1 rs3787554 G G 1   0.313 CYP24A1 rs3787554 A G 0.82 0.59 1.13 0.220 CYP24A1 rs3787554 A A 0.50 0.12 2.07 0.341 CYP24A1 rs3787557 A A 1   0.829 CYP24A1 rs3787557 A G 1.07 0.81 1.42 0.617 CYP24A1 rs3787557 G G 0.89 0.41 1.91 0.760 CYP24A1 rs4809958 A A 1   0.839 CYP24A1 rs4809958 A C 0.99 0.76 1.29 0.953 CYP24A1 rs4809958 C C 1.26 0.58 2.76 0.562 CYP24A1 rs4809959 G G 1   0.789 CYP24A1 rs4809959 A G 1.08 0.82 1.43 0.592 CYP24A1 rs4809959 A A 1.12 0.80 1.58 0.518   214  Gene SNP Genotype OR L95 U95 p CYP24A1 rs4809960 A A 1   0.820 CYP24A1 rs4809960 A G 1.05 0.82 1.35 0.710 CYP24A1 rs4809960 G G 1.16 0.71 1.89 0.564 CYP24A1 rs6022999 A A 1   0.903 CYP24A1 rs6022999 A G 1.05 0.82 1.35 0.695 CYP24A1 rs6022999 G G 1.09 0.63 1.88 0.770 CYP24A1 rs6068816 G G 1   0.804 CYP24A1 rs6068816 A G 0.96 0.71 1.30 0.777 CYP24A1 rs6068816 A A 1.54 0.36 6.60 0.559 CYP24A1 rs6097809 A A 1   0.582 CYP24A1 rs6097809 A G 1.04 0.67 1.62 0.852 CYP24A1 rs6097809 G G 0.48 0.12 1.96 0.308 CYP24A1 rs912505 A A 1   0.432 CYP24A1 rs912505 A G 0.98 0.76 1.26 0.881 CYP24A1 rs912505 G G 0.66 0.35 1.24 0.195 CYP24A1 rs927650 G G 1   0.257 CYP24A1 rs927650 A G 1.21 0.92 1.60 0.174 CYP24A1 rs927650 A A 0.98 0.69 1.40 0.914 Abbreviations  L95: Lower limit of 95% confidence interval; NA: Not applicable (PLINK could not compute e.g. 0 minor allele homozygotes); OR: Odds ratio; SNP: Single nucleotide polymorphism; U95: Upper limit of 95% confidence interval;  N.B. OR for reference group, comprised of major allele homozygotes, is equal to 1 by definition. P -value shown for reference group is the 2 degree of freedom joint test of a model with both the hete rozygous and minor allele homozygous groups. If one or more of these groups are missing, this joint p -value is NA.    215  Table C.3  Additive associations between CYP SNPs and NHL Gene SNP Genotype OR L95 U95 p-trend CYP1A1 CYP_73UTR4v10 nC 1.08 0.79 1.47 0.648 CYP1A1 CYP1A1_109t3x nG 1.09 0.84 1.42 0.528 CYP1A1 CYP1A1_14t1x nG 0.97 0.79 1.17 0.726 CYP1A1 CYP1A1_83t2 nC 1.07 0.82 1.38 0.628 CYP1A1 CYP5UTRNC2v2 nT 1.06 0.77 1.45 0.717 CYP1A1 CYP73UT1_2v9 nG 1.21 0.71 2.06 0.479 CYP1A2 rs2472304 nG 1.02 0.86 1.21 0.807 CYP1A2 rs762551 nC 1.01 0.84 1.22 0.888 CYP1B1 rs1056836 nC 1.00 0.85 1.18 0.974 CYP1B1 rs10916 nC 1.05 0.87 1.28 0.603 CYP1B1 rs1800440 nG 0.91 0.73 1.13 0.380 CYP2A13 rs16974961 nA 0.88 0.64 1.22 0.443 CYP2B6 rs11673270 nC 1.12 0.92 1.36 0.270 CYP2B6 rs16974799 nA 1.11 0.92 1.35 0.274 CYP2B6 rs2099361 nC 0.96 0.81 1.14 0.644 CYP2B6 rs7260329 nA 0.87 0.72 1.04 0.125 CYP2B6 rs8100458 nG 0.92 0.77 1.10 0.379 CYP2B6 rs8192712 nG 0.98 0.74 1.30 0.891 CYP2C9 rs10509679 nA 1.26 1.00 1.60 0.051 CYP2C9 rs1934967 nA 1.16 0.95 1.41 0.154 CYP2C9 rs1934968 nA 0.98 0.75 1.28 0.885 CYP2C9 rs2153628 nG 1.10 0.88 1.36 0.407 CYP2C9 rs2253635 nG 1.04 0.88 1.23 0.662 CYP2C9 rs9325473 nA 0.86 0.61 1.19 0.354 CYP2C9 rs9332172 nG 0.86 0.69 1.07 0.164 CYP2C9 rs9332197 nG 1.37 0.98 1.91 0.066 CYP2C9 rs9332242 nC 0.89 0.69 1.16 0.402 CYP2C19 rs10786172 nG 1.09 0.91 1.29 0.341 216  Gene SNP Genotype OR L95 U95 p-trend CYP2C19 rs11597626 nG 1.03 0.84 1.25 0.811 CYP2C19 rs1322179 nA 1.29 1.00 1.65 0.048 CYP2C19 rs1555474 nG 1.08 0.91 1.28 0.385 CYP2C19 rs4388808 nG 0.90 0.72 1.12 0.350 CYP2C19 rs4917623 nA 0.94 0.78 1.12 0.478 CYP2C19 rs7088784 nG 0.81 0.58 1.13 0.213 CYP2E1 rs2070675 nA 1.05 0.84 1.31 0.655 CYP2E1 rs6413421 nG 0.94 0.63 1.41 0.775 CYP2E1 rs743535 nA 1.22 0.90 1.65 0.196 CYP2E1 rs8192772 nG 1.06 0.76 1.47 0.750 CYP2E1 rs915906 nG 0.96 0.76 1.22 0.745 CYP2E1 rs915908 nA 1.06 0.81 1.37 0.689 CYP3A4 rs2242480 nA 0.87 0.65 1.16 0.343 CYP3A4 rs4646437 nA 0.90 0.68 1.18 0.443 CYP3A5 rs15524 nG 0.92 0.67 1.27 0.627 CYP3A5 rs4646450 nA 0.98 0.78 1.23 0.866 CYP3A43 rs11981167 nT 1.00 0.68 1.48 0.999 CYP3A43 rs472660 nA 1.10 0.84 1.44 0.470 CYP3A43 rs533486 nG 0.98 0.82 1.16 0.778 CYP3A43 rs679320 nG 0.97 0.69 1.36 0.847 CYP7A1 rs11786580 nA 1.02 0.83 1.25 0.845 CYP7A1 rs6997473 nG 1.07 0.73 1.58 0.725 CYP7A1 rs8192877 nG 0.87 0.69 1.10 0.241 CYP7B1 rs1007219 nA 1.18 0.98 1.41 0.073 CYP7B1 rs10112206 nA 1.07 0.90 1.26 0.452 CYP7B1 rs13272719 nG 1.02 0.83 1.25 0.878 CYP7B1 rs13276608 nA 1.04 0.78 1.39 0.790 CYP7B1 rs2884074 nA 0.94 0.77 1.15 0.570 CYP7B1 rs3779874 nG 1.08 0.90 1.30 0.425 217  Gene SNP Genotype OR L95 U95 p-trend CYP7B1 rs4236934 nG 1.12 0.94 1.33 0.190 CYP7B1 rs4258041 nG 0.92 0.75 1.14 0.450 CYP7B1 rs7002672 nG 1.13 0.93 1.38 0.210 CYP7B1 rs7834854 nG 1.12 0.91 1.37 0.288 CYP17A1 rs743572 nG 1.29 1.07 1.55 0.009 CYP19A1 rs1004984 nA 0.96 0.81 1.15 0.689 CYP19A1 rs10459592 nA 1.01 0.86 1.21 0.862 CYP19A1 rs10519295 nG 0.88 0.66 1.18 0.390 CYP19A1 rs12911554 nG 1.01 0.86 1.20 0.875 CYP19A1 rs17523880 nA 1.25 0.98 1.60 0.077 CYP19A1 rs17523922 nG 1.08 0.81 1.43 0.595 CYP19A1 rs17647707 nA 0.68 0.41 1.13 0.133 CYP19A1 rs1870049 nG 0.91 0.70 1.18 0.486 CYP19A1 rs1902584 nA 0.95 0.70 1.29 0.749 CYP19A1 rs2414099 nG 0.93 0.74 1.18 0.562 CYP19A1 rs2445762 nG 0.91 0.76 1.10 0.351 CYP19A1 rs2470152 nA 1.02 0.86 1.20 0.844 CYP19A1 rs2470155 nA 1.07 0.82 1.40 0.635 CYP19A1 rs2470169 nC 0.78 0.55 1.10 0.161 CYP19A1 rs2899472 nA 1.06 0.87 1.28 0.572 CYP19A1 rs3751591 nG 0.90 0.72 1.13 0.362 CYP19A1 rs3751592 nG 1.18 0.97 1.44 0.094 CYP19A1 rs4646 nA 1.04 0.86 1.27 0.683 CYP19A1 rs4774584 nA 1.04 0.88 1.23 0.654 CYP19A1 rs6493494 nA 1.08 0.91 1.28 0.381 CYP19A1 rs6493495 nG 0.89 0.69 1.15 0.362 CYP19A1 rs7172156 nA 0.98 0.82 1.18 0.821 CYP19A1 rs7174997 nA 1.22 0.98 1.53 0.075 CYP19A1 rs730154 nG 0.92 0.73 1.16 0.484 218  Gene SNP Genotype OR L95 U95 p-trend CYP19A1 rs8023263 nC 1.03 0.87 1.22 0.729 CYP19A1 rs9944225 nA 1.05 0.78 1.43 0.737 CYP24A1 rs13038432 nG 0.85 0.59 1.22 0.379 CYP24A1 rs1570669 nG 1.02 0.86 1.22 0.815 CYP24A1 rs2181874 nA 0.89 0.73 1.09 0.258 CYP24A1 rs2244719 nG 1.16 0.97 1.39 0.109 CYP24A1 rs2585428 nA 0.92 0.76 1.10 0.350 CYP24A1 rs2762934 nA 0.95 0.76 1.18 0.633 CYP24A1 rs3787554 nA 0.80 0.59 1.07 0.136 CYP24A1 rs3787557 nG 1.03 0.81 1.30 0.818 CYP24A1 rs4809958 nC 1.03 0.82 1.29 0.805 CYP24A1 rs4809959 nA 1.06 0.89 1.26 0.504 CYP24A1 rs4809960 nG 1.06 0.88 1.29 0.541 CYP24A1 rs6022999 nG 1.05 0.86 1.28 0.653 CYP24A1 rs6068816 nA 0.99 0.75 1.32 0.965 CYP24A1 rs6097809 nG 0.93 0.64 1.34 0.687 CYP24A1 rs912505 nG 0.91 0.74 1.13 0.396 CYP24A1 rs927650 nA 1.01 0.85 1.20 0.898 Abbreviations  L95: Lower limit of 95% confidence interval; OR: Odds ratio; SNP: Single nucleotide polymorphism; U95: Upper limit of 95% confidence interval;   219  Table C.4  Dominant associations between CYP SNPs and NHL  Gene SNP OR L95 U95 p CYP1A1 CYP_73UTR4v10 1.08 0.77 1.51 0.667 CYP1A1 CYP1A1_109t3x 1.14 0.85 1.53 0.379 CYP1A1 CYP1A1_14t1x 0.95 0.74 1.22 0.693 CYP1A1 CYP1A1_83t2 1.13 0.85 1.52 0.403 CYP1A1 CYP5UTRNC2v2 1.07 0.77 1.51 0.682 CYP1A1 CYP73UT1_2v9 1.25 0.71 2.18 0.443 CYP1A2 rs2472304 1.02 0.80 1.29 0.900 CYP1A2 rs762551 0.95 0.75 1.21 0.697 CYP1B1 rs1056836 0.90 0.70 1.17 0.437 CYP1B1 rs10916 1.05 0.82 1.34 0.704 CYP1B1 rs1800440 0.87 0.67 1.12 0.284 CYP2A13 rs16974961 0.85 0.61 1.19 0.348 CYP2B6 rs11673270 1.13 0.89 1.43 0.333 CYP2B6 rs16974799 1.08 0.85 1.37 0.545 CYP2B6 rs2099361 0.94 0.74 1.20 0.619 CYP2B6 rs7260329 0.82 0.64 1.04 0.100 CYP2B6 rs8100458 0.89 0.70 1.13 0.345 CYP2B6 rs8192712 0.94 0.69 1.28 0.701 CYP2C9 rs10509679 1.19 0.92 1.55 0.186 CYP2C9 rs1934967 1.25 0.98 1.59 0.075 CYP2C9 rs1934968 0.96 0.72 1.29 0.794 CYP2C9 rs2153628 1.14 0.87 1.48 0.346 CYP2C9 rs2253635 1.13 0.88 1.44 0.341 CYP2C9 rs9325473 0.88 0.62 1.26 0.498 CYP2C9 rs9332172 0.85 0.66 1.09 0.194 CYP2C9 rs9332197 1.47 1.01 2.12 0.043 CYP2C9 rs9332242 0.86 0.65 1.14 0.297 CYP2C19 rs10786172 1.08 0.85 1.38 0.525 220  Gene SNP OR L95 U95 p CYP2C19 rs11597626 1.03 0.81 1.32 0.799 CYP2C19 rs1322179 1.21 0.92 1.59 0.172 CYP2C19 rs1555474 1.08 0.84 1.39 0.563 CYP2C19 rs4388808 0.83 0.64 1.07 0.140 CYP2C19 rs4917623 0.95 0.71 1.26 0.709 CYP2C19 rs7088784 0.83 0.58 1.19 0.308 CYP2E1 rs2070675 1.07 0.83 1.37 0.628 CYP2E1 rs6413421 0.94 0.63 1.41 0.775 CYP2E1 rs743535 1.19 0.87 1.62 0.288 CYP2E1 rs8192772 1.03 0.74 1.44 0.855 CYP2E1 rs915906 0.96 0.74 1.25 0.751 CYP2E1 rs915908 1.15 0.86 1.55 0.351 CYP3A4 rs2242480 0.88 0.64 1.20 0.418 CYP3A4 rs4646437 0.91 0.67 1.23 0.525 CYP3A5 rs15524 0.86 0.62 1.22 0.404 CYP3A5 rs4646450 0.99 0.76 1.29 0.938 CYP3A43 rs11981167 1.04 0.70 1.56 0.838 CYP3A43 rs472660 1.15 0.86 1.52 0.351 CYP3A43 rs533486 1.01 0.78 1.30 0.969 CYP3A43 rs679320 1.04 0.72 1.49 0.845 CYP7A1 rs11786580 1.08 0.85 1.38 0.535 CYP7A1 rs6997473 1.04 0.70 1.55 0.857 CYP7A1 rs8192877 0.85 0.66 1.11 0.240 CYP7B1 rs1007219 1.22 0.96 1.54 0.110 CYP7B1 rs10112206 1.09 0.83 1.43 0.546 CYP7B1 rs13272719 1.02 0.80 1.30 0.890 CYP7B1 rs13276608 1.01 0.75 1.37 0.938 CYP7B1 rs2884074 0.94 0.74 1.20 0.636 CYP7B1 rs3779874 1.07 0.85 1.36 0.554 221  Gene SNP OR L95 U95 p CYP7B1 rs4236934 1.13 0.89 1.44 0.317 CYP7B1 rs4258041 0.89 0.69 1.16 0.382 CYP7B1 rs7002672 1.27 1.00 1.61 0.053 CYP7B1 rs7834854 1.19 0.93 1.51 0.159 CYP17A1 rs743572 1.28 0.98 1.68 0.070 CYP19A1 rs1004984 0.95 0.74 1.22 0.689 CYP19A1 rs10459592 0.94 0.73 1.22 0.665 CYP19A1 rs10519295 0.89 0.65 1.22 0.472 CYP19A1 rs12911554 0.85 0.65 1.10 0.217 CYP19A1 rs17523880 1.29 0.98 1.71 0.066 CYP19A1 rs17523922 1.11 0.82 1.52 0.499 CYP19A1 rs17647707 0.64 0.38 1.08 0.095 CYP19A1 rs1870049 0.85 0.64 1.13 0.272 CYP19A1 rs1902584 0.92 0.66 1.27 0.594 CYP19A1 rs2414099 0.90 0.69 1.17 0.437 CYP19A1 rs2445762 0.88 0.69 1.12 0.294 CYP19A1 rs2470152 1.00 0.77 1.31 0.974 CYP19A1 rs2470155 1.05 0.79 1.41 0.734 CYP19A1 rs2470169 0.80 0.56 1.17 0.250 CYP19A1 rs2899472 1.02 0.80 1.29 0.904 CYP19A1 rs3751591 0.91 0.70 1.18 0.472 CYP19A1 rs3751592 1.21 0.93 1.56 0.159 CYP19A1 rs4646 1.01 0.79 1.28 0.947 CYP19A1 rs4774584 1.00 0.77 1.29 0.999 CYP19A1 rs6493494 1.01 0.78 1.31 0.932 CYP19A1 rs6493495 0.85 0.64 1.13 0.263 CYP19A1 rs7172156 1.00 0.77 1.31 0.980 CYP19A1 rs7174997 1.29 1.00 1.66 0.053 CYP19A1 rs730154 0.91 0.70 1.18 0.468 222  Gene SNP OR L95 U95 p CYP19A1 rs8023263 0.88 0.68 1.15 0.362 CYP19A1 rs9944225 0.99 0.72 1.38 0.967 CYP24A1 rs13038432 0.78 0.53 1.13 0.191 CYP24A1 rs1570669 0.98 0.77 1.25 0.882 CYP24A1 rs2181874 0.90 0.71 1.15 0.392 CYP24A1 rs2244719 1.28 0.96 1.71 0.095 CYP24A1 rs2585428 0.91 0.68 1.22 0.546 CYP24A1 rs2762934 0.98 0.76 1.25 0.851 CYP24A1 rs3787554 0.80 0.58 1.10 0.166 CYP24A1 rs3787557 1.06 0.81 1.38 0.698 CYP24A1 rs4809958 1.01 0.78 1.31 0.935 CYP24A1 rs4809959 1.09 0.84 1.42 0.517 CYP24A1 rs4809960 1.06 0.84 1.35 0.614 CYP24A1 rs6022999 1.06 0.83 1.34 0.661 CYP24A1 rs6068816 0.97 0.72 1.32 0.861 CYP24A1 rs6097809 0.97 0.64 1.48 0.899 CYP24A1 rs912505 0.94 0.74 1.20 0.635 CYP24A1 rs927650 1.15 0.88 1.49 0.314 Abbreviations  L95: Lower limit of 95% confidence interval; OR: Odds ratio;  SNP: Single nucleotide polymorphism; U95: Upper limit of 95% confidence interval;  N.B. OR for reference group, comprised of major allele homozygotes, is equal to 1 by definition;   223  Table C.5  Recessive associations between CYP SNPs and NHL  Gene SNP OR L95 U95 p CYP1A1 CYP_73UTR4v10 1.19 0.31 4.48 0.802 CYP1A1 CYP1A1_109t3x 0.73 0.27 1.98 0.535 CYP1A1 CYP1A1_14t1x 0.98 0.63 1.52 0.917 CYP1A1 CYP1A1_83t2 0.65 0.26 1.64 0.358 CYP1A1 CYP5UTRNC2v2 0.94 0.23 3.82 0.932 CYP1A1 CYP73UT1_2v9 0.87 0.05 14.2 0.920 CYP1A2 rs2472304 1.06 0.75 1.50 0.755 CYP1A2 rs762551 1.23 0.81 1.86 0.324 CYP1B1 rs1056836 1.15 0.86 1.54 0.351 CYP1B1 rs10916 1.15 0.70 1.87 0.586 CYP1B1 rs1800440 1.02 0.55 1.88 0.951 CYP2A13 rs16974961 Inf 0 Inf 0.999 CYP2B6 rs11673270 1.23 0.75 2.01 0.419 CYP2B6 rs16974799 1.45 0.89 2.36 0.132 CYP2B6 rs2099361 0.96 0.69 1.35 0.827 CYP2B6 rs7260329 0.88 0.58 1.33 0.539 CYP2B6 rs8100458 0.93 0.63 1.37 0.710 CYP2B6 rs8192712 1.77 0.51 6.10 0.369 CYP2C9 rs10509679 3.16 1.30 7.68 0.011 CYP2C9 rs1934967 0.97 0.58 1.65 0.921 CYP2C9 rs1934968 1.19 0.44 3.25 0.734 CYP2C9 rs2153628 1.05 0.59 1.85 0.877 CYP2C9 rs2253635 0.93 0.67 1.29 0.673 CYP2C9 rs9325473 0.34 0.06 1.77 0.199 CYP2C9 rs9332172 0.76 0.39 1.47 0.409 CYP2C9 rs9332197 1.05 0.30 3.70 0.938 CYP2C9 rs9332242 1.31 0.45 3.84 0.625 CYP2C19 rs10786172 1.20 0.84 1.70 0.316 224  Gene SNP OR L95 U95 p CYP2C19 rs11597626 1.02 0.60 1.73 0.937 CYP2C19 rs1322179 4.33 1.41 13.3 0.010 CYP2C19 rs1555474 1.15 0.84 1.57 0.386 CYP2C19 rs4388808 1.45 0.73 2.90 0.292 CYP2C19 rs4917623 0.88 0.65 1.20 0.420 CYP2C19 rs7088784 0.34 0.06 1.77 0.199 CYP2E1 rs2070675 1.03 0.51 2.07 0.940 CYP2E1 rs6413421 NA NA NA NA CYP2E1 rs743535 4.67 0.54 40.8 0.163 CYP2E1 rs8192772 Inf 0 Inf 0.999 CYP2E1 rs915906 0.94 0.39 2.25 0.889 CYP2E1 rs915908 0.47 0.18 1.27 0.138 CYP3A4 rs2242480 0.58 0.17 2.04 0.399 CYP3A4 rs4646437 0.68 0.24 1.94 0.468 CYP3A5 rs15524 2.79 0.55 14.1 0.215 CYP3A5 rs4646450 0.89 0.44 1.80 0.753 CYP3A43 rs11981167 0 0 Inf 0.999 CYP3A43 rs472660 0.62 0.17 2.23 0.466 CYP3A43 rs533486 0.92 0.68 1.25 0.586 CYP3A43 rs679320 0 0 Inf 0.999 CYP7A1 rs11786580 0.78 0.45 1.36 0.388 CYP7A1 rs6997473 Inf 0 Inf 0.999 CYP7A1 rs8192877 0.84 0.40 1.76 0.639 CYP7B1 rs1007219 1.29 0.88 1.90 0.197 CYP7B1 rs10112206 1.09 0.83 1.44 0.532 CYP7B1 rs13272719 1.03 0.59 1.81 0.916 CYP7B1 rs13276608 2.47 0.44 13.8 0.301 CYP7B1 rs2884074 0.88 0.53 1.48 0.637 CYP7B1 rs3779874 1.20 0.77 1.86 0.430 225  Gene SNP OR L95 U95 p CYP7B1 rs4236934 1.23 0.88 1.72 0.235 CYP7B1 rs4258041 0.97 0.57 1.64 0.901 CYP7B1 rs7002672 0.80 0.48 1.33 0.388 CYP7B1 rs7834854 0.91 0.52 1.59 0.736 CYP17A1 rs743572 1.61 1.12 2.31 0.011 CYP19A1 rs1004984 0.96 0.68 1.36 0.825 CYP19A1 rs10459592 1.14 0.84 1.54 0.410 CYP19A1 rs10519295 0.58 0.16 2.09 0.407 CYP19A1 rs12911554 1.29 0.96 1.73 0.094 CYP19A1 rs17523880 1.23 0.51 2.95 0.650 CYP19A1 rs17523922 0.83 0.28 2.51 0.748 CYP19A1 rs17647707 Inf 0 Inf 0.999 CYP19A1 rs1870049 1.88 0.64 5.52 0.251 CYP19A1 rs1902584 1.72 0.42 7.02 0.449 CYP19A1 rs2414099 1.17 0.53 2.58 0.705 CYP19A1 rs2445762 0.94 0.60 1.48 0.800 CYP19A1 rs2470152 1.04 0.79 1.37 0.766 CYP19A1 rs2470155 1.43 0.49 4.21 0.515 CYP19A1 rs2470169 0.21 0.02 1.84 0.158 CYP19A1 rs2899472 1.33 0.82 2.14 0.249 CYP19A1 rs3751591 0.72 0.35 1.48 0.370 CYP19A1 rs3751592 1.33 0.87 2.03 0.184 CYP19A1 rs4646 1.25 0.76 2.06 0.372 CYP19A1 rs4774584 1.13 0.84 1.53 0.428 CYP19A1 rs6493494 1.27 0.93 1.73 0.139 CYP19A1 rs6493495 1.22 0.45 3.25 0.698 CYP19A1 rs7172156 0.92 0.64 1.31 0.637 CYP19A1 rs7174997 1.12 0.55 2.25 0.759 CYP19A1 rs730154 0.93 0.40 2.15 0.870 226  Gene SNP OR L95 U95 p CYP19A1 rs8023263 1.27 0.95 1.69 0.112 CYP19A1 rs9944225 3.48 0.71 17.0 0.124 CYP24A1 rs13038432 Inf 0 Inf 0.999 CYP24A1 rs1570669 1.14 0.79 1.64 0.482 CYP24A1 rs2181874 0.73 0.43 1.26 0.261 CYP24A1 rs2244719 1.16 0.85 1.57 0.352 CYP24A1 rs2585428 0.86 0.63 1.18 0.353 CYP24A1 rs2762934 0.73 0.37 1.40 0.341 CYP24A1 rs3787554 0.52 0.13 2.14 0.365 CYP24A1 rs3787557 0.87 0.41 1.87 0.724 CYP24A1 rs4809958 1.26 0.58 2.76 0.556 CYP24A1 rs4809959 1.07 0.80 1.43 0.666 CYP24A1 rs4809960 1.13 0.70 1.83 0.610 CYP24A1 rs6022999 1.06 0.62 1.82 0.821 CYP24A1 rs6068816 1.56 0.36 6.65 0.551 CYP24A1 rs6097809 0.48 0.12 1.95 0.306 CYP24A1 rs912505 0.67 0.36 1.24 0.198 CYP24A1 rs927650 0.87 0.64 1.17 0.352 Abbreviations  Inf: Infinity; L95: Lower limit of 95% confidence interval;  NA: Not applicable (PLINK could not compute e.g. 0 minor allele homozygotes);   OR: Odds ratio; SNP: Single nucleotide polymorphism; U95: Upper limit of 95% confidence interval;  N.B. OR for reference group, comprised of major allele homozygotes and heterozygotes, is equal to 1 by definition;     227  Table C.6  Dominant associations with additive p-trends between rs743535 (CYP2E1) and NHL subtypes Gene Genotype Controls (n) Cases (n) OR L95 U95 p-trend DLBCL G G 458 115 1   0.174 CYP2E1 A G or A A 89 31 1.41 0.88 2.23  FL G G 458 128 1   0.076 CYP2E1 A G or A A 89 37 1.50 0.97 2.32  Other B-cell G G 458 172 1   0.566 CYP2E1 A G or A A 89 34 0.99 0.63 1.54  All T-cell G G 458 43 1   0.053 CYP2E1 A G or A A 89 <5 0.24 0.06 1.02  Abbreviations  DLBCL: Diffuse large B-cell lymphoma; FL: Follicular lymphoma; L95: Lower limit of 95% confidence interval; OR: Odds ratio; U95: Upper limit of 95% confidence interval;  228  Table C.7  Codominant associations with additive p-trends between rs915906 (CYP2E1) and NHL subtypes  Subtype Genotype Controls (n) Cases (n) OR L95 U95 p-trend DLBCL A A 385 110 1   0.259 CYP2E1 A G 149 36 0.85 0.56 1.31  CYP2E1 G G 10 <5 0.32 0.04 2.58  FL A A 385 111 1   0.390 CYP2E1 A G 149 49 1.11 0.75 1.64  CYP2E1 G G 10 5 1.64 0.54 4.97  Other B-cell A A 385 143 1   0.693 CYP2E1 A G 149 59 1.14 0.78 1.64  CYP2E1 G G 10 <5 0.82 0.25 2.69  All T-cell A A 385 39 1   0.025 CYP2E1 A G 149 <5 0.25 0.09 0.72  CYP2E1 G G 10 <5 1.01 0.12 8.37  Abbreviations  DLBCL: Diffuse large B-cell lymphoma; FL: Follicular lymphoma; L95: Lower limit of 95% confidence interval; OR: Odds ratio; U95: Upper limit of 95% confidence interval;  229  Table C.8  Codominant associations with additive p-trends between rs1007219 (CYP7B1) and NHL subtypes Subtype Genotype Controls (n) Cases (n) OR L95 U95 p-trend DLBCL G G 275 61 1   0.070 CYP7B1 A G 219 69 1.43 0.97 2.12  CYP7B1 A A 52 17 1.50 0.81 2.80  FL G G 275 77 1   0.113 CYP7B1 A G 219 63 1.02 0.70 1.50  CYP7B1 A A 52 25 1.76 1.02 3.04  Other B-cell G G 275 107 1   0.536 CYP7B1 A G 219 84 0.98 0.69 1.38  CYP7B1 A A 52 15 0.77 0.41 1.46  All T-cell G G 275 12 1   0.002 CYP7B1 A G 219 25 2.79 1.36 5.75  CYP7B1 A A 52 8 3.71 1.42 9.69  Abbreviations  DLBCL: Diffuse large B-cell lymphoma; FL: Follicular lymphoma; L95: Lower limit of 95% confidence interval; OR: Odds ratio; U95: Upper limit of 95% confidence interval;  230  Table C.9  Codominant associations with additive p-trends between rs2470152 (CYP19A1) and NHL subtypes Gene Genotype Controls (n) Cases (n) OR L95 U95 p-trend DLBCL G G 155 27 1   0.064 CYP19A1 A G 261 79 1.65 1.02 2.68  CYP19A1 A A 131 41 1.70 0.98 2.93  FL G G 155 50 1   0.456 CYP19A1 A G 261 80 0.98 0.65 1.47  CYP19A1 A A 131 35 0.82 0.50 1.35  Other B-cell G G 155 58 1   0.818 CYP19A1 A G 261 90 0.85 0.57 1.26  CYP19A1 A A 131 58 1.06 0.68 1.65  All T-cell G G 155 20 1   0.053 CYP19A1 A G 261 16 0.46 0.23 0.93  CYP19A1 A A 131 9 0.49 0.21 1.12  Abbreviations  DLBCL: Diffuse large B-cell lymphoma; FL: Follicular lymphoma; L95: Lower limit of 95% confidence interval; OR: Odds ratio; U95: Upper limit of 95% confidence interval;   231  Table C.10  Codominant associations with additive p-trends between rs4774584 (CYP19A1) and NHL subtypes Gene Genotype Controls (n) Cases (n) OR L95 U95 p-trend DLBCL G G 167 52 1   0.062 CYP19A1 A G 277 78 0.87 0.58 1.31  CYP19A1 A A 102 17 0.54 0.30 0.99  FL G G 167 45 1   0.170 CYP19A1 A G 277 80 1.11 0.73 1.68  CYP19A1 A A 102 40 1.43 0.87 2.36  Other B-cell G G 167 66 1   0.535 CYP19A1 A G 277 94 0.87 0.60 1.28  CYP19A1 A A 102 45 1.22 0.77 1.94  All T-cell G G 167 12 1   0.240 CYP19A1 A G 277 21 1.12 0.53 2.36  CYP19A1 A A 102 12 1.69 0.73 3.94  Abbreviations  DLBCL: Diffuse large B-cell lymphoma; FL: Follicular lymphoma; L95: Lower limit of 95% confidence interval; OR: Odds ratio; U95: Upper limit of 95% confidence interval;  232  Table C.11  Codominant associations with additive p-trends between rs7172156 (CYP19A1) and NHL subtypes Subtype Genotype Controls (n) Cases (n) OR L95 U95 p-trend DLBCL G G 189 36 1   0.016 CYP19A1 A G 204 67 1.66 1.05 2.62  CYP19A1 A A 75 28 1.90 1.07 3.38  FL G G 189 61 1   0.112 CYP19A1 A G 204 58 0.86 0.57 1.31  CYP19A1 A A 75 15 0.59 0.31 1.11  Other B-cell G G 189 73 1   0.384 CYP19A1 A G 204 78 0.96 0.65 1.42  CYP19A1 A A 75 24 0.76 0.44 1.32  All T-cell G G 189 19 1   0.294 CYP19A1 A G 204 13 0.56 0.27 1.19  CYP19A1 A A 75 6 0.73 0.27 1.93  Abbreviations  DLBCL: Diffuse large B-cell lymphoma; FL: Follicular lymphoma; L95: Lower limit of 95% confidence interval; OR: Odds ratio; U95: Upper limit of 95% confidence interval;  233  Table C.12  Codominant associations with additive p-trends between rs12911554 (CYP19A1) and NHL subtypes Subtype Genotype Controls (n) Cases (n) OR L95 U95 p-trend DLBCL A A 155 33 1   0.012 CYP19A1 A G 288 72 1.13 0.71 1.79  CYP19A1 G G 101 42 2.00 1.18 3.40  FL A A 155 59 1   0.294 CYP19A1 A G 288 73 0.69 0.46 1.03  CYP19A1 G G 101 31 0.83 0.50 1.37  Other B-cell A A 155 70 1   0.880 CYP19A1 A G 288 86 0.63 0.43 0.92  CYP19A1 G G 101 50 1.06 0.67 1.66  All T-cell A A 155 16 1   0.108 CYP19A1 A G 288 25 0.85 0.44 1.65  CYP19A1 G G 101 <5 0.37 0.12 1.16  Abbreviations  DLBCL: Diffuse large B-cell lymphoma; FL: Follicular lymphoma; L95: Lower limit of 95% confidence interval; OR: Odds ratio; U95: Upper limit of 95% confidence interval;  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0107219/manifest

Comment

Related Items