HEALTHCARE WORKERS AND ANTINEOPLASTIC DRUGS: EVALUATING THE RISKS AND IDENTIFYING DETERMINANTS OF EXPOSURE by Chun-Yip Hon B.Sc., McMaster University, 1994 B.A.Sc., Ryerson University, 1996 M.Sc.(A), McGill University, 1997 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in The Faculty of Graduate Studies (Occupational and Environmental Hygiene) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) June 2012 © Chun-Yip Hon, 2012 ii ABSTRACT Healthcare workers’ exposure to antineoplastic drugs may occur through handling of the drugs and/or via contact with drug-contaminated surfaces. However, studies have been limited to select departments and/or certain job titles. This may lead to an underestimate of the risk as the drugs circulate within a facility known as the hospital medication system (process flow of drugs). This study aimed to answer the following questions related to antineoplastic drugs and the hospital medication system: 1) is contamination found on surfaces located throughout, 2) are workers throughout occupationally exposed (dermal and urinary contamination), and 3) what factors are associated with surface contamination and occupational exposure? Site observations were conducted to identify which surfaces may be contaminated and the job categories that may contact these surfaces. Wipe samples were collected from potentially- contaminated surfaces and the hands of at-risk healthcare workers. Urine samples were collected from these same workers. Participants were asked to complete a questionnaire regarding their knowledge and usual protective habits regarding antineoplastic drugs and surveyed about contact with these agents on their work shift. Drug residual was measurable on surfaces located throughout the hospital medication system. Determinants associated with increased surface contamination were the drug preparation and drug administration stages of the medication system as well as having more job categories responsible for drug transport. Up to 11 job categories per facility may have an exposure risk and the maximum dermal contamination levels for every job category exceeded the limit of detection. Factors associated with increased dermal contamination were working in acute care hospitals, female personnel, working as a porter, nurse, transport, unit clerk or other roles in the drug administration unit and having a duty to handle antineoplastic drugs. Urinary drug iii contamination of participants was higher than in non-hospital controls confirming that exposure is occurring in the workplace. Being a pharmacy receiver, pharmacy technician, porter, nurse, or unit clerk and a facility having more job categories responsible for drug transport were associated with increased urinary contamination. This is believed to be the first study examining environmental contamination and occupational exposure to antineoplastic drugs across the entire hospital medication system. iv PREFACE This dissertation consists of a total of six chapters. There are four research chapters (Chapters 2 to 5) bookended by an introductory and concluding chapter. Chapter 2 has already been published and Chapter 3 has been submitted for publication. Chapters 4 and 5 will be modified to meet journal formatting requirements and submitted in the coming months. I am/will be listed as primary author on each of these journal submissions. In this section, details of my role in the current study are provided as well as the contributions of others to each of the research chapters. I designed the study and served in the capacity as principal author on the grant application to secure funding for the project. An operating grant was awarded by the WorkSafeBC Research Secretariat [RS2008-OG011] with a research budget of $263,325.70. Operationally, I was the coordinator of the project responsible for the following duties: ethics applications and renewals (see below), participant recruitment, development of operating procedures for the various tasks related to data collection and field work, development of all study documents and pilot-testing of the questionnaire, hiring and managing research assistants, purchasing supplies, overseeing all sample collection, and management of the results database as well as the budget. Recruitment of participants, study documents and sample collection methods for the study were approved by the University of British Columbia’s Clinical Research Ethics Board (CREB certificate #H08-01167), as well as by the ethics boards of other participating research sites. The Occupational and Environmental Hygiene laboratory located at the University of British Columbia performed all laboratory analyses. I conducted all the statistical analyses indicated and wrote the body of work presented in this dissertation. v Chapter 2: Identification of surfaces contaminated and job categories exposed throughout the hospital medication system As repeated site observations were necessary, several research assistants, besides myself, were involved in the data collection - Claire Pitcher, Jennifer Shum and Pearl Signaporia. I conducted all data analyses and wrote a majority of the manuscript. Feedback was provided by Kay Teschke, Prescillia Chua, Scott Venners and Lynne Nakashima. My overall contribution: 95%. A version of Chapter 2 has been published: Hon C-Y, Teschke K, Chua P, Venners P and Nakashima L (2011). Occupational exposure to antineoplastic drugs: Identification of job categories potentially exposed throughout the hospital medication system. Safety and Health at Work 2(3): 273-281. http://dx.doi.org/10.5491/SHAW.2011.2.3.273. Copyright permission has been granted from the journal through Creative Commons Attribution Non-Commercial License to include excerpts of the article in this dissertation. Chapter 3: Identification of surface contamination throughout the hospital medication system and identification of determinants of such contamination In addition to myself, several research assistants were responsible for collecting surface wipe samples at the participating facilities: Jennifer Shum, Alexandra Barzan, Louise Hughes-Rhodes, Pearl Siganporia and Sarah Chiarello. I conducted all statistical analyses and wrote most of the chapter. Kay Teschke lent guidance regarding data analysis and provided the majority of feedback about the chapter. Winnie Chu led the team responsible for the laboratory analyses of the wipe samples (with Cris Barzan as the lead chemist) and reviewed the Methods section vi describing the laboratory analyses. Both Paul Demers and George Astrakianiakis provided feedback comments on the chapter. My contribution: 90%. A version of this chapter will be submitted for publication in the next few months. Chapter 4: Evaluation of healthcare workers’ dermal exposure throughout the hospital medication system and identification of determinants of such exposure As in the previous chapter, I had assistance with sample collection. Research assistants involved with dermal wipe sample collection were Jennifer Shum, Pearl Siganporia and Sarah Chiarello. I conducted all statistical analyses and wrote most of the chapter. Kay Teschke provided guidance regarding data analysis, feedback on the questionnaire, and the majority of feedback about the chapter. Winnie Chu led the team responsible for the laboratory analyses of the wipe samples (with Cris Barzan as the lead chemist) and reviewed the Methods section describing the laboratory analyses. Both Scott Venners and Paul Demers provided feedback on the chapter. My contribution: 90%. A version of this chapter will be submitted for publication in the next few months. Chapter 5: Evaluation of healthcare workers’ urinary contamination throughout the hospital medication system and identification of determinants of such exposure The same research assistants involved with dermal wipe sample collection were also involved with urine sample collection: Jennifer Shum, Pearl Siganporia and Sarah Chiarello. Alexandra Barzan helped to dispense the samples as well as document some basic lab measurements e.g. volume of urine, pH and extract temperature readings from probes. I conducted all statistical analyses and wrote most of the chapter. Scott Venners assisted with the handling and interpretation of urinary contamination results as well as provided feedback vii comments regarding the chapter. Kay Teschke provided guidance regarding data analysis and feedback of the chapter. Winnie Chu led the team responsible for the laboratory analyses of the urine samples (with Cris Barzan as the lead chemist) and reviewed the Methods section describing the laboratory analyses. Paul Demers provided constructive comments regarding the chapter as well. My contribution: 85%. A version of this chapter will be submitted for publication in the next few months. viii TABLE OF CONTENTS ABSTRACT ................................................................................................................................................................. ii PREFACE.....................................................................................................................................................................iv TABLE OF CONTENTS .......................................................................................................................................... viii LIST OF TABLES...................................................................................................................................................... xii LIST OF FIGURES ................................................................................................................................................... xiii LIST OF ABBREVIATIONS ..................................................................................................................................... xiv ACKNOWLEDGEMENTS ......................................................................................................................................... xv DEDICATION ......................................................................................................................................................... xvii 1 INTRODUCTION ................................................................................................................................................. 1 1.1 Background .................................................................................................................................................. 1 1.1.1 Release date of safe drug handling guidelines ......................................................................................... 2 1.1.2 Use of cyclophosphamide as marker drug ............................................................................................... 2 1.2 Pilot study conducted to ascertain nature of problem at local facilities ....................................................... 4 1.3 Review of existing literature ........................................................................................................................ 5 1.3.1 Only select departments have been assessed for surface contamination .................................................. 5 1.3.2 Dermal contact of surfaces sampled is unknown ..................................................................................... 7 1.3.3 Underestimation of exposure ................................................................................................................... 9 1.3.4 Dearth of data related to the determinants of contamination or exposure .............................................. 12 1.3.5 Occupational exposure likely to continue and possibly trend upwards ................................................. 14 1.3.6 Lack of consistent and/or effective policies to reduce occupational exposure....................................... 14 1.4 Rationale for current study and research questions .................................................................................... 15 1.5 Dissertation structure ................................................................................................................................. 16 2 IDENTIFICATION OF SURFACES CONTAMINATED AND JOB CATEGORIES EXPOSED THROUGHOUT THE HOSPITAL MEDICATION SYSTEM .................................................................................. 21 2.1 Synopsis ..................................................................................................................................................... 21 2.2 Introduction ................................................................................................................................................ 22 2.3 Methods ...................................................................................................................................................... 23 2.3.1 Selection of participating sites ............................................................................................................... 23 2.3.2 Informant interviews .............................................................................................................................. 24 2.3.3 Site observations .................................................................................................................................... 25 2.3.4 Data analysis .......................................................................................................................................... 26 2.4 Results ........................................................................................................................................................ 26 2.4.1 General hospital medication system ...................................................................................................... 26 2.4.2 Description of participating facilities ..................................................................................................... 27 2.4.3 Contact frequency of work surfaces/objects .......................................................................................... 27 2.4.4 Job categories at risk of exposure .......................................................................................................... 29 2.5 Discussion .................................................................................................................................................. 29 ix 2.5.1 Limitations ............................................................................................................................................. 31 2.5.2 Summary ................................................................................................................................................ 31 3 EVALUATION OF SURFACE CONTAMINATION THROUGHOUT THE HOSPITAL MEDICATION SYSTEM AND IDENTIFICATION OF DETERMINANTS OF SUCH CONTAMINATION ................................. 41 3.1 Synopsis ..................................................................................................................................................... 41 3.2 Introduction ................................................................................................................................................ 42 3.3 Methods ...................................................................................................................................................... 43 3.3.1 Selection of surfaces to be sampled ....................................................................................................... 43 3.3.2 Wipe sample collection .......................................................................................................................... 44 3.3.3 Wipe sample preparation ....................................................................................................................... 44 3.3.4 Wipe sample analysis............................................................................................................................. 45 3.3.5 Supplemental data collection ................................................................................................................. 45 3.3.6 Data analysis .......................................................................................................................................... 46 3.3.7 Reporting of results ................................................................................................................................ 47 3.4 Results ........................................................................................................................................................ 48 3.4.1 Overall summary of surface contamination levels ................................................................................. 48 3.4.2 Surface contamination levels by medication system stage .................................................................... 48 3.4.3 Surface contamination levels based on hospital characteristics ............................................................. 49 3.4.4 Surface contamination levels based on reported CP handling, spills and surface cleaning ................... 50 3.4.5 Surface contamination levels based on attributes of the wipe sample ................................................... 51 3.4.6 Multiple linear regression model ........................................................................................................... 52 3.5 Discussion .................................................................................................................................................. 52 3.5.1 Limitations ............................................................................................................................................. 57 3.5.2 Summary ................................................................................................................................................ 58 4 EVALUATION OF HEALTHCARE WORKERS’ DERMAL EXPOSURE THROUGHOUT THE HOSPITAL MEDICATION SYSTEM AND IDENTIFICATION OF DETERMINANTS OF SUCH CONTAMINATION....... 64 4.1 Synopsis ..................................................................................................................................................... 64 4.2 Introduction ................................................................................................................................................ 65 4.3 Methodology .............................................................................................................................................. 66 4.3.1 Selection of participants ......................................................................................................................... 66 4.3.2 Wipe sampling of hands ........................................................................................................................ 67 4.3.3 Wipe sample preparation and analysis ................................................................................................... 67 4.3.4 Supplemental data collected on site ....................................................................................................... 68 4.3.5 Self-administered questionnaire............................................................................................................. 68 4.3.6 Statistical analysis .................................................................................................................................. 69 4.3.7 Reporting of results ................................................................................................................................ 71 4.4 Results ........................................................................................................................................................ 71 4.4.1 Characteristics of study population ........................................................................................................ 71 4.4.2 Overall summary of dermal contamination levels ................................................................................. 71 4.4.3 Dermal contamination levels by individual factors ............................................................................... 72 x 4.4.4 Dermal contamination levels by cyclophosphamide (CP) contact ......................................................... 73 Dermal contamination levels based on glove usage ............................................................................................ 74 4.4.5 Dermal contamination level based on hand washing practices .............................................................. 75 4.4.6 Multiple linear regression model ........................................................................................................... 78 4.5 Discussion .................................................................................................................................................. 79 4.5.1 Limitations ............................................................................................................................................. 83 4.5.2 Summary ................................................................................................................................................ 84 5 EVALUATION OF HEALTHCARE WORKERS’ URINARY CONTAMINATION LEVELS THROUGHOUT THE HOSPITAL MEDICATION SYSTEM AND IDENTIFICATION OF DETERMINANTS OF SUCH EXPOSURE ..................................................................................................................................................... 91 5.1 Synopsis ..................................................................................................................................................... 91 5.2 Introduction ................................................................................................................................................ 92 5.3 Methodology .............................................................................................................................................. 93 5.3.1 Selection of participants ......................................................................................................................... 93 5.3.2 Selection of control subjects .................................................................................................................. 94 5.3.3 Collection of urine samples ................................................................................................................... 94 5.3.4 Urine sample analysis ............................................................................................................................ 95 5.3.5 Supplemental data collection and questionnaire .................................................................................... 97 5.3.6 Statistical analysis .................................................................................................................................. 97 5.4 Results ........................................................................................................................................................ 98 5.4.1 Characteristics of study population ........................................................................................................ 98 5.4.2 Summary of urinary contamination levels – participants and controls .................................................. 98 5.4.3 Urinary contamination levels by individual factors ............................................................................... 99 5.4.4 Urinary contamination levels by hospital characteristics ..................................................................... 100 5.4.5 Urinary contamination levels by contact with cyclophosphamide (CP) .............................................. 101 5.4.6 Multiple linear regression model ......................................................................................................... 101 5.5 Discussion ................................................................................................................................................ 102 5.5.1 Limitations ........................................................................................................................................... 107 5.5.2 Summary .............................................................................................................................................. 107 6 CONCLUDING CHAPTER .............................................................................................................................. 114 6.1 Study overview and objectives ................................................................................................................. 114 6.2 Key findings ............................................................................................................................................. 115 6.2.1 Chapter 2 (Objective 1) ........................................................................................................................ 115 6.2.2 Chapter 3 (Objective 2) ........................................................................................................................ 116 6.2.3 Chapter 4 (Objective 3) ........................................................................................................................ 116 6.2.4 Chapter 5 (Objective 4) ........................................................................................................................ 117 6.3 Conclusions and implications of research ................................................................................................ 119 6.4 Recommendations to minimize occupational exposure to antineoplastic drugs ...................................... 120 6.5 Knowledge translation ............................................................................................................................. 122 6.6 Study strengths and limitations ................................................................................................................ 123 xi 6.7 Future direction/work ............................................................................................................................... 126 6.7.1 Surface cleaning ................................................................................................................................... 127 6.7.2 Evaluate health risks ............................................................................................................................ 127 6.7.3 Determine mechanism of spread of contamination .............................................................................. 128 6.7.4 Identify determinants for each stage/job category separately .............................................................. 128 6.7.5 Association between environmental monitoring and biological monitoring ....................................... 128 REFERENCES .......................................................................................................................................................... 131 APPENDICES ........................................................................................................................................................... 140 APPENDIX A PASSIVE OBSERVATIONAL CHECKLIST ................................................................................. 141 APPENDIX B DRUG HANDLING AND CLEANING QUESTIONS .................................................................... 143 APPENDIX C LETTER OF INITIAL CONTACT (FOR RAMDOMLY SELECTED PARTICIPANTS) ............. 144 APPENDIX D CONSENT TO CONTACT FORM (FOR PASSIVELY RECRUITED PARTICIPANTS) ............ 145 APPENDIX E CONSENT FORM ............................................................................................................................ 146 APPENDIX F ACTIVITY-RELATED QUESTIONS .............................................................................................. 149 APPENDIX G INTRODUCTION LETTER SENT WITH QUESTIONNAIRES ................................................... 153 APPENDIX H SELF-ADMINISTERED QUESTIONNAIRE ................................................................................. 154 APPENDIX I URINARY CONTAMINATION LEVELS STRATIFIED BY REPORTED CONTACT METHODS DURING WORK SHIFT: SUMMARY STATISTICS AND ONE-WAY ANOVA RESULTS .............................. 162 APPENDIX J WRITTEN INSTRUCTIONS FOR COLLECTING URINE ............................................................. 163 xii LIST OF TABLES Table 2-1 Description of participating facilities ......................................................................................................... 34 Table 2-2 Number of observed job categories and associated tasks with potential for dermal exposure at each healthcare facility ........................................................................................................................................................ 39 Table 2-3 Number of workers per at-risk job category at participating facilities ....................................................... 40 Table 3-1 Surface contamination levels based on categorical variables of hospital characteristics ............................ 61 Table 3-2 Bivariate results of surface contamination levels according to hospital characteristic variables that are continuous in nature ..................................................................................................................................................... 61 Table 3-3 Bivariate results of surface contamination levels based on reported cyclophosphamide (CP) handling and surface cleaning ........................................................................................................................................................... 61 Table 3-4 Summary statistics for cyclophosphamide (CP) surface contamination levels based on employee-reported CP usage and surface cleaning during the work shift .................................................................................................. 62 Table 3-5 Coefficients, standard errors and p-values for multiple linear regression model showing factors related to surface wipe contamination levels (ln-transformed) .................................................................................................... 62 Table 3-6 Reported cyclophosphamide contamination levels in ng/cm2 found in healthcare facilities ...................... 63 Table 4-1 Dermal cyclophosphamide (CP) contamination levels of personnel in the hospital medication system, stratified by demographic variables: summary statistics, one-way ANOVA results and percent of samples less than detection limit .............................................................................................................................................................. 87 Table 4-2 Dermal cyclophosphamide (CP) contamination levels of personnel in the hospital medication system, stratified by reported contact methods during work shift: summary statistics, one-way ANOVA results, and percent of sample less than detection ....................................................................................................................................... 88 Table 4-3 Dermal cyclophosphamide (CP) contamination levels of personnel in the hospital medication system, stratified by protective measures employed by participants: summary statistics, one-way ANOVA and percent of samples less than detection .......................................................................................................................................... 89 Table 4-4 Coefficients, standard errors and p-values for final multiple linear regression model showing variables associated with dermal contamination (ln-transformed) .............................................................................................. 90 Table 4-5 Comparison of cyclophosphamide (CP) dermal contamination levels reported in the current and other published studies, in reverse chronological order ........................................................................................................ 90 Table 5-1 Supplies provided to each participant for collecting urine samples .......................................................... 109 Table 5-2 Urinary contamination levels of cyclophosphamide (CP) and its metabolites of personnel in the hospital medication system, stratified by demographic variables: summary statistics and one-way ANOVA results ............ 110 Table 5-3 Urinary contamination levels of cyclophosphamide (CP) and its metabolites of personnel in the hospital medication system, stratified by categorical hospital characteristics: summary statistics and one-way ANOVA results ......................................................................................................................................................................... 111 Table 5-4 Urinary contamination levels of cyclophosphamide (CP) and its metabolites of personnel in hospital medication system, stratified by continuous variables of hospital characteristics: linear regression results ............. 111 Table 5-5 Coefficients, standard errors and p-values for final multiple linear regression model showing variables associated with urinary contamination levels (ln-transformed) ................................................................................. 112 Table 5-6 Comparison of cyclophosphamide (CP) urinary contamination levels reported in current study and other published studies in reverse chronological order since 2004 ..................................................................................... 113 Table 6-1 Risk assessment for bladder cancer based on Sessink et al. ..................................................................... 130 Table 6-2 Risk assessment for bladder cancer based on Sargent et al. ..................................................................... 130 xiii LIST OF FIGURES Figure 1-1 Structural and molecular formula and relative molecular mass of cyclophosphamide ............................. 19 Figure 1-2 Metabolism of cyclophosphamide. The inactivation pathways are depicted horizontally, while cyclophosphamide activation is shown vertically ........................................................................................................ 20 Figure 2-1 Overview of hospital medication system within participating facilities ................................................... 33 Figure 2-2 Average contact frequency of potentially contaminated surfaces observed per task cycle during drug delivery ........................................................................................................................................................................ 35 Figure 2-3 Average contact frequency of potentially contaminated surfaces observed per task cycle during drug preparation ................................................................................................................................................................... 36 Figure 2-4 Average contact frequency of potentially contaminated surfaces observed per task cycle during drug transport ....................................................................................................................................................................... 37 Figure 2-5 Average contact frequency of potentially contaminated surfaces observed per task cycle during drug administration .............................................................................................................................................................. 38 Figure 3-1 Average cyclophosphamide (CP) contamination levels (in ng/cm2) on the 55 different categories of surfaces sampled stratified by the stage of the hospital medication system ................................................................ 59 Figure 3-2 Box plot and summary statistics of cyclophosphamide (CP) surface contamination levels (in ng/cm2) by stage of hospital medication system ............................................................................................................................ 60 Figure 4-1 Participant recruitment flowchart .............................................................................................................. 86 xiv LIST OF ABBREVIATIONS 4-keto 4-ketocyclophosphamide AM Arithmetic mean BC British Columbia BSC Biological safety cabinet Carboxy Carboxyphosphamide CP Cyclophosphamide DNA Deoxyribonucleic acid Ethyl-CP N-dechloroethylcyclophosphamide GM Geometric mean GSD Geometric standard deviation HPLC-MS/MS High-performance liquid chromatograph- tandem mass spectrometry IV Intravenous LOD Limit of detection NIOSH National Institute for Safety and Health xv ACKNOWLEDGEMENTS Numerous people and agencies need to be recognized for their efforts in providing guidance and support to complete this project. I would like to initially thank all the participants and their respective departments/facilities for agreeing to be part of this study – without their consent, there would be no data. I would like to thank my Thesis Committee members for their guidance and advice throughout the project. Specifically, Kay Teschke for her patience, attention to detail, constructive commentary and positive attitude. To Scott Venners for his insight on biomarkers and for providing a much needed “non-hygiene” perspective. Lastly, to Paul Demers for his knowledge of workplace carcinogens and his understanding as to how the current project relates to the grand scheme of things. Many thanks to the invaluable contributions made by the various research assistants that were involved in some capacity on this project. They were: Claire Pitcher, Jennifer Shum, Alexandra Barzan, Louise Hughes-Rhodes, Pearl Siganporia and Sarah Chiarello. I also need to recognize the following healthcare organizations for lending their support to the project: British Columbia Nurses’ Union (BCNU), Hospital Employees’ Union (HEU), Health Sciences Association of British (HSA) Columbia and Health Employers Association of British Columbia (HEABC). In addition, many individuals contributed to this study so that it could be successfully completed. These key individuals and their role in the study are summarized in the table that follows: xvi Name and organization Contribution Dr. Winnie Chu and analytical laboratory staff, SPPH Laboratory analyses of samples. Dr. George Astrakianakis, SPPH Assisted with securing union support and reviewed grant application. Prescillia Chua, Fraser Health Study liaison at one of the participating health authorities. Dr. Hui Shen and Hind Sbihi, SPPH Provided guidance with respect to statistical analyses. Dr. Robin Ensom, PHSA Initial support of study and reviewed grant application. Dr. Lynne Nakashima, BC Cancer Agency Study liaison and provided insights into pharmacy operations. Christie Hurrell, SPPH Designed the study’s website Dr. Karen Bartlett, SPPH Lent sampling supplies and laboratory space. Miscellaneous healthcare workers Pre-tested study documents. Of course, this work could not be completed without operational funding provided by the WorkSafeBC Research Secretariat (RS2008-OG01). Scholarship support by The Canadian Institutes of Health Research (Doctoral Award), Michael Smith Foundation for Health Research (Junior Graduate Scholarship), WorkSafeBC Research Secretariat (Research Trainee), American Industrial Hygiene Foundation (Robert L. Harris Scholarship and Los Alamos Scholarship), Canadian Centre for Occupational Health and Safety (Dick Martin Scholarship Award) and others were greatly appreciated. Lastly, I would like to thank my fellow colleagues at the School of Population and Public Health (Occupational and Environmental Hygiene Theme) for their support in keeping my spirits positive through the trials and tribulations of pursuing a doctorate degree. In particular, I would like to acknowledge Imelda Wong, Hind Sbihi and Tracy Kirkham. xvii DEDICATION Although my love for knowledge has motivated me to pursue this project, I could not have completed the journey without the two greatest loves in my life, Sonca and Sasha. I therefore dedicate this to the two of you. 1 1 INTRODUCTION 1.1 Background The term antineoplastic is defined as “destroying, inhibiting, or preventing the growth or spread of neoplasms or tumours”.(1) Given these properties, antineoplastic drugs, also referred to as cytotoxic, cytostatic, anticancer or chemotherapy drugs, are primarily utilized for the treatment of cancer. Many of these drugs act by interfering directly with the deoxyribonucleic acid (DNA) or the DNA synthesis of tumour cells and, consequently, the proliferation of these cells decreases. However, because antineoplastic drugs are non-selective in nature, normal healthy cells may also be affected.(2) This results in potential side effects experienced by cancer patients, such as nausea and hair loss(3), but is tolerated because of the therapeutic benefits offered by antineoplastic drugs. The same cannot be said for healthcare workers tasked with providing care and treatment to those undergoing chemotherapy. As such, there is a risk potential to healthcare workers who contact these hazardous agents. Circumstances whereby work-related contact may occur include, but are not limited to, drug preparation, drug administration, and disposal of drug waste. The first study to demonstrate an occupational exposure risk of healthcare workers to antineoplastic drugs was in 1979; it reported elevated frequencies of mutagenicity in urine samples from oncology nurses.(4) Since then, numerous studies have provided evidence that healthcare personnel experience various health impacts, both acute and long-term, due to occupational exposure to antineoplastic drugs. Documented adverse health outcomes amongst exposed healthcare workers include alterations to genetic material (known as mutagenicity), reproductive toxic effects, and cancer.(5)(6)(7) 2 1.1.1 Release date of safe drug handling guidelines In response, a number of agencies have developed safe drug handling guidelines(8)(9)(10) to be used in healthcare facilities to minimize occupational exposure to antineoplastic drugs. One of the most often cited guidelines is the “Alert: Preventing Occupational Exposures to Antineoplastic and Other Hazardous Drugs in Health Care Settings” produced by the National Institute for Occupational Safety and Health (NIOSH) in the United States.(9) The NIOSH Alert was introduced in 2004 and included recommendations to minimize occupational exposure through the use of engineering controls (e.g. biological safety cabinets) during drug preparation, personal protective equipment (e.g. gloves and gowns) when handling antineoplastic drugs, and drug decontamination protocols. Many healthcare facilities, including those in Metro Vancouver, have adopted elements of the NIOSH Alert. Research published prior to 2004 would not reflect the changes in practices and procedures outlined in the NIOSH Alert. The ensuing literature review will therefore primarily refer to findings from articles published from 2004 onwards; however, certain landmark papers released prior to this will still be discussed. 1.1.2 Use of cyclophosphamide as marker drug Almost all studies examining environmental contamination and/or occupational exposure to antineoplastic drugs have measured a small number (usually fewer than five) marker drugs that act as surrogates to represent contamination to all antineoplastic drugs used in the facility. Although there are numerous antineoplastic drugs commercially available to which workers may be exposed, it is not feasible to measure every drug in assessments. Therefore, marker drugs are utilized. The criteria for selecting an appropriate marker drug are based on three factors: 1) the 3 toxicity of the drug, 2) the frequency of use of a drug, and 3) the sensitivity of the available analytical methods.(11) The marker drug most often employed is cyclophosphamide (CP) (Figure 1-1).(12) The reasons for selecting CP as a surrogate are as follows. First, CP is an alkylating drug that forms cross- links with DNA(13) and it has been shown to cause the following malignancies in patients following administration: bladder cancer, acute myeloid leukemia and skin cancer.(14) This lends sufficient evidence that CP is carcinogenic in humans and it is therefore categorized as a known human carcinogen, or Group 1 agent, by the International Agency for Research in Cancer.(15) Secondly, although first introduced almost 50 years ago, CP remains today as one of the most frequently administered chemotherapeutic agents for the treatment of various cancers, including breast, lymphoid and pediatric malignancies. In addition, CP is also employed in bone marrow transplants and for the treatment of different autoimmune conditions.(14) Data indicate that CP was the most widely prescribed antineoplastic drug in the province of British Columbia (BC) in 2009.(16) Its high frequency of use has been noted in other countries as well.(17)(18)(19)(20) Lastly, not only is CP an analyte measured in almost all laboratory methods developed for exposure assessments(21)(22)(23)(24)(25)(26)(27)(28)(29), but recent advances in analytical methods, primarily the use of high-performance instrumentation such as high-performance liquid chromatography-tandem mass spectrometry(30), have resulted in limits of detection for CP as low as 0.05 picogram per square centimetre (pg/cm2) for surface wipes(31) and 0.01 nanogram per milliliter (ng/mL) in urine samples.(30)(31) Given the above, CP was selected as the marker drug in the current study. Therefore, for comparison purposes, only those articles that examined CP will be discussed. Note: where the 4 term “antineoplastic drug(s)” is used in this document, it refers to CP as one of the marker drugs, unless otherwise specified. 1.2 Pilot study conducted to ascertain nature of problem at local facilities Despite the implementation of safe drug handling guidelines, studies have consistently found antineoplastic drug contamination on surfaces in healthcare facilities.(32)(33)(34)(35)(36) (among others) Such widespread contamination presents an opportunity for load transfer of drug residual if an individual contacts the contaminated surface.(37) This is noteworthy because dermal contact, as opposed to other routes of exposure, is considered the main route of exposure of healthcare workers to antineoplastic drugs.(11)(38)(39)(40) Since occupational exposure assessments of healthcare workers in BC had not been performed previously, I designed a pilot study to measure drug contamination on work surfaces and ascertain the potential dermal exposure risk at select hospital pharmacies within Metro Vancouver. The pilot study found that 14 of the 23 surfaces wiped (61%) were contaminated with either CP or methotrexate, another commonly administered antineoplastic drug in BC. Furthermore, even among surfaces that were observed to have been cleaned, some had detectable levels of residual drug; in fact, some samples appeared to have post-cleaning concentrations greater than the corresponding pre-cleaning levels.(41) The pilot study also demonstrated that 28% of hand wipe samples from pharmacy personnel had detectable levels of at least one drug product. Of note, workers who were not responsible for drug preparation on the day of sampling had measurable levels of antineoplastic drugs.(42) Given the findings from the pilot study, I strongly believed that this issue warranted further examination. To assist in the design of a large-scale study, a literature review was conducted to identify some of the knowledge gaps related to occupational exposure to antineoplastic drugs. 5 1.3 Review of existing literature 1.3.1 Only select departments have been assessed for surface contaminationa According to the Association paritaire pour la santé et la sécurité du travail du secteur affaires sociales (ASSTSAS)(10), which released a safe handling guide in 2008, there is a process flow of antineoplastic drugs within a facility known as the hospital medication system. Stages of the hospital medication system that are applicable to local sites include initial delivery of the drugs to the facility (delivery) and then being mixed to the correct dosage in the pharmacy (drug preparation). Subsequently, the prepared drugs, usually in intravenous (IV) form, need to be transported to patient units (transport to ward) where the drugs are administered to cancer patients (drug administration). Lastly, the drug containers, such as the manufacturer vials and IV bags, need to be disposed of in an appropriate manner (disposal). To determine antineoplastic drug contamination levels on work surfaces, wipe samples are often collected.(11) However, a review of the literature indicates that wipe sampling of surfaces has been conducted for select departments within a hospital rather than in all stages of the hospital medication system. Surfaces situated in the drug preparation department have been studied in numerous countries. Some of the more common surfaces examined were biological safety cabinets (inside of which the drugs are prepared), countertops, waste containers, trays and the floors of the drug preparation area.(32)(34)(36)(43)(44)(45)(46)(47) CP was quantifiable on virtually all surfaces examined in the drug preparation area. Of concern is that drug residual was found despite the use of a biological safety cabinet and, in some instances, measurable levels of CP were detected outside of the biological safety cabinet.(20)(35)(48) a Although closed system drug transfer devices (CSDTD) have been shown to be effective in minimizing surface contamination, they were not used at any of the local sites. Therefore, articles related to surface contamination and CSDTD usage were not cited. 6 Drug vials are commonly handled in the drug delivery stage and/or the drug preparation stage. Several studies have demonstrated that both the vials and the outer packaging received from drug manufacturers are contaminated.(32)(49)(50)(51)(52)(53) More locally, Touzin et al. reported contamination on the external surfaces of CP vials from the two manufacturers on the Canadian market.(54) The extent of external contamination may range from a few vials to all vials within a shipment. Connor et al. recommended that the impact of external contamination of vials on the overall contamination levels and the subsequent risk of occupational exposure should be explored further.(51) In a recent study, Hama et al. confirmed that healthcare workers who had just touched a vial were exposed to CP.(52) A smaller number of studies have looked at surface contamination in the patient care units where antineoplastic drugs are administered.b Connor et al. collected wipe samples from areas where the drugs arrived at the nursing stations, carts and trays for transporting or storing drugs, areas where IV bags were hung before use, chairs, tables, and floors in patient rooms, floors in patient washrooms, waste containers, and utility rooms. CP was detected on almost all surface areas.(36) Sottani et al. sampled similar surfaces within patient administration areas in Italian healthcare facilities and reported measurable levels of CP on all surfaces sampled.(44) A Japanese study found CP contamination on 50% of the surfaces including those from the patient ward area.(43) Kromhout et al. reported contamination of patients’ toilets, surfaces near the patients’ beds, utility rooms (urinal washer), and corridors.(39) Cavallo et al. found the average contamination level in the patient care area was higher than in the pharmacy.(55) Hedmer et al. reported the highest surface contamination levels were on the doors in the patient washrooms and the utility rooms.(31) b Surface contamination in patient units is not only due to the act of drug administration but it may also originate from patients undergoing chemotherapy who excrete the drugs via urine and other body fluids. 7 With documented evidence of contamination on the drug vials and in both the drug preparation and administration areas, it is possible that the contamination can spread to other areas of the hospital. Acampora et al. supported this notion by suggesting that the contamination of floors and storage shelves demonstrate the potential for these compounds to spread throughout a facility.(56) Crauste-Manciet et al. theorized that the removal of the drug vial cap could contribute to the spread of contamination.(20) In addition, the authors stated that the outside of the IV bags and workers’ hands could be a major route for spreading drug residual. This is due to external contamination and potential permeation of antineoplastic drugs through gloves, respectively.(20) Another study suggested that drug contamination on floors may be spread by the footwear of workers or during cleaning of the floors and that external contamination of gloves may be transferred to other surfaces/objects.(57) It is therefore conceivable that various surfaces at every stage of the hospital medication system may have drug contamination. However, a review of the literature could not locate studies that examined surface contamination throughout a hospital and its entire medication system. 1.3.2 Dermal contact of surfaces sampled is unknown The previous section provides evidence that many surfaces within a hospital have measurable levels of antineoplastic drug contamination. What is not clear, however, is the actual frequency of skin contact that workers may have with these contaminated surfaces and, in turn, their exposure risk. One of the earliest studies examining surface contamination was in 1992 by Sessink et al. In this study, the authors looked at areas where cytostatic drug contamination ‘may occur’ during preparation due to possible spills and distribution of drugs during handling. Sampling areas chosen included the laminar airflow hood and the floor of the preparation room.(58) In a follow-up study, the authors expanded their scope and, in addition to the 8 pharmacy, examined the preparation and administration areas of the outpatient department and the oncology department. Additional surfaces included the working tray of the hood, floor of the preparation and administration room, floor of oncology department, front of hood, tables, and sinks.(59) Many subsequent studies have followed the sampling scheme established by Sessink et al.(47)(60) Other studies have also selected surfaces based on potential sources of contamination in the drug preparation and/or drug administration units.(31)(32)(46)(56) In some instances, no rationale was provided by the authors for selecting surfaces that were assessed for levels of drug residual.(61)(55)(48)(43)(62) In a study by Schmaus et al., some rationale was provided for selecting the following surfaces: 1) the floor in front of the biological safety cabinet, 2) the floor in the preparation room’s central area, 3) bench-top surfaces on which the drugs and materials were placed before IV mixtures were prepared, 4) surfaces on which IV containers were placed, 5) storage shelves, 6) transport boxes, and 7) waste bins. According to the authors, locations 1 – 3 were chosen because contamination can be found in areas where drug vials are unpacked and disinfected before preparing the infusions; IV bags containing cytotoxic drugs may lead to contamination if they are inappropriately handled and transported; and the surfaces of waste bins are often contaminated when cytotoxic waste is not handled properly. Lastly, the authors surmised that antineoplastic drug contamination has been found on storage shelves that hold contaminated packages of these agents.(24) The lone study that used a sampling scheme reflective of dermal contact was by Castiglia et al. The authors employed an on-the-spot investigation to identify which surfaces may be potentially contaminated and this information was used to form the basis of their subsequent sampling strategy.(63) However, this investigation was conducted in the drug preparation room only. 9 As dermal contact is the main route of occupational exposure, it would appear sensible to determine the contamination levels of those surfaces that workers are likely to contact. In fact, this was suggested by Turci et al. who stated that contamination routes in working areas have to be established and exposure routes, especially skin absorption, in hospital personnel have to be identified.(11) However, the literature does not appear to reflect this notion with surfaces selected based on probability of contamination as opposed to actual contact frequency. 1.3.3 Underestimation of exposure There is a likelihood that the literature has underestimated the number of healthcare workers that are potentially exposed to antineoplastic drugs as well as the actual exposure levels, particularly in urine samples. This section outlines these two shortcomings that currently exist in the literature. Number of workers overall at risk As with the surface contamination studies, occupational exposure studies have also been limited in their scope. A review of the literature indicates that studies examining occupational exposure to antineoplastic drugs have focused on select job categories. Many have looked strictly at pharmacists that directly handle antineoplastic drugs.(48)(52)(62) Other papers have examined nursing personnel as they are tasked with drug administration and also care for patients who undergo chemotherapy.(18)(64)(65) Pharmacy technicians, often responsible for drug preparation, are another cohort that has been cited in the literature.(37) In some cases, combinations of two or more of the aforementioned job categories have been studied.(17)(36)(44). Some studies that examined multiple cohorts also included a medical doctor(43)(35), a surgeon and an anesthetist(66), as well as cleaners (31)(67)(40). 10 Determining the exposure levels of pharmacists, nurses and pharmacy technicians appears warranted because these cohorts handle antineoplastic drugs on a fairly regular basis and often at high concentration/doses. Nevertheless, since it is hypothesized that surface contamination is likely throughout the entire hospital medication system, this does not preclude other job categories from being exposed. Sorsa et al. produced a list of potentially exposed healthcare workers which includes pharmacy personnel preparing the drugs, hospital staff involved with drug administration, physicians and nurses in patient care areas, cleaning and laundry personnel, scientists and laboratory personnel, and drug transport personnel.(68) However, exposure studies of healthcare personnel besides pharmacists, nurses and pharmacy technicians are limited. Fransman et al. demonstrated that cleaning personnel have dermal contamination.(67) Similarly, Kusnetz and Condon concluded that auxillary healthcare personnel, such as care aides, are at risk of exposure to antineoplastic drugs.(69) One way to identify the exposure risks of various healthcare workers is to look at exposures of personnel in different departments. Believed to be one of the first papers looking at several departments, Sessink et al. studied exposure in four departments. The departments were the clinical pharmacy department (preparation), outpatient department (preparation), outpatient department (administration) and oncology department (administration). Unfortunately, the authors examined the exposures of pharmacy technicians and nurses only.(59) Another means to examine exposure risks of various healthcare workers is to identify the risks associated with tasks involved with handling antineoplastic drugs. Fransman et al. measured dermal exposure to CP during the performance of five tasks: 1) preparation of antineoplastic drugs in the hospital pharmacy; 2) decanting patient urine; 3) washing the patient; 4) removing the sheets from the patient’s bed; and 5) cleaning the patients’ toilet on the oncology ward of the 11 hospital. However, the authors’ evaluation was limited to pharmacy technicians, oncology nurses and cleaners.(67) Since surface contamination is conceivable throughout the hospital medication system and not strictly isolated to the drug preparation and drug administration units, the total number of healthcare workers exposed to antineoplastic drugs may be much higher than previously thought.(70) However, a literature review was unable to find studies that simultaneously ascertained the exposure levels of healthcare personnel that work throughout the hospital medication system. Urinary contamination levels The quantification of antineoplastic drugs in urine samples is used to determine occupational exposure across all exposure routes, including dermal, inhalation or ingestion.(36) In particular, measuring CP in urine has proven to be an appropriate biomarker to estimate doses that have been taken up in the body.(71) Many studies have found measurable levels of CP in the urine samples of exposed workers indicating that absorption does indeed take place.(12) However, pharmacokinetic studies suggest that less than 20% of absorbed CP is excreted unchanged in the urine and the rest is metabolized and/or eliminated through other means e.g. feces, expired breath.(72) Metabolism of CP is influenced by factors related to the drug itself (e.g., dosage, route of administration, and drug combination) as well as the person (e.g., age, gender, and hepatic function).(72) This results in both inter- and intra-individual variations in the pharmacokinetics of CP.(71) Given this, if CP alone is measured, the urine levels reported likely underestimate the actual amount absorbed by a worker. To address this shortcoming, both Sottani et al.(44) and Turci et al.(66) suggested that quantifying the amount of parent product, CP, and one or more of its urinary metabolites would 12 be a more accurate reflection of exposure (Figure 1-2). However, a review of the literature did not reveal any studies with more than 20 participants that measured CP and its metabolites. 1.3.4 Dearth of data related to the determinants of contamination or exposure Numerous studies have examined surface contamination levels while others have evaluated occupational exposure levels of antineoplastic drugs through hand wipes or urine samples. Few studies have examined the factors that may be associated with such contamination or exposure. By identifying factors that are associated with elevated or reduced contamination/exposure levels, control measures that specifically target these determinants can be implemented to minimize the risk potential. Surface contamination Only two studies have addressed determinants of surface contamination. One found no association between the number of drug handling events and the proportion of surface wipe samples with measurable drug residual.(36) The other study concluded no apparent correlation between the surface contamination levels and the amount of CP prepared annually, the age of cytotoxic suite or the biological safety cabinet, decontamination and cleaning procedures, or the staff preparing chemotherapy.(34) These studies examined only the drug preparation and drug administration areas. Determinants of contamination for surfaces throughout the hospital medication system have yet to be evaluated. Occupational exposure With respect to occupational exposure in general, a number of studies have indicated various determinants that lead to an elevated risk of exposure or factors that have no association with exposure. Testa et al. concluded that age, gender and the experience of a worker are not likely 13 determinants of exposure risk.(73) In their review, Ritchie et al. found that lack of compliance with safe work practices was associated with an elevated risk of exposure.(6) Factors related specifically to dermal exposure have also been examined. Dermal contamination levels have been found to be higher if a worker handles drug vials.(7)(40) Friese et al. found that organizational factors, such as adequate staffing and a check-and-balance system for drug administration, reduced potential skin exposure.(74) Reported determinants of urinary exposure differ from those for dermal exposure. Schreiber et al. found the following factors influenced uptake of antineoplastic drugs: the amount processed (for workers responsible for drug preparation); the number of preparations handled (for workers responsible for drug preparation and those who assisted in the drug preparation); and hazardous waste that is stored in garbage bins with a lid that can be opened.(75) The importance of the amount handled was supported in a recent study(51) but was also contradicted by Favier et al. who found no correlation between the amount handled and urinary excretion.(76) Rekhadevi et al. found that age, years of exposure and duration of handling antineoplastic drugs per day were positively associated with urinary CP concentrations, though age was the lone variable that was statistically significant.(64) These determinants studies included the following select hospital departments and/or healthcare personnel: pharmacy department(40)(51)(75), patient administration units(40)(64), pharmacists(40)(75), pharmacy technicians(40)(75), nurses(40)(64)(74) and cleaners.(40) Determinants of occupational exposure for all at-risk job categories throughout the hospital medication system have yet to be evaluated. 14 1.3.5 Occupational exposure likely to continue and possibly trend upwards According to the Canadian Cancer Society, the number of cancer cases is expected to steadily rise as the Canadian population increases and ages.(77) With more cancer cases, there will be an increase in the number of patients seeking treatment with antineoplastic drugs. Therefore, healthcare workers’ exposure to these hazardous agents will persist with a strong possibility that it will increase in the near future. In addition, antineoplastic drugs are increasingly being utilized for the treatment of other diseases besides cancer including non-malignant diseases such as aplastic anemia (requiring bone marrow transplants) and arthritis. In turn, the potential for exposure of healthcare providers to these drugs will increase.(7) Furthermore, since the use of antineoplastic drugs has expanded into other specialties, the number of hospital workers who are not properly trained in their safe handling has also increased.(9) 1.3.6 Lack of consistent and/or effective policies to reduce occupational exposure In addition to addressing gaps in the literature, examining the issue of healthcare workers’ exposure to antineoplastic drugs is also likely to have policy implications. Currently, there is no recognized occupational exposure limit for antineoplastic drugs whereby an exposure concentration is deemed to be safe and unlikely to result in toxic effects. In instances where exposure limits are not available for carcinogens like CP, occupational hygienists have adopted the principle of maintaining exposures as low as reasonably achievable (ALARA). However, the challenge lies in defining the ALARA level as this cannot be achieved until an exposure assessment is completed to ascertain current exposure levels. 15 Despite the implementation of control measures indicated in safe handling guidelines, both surface contamination and worker exposure continue to occur. Through the identification of determinants of contamination/exposure, it may be possible to strengthen these existing guidelines in order to reduce the risk. 1.4 Rationale for current study and research questions Based on the results from the pilot study as well as knowledge gaps identified in the literature, a full-scale study to examine healthcare workers’ exposure to antineoplastic drugs was developed and submitted for funding in February 2008. The following were the key research questions that the study was designed to answer: • Is antineoplastic drug contamination found on surfaces located throughout the hospital medication system? • Are healthcare workers throughout the hospital medication system occupationally exposed to antineoplastic drugs? • What are the factors that are associated with surface contamination and occupational exposure (dermal and urinary contamination) throughout the hospital medication system? To answer these questions, this study had the following specific research objectives: Objective 1: Identify surfaces most likely contaminated and the job categories potentially at risk of exposure to antineoplastic drugs throughout the hospital medication system (Chapter 2); Objective 2: Quantify the surface contamination levels throughout the hospital medication system and identify determinants associated with surface contamination (Chapter 3); 16 Objective 3: Assess the dermal CP contamination levels of at-risk healthcare job categories and identify determinants associated with dermal exposure (Chapter 4); and Objective 4: Determine the urinary concentrations of CP and three of its metabolites in at-risk healthcare job categories and identify determinants associated with urinary contamination (Chapter 5). 1.5 Dissertation structure This dissertation consists of six chapters: this introductory chapter, four research chapters and a concluding chapter. The research chapters were written with the intention of submission to peer- reviewed journals. The four research chapters and the concluding chapter are outlined below: Chapter 2 Identification of surfaces contaminated and job categories exposed throughout the hospital medication system The purpose of this phase of the research was to conduct site observations in order to determine the hospital medication system at each participating site and ascertain which surfaces are potentially contaminated with antineoplastic drugs. It also identified the various healthcare job categories that are likely to contact the contaminated surfaces. The overall goal of this phase was to assist in planning an appropriate sampling strategy for the subsequent research chapters. Chapter 3 Evaluation of surface contamination throughout the hospital medication system and identification of determinants of such contamination Wipe samples were collected from those surfaces identified in Chapter 2 as being potentially contaminated and subsequently analyzed to quantify the amount of CP. The goal of this research was to test the theory that surface contamination exists throughout the hospital medication 17 system. Another outcome of this chapter was the identification of hospital-related factors that are associated with surface contamination. Chapter 4 Evaluation of healthcare workers’ dermal exposure throughout the hospital medication system and identification of determinants of such exposure Workers from those job categories identified in Chapter 2 as being potentially exposed were recruited to participate in the study. Utilizing a similar sampling method as employed in Chapter 3, participants’ hands were wiped and subsequently analyzed to measure the amount of CP. This research aimed to determine the dermal contamination levels of healthcare workers at risk of exposure to antineoplastic drugs due to their role in the hospital medication system. Participants were also provided with a self-administered questionnaire regarding knowledge and training and were surveyed at the time of sample collection about known contact with CP. Variables from the survey instruments were used to identify determinants of dermal exposure. Chapter 5 Evaluation of healthcare workers’ urinary contamination level throughout the hospital medication system and identification of determinants of such contamination Dermal samples only provide the level of external contamination whereas urine samples serve to confirm that the drug has been absorbed and, therefore, have the potential to result in chronic adverse health effects. The same workers who provided hand wipe samples were also asked to submit urine samples. Urine samples were analyzed for CP and three of its metabolites. As in the previous chapters, determinants associated with urinary contamination were identified and discussed. 18 Concluding chapter The final chapter provides an overview of the key findings, lists a number of recommendations resulting from the study that may minimize the risk potential, discusses the strengths and limitation of the study and, lastly, summarizes some possible future research arising from the current study. 19 Figure 1-1 Structural and molecular formula and relative molecular mass of cyclophosphamide(15) 20 Figure 1-2 Metabolism of cyclophosphamide. The inactivation pathways are depicted horizontally, while cyclophosphamide activation is shown vertically.(13) 21 2 IDENTIFICATION OF SURFACES CONTAMINATED AND JOB CATEGORIES EXPOSED THROUGHOUT THE HOSPITAL MEDICATION SYSTEM 2.1 Synopsis Antineoplastic drugs are known to circulate within a healthcare facility referred to as the hospital medication system (process flow of drugs from cradle-to-grave) which consists of several different departments involving various healthcare job cohorts. However, virtually all environmental contamination and occupational exposure studies have restricted their assessment to select areas only, namely the drug preparation and the drug administration units. A number of studies have suggested that drug contamination may spread to other areas resulting in exposure to a broader range of healthcare workers. The purpose of this chapter was to identify which surfaces throughout the hospital medication system may be contaminated with antineoplastic drugs and, in addition, ascertain which healthcare job categories are likely to contact these surfaces. Repeated passive observations were conducted at each of the participating sites. During each observation, potentially contaminated surfaces were documented and both the job title and frequency of contact with the surface were recorded. The hospital medication system at each of the sites consisted of five stages: delivery, drug preparation, transport (of drugs) to ward, drug administration and disposal. Various surfaces at every stage were believed to have drug residual because a drug container was placed on it or a worker with contaminated gloves/hands touched the surface. Pharmacy receivers, pharmacy technicians, pharmacists, and nurses were consistently found to be in contact with contaminated surfaces at all participating sites. Up to 11 job categories per site may be at risk of exposure to antineoplastic drugs. 22 2.2 Introduction Antineoplastic (cytotoxic) drugs are widely used agents for the treatment of cancer. Some of these drugs act by interfering directly with the deoxyribonucleic acid (or its synthesis) of tumour cells and thereby interrupt their growth. Unfortunately, antineoplastic drugs are generally non- selective and therefore normal (non-tumour) cells may also be damaged which, in turn, results in toxic effects.(2) Given this, there is a risk to healthcare workers who handle, prepare, and/or administer antineoplastic drugs. Numerous studies have examined antineoplastic drug contamination in healthcare facilities. Studies from several countries have demonstrated surface contamination of biological safety cabinets, countertops, cabinets and floors within the drug preparation area.(59)(60)(61)(24)(19) Detectable levels of environmental drug contamination have also been found in patient care areas where antineoplastic drugs are administered.(55) A recently published summary of surface contamination levels found in the literature reported that cyclophosphamide (CP) ranged in concentration from not detected to 3,834 nanograms per square centimeter (ng/cm2), which suggests that existing control measures are not effective in reducing contamination levels.(33) In the absence of any current occupational exposure limits for CP (and most other antineoplastic drugs), it is therefore important to minimize contamination. A literature review revealed that these surface contamination studies have primarily focused on two departments within a healthcare facility – the pharmacy, where the drugs are prepared, and the administration units where the prepared drugs are given to patients. The emphasis on these two departments is warranted since direct handling of the drugs is expected during both preparation and administration. However, given the fact that the drugs need to be initially delivered to the pharmacy, then transported to the wards and eventually disposed as part of the 23 hospital medication system (process flow of drugs throughout a healthcare facility from cradle- to-grave), it is conceivable that other areas of a facility may have drug residual and, therefore, the number of healthcare workers at risk of exposure is underestimated. The potential for other healthcare workers to be occupationally exposed to antineoplastic drugs is supported in the literature whereby mechanisms of drug contamination spread have been proposed. Surface contamination may arise as early as the facility receiving stage in the hospital medication system as it has been documented that drug vials are often contaminated on the outside.(49)(50)(32) It is also possible that drug residue is spread by the footwear of workers or during cleaning of floors and that external drug contamination on gloves may be transferred to other objects or surfaces.(57) Overall, this suggests a need to examine the healthcare facility as a whole, not just the pharmacy and drug administration units, to determine the extent of antineoplastic drug contamination from the point at which the agents are received at the facility through to disposal or excretion. To our knowledge, no existing literature has investigated this issue. The current study aims to identify surfaces throughout the hospital medication system whereby antineoplastic drug contamination may be possible and to ascertain the various healthcare job categories that may be at risk of dermal exposure to antineoplastic drugs via contact with the contaminated surfaces - not just drug administration nurses and pharmacy personnel. 2.3 Methods 2.3.1 Selection of participating sites Participating sites were selected from healthcare facilities situated within the Metro Vancouver area of British Columbia, Canada that prepare and administer CP, the marker drug in this study. 24 The sites finalized for inclusion were determined by asking a pharmacy member from each participating health administration authority which of their facilities are the largest users of CP on an annual basis (based on overall frequency of compounding). In total, five major acute care hospitals and one cancer treatment centre participated in the study. 2.3.2 Informant interviews Interviews with key informants were conducted in order to understand the site-specific hospital medication system and to predict how and where a worker may be exposed to antineoplastic drugs. Key informants included supervisors/managers, clinical nurse leaders, and team leaders such as the senior pharmacy technician. At all sites, the initial interview was conducted in the pharmacy department as their personnel would be familiar with how the drugs arrive at their department and where the prepared drugs are transported for eventual administration. Subsequently, additional interviews were scheduled with those departments identified by the pharmacy informant as being part of the hospital medication system. All departments at each site involved in the hospital medication system were interviewed with the exception of housekeeping which is operated by the same external contractor at each site; the company declined to participate in the study. The duration of each interview was at least twenty minutes. All interviewees were asked questions related to the shift when antineoplastic drugs are primarily handled, prepared or administered; circumstances under which workers may be exposed to antineoplastic drugs; and the likely job categories which may be at risk of exposure. In addition, pharmacy personnel were questioned about their understanding of the hospital medication system at their site. 25 2.3.3 Site observations Passive (non-intrusive) site observations at each site were conducted by members of the research team to visually establish the hospital medication system and to identify which surfaces/objects may be contaminated with antineoplastic drugs as well as those job categories potentially at risk of dermal exposure via contact with the contaminated surfaces.(78) An employee was considered “at-risk” if they physically handled the drugs, contacted a potentially drug- contaminated surface/object, or used an object previously touched by another worker suspected of having drug-contaminated hands/gloves. A standard observation checklist was developed and used to record: a) the surfaces/objects which came into contact with the drug products, b) the job category of the worker that contacted the drugs and/or the contaminated surfaces and c) the associated frequency of contact of surfaces/objects for each worker (see Appendix A). At all sites, the observations began in the Pharmacy department where antineoplastic drug preparation was observed. We then followed the drugs as they were transported to the unit where they would be administered. Where prepared drugs were delivered to more than one unit, we randomly selected one unit to follow the drug. We conducted observations of the other unit(s) on separate occasions. There was considerable variation with respect to delivery times of the antineoplastic drugs to the healthcare facility. To accommodate this, we scheduled site visits to specifically observe the receiving process but not necessarily review other areas of the hospital medication system. Overall, in order to understand the entire hospital medication system, each site was observed on at least five separate occasions over a course of twelve months starting in June 2009. For each site visit, at least one task cycle was observed in each department. Examples of “one task cycle” include observing a porter picking up a drug order and delivering to the drug administration unit 26 for the drug transport stage; and observing the pharmacy technician compound the chemotherapy drug dose for one patient during drug preparation. By performing repeated observations the sequence of each site’s hospital medication system was determined and the job categories with potential risk of exposure to antineoplastic drugs were identified. 2.3.4 Data analysis The contact frequency of a surface/object was averaged over the number of observation periods at each site and then ranked by order of frequency for each stage of the hospital medication system. For the drug delivery as well as the transport to ward stages, the top three most frequently contacted surfaces were presented in the frequency bar graphs. With respect to the drug preparation and drug administration steps, the top five most frequently contacted surfaces are presented. The rationale for reporting a varying number of surfaces is that drug delivery and transport to the ward are more standard processes with less variability. In addition, both tasks are less likely to be influenced by individual and/or patient factors than either drug preparation or drug administration and therefore fewer surfaces/objects are contacted overall. 2.4 Results 2.4.1 General hospital medication system As shown in Figure 2-1, the general sequence of the hospital medication system for each of the participating facilities has a minimum of the following six stages: 1) delivery of the antineoplastic drugs to the facility, 2) drug preparation, 3) transport to ward, 4) drug administration, 5) disposal and, 6) waste retrieval. As disposal of the waste containers and waste retrieval is performed by private companies that declined to participate in the study, more specific details for these two steps are unavailable. 27 2.4.2 Description of participating facilities Table 2-1 provides a descriptive summary of each of the participating facilities. All six sites were situated in an urban centre and, at every facility, CP is one of the top five most frequently compounded antineoplastic drugs (For example, Site E alone purchased 1,715 grams of CP in 2011). All facilities were acute care hospitals except Site F which is a cancer centre. Sites B, E and F have the drugs initially delivered to the shipping/receiving department whereas the remaining sites have the drugs delivered directly to the pharmacy department. With the exception of Site D, antineoplastic drugs are prepared in dedicated isolation rooms situated in the pharmacy department. At Site D, antineoplastic drugs are prepared adjacent to an area for drugs that are not used for chemotherapy. The prepared drugs are delivered by various job categories to the administration wards, with one site (Site F) having up to three different job categories performing this task. Drugs are either administered in an out-patient clinic or within in-patient wards. The antineoplastic drugs are administered in three wards at sites D and E. In all instances, the antineoplastic drugs are disposed of within the pharmacy (the manufacturers’ vials) as well as in the drug administration wards (the intravenous bags). In the “notes” section of Table 2-1, it can be seen that each facility is designed slightly differently from the others. In most instances, the pharmacy, where drug preparation takes place, is situated on a different floor from where the drug administration takes place. 2.4.3 Contact frequency of work surfaces/objects Figures 2-2 to 2-5 display the frequency of surfaces that came into contact with antineoplastic drugs and/or potentially drug-contaminated surfaces contacted by healthcare workers for each stage of the hospital medication system. According to Figure 2-2, the box cutter is the most 28 frequently contacted object during drug delivery as it is used to open packages of drug shipments. The cart is the second most frequently contacted surface during this stage as it is used to deliver drug shipments from shipping/receiving to pharmacy as well as from pharmacy receiving to the main pharmacy (where the drugs are stored). With respect to drug preparation, Figure 2-3, the biological safety cabinet is the most frequently contacted surface as antineoplastic drugs are prepared in these cabinets at all sites. The next most frequently contacted object during drug preparation is a writing instrument found inside the biological safety cabinet. The writing instrument is used by the pharmacy technician to label the prepared drugs and/or verify drug dosages. Following these two surfaces/objects, there is a great deal of variability with respect to what is contacted and how frequently they are contacted. In the third stage of the hospital medication system, drug transport within the facility, Figure 2-4 shows that the bin/drawer where drugs are held for pick up is the most frequently contacted surface during the drug transport stage. It should be noted that Site B is not represented in this figure because drug transport is not required at this site since the pharmacy is immediately adjacent to the out-patient clinic. During drug administration (Figure 2-5), the intravenous hook/pump is the object that is overwhelmingly contacted by workers because virtually all of the drugs observed were in solution form and had to be administered intravenously using the mechanical pump. Similar to the drug preparation stage, there is a subsequent assortment of surfaces which are contacted across the sites during drug administration. Note that there is no corresponding bar graph for the disposal stage as only a total of three surfaces were found to be contacted by hospital employees (recall that all sites have contracted their housekeeping services to a company that declined to participate in the study). These three 29 surfaces, all from Site E, were an elevator button (to waste holding floor), a door handle (to waste holding room) and a biohazard cart. 2.4.4 Job categories at risk of exposure Table 2-2 summarizes the stages of the hospital medication system with the corresponding job categories that may be at risk of exposure at each of the participating sites. The pharmacy receiver, pharmacy technician, pharmacist, and nurse are potentially exposed to antineoplastic drugs at all sites. During transport of the prepared drugs to the ward, various job categories are at risk of exposure via handling of the antineoplastic drugs. Among the stages of the hospital medication system, drug administration has the most job categories (six) at potential risk of exposure. With respect to disposal of the drug products, Site E had a unique job category known as “biopackers” who are responsible for transferring the sealed waste containers from a holding area to a waste disposal room, where the containers would subsequently be picked-up by a waste disposal company. Based on the information provided by the key personnel during the interviews, it is estimated that over 500 workers at the six participating sites may be occupationally exposed. The distribution of the various job categories at each site that are potentially exposed is shown in Table 2-3. 2.5 Discussion Our study suggests that surface contamination may occur at every stage of the hospital medication system within a healthcare facility. As a result, the potential for occupational exposure to antineoplastic drugs also occurs at every stage. Those job categories most likely 30 exposed are pharmacy receiver, pharmacy technician, pharmacist, and nurse as these cohorts were consistently observed to be in contact with antineoplastic drugs at each of the six participating sites. Up to 11 job categories (not including housekeeping) per site are potentially at risk of exposure. The characteristics of a site that results in the most number of job categories exposed include: a) having the drugs initially delivered to the shipping/receiving department (as opposed to directly to the pharmacy), b) having multiple job categories responsible for transport of prepared drugs to the ward(s) and, c) having more than one drug administration unit. Our study builds on the list of healthcare workers exposed to antineoplastic drugs due to handling activities developed by the National Institute for Occupational Safety and Health (NIOSH).(9) In addition to the cohorts specified by NIOSH, our results suggest that it is reasonable to add the following personnel to the existing list: unit clerk, porter, volunteer, ward aide, dietician, and biopacker. Although our study primarily focused on occupational exposure, it is conceivable that patients, their family members, and their friends are also at risk of exposure as suggested by Sorsa et al.(68) This is because communal objects such as elevator buttons, door handles and patient chair side tables, which may be contacted by non-hospital personnel, were observed to be frequently contacted in our study. As discussed previously, almost all studies related to antineoplastic drug contamination on surfaces have examined either the drug preparation area and/or the drug administration area. Based on a review of the literature, we were only able to find one study which examined surface contamination of multiple departments within a healthcare facility. The authors examined four departments within a facility, the clinical pharmacy, outpatient department (preparation), outpatient department (administration) and oncology department, and found detectable levels of 31 antineoplastic drugs in every department.(59) As our study consisted of multiple sites with repeated observations at each site, we believe that there is likely to be surface contamination in other departments in addition to the drug preparation and drug administration areas. With respect to the risk of exposure, Connor and McDiarmid proposed general handling guidelines to prevent occupational exposure to antineoplastic drugs at each stage of the hospital medication system that was indicated in the current study (except waste retrieval).(7) This implies that an exposure potential is probable throughout the entire medication system. 2.5.1 Limitations Among the study limitations, we were unable to ascertain the frequency of contact associated with housekeeping personnel as they are employees of a company that declined to participate in our study. We can surmise that housekeepers would contact the cytotoxic waste containers but cannot estimate the contact risk for other surfaces. As waste containers, which have been shown to be contaminated in other studies(36)(44), were frequently contacted during drug administration by nursing staff, we can infer that housekeepers do indeed face an exposure potential. Another limitation is the possible absence of certain job categories and/or surfaces despite repeated site observations. Nevertheless, since we performed the site observations on multiple days and observed different individuals, we are confident that we have captured a reasonably representative understanding of the hospital medication system, the job categories commonly at risk of exposure and the potentially contaminated work surfaces at each site. 2.5.2 Summary To our knowledge, this is the first study of its kind to examine the occupational exposure potential to antineoplastic drugs throughout the entire hospital medication system of a healthcare 32 facility. Based on the results of this study, we are now able to develop an appropriate sampling strategy for all job categories considered at risk of exposure. In addition, we have also gained an understanding of which surfaces may be contaminated with antineoplastic drugs and, in turn, allow us to conduct a detailed assessment of environmental drug contamination throughout the entire healthcare facility. This is important in order to validate that existing control measures are not only appropriate in reducing occupational exposure to antineoplastic drugs but that they are comprehensive in scope to protect all at-risk job categories. 33 Figure 2-1 Overview of hospital medication system within participating facilities Reprinted with permission of Safety and Health at Work. 5. DISPOSAL Pharmacy personnel and nurses place drug- contaminated material into designated waste containers. Waste containers removed from department by housekeepers (contracted company). 6. WASTE RETRIEVAL Drug waste is picked up by a hazardous waste contractor (contracted company). Inside healthcare facility Outside healthcare facility 1. DELIVERY Drugs are either delivered to Shipping/Receiving Department or directly to Pharmacy. 2. DRUG PREPARATION Drugs are prepared in biological safety cabinets in an isolated area within the Pharmacy (with exception of Site D) 3. TRANSPORT TO WARD The drugs are delivered by hand. 4. DRUG ADMINISTRATION Half of the sites administer the drug in one ward only. Remaining sites have drug administration in two or more wards. 34 Table 2-1 Description of participating facilities Site A B C D E F Type Hospital Hospital Hospital Hospital Hospital Cancer Centre 1. Drug Delivery (how vials are delivered from manufacturer) Direct to pharmacy Via shipping/ receiving to pharmacy Direct to pharmacy Direct to pharmacy Via shipping/ receiving to pharmacy Via shipping/ receiving to pharmacy 2. Drug Preparation* (where drugs are prepared) In isolated room In isolated room In isolated room In non-isolated room In isolation room In isolation room 3. Transport to Ward (job category tasked with transport) By porter By pharmacist By ward aide By porter or ward aide By porter or nurse By nurse or unit clerk or pharmacy personnel 4. Drug Administration (type of patient ward) 1 out-patient clinic 1 out-patient clinic 1 out-patient clinic 2 in-patient wards; 1 out- patient clinic 1 in-patient wards; 2 out- patient clinics 1 in-patient ward; 1 out- patient clinic 5. Drug Disposal (department where disposal occurs) Pharmacy and out- patient clinic Pharmacy and out- patient clinic Pharmacy and out- patient clinic Pharmacy and drug administration wards Pharmacy and drug administration wards Pharmacy and drug administration wards Notes (unique features of each site) Pharmacy and out- patient clinic are on different floors Pharmacy is adjacent to the out- patient clinic Pharmacy and out- patient clinic are on different floors Drug administration wards are in a separate building from pharmacy on three different floors In-patient ward is on same floor as pharmacy; out-patient clinics are in a separate building from pharmacy Out-patient clinic on same floor as pharmacy; in- patient ward is one floor below the pharmacy * isolated = room is designated strictly for antineoplastic drug preparation; non-isolated = room is open-concept with biological safety cabinets for preparing antineoplastic drugs and other hoods for preparing non-cytotoxic drugs Reprinted with permission of Safety and Health at Work. 35 Figure 2-2 Average contact frequency of potentially contaminated surfaces observed per task cycle during drug delivery 0 5 10 15 20 25 Counter top Box cutter Pen Bins where drugs are stored Cart Door handle Av er ag e co nt ac t f re qu en cy o bs er ve d pe r v is it Surface Surfaces Contacted - Drug Delivery Site A Site B Site C Site D Site E Site F 36 Figure 2-3 Average contact frequency of potentially contaminated surfaces observed per task cycle during drug preparation 0 20 40 60 80 100 120 Av er ag e co nt ac t f re qu en cy o bs er ve d pe r v is it Surface Surfaces Contacted - Drug Preparation Site A Site B Site C Site D Site E Site F 37 Figure 2-4 Average contact frequency of potentially contaminated surfaces observed per task cycle during drug transport 0 5 10 15 Door into/out of pharmacy Bin (to hold drugs for pick-up) Countertop Elevator buttons Stairwell door Pick-up light switch Pushplate to open doors Av er ag e co nt ac t f re qu en cy o bs er ve d pe r v is it Surface Surfaces Contacted - Transport to Ward Site A Site C Site D Site E Site F 38 Figure 2-5 Average contact frequency of potentially contaminated surfaces observed per task cycle during drug administration 0 20 40 60 80 100 Av er ag e co nt ac t f re qu en cy o bs er ve d pe r s ite v is it Surface Surfaces Contacted - Drug Administration Site A Site B Site C Site D Site E Site F 39 Table 2-2 Number of observed job categories and associated tasks with potential for dermal exposure at each healthcare facility Potentially at-risk job categories at each stage of the hospital medication system Participating sites where listed exposure pathway is found Handles drugs* Prepares drugs Administers drugs Contacts drug- contaminated surfaces 1. Delivery Shipper/receiver 3 sites (B, E, F) 0 sites 0 sites 0 sites Pharmacy Receiver All 6 sites 0 sites 0 sites All 6 sites 2. Drug Preparation Pharmacy Technician All 6 sites All 6 sites 0 sites All 6 sites Pharmacist All 6 sites 0 sites 0 sites All 6 sites 3. Transport to Ward Porter 3 sites (A, D, E) 0 sites 0 sites 3 sites (A, D, E) Nurse 2 sites (E, F) 0 sites 0 sites 2 sites (E, F) Pharmacist 2 sites (B, F) 0 sites 0 sites 2 sites (B, F) Unit Clerk One site (F) 0 sites 0 sites One site (F) Ward Aide 2 sites (C, D) 0 sites 0 sites 2 sites (C, D) 4. Drug Administration Nurse All 6 sites 0 sites All 6 sites All 6 sites Volunteer 0 sites 0 sites 0 sites 2 sites (A,B) Unit Clerk 3 sites (D, E, F) 0 sites 0 sites All 6 sites Clinic Pharmacist 1 site (A) 0 sites 0 sites 2 sites (A, C) Dietician 0 sites 0 sites 0 sites 2 sites (A,B) Oncologist 0 sites 0 sites 0 sites 1 site (A) 5. Disposal Nurses All 6 sites 0 sites 0 sites All 6 sites Pharmacist 5 sites (A, B, D, E, F) 0 sites 0 sites 5 sites (A, B, D, E, F) Pharmacy Technician 1 site (C) 0 sites 0 sites 0 sites Biopacker 1 site (E) 0 sites 0 sites 1 site (E) * includes shipments, drug vials, intravenous bags and waste containers Reprinted with permission of Safety and Health at Work. 40 Table 2-3 Number of workers per at-risk job category at participating facilities Department Job Category Site A Site B Site C Site D Site E Site F Subtotals Pharmacy Pharmacy Technician 6 FT; 0 FT; 2 PT 13 FT; 3 PT 7 FT; 2 PT 27 FT; 4 PT 7 FT; 4 PT 60 FT; 15 PT 6.0 FTE 1.8 FTE 15.1 FTE 7.85 FTE 30.2 FTE 9.0 FTE 69.95 FTE Pharmacy Receiver 3 FT 0 FT; 10 PT 2 FT; 4 FT 3 FT 1 FT 13 FT; 10 PT 3.0 FTE 0.25 FTE 2 FTE 4 FTE 3 FTE 1 FTE 13.25 FTE Pharmacist 4 FT; 2 PT 1 FT 11 FT; 9 PT 6 FT; 1 PT 31 FT; 7 PT 7 FT; 5 PT 66 FT; 25 PT 4.8 FTE 1 FTE 16.89 FTE 6.5 FTE 35.7 FTE 9.5 FTE 74.39 FTE Drug Administration Unit Nurse 4 PT 1 FT; 6 PT 6 FT; 1 PT 60 FT; 11 PT 54 FT; 26 PT 26 FT; 16 PT 147 FT; 64 PT 2.60 FTE 4.0 FTE 6.0 FTE 67.2 FTE 73.77 FTE 35.10 FTE 188.67 FTE Pharmacy technician (oncology unit) 1 PT 1 FTE N/A N/A N/A N/A N/A 1 PT 1 FTE Clinical Pharmacist (oncology unit) 1 FT 1 FTE N/A 1 FT 1.0 FTE N/A N/A N/A 2 FT 2 FTE Volunteer 7 PT 0.8 FTE 10 PT 5.0 FTE N/A N/A N/A N/A 17 PT 5.8 FTE Unit Clerk 2FT; 1 PT 3 FT; 1 PT Could not obtain 5 FT; 4 PT 1FT; 1PT 6 FT; 7 PT 17 FT; 14 PT 2.40 FTE 3.0 FTE 7. 83 FTE 1.20 FTE 8.0 FTE 22.43 FTE Oncologist 1 FT; 2 PT 1.8 FTE N/A N/A N/A N/A N/A 1 FT; 2 PT 1.8 FTE Ward Aide N/A N/A 1 FT 1.0 FTE 2 FT 2 FTE N/A 1 FT 1 FTE 4 FT 4.0 FTE Dietitian 2 PT 0.5 FTE 1 PT 0.4 FTE N/A N/A N/A 3 PT 0.9 FTE Patient Transport Porter 8 FT; 3 PT N/A N/A 2 FT; 1 PT 44 FT; 10 PT 54 FT; 14 PT 9.3 FTE 2.8 FTE 51.0 FTE 33.1 FTE Shipping/Receiving Shipper/Receiver N/A Could not obtain N/A Could not obtain 3 FT 3 FTE 4 FT 4.0 FTE 7 FT 7 FTE TOTALS 357 FT; 162 PT 424.29 FTE FT= full time; PT = part-time; FTE= full-time equivalent; N/A = not applicable to the site Reprinted with permission of Safety and Health at Work. 41 3 EVALUATION OF SURFACE CONTAMINATION THROUGHOUT THE HOSPITAL MEDICATION SYSTEM AND IDENTIFICATION OF DETERMINANTS OF SUCH CONTAMINATION 3.1 Synopsis The site observations reported in Chapter 2 indicated that surfaces situated throughout the hospital medication system might be contaminated with antineoplastic drugs. This chapter aimed to quantify the cyclophosphamide (CP) levels on these surfaces. In addition, determinants of surface contamination were identified. A weighted sampling approach was taken whereby more samples were collected from surfaces within the drug preparation and the drug administration stages than other stages. Moistened wipes were used to sample surfaces, which were subsequently analyzed to determine the concentration of CP using high-performance liquid chromatography-tandem mass spectrometry. Both descriptive and inferential statistics were performed. A manual backwards stepwise multiple regression was conducted to identify the determinants. Overall, 229 surfaces were sampled, with duplicates for most, resulting in 438 surface wipes. The mean CP concentration was 0.201 ng/cm2, the geometric mean 0.019 ng/cm2 and the geometric standard deviation 2.54, with a range of less than the limit of detection (LOD) (0.356 ng/wipe; LOD in ng/cm2 varied with surface area) to 26.1 ng/cm2 (64% of samples had concentrations less than the limit of detection). The drug preparation stage had the highest average contamination. Hospitals with more drug transport job categories had higher levels of surface contamination. Reported handling of CP and cleaning of surfaces did not appear to be associated with contamination. The drug preparation and drug administration stages of the hospital medication system and a higher number of job categories responsible for drug transport were factors associated with surface contamination. 42 3.2 Introduction Safe handling guidelines, which often include cleaning and decontamination procedures, have been developed for workplaces where antineoplastic drugs are present. Despite the implementation of these guidelines in healthcare facilities, many studies have demonstrated that antineoplastic drug contamination of various work surfaces still occurs.(56)(36)(31)(34) Such surface contamination means healthcare workers are potentially at risk of dermal contact – which is believed to be the primary route of occupational exposure to these hazardous agents.(11)(67) The literature provides evidence that occupational exposure to antineoplastic drugs may result in genetic damage, adverse reproductive effects, as well as an increased cancer risk.(5)(6)(7) Almost all prior studies have focused on examining surface contamination levels in the pharmacy department where the drugs are prepared and/or in the patient units where the drugs are administered. However, given that the drugs must initially be delivered to the facility, transferred to the pharmacy to be mixed, transported to a patient unit, administered and then disposed, surfaces in other areas of the hospital may also have drug residual.(79) To our knowledge, no single study has examined surface contamination throughout the hospital medication system (process flow of drugs within a facility from cradle-to-grave) and, subsequently, identified determinants of such contamination. In Chapter 2, we identified the hospital medication system at several facilities and conducted observations to identify potentially contaminated surfaces that are frequently touched by healthcare workers.(80) In this chapter, the surface contamination levels are quantified and factors that may be associated with surface contamination in the hospitals are examined. As cyclophosphamide (CP) is frequently administered at the participating sites and has been 43 extensively examined in other occupational exposure studies, it was used as the marker drug in this study. 3.3 Methods 3.3.1 Selection of surfaces to be sampled Five participating sites were acute care hospitals and one was a dedicated cancer treatment hospital. At all sites, antineoplastic drugs were prepared in Class II biological safety cabinets but none of the participating facilities used closed system drug transfer devices, e.g. PhaSeal®, for preparation. Passive (non-intrusive) site observations were performed to identify the hospital medication system for each facility and to identify those surfaces that were most frequently contacted by hands of healthcare workers throughout the hospital medication system; these methods and results are reported in detail elsewhere.(80) The medication system was sub-divided into five separate stages: 1) delivery, 2) drug preparation, 3) transport to ward, 4) drug administration and 5) waste disposal.(80) At each hospital, the five most frequently contacted surfaces from stages 1, 3 and 5 and the eight most frequently contacted surfaces from stage 2 and 4 were selected for sampling. The rationale for sampling a different number of surfaces is that drug preparation and drug administration included more complex tasks with greater variability, such as individual and patient factors, than the other stages and therefore more surfaces/objects were contacted overall. The contracted company that handled waste at all participating hospitals declined to participate in the study. This resulted in relatively fewer samples collected during the waste disposal stage. However, where certain elements of waste disposal were performed by hospital employees, e.g. transfer of cytotoxic drug waste to a holding station, we were able to collect wipe samples. 44 3.3.2 Wipe sample collection Sample collection took place between March 2010 and January 2011. Sampling days were based on availability of research team members as well as consent from all affected departments at each hospital. Duplicate samples of most surfaces were collected with at least a four-month lag between the two collection times (mean 174 days; range 123 to 288 days). A previously described protocol for surface sampling was employed(41) with some minor modifications. Briefly, we used a 10 cm x 10 cm plastic sampling template for flat surfaces having a surface area greater than 100 cm2. For other objects such as door handles and writing instruments, where the overall surface area was less than 100 cm2 and/or was not flat, the entire object was wiped and the surface area estimated by measuring the object’s dimensions. For non- flat objects with an overall surface area greater than 100 cm2, the area most likely to be contacted by an individual’s hands was demarcated for sampling. For instance, the handle of a cart was wiped rather than the entire cart. The research assistant wore a new pair of disposable gloves for each sample collected. A Kimwipe® (Kimberly-Clark Inc., Mississauga, ON) pre-moistened with 0.1 M ammonium acetate (Sigma Aldrich, Oakville, ON) was used to wipe each surface. All collected wipes were placed into separate vials and shipped on ice within 24 hours of sample collection to the analytical lab where they were stored at -20oC until analysis. Both travel and field blanks were collected for quality control purposes. 3.3.3 Wipe sample preparation After thawing the samples, 5.5 mL of 0.1 M ammonium acetate (Sigma Aldrich, Oakville, ON) solution was added to each vial. The wipes were sonicated for 20 minutes to extract the drug 45 residue and then placed into a disposable 10-mL syringe where the solution was squeezed out of the wipe into a 20-mL vial. One mL of the solution was removed and placed into a liquid- chromatography vial with 50 µL of internal standard, D4-cyclophosphamide (Bielefeld University, Bielefeld, Germany). 3.3.4 Wipe sample analysis Wipes were analyzed for CP by high-performance liquid chromatography-tandem mass spectrometry using an Agilent Technologies 6410 with a Zorbax XDB-C18 column (Agilent Technologies, Santa Clara, CA) and the electrospray ionization in the positive ion mode. The mobile phase was a gradient of 5 mM Ammonium Acetate:100% Methanol (A:B) and samples were run at a flow rate of 0.5 mL/min. A 5-point calibration curve was used and, for quality control purposes, a calibration standard was run for every 10 samples. For additional quality control, method blanks were included in the analysis and duplicate analysis took place for every tenth sample. The limit of detection (LOD), established using the 3:1 signal-to-noise ratio, was 0.356 nanogram per wipe (ng/wipe). The method recovery rate was 97% and each resulting sample concentration was adjusted to reflect this. The surface area of each wiped object was calculated and drug contamination levels were reported in nanograms per square centimetre (ng/cm2) as in similar studies.(11) 3.3.5 Supplemental data collection For each surface wipe sample, members of the research team conducted a brief survey of the healthcare worker who was in closest proximity to the surface sampled (see “Drug Handling and Cleaning Questions” in Appendix B). The following questions were asked: 46 • To your knowledge, was cyclophosphamide handled, prepared and/or administered prior to collection of the wipe sample (on this work shift)? • To your knowledge, was there a spill/leak of cyclophosphamide earlier in the day on this surface? • To your knowledge, was the surface/object cleaned prior to collection of wipe sample (end of previous shift until present time)? Response options for all three questions were “yes”, “no”, or “don’t know”. In those instances where a healthcare worker was not readily available, i.e. generally public areas such as elevators and stairwell doors, the response was classified as “unknown” for all three questions. Other data recorded were the characteristics of the hospital and attributes of the collected wipe sample. Hospital characteristics included the hospital type (acute care or cancer treatment), whether the hospital had an isolated drug preparation area (yes or no), the number of job categories responsible for drug transport, the number of drug administration units, the number of job categories working in the drug administration unit, and where drugs were initially delivered to the facility (shipping/receiving department or direct to the pharmacy).(80) Details of each wipe sample included the stage of the hospital medication system where the surface was sampled, the type of surface/object collected, and the relative date of sampling (with the first sampling day designated as day 1). Each one of the aforementioned served as an independent variable in the data analysis. 3.3.6 Data analysis The distribution of surface contamination levels was examined with the data untransformed and ln-transformed. Summary statistics were generated (arithmetic mean (AM), geometric mean (GM), geometric standard deviation (GSD), minimum and maximum, and the proportion less 47 than detection limit) for all wipe samples and stratified by each of the independent variables indicated previously. The response categories “don’t know” and “unknown” were combined for data analysis purposes. Bivariate analyses were performed to examine the relationship between CP contamination level (ln-transformed) and each of the aforementioned independent variables separately, using one-way ANOVA (categorical variables) or simple linear regression (continuous variables). Tukey’s post-hoc test was used to determine which differences in geometric means were significant in ANOVA. A paired t-test was performed to determine if there was a difference in means between the two sampling periods. All independent variables with p < 0.15 in the bivariate analyses were then offered in multiple linear regression, with a random effect for the surface identification number to account for potential correlation within repeated samples of the same surface. A manual backwards stepwise regression was employed to identify those independent variables that were associated with surface contamination levels. Independent variables with p < 0.05 were retained in the final model. A residual plot was generated to determine the appropriateness of the final model. Statistical analyses were performed using SPlus v. 8.0 for Windows (Insightful Corp., Seattle, WA). 3.3.7 Reporting of results A large proportion of surface wipes had contamination levels less than the instrument’s limit of detection (LOD). To alert the reader, in Tables 3-1 and 3-3, we report the proportion of values less than the LOD. To investigate which factors were related to surface contamination, we explored ways to prevent the bias that will occur if samples less than the detection limit are omitted or if inappropriate quantitative values are assigned to them.(81) Laboratory-calculated concentrations below the method limit of detection were available to us. Although this data has 48 a lower signal-to-noise ratio than the data above the detection limit, it is based on the actual measurements and should reflect actual contamination more accurately than a single substitute value for all data below the detection limit, a technique often used for data below detection limits.(82) We therefore used actual analytical data instead of substitute values for observations below the detection limit for all descriptive statistics and inferential analyses. When ln- transformed, the data more closely exhibited a normal distribution; therefore, all statistical analyses utilized the ln-transformed values and the GMs were compared in ANOVA. 3.4 Results 3.4.1 Overall summary of surface contamination levels A total of 229 surfaces were sampled at the participating hospitals, with 209 duplicates, for a total of 438 samples. Overall, 279 of the 438 samples (64%) had concentrations below the limit of detection (0.356 ng/wipe). The AM concentration was 0.201 ng/cm2, the GM 0.019 ng/cm2 and the GSD 2.54, with a range of less than LOD (0.356 ng/wipe; LOD in ng/cm2 varied with surface area) to 26.1 ng/cm2. The number of samples collected per hospital ranged from 47 to 102 samples. The results of the paired t-test did not find any statistically significant difference in average drug contamination levels between the duplicate samples. 3.4.2 Surface contamination levels by medication system stage Figure 3-1 provides an overview of the contamination levels found on the 55 different catgeories of surfaces sampled, stratified by the stage of the hospital medication system. The most contaminated surface at each stage was as follows: 1) delivery – elevator button (0.050 ng/cm2); 2) drug preparation – pen (26.2 ng/cm2); 3) transport to ward – bin for drug pick up (0.106 49 ng/cm2); 4) drug administration – IV pump (0.454 ng/cm2); and 5) waste disposal – elevator button (0.006 ng/cm2). Figure 3-2 indicates that the stage with the highest average contamination levels was drug preparation (AM=0.592 ng/cm2, GM=0.067 ng/cm2). In Tukey post-hoc testing for stage of hospital medication system, drug preparation had a statistically significant higher mean contamination level than both the drug delivery and drug administration stages. Of note, there was a gradual decrease in the proportion of detectable samples in the three hospital medication system stages that followed drug preparation. It should also be noted that the GM at the transport stage was slightly higher than the GM at the drug preparation stage and the stage with the greatest variability in surface contamination levels was drug delivery (GSD=6.68). 3.4.3 Surface contamination levels based on hospital characteristics Categorical variables Table 3-1 summarizes the surface contamination levels by categorical hospital characteristics. Although not statistically significant, the geometric mean concentration at the participating cancer centre was higher than at the acute care hospitals (F(1,436)=2.97, p=0.086; 3.2 times difference in GMs). At hospitals where drugs shipments were received at the shipping/receiving department, the geometric mean surface contamination levels were higher though not statistically significant (F(1,436)=0.217, p=0.64; 1.4 times difference in GMs) than at sites where the drugs were sent directly to the pharmacy department. Lastly, there was no statistically significant difference whether the drugs were prepared in an isolated preparation room or not (F(1,436)=0.039, p=0.84; 0.003 ng/cm2 difference in GMs). 50 Continuous variables Table 3-2 summarizes the surface contamination levels by continuous hospital characteristics. Only one variable was statistically significant: with more drug transport job categories, there was higher surface contamination. Neither the number of drug administration units nor the number of job categories in the drug administration unit observed to contact drug-contaminated surfaces were associated with contamination levels. 3.4.4 Surface contamination levels based on reported CP handling, spills and surface cleaning Table 3-3 is an overview of the surface contamination levels based on reported CP usage and surface cleaning. There were no reported spills or leaks on any of the surfaces. The highest AM and GM contamination levels occurred when CP was reportedly used prior to sampling. However, the highest individual recorded contamination level occurred when CP was reportedly not handled, prepared and/or administered prior to sampling. The difference in contamination levels between the three reported CP usage categories was statistically significant. In the Tukey post-hoc test, reported handling, preparation, or administration of CP had statistically significant higher levels than ‘don’t know’ or ‘unknown’ CP usage. With respect to surface cleaning, the highest GM was when the surface was cleaned, whereas the highest AM and the highest individual contamination level occurred when the surface was reportedly not cleaned. The difference in contamination levels between the three reported surface cleaning categories was not statistically significant (F(2,435)=1.27, p=0.28; 3.8 times difference between minimum and maximum GM). 51 Table 3-4 presents a cross tabulation of surface contamination levels by reported CP handling and surface cleaning. Despite the fact that CP was not reportedly handled, prepared, or administered on the work shift for 245 of all samples collected (56%), detectable levels were still found. The maximum individual contamination measurement was found when no CP was reportedly used and the surface was reported as not cleaned (a pen inside the biological safety cabinet at Site E). The highest arithmetic mean contamination (0.730 ng/cm2) and geometric mean (0.098 ng/cm2) was found when CP was reported as used and the surface reported to have been cleaned. 3.4.5 Surface contamination levels based on attributes of the wipe sample In total, 55 different categories of surfaces/objects were sampled. Differences in the contamination levels between the type of surface/objects were nearly statistically significant in ANOVA (F(54,383)=1.35, p=0.06). In Tukey post-hoc testing, tweezers used during drug preparation had higher contamination levels than 19 other surface categories (biological safety cabinet, IV hook, IV pump, pass through window, bin, box cutter, calculator, cart, countertop, door handle, drawer, elevator button, keyboard, pen, patient bedside table, patient chair side table, refrigerator, sink handles, sticker dispenser and re-sealable plastic bags), and elevator button had significantly lower contamination levels than three other surface categories (marker, pen, and vial). There were no apparent temporal trends in surface contamination as we found no statistically significant difference with respect to the relative sequence by which samples were collected (p=0.12). 52 3.4.6 Multiple linear regression model Six of the twelve independent variables were found to have p < 0.15 following bivariate analyses: 1) the type of hospital, 2) the number of job categories responsible for drug transport, 3) the stage of the hospital medication system, 4) the type of surface/object sampled, 5) whether CP was reportedly handled/prepared/administered on the sampling shift and 6) the relative sampling sequence. All other independent variables had p > 0.20. These six independent variables were offered into a multiple linear regression to predict drug contamination on surfaces with the surface identification number inputted as a random effect. Using backwards stepwise regression as described in the Methods, two independent variables remained in the final model: the number of job categories responsible for drug transport and the stage of the hospital medication system. According to Table 3-5, having more drug transport job categories was positively associated with higher surface contamination levels and the drug preparation and drug administration stages had the highest contamination. A review of the residual plot found that there were larger residuals at higher surface contamination levels resulting in a ‘funnel-like’ distribution. This may be due to the large proportion of samples that were found to be less than the detection limit. 3.5 Discussion Until now, no single study has measured the drug residual levels on surfaces throughout the hospital medication system and, subsequently, identified the determinants of such contamination. Our findings indicate that, although a large proportion of samples did not have detectable contamination, various surfaces at every stage of the hospital medication system had drug residual at the participating facilities. Thus, our study is in agreement with others that there is 53 surface contamination despite the implementation of controls;(36)(34)(62)(45)(40)(83)(46) but extends these findings beyond the drug preparation and patient administration areas. The current study also provides evidence of contamination of surfaces that healthcare workers are likely to contact with their hands through the course of their duties such as door handles, carts and writing instruments (based on site observations in Chapter 2). Other studies have detected contamination on some surfaces that may or may not be contacted by healthcare workers, such as the floor in the drug preparation area.(60)(24)(31) Our review of the literature found only two studies that have identified determinants of surface contamination. Connor et al. found no correlation between the number of handling events and the percentage of surface wipe samples that had measurable levels of antineoplastic drug.(36) Siderov et al. did not detect correlation between surface contamination levels and any of the following variables: 1) amount of CP prepared annually, 2) age of cytotoxic suite and the biological safety cabinet, 3) decontamination and cleaning procedures, and 4) the staff member preparing chemotherapy.(34) These papers examined the drug preparation and patient administration areas only. In our examination of factors which may influence drug contamination of surfaces throughout the hospital medication system, we identified two variables that were statistically significant in the multiple linear regression model: 1) the stage of the hospital medication system and 2) the number of job categories responsible for drug transport. Employee-reported use of CP was not associated with surface contamination levels in the current study. This finding concurs with Hedmer et al.(31) who found measurable levels on surfaces even though drugs were not handled on the sampling dates. 54 It is not surprising that the stage of the hospital medication system is associated with the highest surface contamination levels. One would expect the drug preparation stage to be the most critical because the concentrated drugs are routinely handled and mixed there. Our finding is consistent with the conclusion of others.(36) We and others detected drug contamination on the outside of the manufacturer vials.(51)(53)(32) According to Sessink and Bos, contamination already present on vials before the initiation of preparation will result in further distribution of the drug,(2) which was confirmed in the present study with measurable levels of drug residual in subsequent stages of the hospital medication system. The presence of contamination in stages following drug preparation may be due to worker contact with communal objects such as door handles, trays, and writing instruments.(41) It has been suggested that surfaces in the last three stages of the hospital medication system are generally more porous than those in the drug preparation area and therefore drug residual can diffuse into the pores and may accumulate over time because they are environmentally stable.(31) The potential that the drug preparation stage contributes to contamination in the remaining stages of the hospital medication system is suggested by the decrease in the proportion of detectable samples as one progressed from drug preparation to subsequent stages (see Figure 3-2). With respect to the second determinant, our regression model found that the more job categories involved with drug transport (from the drug preparation area to the drug administration units), the higher the surface contamination levels. We speculate that this may be attributed to more handling events of the prepared drugs by a variety of individuals – a large proportion of whom are not formally trained regarding the hazards of these agents or the safe handling requirements (training history of participants is discussed in Chapters 4 and 5). Interestingly, one study suggested that contamination should not occur during transport because the prepared drugs are 55 protected by a secondary container e.g. paper bag.(20) Given this and the fact that we are the first to identify this determinant, it is important to see if future studies can replicate our result. Although we identified two potential determinants of surface contamination, we were unable to explain the exact mechanism of spread of antineoplastic drugs. The utilization of a tracer may be a way to address this question(84)(85) but it needs to be traceable at the start of the hospital medication system (i.e. at the drug delivery stage not drug preparation) in order to provide a comprehensive account of how these hazardous agents are spread. Table 3-6 compares the CP contamination levels that we found for specific objects at each stage and those reported by others (note that only recent studies that reported contamination levels in ng/cm2 are presented). Overall, the mean contamination levels from the current study were lower than those reported elsewhere. We agree with others who performed multi-site studies that surface contamination levels vary from site-to-site.(36)(34)(44) According to Touzin et al. variability in surface contamination may be related to the layout of the workplace which can impact work techniques (i.e. proximity of drug storage relative to drug preparation) as well as the size of working area where drugs are handled.(33) A review of Table 3-6 shows that our findings were most comparable to those reported by Touzin et al.(33) and Connor et al (2010).(36) Both studies were conducted in facilities that are known to follow the National Institute for Occupational Safety and Health guidelines for the safe handling of hazardous drugs and most of our facilities had similar controls in place, i.e. clean room design, and anterooms between the preparation area and the adjacent areas with pass- through windows for the compounded drug to leave the preparation area. Regardless of these control measures, all three studies had measurable levels of drug residual indicating persistent contamination and worker exposure potential. 56 Of note, the reported cleaning of the surface prior to wipe sampling was not associated with contamination levels. The lack of association may be attributed in part to errors in reporting by the healthcare workers responding to the survey. The majority of the responses to our question “To your knowledge, was the surface/object cleaned prior to collection of wipe sample” was “don’t known/unknown” (238 of 438 samples or 54%). This suggests that cleaning was not frequent or that workers were not aware of cleaning activities, or both. Regardless, others have questioned the cleaning efficacy of surfaces. Sessink et al.(59) suggested drug residual could be spread due to the lack of cleaning effectiveness. Turci et al.(66) noted that contamination was found even before drug preparation activities, implying that decontamination protocols between preparations were incomplete. The apparent lack of temporal variability in surface contamination is consistent with Hedmer et al.’s findings(31) and raises questions of cleaning effectiveness as well. A recent study showed that existing cleaning protocols are not 100% effective in reducing contamination levels and may lead to possible accumulation of drug residual.(86) We therefore agree with Sugiura et al.’s suggestion that effective environmental cleaning needs to be performed to remove drug residual to prevent the introduction of new contamination.(35) Cleaning agents which specifically target the drug products may be more effective than detergents or disinfectants.(87) Although there is no established occupational exposure limit, the United States Pharmacopeia (USP) recently indicated a maximum surface contamination level of 1 ng/cm2 of CP to limit the risks of absorption (uptake) in humans.(33) In our study, eight of the 438 (2%) surfaces sampled exceeded this level. In four (50%) of these samples, there was either no use of CP or use was unknown prior to sampling. Cleaning was not performed or was uncertain for five of the eight surfaces. The surface contamination levels represent a point in time only and are not reflective 57 of the potential for repeated contact by healthcare workers which could conceivably exceed 1 ng/cm2 over the course of a work shift. Given this and the fact that local regulatory bodies have not prescribed the USP threshold, the principle adopted by occupational hygienists in those instances where no limit has been established is to keep exposure levels as low as reasonably achievable. 3.5.1 Limitations It should be noted that we did not take into consideration the wipe efficiencies from each surface sampled; therefore, the concentrations reported in this study may underrepresent the actual amount of CP on the surfaces sampled. [Note that others have also assumed 100% wipe recovery(34)(35)] Additionally, we only analyzed for CP and did not consider potential surface contamination by other antineoplastic drugs. This can also underestimate the surface contamination levels and the resulting exposure risk. Another limitation of the study is that we were unable to confirm whether CP was handled prior to sample collection. Utilization of a diary by healthcare workers may help to overcome this limitation(36) but could be challenging to implement in those areas where multiple people are involved i.e. sites have several porters perform various tasks and no individual porter is dedicated strictly for drug transport. Body fluids from patients undergoing chemotherapy are known to contain antineoplastic drugs.(71) Thus, bed sheets and the patient washrooms may be contaminated. However, we did not sample these surfaces because a) we are uncertain of the effectiveness of our sampling method in recovering drug residual on highly porous material such as linens and b) sampling of patient areas is difficult without compromising patients’ privacy. As such, the number of surfaces contaminated throughout the hospital medication system is likely an underestimate. Lastly, we 58 did not record the cleaning agents employed at each facility; this would be a useful addition in future studies. 3.5.2 Summary In summary, measurable levels of drug residual were found on a variety of surfaces situated throughout all stages of the hospital medication system. Given these findings, we suspect that healthcare job categories involved in some capacity with the hospital medication system are at potential risk for dermal contact with these surfaces. For instance, external contamination on the vials presents an opportunity for exposure as soon as the drug shipment arrives at the healthcare facility.(52) With respect to determinants, we identified the drug preparation and drug administration stages of the hospital medication system and a larger number of transport job categories as being associated with surface contamination in the hospitals studied. This finding provides a basis for targeting interventions to reduce contamination. To our knowledge, this is the first study of its kind to identify determinants of surface contamination throughout the hospital medication system; therefore, further research is recommended to determine the generalizability of the findings. 59 B P m ac hi ne B S C IP A b ot tle IV b ag IV h oo k IV p um p IV tu bi ng O C C P T W bi n bo x cu tte r ca bi ne t ca lc ul at or ca rt ch ai r cl am p cl ip bo ar d co m pu te r m ou se co un te rt op da te s ta m p de liv er y bi nd er de sk do or h an dl e dr aw er el ev at or b ut to n ev ac . c on ta in er gr ea se p en ci l hi gh lig he r ke yb oa rd lig ht s w itc h m ar ke r m ed s up pl ie s ca rt m ed ic at io n la be ls m et al ti n pe n pr in te r pt b ed si de ta bl e pt b in de r pt c ha ir si de ta bl e pt c ov er fo ld er s pu sh p la te re fr ig er at or sh el f si de ta bl e si nk h an dl es so ap d is pe ns er st ic ke r di sp en se r ta bl e te le ph on e tr ay tw ee ze rs vi al vi ta ls m ac hi ne w as te b in zi pl oc b ag 5 20 5 20 5 20 5 20 5 20 C P c on ce nt ra tio n (n g/ cm 2) Stage: 1. Drug Delivery Stage: 2. Preparation Stage: 3. Transport Stage: 4. Drug Admin Stage: 5. Waste N = 60 N = 144 N = 36 N = 192 N = 6 Note: where no bar is present, that surface was not sampled in that stage Figure 3-1 Average cyclophosphamide (CP) contamination levels (in ng/cm2) on the 55 different categories of surfaces sampled stratified by the stage of the hospital medication system 60 1. Drug Delivery 2. Preparation 3. Transport 4. Drug Admin 5. Waste C P co nc en tra tio n (n g/ cm 2) 0.1 0.2 0.3 0.4 0.5 33% > LOD 8% > LOD 28% > LOD 17% > LOD 61% > LOD Drug Delivery Preparation Transport Drug Admin Waste N 60 144 36 192 6 Arithmetic Mean < 0.356 ng/wipe* 0.592 0.010 0.013 < 0.356 ng/wipe* Std Dev 0.0215 3.00 0.021 0.053 0.010 Geometric Mean < 0.356 ng/wipe* 0.067 0.009 0.008 < 0.356 ng/wipe* Geometric Std Dev 6.68 2.65 1.14 1.23 1.09 ANOVA < 0.005; whiskers represent 10th and 90th percentiles * LOD of 0.356 ng/wipe; LOD in ng/cm2 varied with surface area of sampled object Figure 3-2 Box plot and summary statistics of cyclophosphamide (CP) surface contamination levels (in ng/cm2) by stage of hospital medication system 61 Table 3-1 Surface contamination levels based on categorical variables of hospital characteristics Variable N AM (ng/cm2) SD (ng/cm2) GM (ng/cm2) GS D ANOVA (p-value)* % < LOD Hospital Type Cancer Centre 99 0.416 2.42 0.041 2.28 0.086 59.6 Acute Care Hospital 339 0.138 1.49 0.013 2.61 65.2 Isolated preparation room No 87 0.053 0.325 0.016 1.54 0.840 74.7 Yes 351 0.238 1.94 0.019 2.77 61.3 Where drugs are initially delivered Pharmacy 183 0.068 0.544 0.015 1.54 0.642 65.6 Shipping/Receiving 255 0.296 2.23 0.021 3.20 62.7 * where the dependent variable was the ln-transformed surface contamination level Table 3-2 Bivariate results of surface contamination levels according to hospital characteristic variables that are continuous in nature Variable Intercept Coefficient SE p-value* No. of transport job categories -2.17 0.130 0.060 0.029 No. of drug administration units -2.06 0.063 0.051 0.214 No. of drug admin job categories -1.80 -0.037 0.039 0.337 * where the dependent variable was the ln-transformed surface contamination level Table 3-3 Bivariate results of surface contamination levels based on reported cyclophosphamide (CP) handling and surface cleaning Variable N AM (ng/cm2) SD (ng/cm2) GM (ng/cm2) GSD ANOVA (p-value)* % < LOD Reported CP usage Don’t know/Unknown 123 0.254 2.00 0.007 4.45 0.047 70.7 No 245 0.135 1.68 0.015 1.58 67.8 Yes 70 0.339 1.45 0.060 2.37 38.6 Reported surface cleaning Don’t know/Unknown 238 0.166 1.51 0.010 3.03 0.281 68.5 No 132 0.250 2.29 0.025 1.86 61.4 Yes 68 0.227 1.25 0.039 2.05 52.9 * where the dependent variable was the ln-transformed surface contamination level 62 Table 3-4 Summary statistics for cyclophosphamide (CP) surface contamination levels based on employee-reported CP usage and surface cleaning during the work shift Was cyclophosphamide handled/prepared/administered prior to sampling? Don’t know/ Unknown No Yes W as th e su rf ac e cl ea ne d pr io r to sa m pl in g? Don’t know /Unknown N=107 Mean = 0.284 GM = 0.003 Max = 21.1 N = 110 Mean = 0.010 GM = 0.007 Max = 0.41 N = 21 Mean = 0.377 GM = 0.075 Max = 6.72 No N = 7 Mean = 0.025 GM = 0.020 Max = 0.12 N = 95 Mean = 0.325 GM = 0.024 Max = 26.1 N = 30 Mean = 0.065 GM = 0.030 Max = 1.15 Yes N = 9 Mean = 0.077 GM = 0.056 Max = 0.30 N = 40 Mean = 0.022 GM = 0.013 Max = 0.48 N = 19 Mean = 0.730 GM = 0.098 Max = 9.71 Mean, GM and Max values reported in ng/cm2 Table 3-5 Coefficients, standard errors and p-values for multiple linear regression model showing factors related to surface wipe contamination levels (ln-transformed) Variable Coefficients Standard Error p-value Intercept -3.015 0.241 <0.0001 No. of transport job categories 0.213 0.088 0.017 Stage 1 Drug Delivery Reference Stage 2 Preparation 0.869 0.205 <0.0001 Stage 3 Transport 0.397 0.287 0.169 Stage 4 Drug Admin 0.488 0.197 0.014 Stage 5 Waste 0.382 0.582 0.512 63 Table 3-6 Reported cyclophosphamide contamination levels in ng/cm2 found in healthcare facilities Stage Object Mean concentration (range) Author and Country Drug Delivery Unable to find comparable results in the literature Drug Preparation Biological safety cabinet (hood work surface) 0.056 (0.007-0.130) Touzin et al., Canada (2009)(33) 0.101 (ND – 0.38) Siderov et al., Australia (2009)(34) 18.1 Connor et al., USA (2010)(36) 0.009 (0.003 – 0.2) Hedmer et al., Sweden (2005)(32) 0.01 Sugiura et al., Japan (2010)(35) 47.1 Sottani et al., Italy (2011)(44) 0.01 Tanimura et al., Japan (2009)(48) 0.01 Sugiura et al., Japan (2010)(43) 0.052 (ND – 0.0321) Current study Drug vial 9 (ND – 912) Schierl et al., Germany (2010)(53) 0.127 (ND – 0.53) Hama et al., Japan (2011)(52) 0.1 Favier et al., France (2003)(49) 1.28 (ND – 9.71) Current study Pass through 0.200 Connor et al., USA (2010)(36) 0.257 Sottani et al., Italy (2011)(44) 0.019 (ND – 0.063) Current study Pharmacy checking counter 0.013 (0.005-0.029) Touzin et al., Canada (2009)(33) 0.133 (ND – 0.67) Siderov et al., Australia (2009)(34) 0.039* Schmaus et al., Germany (2002)(24) 1.7 Connor et al., USA (2010)(36) 4.63 Sottani et al., Italy (2011)(44) 0.008 (ND – 0.041) Current study Transport Cart 0.534 (0.01 – 1.94) Connor et al., USA (1999)(60) ND Connor et al., USA (2010)(36) 0.002 (ND – 0.038) Current study Drug Administration Bed side table 2.65 Sugiura et al., Japan (2010)(43) 0.052 (ND - 0.328) Current study Handle of fridge ND Ziegler et al., United Kingdom (2002)(88) GM 0.0002 Hedmer et al., Sweden (2008)(31) 0.039 (ND - 0.191) Current study Nursing station countertop 27 McDevitt et al., USA (1993)(19) 0.03 Connor et al., USA (2010)(36) 0.014 (ND - 0.406) Current study Sink 8.5 McDevitt et al., USA (1993)(19) 0.0006 (ND - 0.0170) Current study IV bags N/A (0.1 – 533) Martins et al., Brazil (2009)(46) 0.91(Centre 1); 0.07 (Centre 2) Crauste-Manicet et al., France (2005)(20) 0.206 Sottani et al., Italy (2011)(44) 0.022 (ND - 0.104) Current study Waste Waste container 0.08 Connor et al., USA (2010)(36) 0.042* Schmaus et al., Germany (2002)(24) 0.010 (ND – 0.110) Current study * median levels presented; ND = not detected 64 4 EVALUATION OF HEALTHCARE WORKERS’ DERMAL EXPOSURE THROUGHOUT THE HOSPITAL MEDICATION SYSTEM AND IDENTIFICATION OF DETERMINANTS OF SUCH CONTAMINATION 4.1 Synopsis Chapter 3 demonstrated that antineoplastic drug contamination is found on surfaces located throughout the hospital medication system. This is of concern as dermal contact is known to be the primary route of occupational exposure for healthcare workers. The aim of this chapter was to ascertain the dermal contamination levels of those healthcare job categories observed in Chapter 2 that have a high probability of contacting the surfaces that have drug residual. In addition, determinants associated with dermal contamination were identified. Using similar methods employed in Chapter 3, wipe samples were collected from healthcare workers’ hands and subsequently quantified for CP using high-performance liquid chromatography-tandem mass spectrometry. Participants were also asked to answer a questionnaire regarding knowledge and behaviours with respect to antineoplastic drugs as well as complete a brief survey regarding contact with CP on the work shift. Up to three representatives from each job category at each site were invited to participate. In total, 225 wipe samples were collected (115 participants with 110 providing a duplicate). Overall, the dermal concentration AM was 0.360 ng/wipe, the GM < LOD, the GSD 1.98, and the range less than 0.356 to 22.8 ng/wipe (80% of samples were less than the detection limit). When stratified by job title, every maximum concentration exceeded the detection limit with the job group “other workers in the drug administration unit” (volunteers, oncologists, ward aide and dieticians) having the highest mean contamination level. Participants’ hand washing practices did not appear to impact the dermal contamination levels. Four factors were found to be associated with increased dermal contamination: working in an 65 acute care hospital, working as a porter, nurse, transport, unit clerk or other role in the drug administration unit, being female, and having had a duty to handle antineoplastic drugs. 4.2 Introduction A number of studies have documented that occupational exposure to antineoplastic (cytotoxic) drugs can result in adverse health outcomes including genetic damage(89)(55), which could lead to cancer, and reproductive effects.(5)(65) The literature lends evidence that occupational exposure to antineoplastic drugs occurs through the transfer of surface load of these drugs to the skin(37) and therefore the primary route of exposure is via dermal contact.(38)(40)(90) This may occur by contacting the drugs or their containers directly (i.e., handling of manufacturers’ vials and/or prepared drug solutions in intravenous bags) or by indirect contact via touching of drug-contaminated surfaces. We conducted a pilot study to assess the potential dermal contamination of personnel at select British Columbian hospital pharmacies and confirmed that both pharmacists and pharmacy technicians have detectable levels of drug contamination on their hands.(42) In Chapter 3, we showed that drug contamination of work surfaces is found throughout the hospital medication system and is not limited to the drug preparation and drug administration areas.(80) As such, healthcare workers who are involved in some capacity with the hospital medication system, such as porters, unit clerks and receivers, may contact contaminated surfaces and are therefore at potential risk of exposure. Based on our review of the literature, no single study has simultaneously examined the dermal contamination levels of multiple job categories that are potentially exposed to antineoplastic drugs due to their role in the hospital medication system. The aim of the current study was to quantify dermal contamination levels of healthcare workers in job categories identified as part of the hospital medication system.(80) In addition, we 66 examined various factors, including glove usage and hand washing practices, to identify potential determinants of dermal contamination. To our knowledge, this is the first study to ascertain occupational dermal exposure of a broad range of at-risk healthcare job categories and the determinants of such exposure. For the reasons described in the introductory chapter, cyclophosphamide (CP) was used as the marker drug for exposure. 4.3 Methodology 4.3.1 Selection of participants Study participants were employed at the six healthcare facilities located in Metro Vancouver that participated in the earlier phases of this study (Chapters 2 and 3) – five sites were acute care hospitals and one site was a cancer treatment facility. Healthcare workers were selected from job categories potentially at-risk of exposure to antineoplastic drugs according to site observations described in Chapter 2.(80) Eligible participants were selected by either active recruitment via a mailed letter of invitation (Appendix C) or passive recruitment through the distribution of consent to contact forms (Appendix D) at departmental meetings. The recruitment methods were dictated by the requirements of the participating hospitals’ research ethics boards. Up to three representatives of each job category at each site were invited to participate. An overview of the recruitment process is found in Figure 4-1. Upon receiving consent from workers to participate in the study (Consent Form, Appendix E), members of the research team contacted each participant via email or telephone to arrange a mutually convenient time to meet at their place of work to collect hand wipe samples. Participants were given a cash honorarium of $10 for providing a hand wipe sample. 67 4.3.2 Wipe sampling of hands Participants had their hands sampled at their convenience during their work shift. Sampling dates were selected based on availability of research team members and participants’ work schedules. Participants who were wearing gloves when our team member arrived to collect the samples were asked to remove them prior to the hand wipe. This was done in order to assess any permeation of drugs through the gloves. The hand wipe sample collection method is similar to that described previously for measuring surface contamination.(80) A new pair of disposable gloves was worn by the research team member to collect hand wipes from each participant. In summary, the front and back of both hands of each participant was wiped with a Kimwipe® (Kimberly-Clark, Mississauga, ON) that was pre-moistened with 1.0 mL of 0.1 M ammonium acetate solution (Sigma Aldrich, Oakville, ON). The Kimwipe® was then placed in a 20 mL vial and kept in a portable cooler with ice packs. All collected samples were shipped on ice to the analytical lab within 24 hours of sample collection and stored at -20oC until analysis. Both travel and field blanks were collected for quality control purposes. Sample collection took place between June 2010 and February 2011. Duplicate dermal wipe samples were collected from most participants with at least three weeks’ lag between collection times (average of 97 days; range 22 to 188 days). 4.3.3 Wipe sample preparation and analysis The methods employed for preparation and analysis of samples in the current study were the same as the surface wipes (Chapter 3). All results were corrected for blanks. Hand wipe 68 contamination levels were reported in nanograms per wipe (ng/wipe). The limit of detection (LOD) for the dermal wipes was 0.356 ng/wipe. 4.3.4 Supplemental data collected on site After the hand wipe was collected, all participants were surveyed regarding the following: a) the types and frequency of contact with CP during the current work shift (e.g. mixing drugs, administering drugs, disconnecting IV line, providing physical care to patients, disposing of CP- contaminated body fluids, handling a container of CP (i.e. vial or IV bag), disposing of vial or IV bag, cleaning up leak/spill of CP, touching a CP-contaminated surface/object, and consuming food/drink in an area where CP is mixed/handled/administered), b) number of gloves worn immediately prior to sample collection, and c) hand hygiene practices immediately before sample collection as well as during the current work shift. The latter included the participants’ most recent hand wash prior to sample collection, what cleaning agent they used for their most recent hand wash and the number of times they washed their hands during the work shift up to the time of sample collection. See Appendix F for a copy of “Activity-Related Questions.” 4.3.5 Self-administered questionnaire Instrument overview Participants were also provided with a self-administered questionnaire (Appendices G and H) which was divided into nine sections. The following sections of the questionnaire were included for data analyses: a) demographic data such as age and work experience; b) degree of contact with antineoplastic drugs; c) previous education/training related to antineoplastic drugs; d) usual hand hygiene practices and use of personal protective equipment, and e) details of known previous occupational exposure to antineoplastic drugs. 69 The majority of the questions incorporated for analyses were in a closed-ended categorical answer format. An open-ended question was utilized for respondents to describe any direct, unintended contact to antineoplastic drugs that they may have experienced. Theory and construct of instrument Several sources were employed to develop the questionnaire. The questions ascertaining the type of training received by the participant were based on the Occupational Safety and Health Administration’s training and information dissemination requirements for controlling occupational exposure to hazardous drugs.(91) The survey used by Geer et al. (92) in their study examining workers’ knowledge, attitudes and perceptions related to dermal exposure served as a basis for questions related to hand hygiene practices and personal protective equipment usage. Pre-testing of instrument A total of 28 individuals at non-participating hospitals in various healthcare job categories that were representative of the study jobs were asked to answer the original draft version of the questionnaire and provide feedback on wording and clarity of the questions. Modifications to the questionnaire were made in response to the feedback and the working version of the questionnaire received research ethics approval prior to dissemination. 4.3.6 Statistical analysis Both untransformed and ln-transformed data were used to examine the distribution of dermal contamination levels. Summary statistics (arithmetic mean (AM), geometric mean (GM), geometric standard deviation (GSD), minimum and maximum, and proportion less than limit of detection) were used to describe all hand wipe samples. When ln-transformed, the measurements more closely approximated a normal distribution than the untransformed data. As 70 such, all statistical analyses utilized the ln-transformed values and the GMs were compared in ANOVA. Summaries of the dermal contamination levels were stratified by various independent variables from the survey instruments. “No” and “don’t know” responses to pertinent questions from either of the two survey tools were combined into one category and compared with “yes” respondents of the same question. Bivariate analyses were performed to examine the relationship between CP contamination level (ln-transformed) and each of the independent variables separately, using one-way ANOVA (categorical variables) or simple linear regression (continuous variables). Tukey’s post-hoc tests were performed following ANOVA to determine which differences in geometric means within a categorical variable were significantly different. A paired t-test was performed to determine if there was a difference in means between the duplicate samples. All independent variables with p < 0.20 from the bivariate analyses and those with a strong a priori hypothesis for exposure were then offered into a multiple linear regression with the ln- transformed dermal contamination level as the dependent variable. The regression model also included a random effect with independent covariance structure for subject to account for potential correlation within repeated samples of the same participant. A manual backwards stepwise approach was employed to identify those independent variables that were significantly associated with the dependent variable by removing the independent variable with the largest p value at each step. Independent variables in which one or more categories had p < 0.05 were retained in the final model. A residual plot was generated to assess the appropriateness of the final model. Statistical analyses were performed using R version 2.13.1 (The R Foundation for Statistical Computing). 71 4.3.7 Reporting of results Similar to the surface wipe results, a large proportion of hand wipe samples had contamination levels less than the limit of detection. To alert the reader, in Tables 4-1, 4-2 and 4-3, we report the proportion of values less than the LOD. As with the surface contamination levels, laboratory-calculated concentrations below the method limit of detection were available to us. The rationale for using all laboratory-calculated concentrations for analyses is found in the Methods section of Chapter 3. 4.4 Results 4.4.1 Characteristics of study population In total, 115 workers agreed to participate in the study. It was not possible to calculate a true response rate because of the constraints of the recruitment methods dictated by hospital ethics board; however, the proportions who participated of those contacted by each method ranged from 55% to 76% (Figure 4-1). Table 4-1 presents a summary of the characteristics of the study population. Study participants had been in their current job positions for an average of 103 months (range 0 to 433 months) and workers who had a duty to handle antineoplastic drugs (N=91) had been handling these agents for an average of 82 months (range 0 to 336 months). The results of the paired t-test found that there was a statistically significant difference in means between the two samples with wipes collected on the second occasion having a higher overall mean (less than 0.356 ng/wipe in round #1 compared to 1.09 ng/wipe in round #2). 4.4.2 Overall summary of dermal contamination levels Of the 115 participants, 110 supplied a duplicate hand wipe sample, resulting in a total of 225 dermal wipe samples. Only 44 of all samples (20%) were above the LOD of 0.356 ng/wipe. 72 Overall, the dermal concentration AM was 0.360 ng/wipe, the GM < LOD, the GSD 1.98, and the range less than 0.356 to 22.8 ng/wipe. The dermal contamination levels among female participants were higher than male participants and the difference between the sexes was statistically significant. Dermal samples from acute care hospitals had a higher average contamination level than those from the cancer treatment hospital and this difference was statistically significant (Table 4-1). 4.4.3 Dermal contamination levels by individual factors Although participants from the drug administration units had the highest AM contamination levels, the difference in GM contamination levels between departments was not statistically significant (F(2, 222) = 2.06, p=0.13; GMs less than detection limit). When stratified by job title, the highest average contamination level was for other workers in the drug administration unit (volunteers, oncologists, ward aide and dieticians). The highest maximum dermal concentration of 22.8 ng/wipe was recorded for a registered nurse. The difference in GM contamination levels between job titles was not statistically significant (F(7,217) = 1.42, p=0.20) with only two job categories, registered nurses and workers in patient units not responsible for drug administration (volunteer, oncologist, dietician, ward aide), having means greater than the limit of detection (see Table 4-1). Those workers who reported that they had a duty to handle antineoplastic drugs had a higher average contamination level than those who did not have this responsibility and this difference was nearly statistically significant (F(2,223) = 3.24, p=0.07). There was no statistically significant difference in dermal contamination levels between workers who had ever received health and safety training related to antineoplastic drugs and those who never received training (F(1,223) = 0.495, p=0.48; GMs less than limit of detection). 73 Neither the duration of participants’ tenure in their current position nor the length of time that an individual had been handling antineoplastic drugs had a significant association with dermal contamination levels (p=0.794 and p=0.605, respectively). 4.4.4 Dermal contamination levels by cyclophosphamide (CP) contact Sixty-three of all participants (28%) reported that they handled or came into contact with CP at least once during their work shift on the day the hand wipe sample was collected. Every participant was asked if they were potentially exposed to CP via ten different modes of contact; the resulting average reported number of CP contact methods on the shift was 1, with a minimum of 0 and a maximum of 6 contact methods. The individual who had the maximum measured contamination level reported having no contact with CP during the work shift. When the results were stratified by the different CP contact methods, the AM contamination level of those who indicated that they had contacted CP was higher than those who had indicated no contact (Table 4-2). Regardless of the CP contact method, in all instances where a participant reported having contacted CP during the work shift, a higher proportion of the dermal measurements were above the detection limit. However, for every contact method examined in this study, in ANOVA none of these differences were statistically significant, likely due to the small number of respondents who reported contact. When the number of contact methods was tallied and treated as a continuous variable, the results of the linear regression suggest that for every additional mode of contact, the level of dermal contamination increased by 0.05 ng/wipe (p=0.13). 74 Dermal contamination levels based on glove usage Frequency of glove usage when handling antineoplastic drugs Of those tasked with handling antineoplastic drugs, 49 of 93 (53%) participants indicated that they wore gloves all the time when handling. There was no statistically significant difference in contamination levels between those who reported using gloves all the time and those who claimed to use gloves some of the time (F(2,222) = 1.08, p=0.34; GMs less than detection limit) (Table 4-3). Of the various job titles, nurses were most likely to report wearing gloves all the time when handling hazardous drugs. However, the frequency of glove usage when handling antineoplastic drugs did not appear to have any correlation with either job title or with the percentage of time handling antineoplastic drugs. Thirty individuals stated that they did not wear gloves all the time when handling antineoplastic drugs despite the fact that they received health and safety training related to these hazardous agents. In addition, of those who indicated that they had previous direct, unintended contact with antineoplastic drugs while at work, less than 50% (17 of 35) claimed to use gloves all the time when handling antineoplastic drugs. Frequency of glove usage while in an area where antineoplastic drugs are handled When the participant was simply in an area where antineoplastic drugs are handled but not assigned to manipulate the agents, the percentage of workers who wore gloves all the time decreased to 27% (31 of 115 participants). Those who indicated they wore gloves all the time when in an area where antineoplastic drugs are handled had the highest level of contamination of all response categories (Table 4-3), but the difference was not statistically significant (F(2,222) = 75 0.951, p=0.39; GMs less than detection limit). There was no association between job title and the frequency of glove usage when in an area where antineoplastic drugs are handled. Number of gloves worn We sampled workers at a time that was convenient according to their work schedules, and at those times, 22 of all participants (10%) were observed wearing gloves (19 wore one pair of gloves and three wore double gloves). The participants who were wearing double gloves had just finished mixing drugs and/or handling a drug vial prior to sample collection. It is not surprising then that these three individuals had a higher mean level of dermal contamination than participants who wore one pair or no gloves. However, this difference was not statistically significant (F(2,221) = 0.607, p=0.55) likely due to the fact that the sample size was small (see Table 4-3). 4.4.5 Dermal contamination level based on hand washing practices Hand washing practices after glove usage There was no statistically significant difference in dermal contamination levels between those who washed their hands all the time and some of the time after glove use (F(1,223)=0.968, p = 0.33); GMs less than limit of detection). The sample with the maximum reported dermal contamination belonged to a participant who reported washing his/her hands all the time after glove usage. However, at the time of sample collection, this worker was not wearing gloves, indicated no known contact with CP on the shift and his/her most recent hand wash was less than 10 minutes prior to sample collection. Of those workers who reported handling antineoplastic drugs more than 25% of the time, only 19 of 73 (26%) reported washing their hands all the time after glove use (see Table 4-3). 76 Twenty-five of the 35 (71%) participants who reported having previous, unintended contact with antineoplastic drugs while at work did not wash their hands all the time after glove usage. Similarly, a majority (36 of 60; 60%) of those who had received health and safety training related to antineoplastic drugs stated that they did not wash their hands all the time after glove usage. Hand washing after being in an area where antineoplastic drugs are handled There was no statistically significant difference in hand contamination between those who washed their hands all the time and some of the time after being in an area where antineoplastic drugs are handled (F(1,223) = 0.227, p=0.63) (see Table 4-3). Fourteen of the 35 (40%) participants who reported having previous direct, unintended contact with antineoplastic drugs while at work stated that they washed their hands all the time after being in an area where antineoplastic drugs are handled. Of those who had received health and safety training related to antineoplastic drugs, only 20 of 60 (33%) stated that they washed their hands all the time after being in an area where antineoplastic drugs were handled. Hand wash duration A large proportion (60 of 74; 81%) of those who handled antineoplastic drugs for more than 25% of their time reported washing their hands for less than 15 seconds. The maximum reported concentration was for a participant who stated that they usually washed their hands for 15 seconds or less, but there was no statistically significant difference in dermal contamination levels based on hand wash duration – less than 15 seconds or at least 15 seconds (F(1,223)=0.665, p=0.42) (see Table 4-3). 77 Hand wash cleaning method We asked participants to indicate their usual hand wash cleaning method and also asked about their most recent hand wash cleaning method. In both instances, there was no statistically significant difference in contamination levels between the different types of cleaning agents (water only, soap and water, alcohol hand gel and other) used for hand washing (p=0.937 and p=0.978, respectively) (see Table 4-3). Most recent hand wash The maximum dermal concentration level was from a participant who reported washing their hands less than ten minutes prior to the hand wipe. Collectively, individuals who reported washing their hands less than ten minutes prior to collection of the hand wipe sample had the highest average concentration but the difference was not statistically significant (F(3,221)=1.48, p=0.22; 9.1 times difference in GMs). It should be noted that a larger proportion of those who reported having had contact with CP by at least five methods on the shift indicated washing their hands less than ten minutes prior to sample collection compared to those participants with less than five CP contact methods (see Table 4-3). Hand wash frequency Individuals who reported washing their hands five or more times on the shift had the highest average contamination levels. This finding was statistically significant and the Tukey post-hoc test revealed that the average contamination level for those who washed five or more times on the current shift was significantly greater than those who only washed their hands one or two times on the shift. It is notable that 100% of those who indicated they had at least five different CP contact methods on the shift washed their hands at least three times before the collection of 78 the wipe sample. This represents a higher proportion of frequent hand washes than those with fewer CP contact methods. However, after controlling for the number of CP contact methods in linear regression, the association between dermal contamination levels and frequency of hand washing was reduced and no longer statistically significant. 4.4.6 Multiple linear regression model The results of bivariate analyses found nine variables that had p < 0.20: hospital type; department; job title; sex; whether the worker had a duty to handle antineoplastic drugs; whether the worker contacted CP during the work shift; number of times hands were washed on the work shift; number of CP contact methods on the work shift; and number of job categories responsible for drug transport at the hospital. Based on a priori hypotheses regarding potential exposure, another five variables were selected to be offered in the regression model: whether the worker mixed CP on the shift; whether the worker administered CP on the shift; whether the worker handled a container of CP on the shift; whether the worker disposed of waste containing CP on the shift; and whether the worker touched a surface potentially-contaminated with CP on the shift. All 14 of these independent variablesc were offered in a multiple linear regression model to predict dermal contamination. Four independent variables remained in the final model: the type of hospital, job title, sex, and whether the worker had a duty to handle antineoplastic drugs. A review of the residual plot of the final model suggested a random distribution. Table 4-4 provides an overview of the multiple linear regression results. c Department was found to be in linear combination with one or more other categorical variables and therefore was the first independent variable removed from the regression model 79 4.5 Discussion This study examined the dermal contamination levels of healthcare workers throughout the hospital antineoplastic drug medication system. Although a majority of the samples (80%) were below the analytical detection limit, the maximum measured dermal contamination levels in each of the eight different job categories, including those not responsible for drug preparation and/or drug administration, were all above the LOD of 0.356 ng/wipe. This confirms our hypothesis that, since drug contamination is present on surfaces throughout the hospital medication system, numerous job categories are at risk of exposure via dermal contact (Chapter 3). This is reinforced by our finding that the maximum reported level of dermal contamination was from a worker who had no known contact with CP during his/her work shift and the second highest overall contamination level was from an individual in a drug administration unit job category not responsible for the actual administration (volunteers, oncologists, ward aides and dieticians) Our results found the following four variables to be significant determinants of dermal exposure: the type of hospital, job title, sex, and whether the worker had a duty to handle antineoplastic drugs. Other studies examining determinants of dermal exposure are scarce. Both Connor et al. (7) and Fransman et al.(40) suggested that direct handling of vials is likely to result in higher dermal contamination levels. Such work is mainly performed by workers in the drug preparation area. In a recent study, Friese et al. found that organizational factors, such as adequate staffing and a check-and-balance system for drug administration, reduced the potential for skin exposure among oncology nurses; no other job categories were examined.(74) Testa et al. concluded that employment duration does not influence exposure, similar to the result we found that level of work experience was not associated with dermal contamination levels.(73) 80 The results of the multiple linear regression indicate that those who worked in acute care hospitals had higher dermal contamination levels than those employed at a cancer treatment hospital. This difference may be because the cancer treatment hospital has more stringent education/training with a broader range of workers being educated due to the specialized nature of the facility. However, with only one cancer treatment hospital participating in our study, additional studies are needed before we can confidently generalize this result. We do not believe that there were any differences in participant selection between the two types of hospitals; this matter is discussed in more detail in the concluding chapter. The regression model suggests that men were less likely to have dermal contamination than women. This is consistent with the findings reported by Geer et al. with respect to demographic factors associated with occupational dermal exposures.(93) Although we are unable to provide a physiological explanation for this phenomenon, we speculate that this may be because women generally have smaller hands than men do; therefore, if both sexes were to touch the same contaminated surface, the resulting drug concentration (in ng/cm2) would be greater on the woman’s hands. In addition, women generally tend to use hand lotions more frequently than men and the lotions may cause the drugs to adhere to the hands more so than individuals who do not apply lotions. Both job title and whether the worker had a duty to handle antineoplastic drugs were found to impact dermal contamination levels. This is not unexpected as certain job titles, based on their respective roles and responsibilities, have been deemed to be at higher risk of exposure to antineoplastic drugs. However, in our model, “pharmacist” served as the reference job category and all seven other job categories examined had positive coefficients with four job categories having significantly higher hand contamination: porters; registered nurses; transport staff; and 81 other job categories based in the drug administration unit (volunteers, oncologists, ward aide and dieticians). The result that porters, transport staff and other job categories within the drug administration unit have higher dermal contamination is a novel finding as pharmacists have long been considered a high-risk cohort for occupational exposure to antineoplastic drugs due performing duties within the drug preparation unit.(94) It is important to note that at all participating hospitals, pharmacy technicians, not pharmacists, were tasked with drug preparation. Those who had a duty to handle antineoplastic drugs had higher levels of dermal contamination than those who did not and, since a majority of pharmacy and nursing personnel have this duty, the hand contamination of those employees is likely reflected in this variable rather than the variable of job title. The dermal contamination levels measured in the current study were lower than those found in the literature (see Table 4-5), with the maximum reported level consistent with the mean dermal contamination levels indicated in other studies. The reason for the difference may be that several of the published studies collected dermal samples immediately after participants performed tasks involving CP(67)(40), which would likely result in elevated contamination levels. The fact that 20% of the dermal samples in the current study had some measureable levels of drug contamination suggests that further improvements can be made to minimize occupational exposure. This includes improved compliance with glove usage when handling as well as appropriate hand hygiene after handling or being in an area where the drugs are stored/mixed/administered (discussed later on). A review of the literature found a proposed dermal occupational exposure limit (DOEL) for CP of 4 ng/cm2 (95) or 3360 ng/wipe using an average surface area for both hands of 840 cm2.(96) None of the wipe samples in the current study exceeded the DOEL. However, it is important to 82 mention that the DOEL was calculated based on absorption of CP on a daily basis. Our study design was cross-sectional and therefore we are unable to speculate on a worker’s exposure over a full day. The fact that drug residual was found even though workers wore gloves immediately before sample collection confirms previous studies which concluded that permeation of gloves is possible.(97)(98) Although a controlled laboratory study found that various glove materials are impermeable to antineoplastic drugs(99), under actual working conditions, the durability of the gloves may change resulting in leakage.(100) It is therefore recommended that various combinations of glove material be evaluated during the course of normal duties to ensure that users are adequately protected when wearing gloves. It is interesting to note that, of the various cleaning agents employed for washing hands at the participating facilities, none appeared to be more effective than any other in removing drug residual. We asked about cleaning agents in both the self-administered questionnaire and in the short survey of participants immediately before sample collection took place. These results are consistent with our finding that, even though a surface was reportedly cleaned, it had little impact on the residual drug contamination levels (Chapter 3). We recommend that hand cleaning protocols are reviewed to ensure that they are indeed effective in removing drug residual. Our results also indicated that individuals had dermal contamination at the start of their work shift (i.e. 6 participants provided hand wipe samples before they officially started their shift). This suggests dermal exposure on a previous shift with measurable drug residual due to ineffective hand washing and/or that load transfer of drug is occurring prior to actual handling 83 activities. Additional studies examining the phenomenon of skin contamination before the start of work shift is suggested. We found that glove usage and hand washing practices amongst healthcare workers appears to vary from person-to-person. Whether a worker had a duty to handle antineoplastic drugs, had ever received training related to the safe handling of antineoplastic drugs or had previous direct, unintended contact with antineoplastic drugs while at work were all inconsequential to complying with glove usage guidelines. The lack of adherence to hand protection is not due to availability of gloves as a majority of participants (N=196; 87%) agreed or strongly agreed that they were readily available. Nor is it likely due to the lack of knowledge surrounding glove usage as a large proportion of all participants (N=180; 81%) agreed or strongly agreed that they believed they could use the required gloves properly. Our findings support those of others who concluded that healthcare workers do not always comply with personal protective equipment usage and optimal hand hygiene practices.(101) As all agencies that have developed safe handling guidelines have indicated using gloves for protection (8)(9)(10), it is therefore recommended that barriers to compliance with glove usage and hand hygiene protocols be investigated for all at-risk job categories within the hospital medication system. 4.5.1 Limitations Limitations of the current study need to be addressed. When the results were stratified by the independent variables, in many instances select categories within each variable had relatively small sizes and therefore we had limited power to detect differences in these groups. Another limitation is that a large proportion of the dermal contamination levels were less than the detection limit. 84 At least one job cohort was not represented that has potential for exposure to antineoplastic drugs – housekeepers. Housekeepers at the participating sites are responsible for removing the cytotoxic waste bins from each facility. Unfortunately, as the contract company that employed the housekeepers declined to participate, we were unable to recruit this job category. As this study was cross-sectional in nature, the findings are limited to the point in time when samples were collected; cumulative occupational exposure to antineoplastic drugs, if any, cannot be elucidated. Another temporal limitation is the fact that some of the independent variable data were collected on site while other data were gathered from the self-administered questionnaire. We were therefore unable to make certain comparisons, such as the actual frequency of glove usage when there was potential CP contact during the sampling work shift. There is a possibility that the reported contamination levels were an underestimate of actual exposure, as CP may have been absorbed through the skin.(102) In addition, we analyzed for only one antineoplastic drug. We were also unable to quantify the amount of CP handled and/or contacted. Although we attempted to ask this information from participants, the reported amounts were not considered reliable because a majority of participants were unable to answer this question and those who did respond merely provided an estimate. Lastly, many of the independent variables were gathered via self-reports. This can result in misclassification of responses and, depending on the response, could bias the outcomes of the statistical analyses in either direction; though non-differential misclassification and bias to the null are most likely. 4.5.2 Summary To our knowledge, this is the first study of its kind demonstrating dermal contamination to antineoplastic drugs of job categories throughout the hospital medication system. The findings are consistent with our earlier suggestion that there is likely an underestimate of the total number 85 of healthcare workers at risk of exposure to antineoplastic drugs.(80) We found four variables were significantly associated with dermal contamination. The determinants that were associated with increased exposure were: 1) employment at an acute care hospital, 2) work as a porter, nurse, transport staff or in the drug administration unit, 3) being female, and 4) having a duty to handle antineoplastic drugs. As this is the first study of its kind, additional studies are recommended to confirm or refute the reported associations. 86 * Reasons for refusal include fear of results, lack of time, perceived sample collection method as intrusive, and lack of interest in study. ** Reasons for non-contact include long-term illness, moved, transferred to another job within organization, did not respond to repeated call backs, or retired. Figure 4-1 Participant recruitment flowchart 87 Table 4-1 Dermal cyclophosphamide (CP) contamination levels of personnel in the hospital medication system, stratified by demographic variables: summary statistics, one-way ANOVA results and percent of samples less than detection limit Variable Subcategory N AM SD GM GSD ANOVA (p-value)¥ % < LOD (ng/wipe) (ng/wipe) (ng/wipe) Hospital type Cancer centre 22 0.014* 1.42 -0.863* 3.98 0.023 72.7 Acute care hospital 93 0.444* 2.65 0.135* 1.40 82.3 Department Pharmacy 46 -0.007* 1.36 -0.518 2.67 0.130 84.6 Drug Administration 59 0.668 3.18 0.250* 1.46 75.7 Other 10 0.285* 1.04 0.208* 1.18 87.0 Job title Pharmacist 20 -0.318* 1.08 -1.123* 4.15 0.198 90.0 Pharmacy Receiver 6 0.239* 0.395 0.224* 1.09 75.0 Pharmacy Technician 23 0.232* 1.63 -0.001* 1.42 82.2 Porter 6 0.404 1.37 0.280* 1.25 90.9 Registered Nurse (includes LPN)§ 33 0.767 3.13 0.363 1.46 73.4 Transport (biopacker, shipper/receiver, and transporter) 4 0.121* 0.212 0.117* 1.05 87.5 Unit clerk 12 -0.074* 0.978 -0.203* 1.31 83.3 Other workers in drug admin unit (volunteer, oncologist, dietician, ward aide) 11 1.32 4.93 0.504 1.64 71.4 Sex Female 92 0.455 2.69 0.118* 1.43 0.038 77.8 Male 23 -0.013* 1.18 -0.792* 3.85 91.1 Age 20 to 29 12 0.254* 1.17 0.116* 1.30 0.679 79.2 30 to 39 32 0.133* 1.37 -0.507* 3.18 77.4 40 to 49 33 0.586 3.06 0.175* 1.48 80.3 50 to 59 28 0.472 3.25 0.063* 1.45 83.6 60+ 10 0.119* 1.15 -0.137* 1.55 83.3 Do you directly handle, prepare and/or administer antineoplastic drugs as part of your normal duties? Yes 91 0.433 2.70 0.096* 1.43 0.073 80.3 No 24 0.086* 1.25 -0.689* 3.75 80.9 Percentage of time spent handling, preparing and/or administering antineoplastic drugs? < 25% 53 0.155* 1.18 -0.006* 1.35 0.352 82.1 > 25% 38 0.524 2.99 0.141* 1.46 78.4 Never 24 0.575 3.56 -0.580* 3.99 80.0 * Reported values are less than the limit of detection of 0.356 ng/wipe ¥ Where the dependent variable was the ln-transformed dermal contamination level § LPN = licensed practical nurse 88 Table 4-2 Dermal cyclophosphamide (CP) contamination levels of personnel in the hospital medication system, stratified by reported contact methods during work shift: summary statistics, one-way ANOVA results, and percent of sample less than detection Contact query Response N AM SD GM GSD ANOVA (p-value)¥ % < LOD (ng/wipe) (ng/wipe) (ng/wipe) All participants N/A 225 0.360 2.46 -0.081* 1.98 N/A 80.4 Did you contact CP (by any means)? No§ 162 0.291* 2.69 -0.235* 2.20 0.184 82.2 Yes 63 0.538 1.72 0.342* 1.41 74.6 Did you mix/compound CP? No 221 0.314* 2.40 -0.113* 1.98 0.211 81.0 Yes 4 2.92 4.38 2.11 1.72 50.0 Did you physically handle a container of CP? No 192 0.343* 2.54 -0.125* 2.07 0.575 80.7 Yes 33 0.463 1.99 0.183* 1.41 78.8 Did you dispose of waste (IV bag or vial) containing CP? No 210 0.343* 2.49 -0.112* 2.02 0.546 80.5 Yes 15 0.602 2.01 0.369 1.34 80.0 Did you touch a surface which came into contact with a container / package of CP? No 176 0.311 2.59 -0.176* 2.13 0.335 81.8 Yes 49 0.538 1.94 0.279* 1.38 75.5 Did you administer CP to a patient? No 213 0.326 2.47 -0.121* 2.01 0.373 81.7 Yes 12 0.962 2.17 0.694 1.36 58.3 Did you disconnect an IV line containing CP? No 216 0.330* 2.49 -0.112 2.00 0.425 81.0 Yes 9 1.08 2.52 0.729 1.43 66.7 Did you provide physical care for a patient on CP? No 222 0.333* 2.43 -0.100* 1.99 0.406 80.6 Yes 3 2.37 4.45 1.52 1.86 66.7 Did you dispose of body fluids from a patient on CP? No 225 0.360 2.46 -0.081* 1.98 N/A 80.4 Did you clean up a spill/leak containing CP? No 225 0.360 2.46 -0.081* 1.98 N/A 80.4 Did you eat/drink in an area where CP is mixed, handled or administered? No 197 0.335* 2.57 -0.145* 2.07 0.369 81.7 Yes 28 0.538 1.51 0.395 1.25 71.4 * Reported values are less than the limit of detection of 0.356 ng/wipe ¥ Where the dependent variable was the ln-transformed dermal contamination level § All “No” responses are actually No/Don’t know combined 89 Table 4-3 Dermal cyclophosphamide (CP) contamination levels of personnel in the hospital medication system, stratified by protective measures employed by participants: summary statistics, one-way ANOVA and percent of samples less than detection Variable Subcategory N AM SD GM GSD ANOVA % < LOD (ng/wipe) (ng/wipe) (ng/wipe) (p-value)¥ All participants N/A 225 0.360 2.46 -0.081* 1.98 N/A 80.4 Frequency of glove use when handling antineoplastic drugs Some of the time 43 0.155* 1.51 -0.140* 1.36 0.342 84.7 All the time 49 0.442 2.53 0.140* 1.42 76.8 Not applicable 23 0.575 3.56 -0.580* 3.99 80.0 Frequency of glove use when in an area where antineoplastics are handled Never 29 -0.10* 0.939 -0.138* 1.31 0.388 86.0 Some of the time 55 0.383 2.65 -0.280* 2.56 79.6 All the time 31 0.671 3.02 0.361 1.34 76.7 Number of gloves worn immediately prior to sample collection* 0 (none) 202 0.363 2.48 -0.066* 2.01 0.546 81.2 1 pair 19 -0.057* 1.39 -0.465* 1.76 73.7 2 pairs 3 2.91 5.53 1.71 2.05 66.7 Frequency of hand wash after glove usage Not all the time 78 0.188* 1.40 -0.208* 2.19 0.326 79.6 All the time 37 0.720 3.81 0.195* 1.51 82.2 Frequency of hand wash after being in an area where antineoplastics are handled Not all the time 83 0.412 2.83 -0.137* 2.19 0.634 82.7 All the time 32 0.226* 1.02 0.066* 1.37 74.6 Average duration of usual hand wash < 15 sec 87 0.343* 2.68 -0.167* 2.15 0.416 80.7 15+ sec 28 0.415 1.59 0.202* 1.38 79.6 Usual hand wash method Water only 1 0.000* 0.000 0.000 1.00 0.937 100.0 Soap and water 87 0.379 2.72 -0.132* 2.15 82.2 Alcohol hand gel 18 0.272* 1.68 -0.044* 1.53 77.8 Other 9 0.404 0.857 0.336* 1.19 66.7 Time since most recent hand wash (prior to sample collection)§ < 10 min ago 77 0.929 3.76 0.403 1.50 0.220 75.3 10 to 30 min ago 71 0.029* 1.21 -0.557* 2.99 88.6 31 to 60 min ago 43 0.083* 1.72 -0.178* 1.44 79.1 60+ min ago 34 0.116* 0.642 0.061* 1.19 79.4 Most recent hand wash method (prior to sample collection)§ Water only 8 0.126* 0.720 0.080* 1.16 0.930 87.5 Soap and water 189 0.370 2.61 -0.131* 2.10 81.0 Alcohol hand gel 22 0.520 1.76 0.302* 1.35 68.2 Other 6 -0.010* 0.190 -0.075* 1.05 100.0 Number of times hands washed on shift (prior to sample collection)§ 1 to 2 43 -0.197* 1.19 -1.036* 4.04 0.024 81.4 3 to 4 64 0.002* 1.11 -0.172* 1.38 84.4 5+ 112 0.768 3.25 0.382 1.39 80.4 0 (start of shift) 6 0.573 0.817 0.521 1.17 33.3 * Values reported are less than the limit of detection of 0.356 ng/wipe ¥ Where the dependent variable was the ln-transformed dermal contamination level § Information collected from participants while on site. Total N is 225 for these questions because responses may vary between the two sampling dates 90 Table 4-4 Coefficients, standard errors and p-values for final multiple linear regression model showing variables associated with dermal contamination (ln-transformed) Explanatory Variable Subcategory N Coefficient Std Error p-value Intercept 0.665 0.185 0.00 Type of hospital Cancer Treatment 44 Ref Acute Care 181 0.272 0.118 0.02 Job title Pharmacist 40 Ref Pharmacy receiver 12 0.309 0.220 0.16 Pharmacy technician 45 0.200 0.147 0.17 Porter 11 0.583 0.236 0.01 Registered nurse 64 0.284 0.137 0.04 Transport 8 0.874 0.286 0.00 Unit clerk 24 0.294 0.177 0.10 Other workers in drug admin unit 21 0.625 0.194 0.00 Sex Female 180 Ref Male 45 -0.307 0.128 0.02 Duty to handle antineoplastic drugs? No 178 Ref Yes 47 0.393 0.142 0.01 Table 4-5 Comparison of cyclophosphamide (CP) dermal contamination levels reported in the current and other published studies, in reverse chronological order Concentration of CP Author Comments AM = 0.360 ng/wipe (range of < LOD to 22.80 ng/wipe) Current study GM = 0.98 ng/wipe (range of < LOD to 3.96 ng/wipe) Hon et al. (2011)(42) Pilot study; small sample size (N=9) Concentrations in the µg range Turci et al.(2010)(66) Used pads for collection which are known to overestimate exposure(40) 25.2 ng/wipe1 van Wendel de Joode et al.(2005)(103) Samples collected from nurses, technicians and cleaners – not stratified by job title Not detected Fransman et al. (2006) (104) Small sample size (N=3); from workers in an industrial laundry facility handling hospital bed linen 13.6 ng/task - pharmacy technicians; 28.6 ng/task - nurses handling urine; 154.0 ng/task - nurses washing patients; 57.8- nurses removing bed sheets; 12.5 ng/task - cleaning personnel cleaning toilets Fransman et al. (2005)(40) Arithmetic means presented; units are task-based and therefore presents challenges for direct comparison 37,800 ng – Pharmacy technicians; 163.8 ng – oncology nurses2 Fransman et al. (2004) (67) Samples collected immediately after the completion of tasks involving cyclophosphamide 20 to 63,400 ng/glove (on the internal side of gloves) Minoia et al.(1998)(61) Study conducted before NIOSH guidelines released; potential that gloves were contaminated during doffing resulting in elevated levels(105) 1 calculated using 840cm2 as surface area of both hands(96) 2 calculated using mean duration of task as per Table 3 of reference #(67) 91 5 EVALUATION OF HEALTHCARE WORKERS’ URINARY CONTAMINATION LEVELS THROUGHOUT THE HOSPITAL MEDICATION SYSTEM AND IDENTIFICATION OF DETERMINANTS OF SUCH EXPOSURE 5.1 Synopsis Chapter 4 found evidence that healthcare workers involved in some capacity with the hospital medication system are at risk of dermal exposure to antineoplastic drugs. However, dermal wipe levels may not be reflective of absorbed dose nor do they not take into account other possible routes of exposure. The objective of this chapter was to measure drug contamination in the urine of healthcare workers, to overcome the shortcomings of dermal samples. As in the previous chapters, determinants associated with urinary contamination were also identified. Workers who provided hand wipe samples also participated in urine monitoring. Participants were asked to provide 24-hour urine samples which were subsequently analyzed for CP and three of its urinary metabolites using high-performance liquid chromatography-tandem mass spectrometry. Urine samples from non-hospital control subjects were collected to confirm that exposure to hospital participants occurs. Independent variables from the same survey instruments used in Chapter 4 were offered into a multiple linear regression to identify determinants associated with urinary contamination. A total of 223 urine samples were collected (115 participants with 108 providing a second sample). The mean urinary CP concentration of study participants was 17.4 nmol/L, the GM 4.16 nmol/L, and the GSD 3.60, with a range from 0.1 to 839 nmol/L. The mean, GM and maximum value from hospital participants were all greater than the corresponding values from non-hospital controls. When stratified by job title, pharmacy receivers had the highest mean contamination level as well as the maximum reported level overall. Two variables were found to be statistically significant following bivariate analysis - the number of drug transport job categories and the number of drug administration units within a site. We identified two 92 determinants associated with urinary contamination – jobs as a pharmacy receiver, pharmacy technician, porter, nurse, or unit clerk and having a larger number of drug transport job categories at the facility. 5.2 Introduction The primary route of healthcare workers’ exposure to antineoplastic drugs is considered to be through dermal contact.(38)(40) This can occur either: a) directly through handling of the drug vials or IV preparations or b) indirectly by contacting drug-contaminated surfaces/objects. We reported in Chapter 4 that healthcare workers are indeed at risk of exposure to antineoplastic drugs by confirming the presence of drug contamination on workers’ hands. In addition, we demonstrated that an assortment of job categories throughout the hospital medication system (process flow of drugs within a facility from cradle-to-grave) have measurable levels of dermal contamination suggesting that previous occupational exposure studies have underestimated both the number and variety of healthcare workers at risk. However, the use of dermal wipe samples is limited by the fact that they are only representative of the level of external contamination - they cannot estimate the amount that may be absorbed by an individual. The absorbed dose is important because most antineoplastic drugs, including the marker drug in our study cyclophosphamide (CP), are initially inactive until metabolized and can subsequently exhibit cytostatic effects.(13) Furthermore, dermal wipe samples do not take into account other potential routes of exposure such as inhalation or ingestion and, therefore, may not be the most reliable or accurate estimate of exposure. Given this, a number of occupational exposure studies have collected urine samples as a biomarker to estimate the dose of antineoplastic drug exposure.(106)(75)(107)(36) 93 Since we demonstrated that various job categories within the hospital medication system have antineoplastic drug contamination on their hands, it would also be sensible to determine the body burden of these hazardous agents in the same workers. Based on our review of the literature, no single study has simultaneously measured urinary drug contamination levels of all healthcare job categories across the hospital medication system potentially exposed to antineoplastic drugs. The aim of this study was to assess the urinary contamination levels of the same job categories that were evaluated for dermal exposure. In addition to quantifying the levels of CP in urine, we also measured three of its more stable urinary metabolites, carboxyphosphamide, 4- ketocyclophosphamide, and N-dechloroethylcyclophosphamide, and summed the four analytes to ascertain contamination levels. This was done in order to obtain a more accurate indication of overall urinary contamination because a) less than 20% of the administered dose of CP is eliminated unchanged in the urine(13) and b) there is a large inter-individual variability in metabolism of CP, so the percentage of parent product excreted varies between individuals.(72)(108) In addition to quantifying the drug output, we sought to identify individual and workplace factors that may be associated with urinary contamination levels. To our knowledge, this is the first study of its kind to examine urinary contamination of a broad range of at-risk healthcare job categories and the determinants of such exposure. 5.3 Methodology 5.3.1 Selection of participants Participants enrolled in the current study were the same individuals that provided hand wipe samples in Chapter 4. Every participant was given an honorarium of $25 each for providing a 24-hr urine sample. 94 5.3.2 Selection of control subjects One control subject was recruited from a local university for approximately every 10 study participants. Control subjects were representative of the study population with respect to age and sex. A potential control subject was excluded from the study if he/she worked at or frequently visited a healthcare facility that prepares and/or administers antineoplastic drugs. Fourteen control subjects were recruited, of whom eight provided a second urine sample, for a total of 22 urine samples from control subjects. 5.3.3 Collection of urine samples Urine sample collection took place immediately after a participant provided the hand wipe sample as outlined in the previous chapter. Participants were asked to provide 24-hour urine samples as this sampling period is recommended by the U.S. Environmental Protection Agency for biological monitoring of toxic substances(109) and is also consistent with other urinary biomarker studies of anticancer agents.(12) In addition, urinary elimination of CP and its metabolites is almost complete 24 hours after uptake.(13) (Note that the standardized collection time of 24-hours eliminates the need for creatinine correction of urine samples). Table 5-1 lists the supplies that were provided to each participant for the collection of urine samples. Every collection jar was pre-filled with 50 mL of phosphate buffered saline solution (137 mM NaCl, 2.7 mM KCl, 1.76 mM KH2PO4 and 10 mM Na2HPO4) (Sigma Aldrich, Oakville, ON) before being furnished to participants. The buffer served to minimize the effect of pH on the stability of the urinary metabolites. To prevent potential degradation, participants were asked to keep the urine samples refrigerated whenever practical and, where a refrigerator was unavailable, it was recommended to participants to use the instant cold packs provided to facilitate the maintenance of a cool storage environment. 95 Due to the size limitations of the collapsible cooler, only nine polypropylene collection jars could be placed inside each cooler. To minimize the probability of participants using all the supplied collection jars, research team members collected the first urine sample of the set from every participant while on site (i.e. every participant had a total of ten jars). This first sample jar from every participant was placed into a separate portable cooler with ice packs and at the end of the sampling day was transported to the analytical laboratory where it was stored in a walk-in refrigerator (4oC). Once the 24-hour sampling period was completed, participants notified a local courier company to have the samples delivered to the laboratory. Upon receipt of the remaining samples, all collected urine from a participant was pooled, the total volume of urine estimated and three 5-mL aliquots of every participant’s urine sample were placed into cryogenic tubes and subsequently stored at -80oC until analysis. Sample collection took place between June 2010 and February 2011. A second set of 24-hour urine samples were collected from most participants with at least three weeks’ lag between collection times (average of 97 days; range 22 to 188 days). 5.3.4 Urine sample analysis For every participant’s urine sample, one of the 5 mL tubes was allowed to thaw. After thawing, contents were transferred to a scintillation vial and an internal standard, 50 µL of D4-CP (0.05 ng/µl) (University of Bielefeld, Bielefeld, Germany), was added. Ethyl acetate solvent (Sigma Aldrich, Oakville, ON) was added to separate the organic matter from the aqueous layer; this step was repeated three times until virtually all the organic matter could be extracted. The culture tube containing organic matter was then allowed to dry under a gentle stream of nitrogen 96 gas. Once dry, the residual was reconstituted in 1.0 mL 0.1M ammonium acetate (Sigma Aldrich, Oakville, ON) and this amount was transferred to liquid chromatography vials. The urine was analyzed for CP and three of its metabolites, 4-ketocyclophosphamide (4-keto), carboxyphosphamide (carboxy), and N-dechloroethylcyclophosphamide (ethyl-CP) (Bielefeld University, Bielefeld, Germany), by high-performance liquid chromatography-tandem mass spectrometry using an Agilent Technologies 6410 Triple Quadruple LC-MS/MS (Santa Clara, CA). The instrument utilized a Zorbax XDB-C18 column (Agilent Technologies, Santa Clara, CA) with a gradient mobile phase of 5 mM Ammonium Acetate:100% Methanol (A:B) (Sigma Aldrich, Oakville, ON) and samples were run at a flow rate of 0.5 mL/min. A 10-point calibration curve was used and, for quality control purposes, a calibration standard was run for every 10 samples. For additional quality control, QC spike samples, urine spike samples and blanks were included in the analysis. Ten percent of the samples in each batch were run in duplicate and a duplicate response that varied by more than 10% prompted reanalysis of the batch. The limit of detection (LOD) for each analyte was as follows: 4-keto 0.044 ng/mL, carboxy 0.0373 ng/mL, CP 0.0522 ng/mL and ethyl-CP 0.035 ng/mL. Any analyte concentrations that were less than the corresponding limit of detection (LOD) were substituted using LOD/2 as commonly performed on occupational exposure data.(82) The recovery rate of each analyte was calculated, averaged following multiple trials, and the corresponding urinary concentration values were adjusted to reflect the recovery rate (4-keto 85.2%, carboxy 4.1%, CP 111% and ethyl-CP 39.9%). In addition, the molar weight of each analyte was factored into the urinary concentration levels (4-keto MW=275.07 g/mol, carboxy MW=293.08 g/mol, CP MW=261.09 g/mol, and ethyl-CP MW=198.59 g/mol). The sum of the molar concentrations of the four analytes for each sample was calculated and urinary contamination levels were reported in nanomol per liter (nmol/L). 97 5.3.5 Supplemental data collection and questionnaire As detailed in Chapter 4, all participants were interviewed on-site using the “Activity-Related Questionnaire” and were asked to complete the self-administered “Cytotoxic (Antineoplastic) Drug Exposure Questionnaire”. 5.3.6 Statistical analysis Both untransformed and ln-transformed data were used to examine the distribution of urinary contamination levels. Summary statistics (arithmetic mean (AM), geometric mean (GM), geometric standard deviation (GSD), minimum and maximum) were used to describe all urine samples. When ln-transformed, the measurements more closely approximated a normal distribution than the corresponding untransformed data. As such, all statistical analyses utilized the ln-transformed values and the GMs were compared in ANOVA. Summaries of the urinary contamination levels were stratified by various independent variables from the survey instruments. “No” and “don’t know” responses to pertinent questions from either of the two survey tools were combined into one category and compared with “yes” respondents of the same question. A paired t-test was conducted to determine if there was a statistical difference in mean urinary contamination between the two sample collection times. To determine which factors were associated with urinary contamination levels, a multiple linear regression analysis was conducted in several steps. First, we considered whether there was plausible support for a relationship between the factor and exposure. Second, we examined whether the variables were associated with exposure (ln-transformed urinary contamination level) in bivariate analyses (one-way ANOVA for categorical variables and simple linear regression for continuous variables). Tukey’s post-hoc tests were performed following ANOVA to determine which differences in geometric means within a categorical variable were 98 significantly different. Those independent variables with p < 0.20 in the bivariate analyses and for which the direction of association could be logically interpreted, as well as variables with a strong a priori hypotheses for exposure were then offered into a multiple linear regression model with the ln-transformed molar urinary contamination of all four analytes combined serving as the dependent variable. The model also included a random effect with independent covariance structure for subject to account for potential correlation within repeated samples of the same subject. A manual backwards stepwise approach was utilized to identify those independent variables with p < 0.10; these were retained in the final model. A residual plot was generated to determine the appropriateness of the final model. Statistical analyses were performed using R version 2.13.1 (The R Foundation for Statistical Computing). 5.4 Results 5.4.1 Characteristics of study population The participants for the current study were the same as those who provided dermal wipe samples from Chapter 4. 5.4.2 Summary of urinary contamination levels – participants and controls Of the 115 participants, 108 provided a second sample, resulting in a total of 223 urine samples. The mean urinary concentration of CP and its metabolites in study participants was 17.4 nmol/L, the GM 4.16 nmol/L, and the GSD 3.60, with a range from 0.1 to 839 nmol/L.d In comparison, the mean urinary contamination for control subjects was 6.80 nmol/L, the GM 3.98 nmol/L, and the GSD 2.58 with a range from 1 to 45.8 nmol/L. The paired t-test found a statistically d Our dataset had one very large outlier that skewed the summary statistics. For quality assurance purposes, this sample was re-analyzed along with other randomly-selected samples. All re-analyzed results reflected the original urinary concentrations except for this outlier. In addition, the re-analyzed result of the outlier was found to be more consistent with the other sample concentrations as well as with the second sample from the same participant. Therefore, the re-analyzed result for this one sample was employed for all statistical analyses. 99 significant difference between the two sample collection periods with the second collection period having a higher mean (3.73 nmol/L in round #1 compared to 31.9 nmol/L in round #2). 5.4.3 Urinary contamination levels by individual factors Table 5-2 presents a summary of the urinary contamination levels grouped by various characteristics of the study population. When stratified by job title, the highest mean concentration belonged to the pharmacy receivers, one of whom had the maximum reported concentration. The difference in contamination levels between job titles was not statistically significant (F(7,215)=1.74, p=0.10; largest difference in job means was 2.7 times). As expected, participants who reportedly handled antineoplastic drugs as part of their normal duties had higher urinary contamination levels (both AM and GM) than those who did not (F(1,221)=2.58, p=0.11; difference in means of 1.28 nmol/L). Similarly, those who regularly worked in an area where antineoplastic drugs were handled, prepared or administered had higher urinary contamination levels (F(1,221)=0.10, p=0.75; difference in means of 0.250 nmol/L). Those who handled antineoplastic drugs less than 25% of the time had higher urinary contamination levels (both AM and GM) than those who handled the drugs more frequently (F(1,176)=0.29, p=0.06; difference in means of 0.465 nmol/L). The cohort that handled antineoplastic drugs less than 25% of the time had a smaller proportion of individuals who received training related to safe handling of antineoplastic drugs (21% vs. 61%) (see below). None of the ANOVA results of the three drug-handling variables were statistically significant. Participants who indicated they received training related to antineoplastic drugs had lower levels of urinary contamination. The difference between those who received training compared to those who did not was nearly statistically significant (F(1,221)=3.23, p=0.07; difference in means of 1.29 nmol/L). 100 Neither the duration of participants’ tenure in their current position nor the length of time that participants had been handling antineoplastic drugs had a significant association with urinary contamination levels (p=0.98 and p=0.49, respectively) 5.4.4 Urinary contamination levels by hospital characteristics Table 5-3 is a summary of urinary contamination levels based on the categorical variables associated with characteristics of the hospital. A review of this table indicates that, although the maximum recorded contamination level was from an individual working at an acute care hospital, participants from the cancer treatment hospital had higher AM and GM urinary contamination levels than participants from acute care hospitals. This difference was not statistically significant (F(1,221)=2.23, p=0.14; difference in means of 1.50 nmol/L). An individual from the Pharmacy department had the highest overall urinary contamination level. However, there was no statistically significant difference in GM contamination levels between departments (F(2,220)=1.18; p=0.31; largest difference in groups means was 1.96 nmol/L). Similarly, there was no statistically significant difference in urinary contamination levels between sites that had an isolated room to prepare antineoplastic drugs compared to the site that had no isolated room (F(1,221)=0.38, p=0.54; difference in means of 0.573 nmol/L). Both the AM and GM were higher at facilities where antineoplastic drugs were initially delivered to the shipping/receiving department; this difference was not statistically significant (F(1,221)=2.51; p=0.11; difference in means of 1.123 nmol/L). Table 5-4 presents the results for those hospital characteristics that are continuous in nature and have a logical association with urinary contamination levels. Urinary contamination levels increased with increases in either the number of job categories responsible for drug transport or 101 the number of drug administration units in a facility. Both of these variables were statistically significant. 5.4.5 Urinary contamination levels by contact with cyclophosphamide (CP) We asked participants regarding their potential CP contact method within the past 24 hours (to allow for absorption and metabolism of CP) as well as on the current shift. There was no apparent association between urinary contamination levels and the methods by which an individual came into contact with CP within the past 24 hours and on the work shift (see table in Appendix I). 5.4.6 Multiple linear regression model The results of bivariate analyses identified ten variables with p < 0.20 and a logical direction of association: health authority; hospital type; job title; whether the worker had a duty to handle antineoplastic drugs; frequency of gloves use when handling antineoplastic drugs; where the drugs were initially delivered to the facility; whether the worker ever received training related to workplace health and safety of antineoplastic drugs; whether the individual had previous, unintended contact with antineoplastic drugs while at work; the number of job categories responsible for drug transport within a facility; and the number of drug administration units at the facility. Based on a priori hypotheses regarding potential exposure, another two variables were selected to be offered in the regression model: dermal contamination levels (from Chapter 4); and the number of chemotherapy patients with whom the participant was directly involved with in the past seven days. 102 All 12 of the aforementioned independent variables were offered into a multiple linear regression model to predict urinary contamination levels.e Two variables remained in the final model: job title and the number of job categories responsible for drug transport. With respect to the job title variable, five categories exhibited relatively large differences in urinary contamination levels compared with pharmacists (reference category): pharmacy receivers; pharmacy technicians; porters; registered nurses; and unit clerks. A review of the residual plot suggested a random distribution. Table 5-5 provides an overview of the multiple linear regression results. 5.5 Discussion Until now, little attention has been paid to the body burden of healthcare workers who are part of the hospital medication system other than pharmacy personnel and drug administration nurses. Our results demonstrate that, collectively, the healthcare workers examined have increased levels of CP and its metabolites in their urine compared to control subjects. When stratified by job title, pharmacy receivers (individuals responsible for ordering and stocking drugs) had the highest mean exposure and the maximum urinary contamination level measured. Twelve independent variables were offered into the regression model to identify determinants of urinary exposure for healthcare workers at risk throughout the hospital medication system. Two variables, job title and the number of job categories responsible for drug transport within a facility, proved to be statistically significant at p < 0.10. Of note, job title was found to be significantly associated with dermal contamination as well as in this urinary contamination study. In the urinary contamination model, “pharmacists” served as the reference job category and all seven of the other categories had positive coefficients with five job categories having relatively large differences – pharmacy receivers, pharmacy technicians, porters, registered e Number of drug administration units was identified to be in linear combination with one or more other variables and therefore was the first independent variable removed from the regression model 103 nurses and unit clerks. It was surprising that all other job categories had positive coefficients relative to pharmacists as the latter have long been considered a high-risk cohort with respect to occupational exposure to antineoplastic drugs.(94) However, unlike other studies, pharmacy technicians, not pharmacists, at participating sites were responsible for preparing the drugs. Therefore, it was not unexpected to find that pharmacy technicians had statistically significant elevated urinary contamination levels compared with pharmacists. Other studies have had varying conclusions regarding the urinary CP levels of pharmacists. Connor et al. reported urinary contamination was found in pharmacists only, with no measurable levels in the urine of nurses, pharmacy technicians or nurse assistants.(36) Sugiura et al. did not detect CP in the urine of any pharmacists and observed that urinary contamination levels varied considerably amongst the healthcare workers that participated in their study.(35) With respect to the second determinant identified in the current study, it is not clear why the number of job categories responsible for drug transport within a facility was associated with urinary contamination levels. This same variable was found to be associated with surface contamination levels (Chapter 3). As with the surface contamination, we speculate that the more job categories involved, the more drug handling activities are performed by various individuals – a large proportion of whom are not formally trained regarding the hazards of these agents or the associated safe handling requirements (based on questionnaire responses). This is supported by the fact that porters and unit clerks (proportion not trained was 82% and 74%, respectively), two of the three job categories responsible for drug transport at participating sites, had the highest GM urinary contamination levels when stratified by job title. As we are the first to identify this determinant, we suggest that future studies examine this phenomenon more closely. There have been other determinants of urinary contamination reported in the literature. Rekhadevi et al. found that age, years of exposure and duration of handling antineoplastic drugs 104 per day were positively associated with urinary CP concentrations, though age was the lone variable that was statistically significant. None of the exposed nurses in their study employed protective measures such as gloves nor did they work within safety cabinets, whereas nurses in the current study do not mix drugs and are required to wear gloves when administering.(64) Another study concluded that the amount of CP compounded was significantly correlated with urinary CP concentration despite the implementation of a revised compounding standard operating procedure at the facility(48) (based in part on the National Institute for Occupational Safety and Health (NIOSH) “Alert” aimed at protecting healthcare workers from exposure to cytotoxic compounds(9)). This same study also observed measureable levels of CP in morning urine samples even before pharmacists began compounding and, according to the authors, suggests that these workers may have been exposed through skin contact due to environmental contamination from compounding activities the day before.(48) This finding supports the conclusions from our two earlier chapters that surface contamination is widespread and, in turn, the drug residual can lead to dermal exposure. Lastly, Schreiber et al. concluded that the amount processed and the number of CP preparations resulted in an increase in the number of reportable urinary CP results for those tasked with drug preparation.(75) In addition, the authors found the use of a garbage can with a removable lid increased the number of positive samples of urinary CP.(75) We did not record the type of waste containers so could not examine this variable. Other studies examined determinants within more restricted cohorts and could not examine differences between the range of jobs involved in the hospital medication system. Rekhadevi et al. studied nurses only (tasked with drug preparation, drug administration and disposal of drug- contaminated body fluids)(64), Tanimura et al. looked solely at pharmacists responsible for compounding(48), and Schreiber et al. examined pharmacy technicians and pharmacists.(75) 105 The current study reported urinary contamination levels in molar concentration of CP and three of its urinary metabolites. However, the literature primarily presents urinary concentration of the parent product, CP, only. Given this, Table 5-6 compares the CP urinary levels in the current study to the results reported by others since 2004, the year that NIOSH initially released their “Alert.” We therefore believe that most of these studies would have comparable control measures implemented at the participating sites.(9) A review of the Table indicates that the urinary contamination levels in the current study are consistent with the findings of others; however, for all studies cited, there is a great deal of variability with respect to the proportion of samples above detection limits as well as the range of reported contamination levels. Differences in urinary CP concentrations within and between studies are likely due to variations in handling practices including personal protective equipment usage as well variability in metabolic rates amongst individuals. The latter is suspected to be due, in part, to genetic polymorphisms of the enzymes responsible for metabolizing xenobiotics.(17) This variability is supported by the low correlation between CP and sum of molar concentrations found in the current study (Pearson r = 0.22). In addition, liver function may influence the metabolism of CP. Where liver function is impaired, a smaller fraction of the CP dose is metabolized and a higher fraction of CP may be eliminated in the urine.(71) Of concern is the fact that we detected drug residual despite workers reporting not having any contact with CP. Long-term urinary excretion of CP (continuously positive urine samples) may signify continuous absorption of CP.(43) This may be due, in part, to the fact that the skin has a reservoir function whereby internal exposure may continue despite the cessation of external exposure.(95) 106 Sargent et al. indicated a no-significant-risk level (NSRL) of 1000 ng/day for CP,f the level at which no increase in cancer risk and no adverse effects on organ systems or the developing fetus are expected.(110) A review of our data shows that ten participants (4.5%) had exposure levels which exceeded the NSRL. Of these, four were nurses, two were unit clerks, two were pharmacy technicians, one was a pharmacist and the final sample was from an individual who worked in the drug administration unit but was not responsible for drug administration. Our results found no correlation between the dermal contamination levels (from Chapter 4) and urinary CP contamination. This could be due to the fact that we collected hand wipes at a single point in time which is not reflective of the day’s exposure. It could also suggest that other routes of exposure, such as inhalation and accidental ingestion, may play a role in the body burden of antineoplastic drugs. We performed bivariate analyses (ANOVA or linear regression) on 65 variables and only ten variables had a p < 0.20 that also demonstrated a logical direction of association. These 65 variables included environmental and organizational factors which have been associated with better safety practices in the workplace and, presumably, lower exposure levels.(111) However, we were unable to detect any association between the environmental and organizational factors examined and the urinary contamination levels. We believe that the lack of statistically significant results following bivariate analysis is related to the small numbers in each category and/or small differences in the exposure concentrations between groups. Further discussion on this matter is found in the concluding chapter. f Assumed that “day” is equivalent to 24 hours 107 5.5.1 Limitations Additional limitations of this study are described here. We were unable to accurately determine the amount of antineoplastic drug handled by each participant; therefore, we could not evaluate if this factor is associated with urinary CP contamination levels. We did not obtain information on participants’ contact with CP during the 24-hour sample collection period; this information may have allowed us to identify other variables associated with exposure. This study was cross- sectional in nature and therefore cumulative exposures to antineoplastic drugs cannot be determined. We only examined CP but other antineoplastic drugs are handled at the sites and we measured three urinary metabolites of CP but there are other, less stable metabolites that may have been present(13). Both of these factors likely lead to an underestimate of the true exposure levels. Lastly, a multitude of the independent variables were gathered via self-reports. This can result in misclassification of responses, and depending on the response, could bias the outcomes of the statistical analyses in either direction; though non-differential misclassification and bias to the null are most likely. 5.5.2 Summary To our knowledge, this is the first study of its kind demonstrating urinary contamination with antineoplastic drugs among individuals with jobs throughout the hospital medication system. From our regression model, job title and the number of drug transport job categories were found to be correlated with urinary contamination levels. Of note, the regression model suggests that other job categories may be more at risk than pharmacists – traditionally believed to be a high- risk cohort. Future studies are recommended to ensure that all at risk job categories throughout the hospital medication system are examined independently to identify specific determinants of exposure to antineoplastic drugs applicable for each category. In addition, we are in agreement 108 with others(44)(66) that measuring for CP and one or more of its urinary metabolites, as opposed to quantifying CP only, is a more accurate reflection of body burden.g g Based on the laboratory results, the proportion of analytes in the current study was as follows: ethyl-CP > 4-keto > CP > carboxy 109 Table 5-1 Supplies provided to each participant for collecting urine samples Number of units Supplied item (Manufacturer/Supplier) 1 Written instructions for collecting and returning urine sample (Appendix J) 10 250mL straight-side wide-mouth polypropylene jars with lids (Thermo Scientific) 1 12-can capacity collapsible cooler (Escort) 2 Ice gel liners (Cole-Palmer) 1 5” x 8” ice pack (Rapid-Aide) 1 Permanent marker 3 Large re-sealable plastic storage bags 4 Instant cold packs (Ardes Medical) 110 Table 5-2 Urinary contamination levels of cyclophosphamide (CP) and its metabolites of personnel in the hospital medication system, stratified by demographic variables: summary statistics and one-way ANOVA results Variable Subcategory N AM (nmol/L) SD (nmol/L) Max (nmol/L) GM (nmol/L) GSD ANOVA (p-value)* Overall 223 17.3 75.2 839 4.16 3.60 Job titles Pharmacist 20 3.83 3.41 18.3 2.57 2.74 0.10 Pharmacy Receiver 6 79.4 240 839 5.83 6.97 Pharmacy Technician 23 13.7 49.0 332 4.47 3.24 Porter 6 14.4 29.9 105 6.84 2.59 RN (includes LPN) 33 24.3 82.5 565 4.26 4.41 Transport (biopacker, transporter, and shipper/receiver) 4 5.54 2.76 9.84 5.01 1.60 Unit clerk 12 12.9 24.4 114 6.26 2.91 Other workers in drug admin unit (volunteer, oncologist, dietitian, and ward aide) 11 4.98 4.85 22.8 3.07 3.42 Sex Female 92 18.1 80.1 839 4.10 3.72 0.72 Male 23 14.4 51.3 332 4.43 3.16 Age 20 to 29 12 47.3 132 565 6.74 5.45 0.24 30 to 39 32 21.7 107 839 4.39 3.54 40 to 49 33 9.63 36.4 299 3.80 2.92 50 to 59 28 13.8 32.5 172 4.13 4.13 60+ 10 3.59 2.18 9.35 2.75 2.44 Do you directly handle, prepare and/or administer antineoplastic drugs as part of your normal duties? Yes 91 20.3 84.1 839 4.47 3.67 0.11 No 24 6.39 14.9 105 3.19 3.26 Percentage of time handling spent handling, preparing and/or administering antineoplastic drugs? < 25% 54 23.1 98.8 839 4.58 3.85 0.59 > 25% 38 15.5 53.7 332 4.11 3.48 Do you regularly work in an area where antineoplastics are handled, prepared and/or administered? Yes 82 19.5 84.4 839 4.23 3.74 0.75 No 33 11.9 42.9 332 3.98 3.29 Have you ever received training related to workplace health and safety of antineoplastic drugs? Ever 60 15.5 61.7 565 3.59 3.71 0.07 Never 55 19.5 88.0 839 4.89 3.44 Have you ever had previous, unintended contact with antineoplastic drugs while at work? Yes 18 26.3 60.9 299 5.48 5.03 0.17 No 97 15.7 77.6 839 3.95 3.35 * where the dependent variable was the ln-transformed urinary concentration level 111 Table 5-3 Urinary contamination levels of cyclophosphamide (CP) and its metabolites of personnel in the hospital medication system, stratified by categorical hospital characteristics: summary statistics and one-way ANOVA results Categorical variable Subcategory N AM (nmol/L) SD (nmol/L) Max (nmol/L) GM (nmol/L) GSD ANOVA* (p-value) Overall N/A 223 17.4 75.2 839 4.16 3.60 Hospital A 18 7.37 21.5 128 2.71 3.66 0.11 B 16 6.97 14.3 80.5 3.33 3.18 C 13 6.19 7.79 34.0 3.38 3.23 D 23 12.7 29.2 172 4.63 3.48 E 23 31.5 130 839 5.25 3.59 F 22 29.0 98.7 565 5.40 3.99 Health Authority 1 31 6.87 17.0 128 2.98 3.46 0.10 2 23 12.7 29.2 172 4.63 3.48 3 39 21.7 101 839 4.37 3.47 4 22 29.0 98.7 565 5.40 3.99 Hospital type Cancer centre 22 29.0 98.7 565 5.40 3.99 0.14 Acute care hospital 93 14.6 68.4 839 3.91 3.50 Department Pharmacy 46 18.6 94.5 839 3.64 3.55 0.31 Drug Administration 57 18.0 64.2 565 4.36 3.92 Other 12 9.78 20.8 105 5.60 2.26 Does the facility have an isolated room for antineoplastic drug preparation? Yes 92 18.6 82.7 839 4.05 3.64 0.54 No 23 12.7 29.2 172 4.63 3.48 Where are the drugs received at the facility? Pharmacy 54 9.35 23.0 172 3.59 3.52 0.11 Shipping/receiving 61 24.3 99.9 839 4.72 3.65 * where the dependent variable was the ln-transformed urinary concentration level Table 5-4 Urinary contamination levels of cyclophosphamide (CP) and its metabolites of personnel in hospital medication system, stratified by continuous variables of hospital characteristics: linear regression results Variable Intercept β Std Error p-value* No. transport job categories (range 1 – 3) -6.03 0.301 0.114 0.007 No. drug admin units (range 1-3) -5.95 0.233 0.095 0.014 * where the dependent variable was the ln-transformed urinary contamination level 112 Table 5-5 Coefficients, standard errors and p-values for final multiple linear regression model showing variables associated with urinary contamination levels (ln-transformed) Explanatory variable Subcategory N Coefficient Std Error p-value Intercept -6.47 0.312 < 0.001 Job title Pharmacist 39 Ref Pharmacy receiver 12 0.666 0.452 0.14 Pharmacy technician 45 0.580 0.296 0.05 Porter 11 0.772 0.470 0.10 Registered nurse 64 0.455 0.277 0.10 Transport 8 0.327 0.544 0.55 Unit clerk 23 0.597 0.375 0.11 Other workers in drug administration unit 21 0.053 0.392 0.89 Number of drug transport categories 0.240 0.132 0.07 113 Table 5-6 Comparison of cyclophosphamide (CP) urinary contamination levels reported in current study and other published studies in reverse chronological order since 2004 Proportion of samples > LOD (LOD) Cyclophosphamide concentration Comments Author and Country Mean in µg/L (range) Mean in ng/24hrs (range) 202/223 (0.052 ng/mL) 0.14 (< LOD – 2.37) 232.1 (< LOD – 7069) 24-hr urine samples from at-risk job categories throughout hospital medication system Current study 0/36 (0.2 ng/mL) None reported Pre- and post-shift urine samples from pharmacy technicians and nurses Sottani (2011); Italy(44) 1/1 (0.01 ng/mL) N/A 13.5 24-hr urine sample from pharmacist (not wearing PPE) Hama (2011)(52); Japan 7/40 (0.1 ng/mL) (0.1–1.2) N/A Post-shift urine samples from pharmacy techs, nurses and attendants Villarini (2011)(17); Italy 90/226 (None reported) N/A ( 2.7 to 462.8) Examined physicians, pharmacists and nurses. Sugiura (2011)(43); Japan 11/62 (0.01 ng/mL) N/A 29.3 Positive results from nurses administering drugs and a medical doctor. Sugiura (2011)(35); Japan 3/17 (0.1 ng) N/A (6.7 to 52) Sampled pharmacists responsible for mixing and checking Yoshida (2011)(62); Japan 0/35 (0.4 ng/mL) None reported Varying urine sampling times; urine samples from 6 pharmacists and 2 nurses. Maeda (2010)(112); Japan 2/67 (0.015 ng/mL) (0.000043 - 0.000079) N/A Results from two pharmacists that prepared drugs. Connor (2010)(36); USA 4/4 (None reported) N/A 165.3 [before changes to SOP]; 47.4 [after changes to SOP] Examined pharmacists responsible for compounding Tanimura (2009)(48); Japan 0/22 (0.01 ng/mL) None reported Pre- and post-shift samples from pharmacy workers, nurses, assistant nurses and cleaners. Hedmer (2008)(31); Sweden 42/52 (0.04 ng/mL) 440 (80 – 900) N/A Pre-shift samples from nurses exposed to antineoplastic drugs (not wearing PPE) Rekhadevi (2007)(64); India 0/50 (~ 1 nM) None reported Pre- and post-shift samples from pharmacy personnel Mason (2005)(45); UK LOD = limit of detection; N/A= not applicable; PPE = personal protective equipment 114 6 CONCLUDING CHAPTER 6.1 Study overview and objectives A review of the literature regarding healthcare workers’ exposure to antineoplastic drugs reveals that the drug preparation and drug administration units are most often examined. However, there have been evidence suggesting that drug contamination of surfaces is found beyond these two areas(20)(57) while others have argued that additional healthcare cohorts, besides pharmacy personnel and drug administration nurses, are potentially exposed to antineoplastic drugs, such as drug transport personnel and other hospital staff in the drug administration units.(68)(70) These theories are plausible because antineoplastic drugs follow a site-specific pathway known as the hospital medication system (process flow of drugs from receipt to disposal).(79)(10) In order to test the theory that surface contamination and occupational exposure occurs throughout the hospital medication system, the current study was undertaken with cyclophosphamide (CP) as the marker drug of exposure. The following is an outline of the four objectives of the study: Objective 1: Identify surfaces most likely contaminated and the job categories potentially at risk of exposure to antineoplastic drugs throughout the hospital medication system; Objective 2: Quantify the surface contamination levels throughout the hospital medication system and identify determinants associated with surface contamination; Objective 3: Assess the dermal CP contamination levels of at-risk healthcare job categories and identify determinants associated with dermal exposure; and Objective 4: Determine the urinary contamination levels of CP and three of its metabolites of at- risk healthcare job categories and identify determinants associated with urinary contamination. 115 6.2 Key findings The study took place at six hospitals (five were acute care facilities and one was a cancer treatment centre) within Metro Vancouver in the province of British Columbia. Every participating site frequently prepares and administers CP, a known human carcinogen.(15) The study objectives and the respective results of each were presented in separate chapters and below is a review of the key findings from each chapter. 6.2.1 Chapter 2 (Objective 1) Using repeated site observations, I identified 55 potentially drug-contaminated categories of surfaces located throughout the hospital medication system that were contacted frequently by healthcare personnel. Some of the more frequently contacted surfaces include a box cutter (delivery stage), writing instrument (drug preparation, delivery and drug administration stages), bin where drugs are held for pick up (transport to ward stage) and IV pump (drug administration stage). For the purposes of the study, exposure potential was defined as physically manipulating the drugs, contacting a potentially contaminated surface/object, or using an instrument that was touched by someone suspected of having drug-contaminated hands/gloves. Site observations found that an exposure potential exists at all five stages of the hospital medication system. Up to 11 job categories per site were potentially at risk of exposure. Job categories that were found to be at risk of exposure common to all participating sites were: pharmacy receivers, pharmacy technicians, pharmacists and nurses. In addition to pharmacy receivers, other novel job categories that were identified to be at risk that have not been mentioned in the literature or guidelines include volunteers, unit clerks, dieticians (in the drug administration stage), porters (transport to ward stage) and biopackers (waste disposal).(9) 116 6.2.2 Chapter 3 (Objective 2) Wipes samples were collected from those surfaces identified in Chapter 2 as potentially contaminated and analyzed using high-performance liquid chromatography-tandem mass spectrometry to quantify the level of CP. Four-hundred thirty-eight samples were collected (229 surfaces with 209 duplicates). The mean CP surface concentration was 0.201 nanograms per square centimeter (ng/cm2), the geometric mean 0.019 ng/cm2 and the geometric standard deviation 2.54 with a range of less than detection (0.356 ng/wipe; LOD in ng/cm2 varied with surface area) to 26.1 ng/cm2 (64% of surface wipe samples were less than the limit of detection). Measureable levels of CP were found at every stage of the hospital medication system with the highest average contamination found in the drug preparation stage. Of note, reported handling of CP and cleaning of surfaces did not appear to be associated with contamination. I identified two factors associated with surface contamination: the stage of the hospital medication system and the number of job categories responsible for drug transport. In particular, the drug preparation and drug administration stages had significantly higher surface contamination than the delivery stage and the more job categories tasked with drug transport at a facility, the higher the levels of surface contamination. 6.2.3 Chapter 4 (Objective 3) As surface contamination was detected at all stages of the hospital medication system, it was deemed important to confirm if an occupational exposure potential exists. Personnel selected for inclusion in the study were those job categories found in Chapter 2 that had a high probability of touching either the drugs directly or drug-contaminated surfaces. One-hundred fifteen workers agreed to participate with 110 able to provide a duplicate sample. The mean dermal CP concentration was 0.360 ng/wipe, the geometric mean was less than limit of detection (0.356 117 ng/wipe), the geometric standard deviation was 1.98 with a range of less than 0.356 to 22.8 ng/wipe (80% of dermal wipe samples were less than the detection limit). Female participants had statistically significant higher dermal contamination levels than males. When stratified by department, personnel working in the drug administration unit had the highest average contamination levels. When stratified by job titles, the maximum value in each job category exceeded the detection limit and the cohort with the highest average dermal CP concentrations were other workers in the drug administration unit (volunteers, oncologists, ward aides and dieticians). Despite glove use, measurable levels of drugs were found on the hands of workers. This could be due to contamination prior to donning of gloves or permeation of drugs through gloves or both. Variables that were not associated with dermal contamination included previous safe drug handling training, job tenure and experience handling antineoplastic drugs. With respect to determinants associated with dermal contamination, we identified four factors: type of hospital, job title, sex and whether the worker had a duty to handle antineoplastic drugs. Specifically, workers in acute care facilities had higher levels of contamination than workers in cancer treatment facilities; all examined job categories had higher CP concentration levels on their hands relative to pharmacists; males had lower dermal contamination levels; and having a duty to handle antineoplastic drugs resulted in higher CP concentrations on one’s hands. Interestingly, hand washing practices and contact with CP on the shift were not found to be significant variables in the final model for dermal contamination. In addition, the surface contamination levels from Chapter 3 were not correlated with dermal contamination. 6.2.4 Chapter 5 (Objective 4) Although measurable levels of CP were detected on the hands of individuals in jobs across all stages of the hospital medication system, whether individuals absorbed CP remained unknown. 118 Therefore, urine samples were collected from the same workers who provided dermal wipes to ascertain if a body burden exists. As approximately 20% of CP remains unchanged in the urine(13), we also measured for three of its metabolites and tallied all four analytes to determine urinary contamination levels. Non-hospital subjects were enrolled as controls for evaluating exposure potential of healthcare workers. A total of 223 urine samples were collected from participants (115 participants, 108 of whom provided a second sample). The mean urinary concentration for study participants was 17.4 nanomols per liter (nmol/L), the geometric mean was 4.16 nmol/L, the geometric standard deviation was 3.60 with a range of 0.1 to 839 nmol/L. In comparison, the non-hospital control subjects had lower values for all corresponding summary statistics providing evidence that the study participants were occupationally exposed to antineoplastic drugs. When stratified by job title, pharmacy receivers had the highest arithmetic mean urinary CP contamination levels. Both the number of job categories responsible for drug transport and the number of drug administration units within a facility were found to be associated with urinary contamination levels. Reported contact with CP did not appear to influence CP levels in urine. Variables that were not associated with urinary contamination included job tenure and experience handling antineoplastic drugs. With respect to determinants associated with urinary contamination, the job title and the number of job categories responsible for drug transport were significant. More precisely, all evaluated job titles had higher urinary contamination levels relative to pharmacists and the more job categories responsible for drug transport within a facility, the higher the urinary contamination levels. Of note, dermal contamination levels from Chapter 4 were not correlated with urinary contamination levels. 119 6.3 Conclusions and implications of research To date, no other body of work (i.e. single report/document) has examined drug contamination on work surfaces or occupational exposure of at-risk healthcare workers throughout all stages of the hospital medication system. Specifically, the drug delivery, drug transport and waste disposal stages of the hospital medication system had not been evaluated concurrently with the drug preparation and administration stages. In the current study, although levels were generally low, I found evidence that environmental contamination is present on various surfaces throughout the hospital medication system. As a result of this widespread contamination, I demonstrated that several healthcare job categories had measurable levels of dermal and urinary drug contamination. This implies that the number of healthcare workers at risk of exposure to antineoplastic drugs is higher than previously believed as the list of occupations exposed should, based on the current study, be extended from pharmacy and nursing personnel to include unit clerks, pharmacy receivers, porters, volunteers, biopackers and dieticians.(70) A national carcinogen exposure surveillance project, CAREX Canada, recently estimated that 17,000 Canadians are occupationally exposed to antineoplastic drugs; but this estimate considered only nurses and pharmacists as potentially exposed.(16) The results found here indicate that this estimate will need to be updated to include the additional at-risk job categories identified in the current study. Based on our estimates, the number of Canadian healthcare workers at risk of exposure to antineoplastic drugs could exceed 100,000. Furthermore, the study suggests that current control measures are not 100% effective in minimizing contamination and/or occupational exposure and that additional control efforts, through the identification of factors associated with contamination or exposure, are necessary to reduce the risk. 120 With respect to the risks associated with the contamination levels found, only 2% of the surfaces sampled exceeded the United States Pharmacopeia maximum suggested level of 1 ng/cm2(33) and none of the hand wipe samples exceeded a published dermal occupational exposure limit of 4 ng/cm2 per day.(95) However, as we collected wipe samples at a point in time only, it cannot be confirmed if cumulative exposure is likely to exceed either referenced threshold level. Ten participants (4.5%) had urinary CP contamination levels that exceeded a published no- significant-risk level (NSRL) of 1000 ng/day.(110) Individuals with drug levels greater than the NSRL have an increased risk of developing cancer. Based on the risk assessment reported by Sessink et al.(108), an individual exposed to the average urinary CP contamination levels in our study of 232 ng/24 hours has a lifetime cancer risk of 7.78 per million (see Table 6-1). Alternatively, if one were to use the cancer slope factor referenced by Sargent et al.(110), then the lifetime cancer risk would be 1.89 per million using our average urinary CP concentration (see Table 6-2). Regardless of the quantitative risk assessment method used for the calculations, both models suggest a slight increase in the number of cancer cases among healthcare workers who have absorbed CP. [For comparison purposes, urinary CP levels in patients 24-hours after treatment ranged from 1.75 ng/day to 190,000 ng/day(71), which translates to a maximum cancer risk of 636 per 100,000 using Sessink et al.’s calculation.(108)] 6.4 Recommendations to minimize occupational exposure to antineoplastic drugs The findings in the current study are in agreement with others that an exposure potential exists despite the implementation of control measures.(34)(36)(43)(44)(47) The objective of determining factors associated with surface contamination and occupational exposure was to allow for the identification of control initiatives that specifically address these factors. 121 With respect to determinants associated with surface contamination, reducing the number of job categories responsible for drug transport may minimize surface contamination levels. This can be achieved through configuration of the facility so that the drug preparation unit is closer to the patient administration areas or employing only dedicated staff for drug transport. I identified that the drug preparation area was the most contaminated stage in the hospital medication system at the participating sites. This area is likely the primary source of contamination for the remaining hospital medication stages of transport, drug administration and waste disposal. The use of closed system drug transfer devices during drug preparation has proven to be effective in other studies and it is recommended that they be adopted at the participating facilities.(113)(114)(115) Given that external contamination on vials was detected, another recommendation is to clean the drug vials prior to handling to minimize the level of external contamination.(51)(54) It is also suggested that every site perform routine surface monitoring to confirm reduction in contamination. This can be achieved using a recently developed rapid analytical technique.(116) Regarding factors that were identified as being associated with dermal contamination (hospital type, job title, sex, and whether the worker had a duty to handling antineoplastic drugs), little can be done to actually modify any of these factors. However, these factors do indicate the jobs and facilities where monitoring and training should be ongoing, and where effective control measures need to be sought. As I found that compliance with personal protective equipment and hand hygiene varied from individual-to-individual, enhancement of the safety culture/climate, as suggested by McDiarmid and Condon(70), may improve workers’ compliance with safe drug handling guidelines which advocates glove usage and regular hand washing. Whether increased compliance with these recommendations leads to reduced exposure will need to be tested. 122 The two determinants associated with urinary contamination, job title and number of job categories responsible for drug transport, were also found to be factors for dermal contamination and surface contamination, respectively. The same recommendations suggested earlier for these two factors apply for urinary contamination as well. 6.5 Knowledge translation Research findings will be sent to all participants who requested this information on the consent form. Individual dermal and urinary contamination levels will be reported in relation to other participants as well as to comparable results in the literature. Participants will also be provided with a summary of the key study findings and a website address for additional information regarding the study. A non-scientific report will be drafted and provided to the funding agency (WorkSafeBC Research Secretariat); after external review, the report will be posted on their website. Study results will also be shared in presentations with the affected departments, health and safety departments and joint health and safety committees at the six participating hospitals. Other study stakeholders will be sent a summary of the key study findings, the study’s website address and, in addition, be given the option to request a presentation. These stakeholders include the following groups: Health Sciences Association of British Columbia (HSA), British Columbia Nurses’ Union (BCNU), Hospital Employees’ Union (HEU), Health Care Safety Professionals Association of British Columbia (HCSPA), and Health Employers’ Association of British Columbia (HEABC). Stakeholders can use the study results to raise awareness of the issue of healthcare workers’ exposure to antineoplastic drugs and, in turn, lead to discussions surrounding changes to practices and policies. In addition, the drug levels found in this 123 document can serve as a baseline to gauge the effectiveness of any future control initiatives aimed at reducing surface contamination and/or occupational exposure. 6.6 Study strengths and limitations The predominant strength of the study was examination of surfaces and healthcare workers throughout the hospital medication system to ascertain the extent of environmental contamination and exposure potential within a hospital. This, in turn, allowed us to identify determinants of surface contamination as well as occupational exposure throughout the hospital medication system. Another strength of the study was the sample size. To my knowledge, this study had the largest number of surface wipes and personal samples compared to other similar studies.(36) This study also evaluated surfaces whereby the probability of worker contact is known. In comparison, published studies sampled surfaces where frequency of contact is unknown e.g. floors (59)(60) while in other instances no rationale was provided for inclusion of surfaces.(43)(55)(48) Furthermore, I collected duplicate samples to ascertain the variability of contamination/exposure over two sampling periods. For the laboratory analyses, highly sensitive instrumentation was used resulting in detection limits in the nanogram per wipe (surface and dermal) or nanogram per milliliter (urine) range. Regarding the urine analyses, we measured for the parent product as well as three of its urinary metabolites. This results in a more accurate reflection of urinary contamination levels as the current study found that 25% of the absorbed CP was excreted unchanged (similar to literature reports of 20% of parent product excreted unchanged (72)). 124 This study is not without its limitations, however. This study was cross-sectional in nature and, therefore, we are unable to comment about cumulative exposures. With only CP being measured, the results are likely to be an underestimate of contamination/exposure to all antineoplastic drugs that are found at the participating facilities. I was unable to obtain accurate quantities of CP handled/contacted by participants on the work shift (asked during collection of dermal wipes) and during the 24-hour urine sampling period. Although various types of surfaces were sampled, I only had wipe recovery information for stainless steel surfaces (recovery rate of 99.65%). Due to the fact that cleaning services at participating facilities were provided by a contractor who declined to participate in the study, we did not have representation from housekeepers which are known to be at risk of exposure.(31)(67)(40) The relevant ethics boards dictated the manner in which workers were recruited for inclusion in the study. Some workers were actively recruited via random sampling from employee lists, while others were passively recruited via volunteering their contact information after formal and informal meetings at the hospital site (non-random). As such, whether the differing recruitment methods resulted in a bias of the findings needs to be explored. One of the recruitment methods was based on potential participants agreeing to be contacted after introduction to the study via department or face-to-face meetings (passive recruitment). This might result in a volunteer effect, where individuals who are more likely to be exposed (those who are concerned) might preferentially participate. However, it is unlikely that any of the potential participants had concerns related to previous measurement data, because none of the healthcare workers at any of the participating sites have ever been asked to provide samples for exposure assessment to antineoplastic drugs. There was still potential for people who routinely handle antineoplastic drugs to be more concerned and to preferentially volunteer. I was able to examine this potential, 125 because 63% of participants were recruited actively (i.e. randomly selected based on staff lists of job titles that were considered at potential risk of exposure). I compared contamination levels between those who were actively versus passively recruited and found no statistically significant difference (t-test results of p=0.46 and p=0.11 for dermal and urine contamination levels, respectively). Breaking down the comparison even further, we examined the results between the recruitment methods for two specific independent variables – hospital and job title. A majority (95%) of the 20 t-tests conducted were found to be not statistically significant and no trends in mean contamination levels were evident. I also compared the risk perception scores (with respect to hazards associated with antineoplastic drugs) of those recruited by the two methods and did not find a statistically significant difference (p=0.42). These comparisons provide some comfort that the participant recruitment method was not likely to have biased the contamination levels. Throughout this study, results of analyses examining factors that might be associated with exposure levels were often not statistically significant. This issue was explored further to determine whether important differences in ln of the exposure concentration (dependent variable) were missed because of the study design. Based on a significance level of 0.05 and a power of 80%, the difference in ln of the exposure concentration that we could detect using a one-way ANOVA balanced design was 0.257 (for 8 categories with 28 participants in each) or 0.188 (for 2 categories with 112 participants in each). [Note 8 was chosen because it represents the most categories of a variable in the current study i.e. job title]. This represents a 1.29 or 29% (e0.257 = 1.29) and 1.21 or 21% (e0.188 = 1.21) times difference in GM concentrations, respectively. This suggests that study power was very good, more than adequate to detect a doubling of exposure levels. Reviewing the dermal results, a majority of the GM values were less than the detection 126 limit. As a result, many of the one-way ANOVAs were found to be not statistically significant. With respect to the one-way ANOVAs for the urine contamination levels, some categories had relatively small numbers (<10) and/or the difference in contamination levels were less than 21% resulting in non-statistically significant findings. Another way to consider whether the study design, in particular the sensitivity of the laboratory analyses, was sufficient to identify differences in the dependent variable is to compare the lowest concentration values and the referenced exposure thresholds. The dermal limit of detection was 0.356 ng/wipe and a proposed dermal occupational exposure limit is 3360 ng/wipe, almost a 10,000X difference. For urine samples, the detection limit for CP was 0.052 ng/mL or 91 ng/24 hrs (assuming an average urinary volume of 1.75 L per day) and a proposed limit is 1000 ng/day. This represents more than a 10X difference. Based on these calculations, I believe that the study design was appropriate to detect meaningful differences in the ln of the exposure concentrations. Lastly, there exists the possibility of misclassification of measurements of contamination levels and responses to questionnaires. Any analytical measurement error in drug contamination levels is expected to be non-differential and therefore bias towards the null. With respect to self- reported responses to questionnaires, depending on the response, the bias either could be in either direction, though non-differential misclassification and bias to the null are most likely. 6.7 Future direction/work As mentioned earlier, novel findings of the current study include the fact that measureable levels of drug contamination were found on surfaces throughout the hospital medication system and that more job categories than previously believed face an exposure potential. Despite these novel findings, questions remain surrounding the matter of surface contamination and 127 occupational exposure to antineoplastic drugs. These outstanding questions should be addressed in future work as outlined below. 6.7.1 Surface cleaning In Chapter 3 I found that a majority of participants were unaware if a potentially-contaminated surface/object was cleaned, making it difficult to determine the impact of surface cleaning. However, even where cleaning was reported, measurable levels of drug residual were still found. This suggests that cleaning of surfaces may not be effective, and raises the possibility that accumulation of contamination may occur, as suggested by Touzin et al.(86) This is noteworthy as antineoplastic drugs have been found to be environmentally stable for up to two months.(117) Therefore, in order to ensure the efficacy of surface cleaning, we agree with others that the type of cleaning agent used(87) and the time of action of the cleaning agent(86) need to be examined. 6.7.2 Evaluate health risks This was an occupational exposure study whereby we discovered novel information that a number of healthcare job categories, besides pharmacy and nursing personnel, have measurable levels of drug contamination on their hands as well as in their urine. The next logical step would be to ascertain if these newly-identified job cohorts face health risks due to this exposure. This includes damage to genetic material (17)(118), reproductive effects (including spontaneous abortions, still births and ectopic pregnancies) as well as cancer.(5) Information about chronic health risks, especially cancer, in conjunction with the exposure levels reported in the current study can then lead to the establishment of an updated quantitative risk assessment for CP, the most recent of which was in 2002.(110) 128 6.7.3 Determine mechanism of spread of contamination Although we identified that drug residual was found on surfaces situated throughout the hospital medication system, we are unable to explain the mechanism of spread. By understanding the mechanism, one can then apply specific control measures to prevent dissemination of antineoplastic drugs. One method to track the spread of contamination is to use a tracer (119)(85) and, concurrently, systematically analyze tasks/processes where contamination may occur (i.e. critical control points) as suggested by Bonan et al.(120) 6.7.4 Identify determinants for each stage/job category separately Although we have identified determinants associated with contamination or exposure throughout the hospital medication system, it may be worthwhile to break down this process and ascertain the determinants by stage of the medication system and/or healthcare job categories at risk of exposure. This is suggested for two reasons: a) when stratified by job title, some of the categories had relatively small sample sizes (N < 10) and b) different processes take place in each stage and therefore certain determinants may only be applicable to specific tasks. By examining each stage separately, it may provide some additional insight into the matter of environmental contamination and/or occupational exposure. 6.7.5 Association between environmental monitoring and biological monitoring There was a time lag between the collection of surface wipes and the collection of personal samples (dermal wipe and urine samples). This may be one reason we did not find an association between the two matrices. It is therefore recommended to collect the environmental samples concurrently with the personal samples as the former can demonstrate how, where and 129 possibly when contamination occurred while the biological samples indicates if exposure occurred.(66) 130 Table 6-1 Risk assessment for bladder cancer based on Sessink et al. Variable Calculation Value A Lowest dose with significant increase in tumors 1.25 (mg/kg)/day B Bladder tumour incidence 14.3% C Exposure period (Median survival period – age when started x days/week exposed) (646-100) x 5/7 390 days D Total cumulative dose (A x C) 1.25 x 390 487.5 mg/kg E Cancer risk (B / C) 0.143/487.5 293x10-6 per mg CP F Average cumulative dose from urine chapter (average dose in mg x 200 days/yr x 40 yrs)* 2.32x10-4 x 200 x 40 1.86 mg G Lifetime cancer risk (1/average mass x E x F) 1/70 x 293x10-6 x 1.86 7.78 per million H Population at risk of exposure** 6,893,979 I Excess cases of cancer in this population (G x H) 7.78x10-6 x 6,893,979 54 * time periods used are those utilized by the author ** based on number of US healthcare workers estimated by McDiarmid and Condon(70); note that not all job categories found to be at risk of exposure in the current study are included in the equation Table 6-2 Risk assessment for bladder cancer based on Sargent et al. Variable Calculation Value A Cancer slope factor 0.57 (mg/kg)/day B Lifetime average daily dose (average does in mg/70kg) 2.32x10-4 / 70 3.31x10-6 (mg/kg)/day C Population at risk of exposure± 6,893,979 D Lifetime cancer risk (A x B) 0.57 x 3.31x10-6 1.89 per million E Excess cases of cancers in this population (C x D) 1.89x10-6 x 6,893,979 13 ±based on number of US healthcare workers estimated by McDiarmid and Condon(70); note that not all job categories found to be at risk of exposure in the current study are included in the equation 131 REFERENCES 1. the definition of antineoplastic [Internet]. Dictionary.com. [cited 2011 Dec 22]. Available from: http://dictionary.reference.com/browse/antineoplastic 2. Sessink PJ, Bos RP. Drugs hazardous to healthcare workers. Evaluation of methods for monitoring occupational exposure to cytostatic drugs. Drug Saf. 1999 Apr;20(4):347–59. 3. Nussbaumer S, Bonnabry P, Veuthey J-L, Fleury-Souverain S. Analysis of anticancer drugs: A review. Talanta. 2011 Oct;85(5):2265–89. 4. Falck K, Gröhn P, Sorsa M, Vainio H, Heinonen E, Holsti LR. Mutagenicity in urine of nurses handling cytostatic drugs. Lancet. 1979 Jun 9;1(8128):1250–1. 5. Dranitsaris G, Johnston M, Poirier S, Schueller T, Milliken D, Green E, et al. Are health care providers who work with cancer drugs at an increased risk for toxic events? A systematic review and meta-analysis of the literature. J Oncol Pharm Pract. 2005 Jun;11(2):69–78. 6. Ritchie MA, McAdams C, Fritz N. Exposure Risk in the Handling and Administration of Chemotherapy Agents: A Review and Synthesis of the Literature. Online J Knowl Synth Nurs. 2000 Feb 2;7:Document No. 4. 7. Connor TH, McDiarmid MA. Preventing occupational exposures to antineoplastic drugs in health care settings. CA: a cancer journal for clinicians. 2006 Dec;56(6):354–65. 8. Anon. ASHP Technical Assistance Bulletin on Handling Cytotoxic and Hazardous Drugs. Am J Hosp Pharm. 1990;47(5):1033. 9. CDC - NIOSH Publications and Products - Preventing Occupational Exposure to Antineoplastic and Other Hazardous Drugs in Health Care Settings (2004-165) [Internet]. [cited 2011 Nov 30]. Available from: http://www.cdc.gov/niosh/docs/2004-165/ 10. Prevention Guide - Safe Handling of Hazardous Drugs [Internet]. [cited 2011 Nov 30]. Available from: http://www.irsst.qc.ca/en/-irsst-publication-prevention-guide-safe-handling-of- hazardous-drugs-cg-002.html 11. Turci R, Sottani C, Spagnoli G, Minoia C. Biological and environmental monitoring of hospital personnel exposed to antineoplastic agents: a review of analytical methods. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 2003 Jun 15;789(2):169–209. 12. Suspiro A, Prista J. Biomarkers of occupational exposure do anticancer agents: A minireview. Toxicology Letters [Internet]. 2011 Sep [cited 2011 Sep 22]; Available from: http://linkinghub.elsevier.com/retrieve/pii/S0378427411015098 13. de Jonge ME, Huitema ADR, Rodenhuis S, Beijnen JH. Clinical pharmacokinetics of cyclophosphamide. Clin Pharmacokinet. 2005;44(11):1135–64. 14. Emadi A, Jones RJ, Brodsky RA. Cyclophosphamide and cancer: golden anniversary. Nature Reviews Clinical Oncology. 2009 Sep 29;6(11):638–47. 15. Cyclophosphamide (IARC Mongraphs on the Evaluation of Carcinogenic Risks to Huans 100A, 2012) [Internet]. [cited 2012 Jan 7]. Available from: http://monographs.iarc.fr/ENG/Monographs/vol100A/index.php 132 16. Antineoplastic Agents — Occupational Exposure Estimates — Phase I — CAREX Canada [Internet]. [cited 2012 Jan 11]. Available from: http://www.carexcanada.ca/en/antineoplastic_agents/occupational_exposure_estimates/phase_2/ 17. Villarini M, Dominici L, Piccinini R, Fatigoni C, Ambrogi M, Curti G, et al. Assessment of primary, oxidative and excision repaired DNA damage in hospital personnel handling antineoplastic drugs. Mutagenesis. 2011 May;26(3):359–69. 18. Fransman W, Peelen S, Hilhorst S, Roeleveld N, Heederik D, Kromhout H. A pooled analysis to study trends in exposure to antineoplastic drugs among nurses. Ann Occup Hyg. 2007 Apr;51(3):231–9. 19. McDevitt JJ, Lees PS, McDiarmid MA. Exposure of hospital pharmacists and nurses to antineoplastic agents. J Occup Med. 1993 Jan;35(1):57–60. 20. Crauste-Manciet S, Sessink PJ, Ferrari S, Jomier JY, Brossard D. Environmental contamination with cytotoxic drugs in healthcare using positive air pressure isolators. Ann Occup Hyg. 2005 Oct;49(7):619–28. 21. Sabatini L, Barbieri A, Tosi M, Violante FS. A new high-performance liquid chromatographic/electrospray ionization tandem mass spectrometric method for the simultaneous determination of cyclophosphamide, methotrexate and 5-fluorouracil as markers of surface contamination for occupational exposure monitoring. J Mass Spectrom. 2005 May;40(5):669–74. 22. Barbieri A, Sabatini L, Indiveri P, Bonfiglioli R, Lodi V, Violante FS. Simultaneous determination of low levels of methotrexate and cyclophosphamide in human urine by micro liquid chromatography/electrospray ionization tandem mass spectrometry. Rapid Commun Mass Spectrom. 2006;20(12):1889–93. 23. Larson RR, Khazaeli MB, Dillon HK. Development of an HPLC method for simultaneous analysis of five antineoplastic agents. Appl Occup Environ Hyg. 2003 Feb;18(2):109–19. 24. Schmaus G, Schierl R, Funck S. Monitoring surface contamination by antineoplastic drugs using gas chromatography-mass spectrometry and voltammetry. Am J Health Syst Pharm. 2002 May 15;59(10):956–61. 25. Sottani C, Rinaldi P, Leoni E, Poggi G, Teragni C, Delmonte A, et al. Simultaneous determination of cyclophosphamide, ifosfamide, doxorubicin, epirubicin and daunorubicin in human urine using high-performance liquid chromatography/electrospray ionization tandem mass spectrometry: bioanalytical method validation. Rapid Commun. Mass Spectrom. 2008 Sep;22(17):2645–59. 26. Larson RR, Khazaeli MB, Dillon HK. Monitoring method for surface contamination caused by selected antineoplastic agents. Am J Health Syst Pharm. 2002 Feb 1;59(3):270–7. 27. Sottani C, Turci R, Schierl R, Gaggeri R, Barbieri A, Violante FS, et al. Simultaneous determination of gemcitabine, taxol, cyclophosphamide and ifosfamide in wipe samples by high- performance liquid chromatography/tandem mass spectrometry: protocol of validation and uncertainty of measurement. Rapid Commun Mass Spectrom. 2007;21(7):1289–96. 133 28. Pretty JR, Connor TH, Spasojevic I, Kurtz KS, McLaurin JL, B’ Hymer C, et al. Sampling and mass spectrometric analytical methods for five antineoplastic drugs in the healthcare environment. J Onc Pharm Pract. 2012 Mar; 18(1): 23-36. 29. Nussbaumer S, Fleury-Souverain S, Antinori P, Sadeghipour F, Hochstrasser DF, Bonnabry P, et al. Simultaneous quantification of ten cytotoxic drugs by a validated LC–ESI– MS/MS method. Anal Bioanal Chem. 2010 Oct 7;398(7-8):3033–42. 30. Turci R, Minoia C. Residual hazard assessment related to handling of antineoplastic drugs: safety system evolution and quality assurance of analytical measurement. Ann NY Acad Sci. 2006 Sep;1076:649–56. 31. Hedmer M, Tinnerberg H, Axmon A, Joensson BAG. Environmental and biological monitoring of antineoplastic drugs in four workplaces in a Swedish hospital. Int Arch Occup Environ Health. 2008;81(7):899–911. 32. Hedmer M, Georgiadi A, Bremberg ER, Jönsson BAG, Eksborg S. Surface contamination of cyclophosphamide packaging and surface contamination with antineoplastic drugs in a hospital pharmacy in Sweden. Ann Occup Hyg. 2005 Oct;49(7):629–37. 33. Touzin K, Bussieres JF, Langlois E, Lefebvre M. Evaluation of surface contamination in a hospital hematology–oncology pharmacy. J Onc Pharm Pract. 2009 Mar;15(1):53–61. 34. Siderov J, Kirsa S, McLauchlan R. Surface Contamination of Cytotoxic Chemotherapy Preparation Areas in Australian Hospital Pharmacy Departments. J Pharm Pract Res. 2009;39(2):117. 35. Sugiura S, Asano M, Kinoshita K, Tanimura M, Nabeshima T. Risks to health professionals from hazardous drugs in Japan: A pilot study of environmental and biological monitoring of occupational exposure to cyclophosphamide. J Onc Pharm Pract. 2011 Mar;17(1):14–9. 36. Connor TH, DeBord DG, Pretty JR, Oliver MS, Roth TS, Lees PS, et al. Evaluation of antineoplastic drug exposure of health care workers at three university-based US cancer centers. J Occup Environ Med. 2010 Oct;52(10):1019–27.. 37. Sottani C, Porro B, Comelli M, Imbriani M, Minoia C. An analysis to study trends in occupational exposure to antineoplastic drugs among health care workers. J Chromatography B 2010 Oct 1;878(27):2593–605. 38. Sessink PJ, Van de Kerkhof MC, Anzion RB, Noordhoek J, Bos RP. Environmental contamination and assessment of exposure to antineoplastic agents by determination of cyclophosphamide in urine of exposed pharmacy technicians: is skin absorption an important exposure route? Arch Environ Health. 1994 Jun;49(3):165–9. 39. Kromhout H, Hoek F, Uitterhoeve R, Huijbers R, Overmars RF, Anzion R, et al. Postulating a dermal pathway for exposure to anti-neoplastic drugs among hospital workers. Applying a conceptual model to the results of three workplace surveys. Ann Occup Hyg. 2000 Oct;44(7):551–60. 40. Fransman W, Vermeulen R, Kromhout H. Dermal exposure to cyclophosphamide in hospitals during preparation, nursing and cleaning activities. Int Arch Occup Environ Health. 2005 Jun;78(5):403–12. 134 41. Chu WC, Hon C-Y, Danyluk Q, Chua PPS, Astrakianakis G. Pilot assessment of the antineoplastic drug contamination levels in British Columbian hospitals pre- and post-cleaning. J Onc Pharm Pract. 2012 Mar;18(1):46-51 42. Hon C-Y, Astrakianakis G, Danyluk Q, Chu W. Pilot Evaluation of Dermal Contamination by Antineoplastic Drugs among Hospital Pharmacy Personnel. Can J Hosp Pharm. 2011 Sep-Oct;64(5):327-332. 43. Sugiura S, Nakanishi H, Asano M, Hashida T, Tanimura M, Hama T, et al. Multicenter study for environmental and biological monitoring of occupational exposure to cyclophosphamide in Japan. J Onc Pharm Pract. 2011 Mar;17(1):20–8. 44. Sottani C, Porro B, Imbriani M, Minoia C. Occupational exposure to antineoplastic drugs in four Italian health care settings. Toxicology Letters [Internet]. 2011 Apr [cited 2011 Dec 6]; Available from: http://linkinghub.elsevier.com/retrieve/pii/S0378427411001275 45. Mason HJ, Blair S, Sams C, Jones K, Garfitt SJ, Cuschieri MJ, et al. Exposure to antineoplastic drugs in two UK hospital pharmacy units. Ann Occup Hyg. 2005 Oct;49(7):603– 10. 46. Martins I, Apostoli P, Della Rosa HV. Cyclophosphamide Levels in Sites of Preparation and Administration of Antineoplastic Drugs. Lat Am J Pharm. 2008;27(2):217–23. 47. Bigelow S, Schulz H, Dobish R, Chambers CR. Antineoplastic agent workplace contamination study: the Alberta Cancer Board Pharmacy perspective Phase III. J Onc Pharm Pract. 2009 Jan 26;15(3):157–60. 48. Tanimura M, Yamada K, Sugiura S, Mori K, Nagata H, Tadokoro K, et al. An Environmental and Biological Study of Occupational Exposure to Cyclophosphamide in the Pharmacy of a Japanese Community Hospital Designated for the Treatment of Cancer. J Health Sci 2009;55:750–6. 49. Favier B, Gilles L, Ardiet C, Latour JF. External contamination of vials containing cytotoxic agents supplied by pharmaceutical manufacturers. J Onc Pharm Pract. 2003 Mar 1;9(1):15–20. 50. Mason HJ, Morton J, Garfitt SJ, Iqbal S, Jones K. Cytotoxic drug contamination on the outside of vials delivered to a hospital pharmacy. Ann Occup Hyg. 2003 Nov;47(8):681–5. 51. Connor TH, Sessink PJM, Harrison BR, Pretty JR, Peters BG, Alfaro RM, et al. Surface contamination of chemotherapy drug vials and evaluation of new vial-cleaning techniques: results of three studies. Am J Health Syst Pharm. 2005 Mar 1;62(5):475–84. 52. Hama K, Fukushima K, Hirabatake M, Hashida T, Kataoka K. Verification of surface contamination of Japanese cyclophosphamide vials and an example of exposure by handling. Journal of Oncology Pharmacy Practice [Internet]. 2011 Sep 23 [cited 2011 Dec 6]; Available from: http://opp.sagepub.com/cgi/doi/10.1177/1078155211419543 53. Schierl R, Herwig A, Pfaller A, Groebmair S, Fischer E. Surface contamination of antineoplastic drug vials: comparison of unprotected and protected vials. Am J Health Syst Pharm. 2010 Mar 15;67(6):428–9. 135 54. Touzin K, Bussieres J-F, Langlois E, Lefebvre M, Gallant C. Cyclophosphamide Contamination Observed on the External Surfaces of Drug Vials and the Efficacy of Cleaning on Vial Contamination. Ann Occup Hyg. 2008 Sep 23;52(8):765–71. 55. Cavallo D, Ursini CL, Perniconi B, Francesco AD, Giglio M, Rubino FM, et al. Evaluation of genotoxic effects induced by exposure to antineoplastic drugs in lymphocytes and exfoliated buccal cells of oncology nurses and pharmacy employees. Mutat Res. 2005 Nov 10;587(1-2):45–51. 56. Acampora A, Castiglia L, Miraglia N, Pieri M, Soave C, Liotti F, et al. A case study: surface contamination of cyclophosphamide due to working practices and cleaning procedures in two Italian hospitals. Ann Occup Hyg. 2005 Oct;49(7):611–8. 57. Zeedijk M, Greijdanus B, Steenstra F, Uges D. Monitoring exposure of cytostatics on the hospital ward. Eur J Hosp Pharm Sci. 2005;11(1):18 – 22. 58. Sessink PJ, Anzion RB, Van den Broek PH, Bos RP. Detection of contamination with antineoplastic agents in a hospital pharmacy department. Pharm Weekbl Sci. 1992 Feb 21;14(1):16–22. 59. Sessink PJ, Boer KA, Scheefhals AP, Anzion RB, Bos RP. Occupational exposure to antineoplastic agents at several departments in a hospital. Environmental contamination and excretion of cyclophosphamide and ifosfamide in urine of exposed workers. Int Arch Occup Environ Health. 1992;64(2):105–12. 60. Connor TH, Anderson RW, Sessink PJ, Broadfield L, Power LA. Surface contamination with antineoplastic agents in six cancer treatment centers in Canada and the United States. Am J Health Syst Pharm. 1999 Jul 15;56(14):1427–32. 61. Minoia C, Turci R, Sottani C, Schiavi A, Perbellini L, Angeleri S, et al. Application of high performance liquid chromatography/tandem mass spectrometry in the environmental and biological monitoring of health care personnel occupationally exposed to cyclophosphamide and ifosfamide. Rapid Commun Mass Spectrom. 1998;12(20):1485–93. 62. Yoshida J, Koda S, Nishida S, Yoshida T, Miyajima K, Kumagai S. Association between occupational exposure levels of antineoplastic drugs and work environment in five hospitals in Japan. J Onc Pharm Pract. 2011 Mar;17(1):29–38. 63. Castiglia L, Miraglia N, Pieri M, Simonelli A, Basilicata P, Genovese G, et al. Evaluation of occupational exposure to antiblastic drugs in an Italian hospital oncological department. J Occup Health. 2008 Jan;50(1):48–56. 64. Rekhadevi PV, Sailaja N, Chandrasekhar M, Mahboob M, Rahman MF, Grover P. Genotoxicity assessment in oncology nurses handling anti-neoplastic drugs. Mutagenesis. 2007 Nov;22(6):395–401. 65. Fransman W, Roeleveld N, Peelen S, de Kort W, Kromhout H, Heederik D. Nurses with dermal exposure to antineoplastic drugs: reproductive outcomes. Epidemiology. 2007 Jan;18(1):112–9. 136 66. Turci R, Minoia C, Sottani C, Coghi R, Severi P, Castriotta C, et al. Occupational exposure to antineoplastic drugs in seven Italian hospitals: the effect of quality assurance and adherence to guidelines. J Onc Pharm Pract. 2011 Dec;17(4): 320-32. 67. Fransman W, Vermeulen R, Kromhout H. Occupational dermal exposure to cyclophosphamide in Dutch hospitals: A pilot study. Ann Occup Hyg. 2004 Apr;48(3):237–44. 68. Sorsa M, Hämeilä M, Järviluoma E. Handling anticancer drugs: from hazard identification to risk management? Ann NY Acad Sci. 2006 Sep;1076:628–34. 69. Kusnetz E, Condon M. Acute effects from occupational exposure to antineoplastic drugs in a para-professional health care worker. Am J Ind Med. 2003 Jul;44(1):107–9. 70. McDiarmid MA, Condon M. Organizational safety culture/climate and worker compliance with hazardous drug guidelines: lessons from the blood-borne pathogen experience. J Occup Environ Med. 2005 Jul;47(7):740–9. 71. Hedmer M, Höglund P, Cavallin-Ståhl E, Albin M, Jönsson BAG. Validation of urinary excretion of cyclophosphamide as a biomarker of exposure by studying its renal clearance at high and low plasma concentrations in cancer patients. Int Arch Occup Environ Health. 2008 Jan;81(3):285–93. 72. Zhang J, Tian Q, Zhou S-F. Clinical pharmacology of cyclophosphamide and ifosfamide. Curr Drug Ther. 2006;1(1):55–84. 73. Testa A, Giachelia M, Palma S, Appolloni M, Padua L, Tranfo G, et al. Occupational exposure to antineoplastic agents induces a high level of chromosome damage. Lack of an effect of GST polymorphisms. Toxicol Appl Pharmacol. 2007 Aug 15;223(1):46–55. 74. Friese CR, Himes-Ferris L, Frasier MN, McCullagh MC, Griggs JJ. Structures and processes of care in ambulatory oncology settings and nurse-reported exposure to chemotherapy. BMJ Quality & Safety [Internet]. 2011 Aug 16 [cited 2011 Dec 22]; Available from: http://qualitysafety.bmj.com/lookup/doi/10.1136/bmjqs-2011-000178 75. Schreiber C, Radon K, Pethran A, Schierl R, Hauff K, Grimm C-H, et al. Uptake of antineoplastic agents in pharmacy personnel. Part II: study of work-related risk factors. Int Arch Occup Environ Health. 2003 Feb;76(1):11–6. 76. Favier B, Gilles L, Desage M, Latour J-F. [Analysis of cyclophosphamide in the urine of antineoplastic drugs handlers]. Bull Cancer. 2003 Oct;90(10):905–9. 77. Canadian Cancer Society Statistics PDF 2008_614137951.ashx [Internet]. [cited 2012 Jan 19]. Available from: http://www.cancer.ca/~/media/CCS/Canada%20wide/Files%20List/English%20files%20heading /pdf%20not%20in%20publications%20section/Canadian%20Cancer%20Society%20Statistics% 20PDF%202008_614137951.ashx 78. Shih T-S, Lu P-Y, Chen C-H, Soo J-C, Tsai C-L, Tsai P-J. Exposure profiles and source identifications for workers exposed to crystalline silica during a municipal waste incinerator relining period. J Haz Materials. 2008 Jun;154(1-3):469–75. 137 79. Garlantezec PL, Rizzo-Padoin N, Lamand V, Aupée O, Broto H, Alméras D. Manipulation des médicaments anticancéreux à l’hôpital : le point sur l’exposition et sur les mesures de prévention. Archives des Maladies Professionnelles et de l’Environnement. 2011;72(1):24 – 35. 80. Hon C-Y, Teschke K, Chua P, Venners S, Nakashima L. Occupational exposure to antineoplastic drugs: Identification of job categories potentially exposed throughout the hospital medication system. Saf Health Work. 2011 Sep 30;2:273–81. 81. Helsel D. Much Ado about next to nothing: Incorporating non-detects in science. Ann Occup Hyg. 2009 Dec;54(3):257–62. 82. Hornung R, Reed L. Estimation of average concentration in the presence of nondetectable values. App Occup Env Hyg. 1990;5(1):46 - 51. 83. Schierl R, Böhlandt A, Nowak D. Guidance values for surface monitoring of antineoplastic drugs in German pharmacies. Ann Occup Hyg. 2009 Oct;53(7):703–11. 84. Stellman JM. The spread of chemotherapeutic agents at work: assessment through simulation. Cancer Invest. 1987;5(2):75–81. 85. Nygren O, Gustavsson B, Eriksson R. A test method for assessment of spill and leakage from drug preparation systems. Ann Occup Hyg. 2005 Nov;49(8):711–8. 86. Touzin K, Bussieres JF, Langlois E, Lefebvre M, Metra A. Pilot study comparing the efficacy of two cleaning techniques in reducing environmental contamination with cyclophosphamide. Ann Occup Hyg. 2010 Jan;54(3):351–9. 87. Barzan CV. Towards the reduction of occupational exposure to cytotoxic drugs [Internet]. 2010 [cited 2011 Aug 5]. Available from: https://circle.ubc.ca/handle/2429/29487 88. Ziegler E, Mason HJ, Baxter PJ. Occupational exposure to cytotoxic drugs in two UK oncology wards. Occup Environ Med. 2002 Sep;59(9):608–12. 89. Sasaki M, Dakeishi M, Hoshi S, Ishii N, Murata K. Assessment of DNA damage in Japanese nurses handling antineoplastic drugs by the comet assay. J Occup Health. 2008 Jan;50(1):7–12. 90. Connor TH. Hazardous anticancer drugs in health care: environmental exposure assessment. Ann NY Acad Sci. 2006 Sep;1076:615–23. 91. Occupational Safety & Health Administration. Hazard Communication. - 1910.1200 [Internet]. OSHA 1910.1200. [cited 2011 Oct 17]. Available from: http://www.osha.gov/pls/oshaweb/owadisp.show_document?p_table=STANDARDS&p_id=100 99 92. Geer LA, Curbow BA, Anna DH, Lees PSJ, Buckley TJ. Development of a questionnaire to assess worker knowledge, attitudes and perceptions underlying dermal exposure. Scand J Work Environ Health. 2006 Jun;32(3):209–18. 93. Geer LA, Anna D, Curbow B, Diener-West M, van Wendel de Joode B, Mitchell C, et al. Survey assessment of worker dermal exposure and underlying behavioral determinants. J Occup Environ Hyg. 2007 Sep 24;4:809–20. 138 94. McDiarmid MA, Oliver MS, Roth TS, Rogers B, Escalante C. Chromosome 5 and 7 abnormalities in oncology personnel handling anticancer drugs. J Occup Environ Med. 2010 Oct;52(10):1028–34. 95. Bos PM, Brouwer DH, Stevenson H, Boogaard PJ, de Kort WL, van Hemmen JJ. Proposal for the assessment of quantitative dermal exposure limits in occupational environments: Part 1. Development of a concept to derive a quantitative dermal occupational exposure limit. Occup Environ Med. 1998 Dec;55(12):795–804. 96. Lee J-Y, Choi J-W, Kim H. Determination of Hand Surface Area by Sex and Body Shape using Alginate. J Physiol Anthropol. 2007;26(4):475–83. 97. Wallemacq PE, Capron A, Vanbinst R, Boeckmans E, Gillard J, Favier B. Permeability of 13 different gloves to 13 cytotoxic agents under controlled dynamic conditions. Am J Health Syst Pharm. 2006 Mar 15;63:547–56. 98. Connor TH. Permeability of nitrile rubber, latex, polyurethane, and neoprene gloves to 18 antineoplastic drugs. Am J Health Syst Pharm. 1999 Dec 1;56(23):2450–3. 99. Klein M, Lambov N, Samnev N, Carstens G. Permeation of cytotoxic formulations through swatches from selected medical gloves. Am J Health Syst Pharm. 2003;60(10):1006-11. 100. Kerr LN, Chaput MP, Cash LD, O’Malley LG, Sarhrani EM, Teixeira JC, et al. Assessment of the durability of medical examination gloves. J Occup Environ Hyg. 2004 Sep;1(9):607–12. 101. Gammon J, Morgan-Samuel H, Gould D. A review of the evidence for suboptimal compliance of healthcare practitioners to standard/universal infection control precautions. J Clin Nur. 2008 Jan; 17(2): 157-67. 102. Hirst M, Tse S, Mills DG, Levin L, White DF. Occupational exposure to cyclophosphamide. Lancet. 1984 Jan 28;1(8370):186–8. 103. van Wendel de Joode B. Accuracy of a semiquantitative method for Dermal Exposure Assessment (DREAM). Occup Environ Med. 2005 Sep 1;62:623–32. 104. Fransman W, Huizer D, Tuerk J, Kromhout H. Inhalation and dermal exposure to eight antineoplastic drugs in an industrial laundry facility. Int Arch Occup Environ Health. 2006 Oct;80(5):396–403. 105. Eisenberg S. Safe handling and administration of antineoplastic chemotherapy. J Infus Nurs. 2009 Feb;32(1):23–32. 106. Turci R, Sottani C, Ronchi A, Minoia C. Biological monitoring of hospital personnel occupationally exposed to antineoplastic agents. Toxicol Lett. 2002 Aug 5;134(1-3):57–64. 107. Karahalil B, Akkoyunlu KI. Determination of urinary cyclophosphamide in oncology nurses handling antineoplastic drugs by gas chromatography-mass spectrometry. FABAD J Pharm Sci. 2003;28(3):125–30. 108. Sessink PJ, Kroese ED, van Kranen HJ, Bos RP. Cancer risk assessment for health care workers occupationally exposed to cyclophosphamide. Int Arch Occup Environ Health. 1995;67(5):317–23. 139 109. Series 875| Chemical Safety and Pollution Prevention | US EPA [Internet]. [cited 2011 Oct 26]. Available from: http://www.epa.gov/ocspp/pubs/frs/publications/Test_Guidelines/series875.htm 110. Sargent EV, Naumann BD, Dolan DG, Faria EC, Schulman L. The importance of human data in the establishment of occupational exposure limits. Hum Ecol Risk Assess. 2002 Oct 1;8:805–22. 111. Yassi A, Moore D, Fitzgerald JM, Bigelow P, Hon C-Y, Bryce E. Research gaps in protecting healthcare workers from SARS and other respiratory pathogens: an interdisciplinary, multi-stakeholder, evidence-based approach. J Occup Environ Med. 2005 Jan;47(1):41–50. 112. Maeda S, Miyawaki K, Matsumoto S, Oishi M, Miwa Y, Kurokawa N. Evaluation of environmental contaminations and occupational exposures involved in preparation of chemotherapeutic drugs. Yakugaku Zasshi. 2010 Jun;130(6):903–10. 113. Siderov J, Kirsa S, McLauchlan R. Reducing workplace cytotoxic surface contamination using a closed-system drug transfer device. J Oncol Pharm Pract. 2010 Mar;16(1):19–25. 114. Yoshida J, Tei G, Mochizuki C, Masu Y, Koda S, Kumagai S. Use of a closed system device to reduce occupational contamination and exposure to antineoplastic drugs in the hospital work environment. Ann Occup Hyg. 2009;53(2):153–60. 115. Sessink PJ, Connor TH, Jorgenson JA, Tyler TG. Reduction in surface contamination with antineoplastic drugs in 22 hospital pharmacies in the US following implementation of a closed-system drug transfer device. J Oncol Pharm Pract 2011 Mar;17(1):39–48. 116. Fabrizi G, Fioretti M, Mainero Rocca L, Curini R. DESI-MS2: a rapid and innovative method for trace analysis of six cytostatic drugs in health care setting. Analytical and Bioanalytical Chemistry [Internet]. 2011 Dec 27 [cited 2012 Jan 23]; Available from: http://www.springerlink.com/index/10.1007/s00216-011-5626-7 117. Connor TH, Anderson RW, Sessink PJ, Spivey SM. Effectiveness of a closed-system device in containing surface contamination with cyclophosphamide and ifosfamide in an i.v. admixture area. Am J Health Syst Pharm. 2002 Jan 1;59(1):68–72. 118. Halsen G, Krämer I. Assessing the risk to health care staff from long-term exposure to anticancer drugs--the case of monoclonal antibodies. J Oncol Pharm Pract. 2011 Mar;17(1):68– 80. 119. Queruau-Lamerie T, Carrez L, Decaudin B, Bouchoud L, Goossens J-F, Barthelemy C, et al. Multiple-test assessment of devices to protect healthcare workers when administering cytotoxic drugs to patients. Journal of Oncology Pharmacy Practice [Internet]. 2011 Aug 23 [cited 2012 Jan 31]; Available from: http://opp.sagepub.com/cgi/doi/10.1177/1078155211416531 120. Bonan B, Martelli N, Berhoune M, Maestroni M-L, Havard L, Prognon P. The application of hazard analysis and critical control points and risk management in the preparation of anti-cancer drugs. Int J Qual Health Care. 2009 Feb 1;21(1):44–50. 140 APPENDICES 141 APPENDIX A PASSIVE OBSERVATIONAL CHECKLIST 142 143 APPENDIX B DRUG HANDLING AND CLEANING QUESTIONS 144 APPENDIX C LETTER OF INITIAL CONTACT (FOR RAMDOMLY SELECTED PARTICIPANTS) 145 APPENDIX D CONSENT TO CONTACT FORM (FOR PASSIVELY RECRUITED PARTICIPANTS) 146 APPENDIX E CONSENT FORM 147 148 149 APPENDIX F ACTIVITY-RELATED QUESTIONS 150 151 152 153 APPENDIX G INTRODUCTION LETTER SENT WITH QUESTIONNAIRES 154 APPENDIX H SELF-ADMINISTERED QUESTIONNAIRE 155 156 157 158 159 160 161 162 APPENDIX I URINARY CONTAMINATION LEVELS STRATIFIED BY REPORTED CONTACT METHODS DURING WORK SHIFT: SUMMARY STATISTICS AND ONE- WAY ANOVA RESULTS Method of contact Time Response N AM (nmol/L) Max (nmol/L) GM (nmol/L) GSD Contact with CP by any means… past 24 hrs Yes 81 20.79 839.14 3.83 3.73 No/DK 142 15.44 565.29 4.36 3.54 on shift Yes 64 20.10 839.14 3.74 3.47 No/DK 159 16.29 565.29 4.34 3.66 Mix CP… past 24 hrs Yes 10 9.90 32.32 6.27 2.74 No/DK 213 17.74 839.14 4.08 3.64 on shift Yes 4 11.85 22.56 9.20 2.38 No/DK 219 17.49 839.14 4.10 3.61 Administer CP… past 24 hrs Yes 17 20.48 298.87 3.34 3.89 No/DK 206 17.13 839.14 4.24 3.59 on shift Yes 12 3.23 5.10 2.81 1.90 No/DK 211 18.19 839.14 4.25 3.69 Disconnect IV line containing CP… past 24 hrs Yes 14 24.09 298.87 3.33 4.47 No/DK 209 16.94 839.14 4.22 3.56 on shift Yes 9 3.11 5.10 2.64 2.03 No/DK 214 17.99 839.14 4.24 3.66 Provide physical care to a patient on CP… past 24 hrs Yes 5 62.43 298.87 6.72 9.74 No/DK 218 16.35 839.14 4.11 3.52 on shift Yes 3 4.21 5.10 4.16 1.21 No/DK 220 17.56 839.14 4.16 3.63 Handle a container of CP… past 24 hrs Yes 49 12.32 298.87 3.70 3.55 No/DK 174 18.81 839.14 4.30 3.63 on shift Yes 32 6.61 50.37 3.82 2.70 No/DK 191 19.19 839.14 4.22 3.77 Dispose of waste containing CP… past 24 hrs Yes 27 15.97 298.87 3.86 3.52 No/DK 196 17.58 839.14 4.20 3.63 on shift Yes 14 4.21 18.28 3.06 2.28 No/DK 209 18.27 839.14 4.25 3.69 Touch a CP- contaminated surface… past 24 hrs Yes 66 22.73 839.14 3.77 3.84 No/DK 157 15.14 565.29 4.33 3.51 on shift Yes 48 21.97 839.14 3.48 3.37 No/DK 175 16.13 565.29 4.37 3.66 Consume food/water… past 24 hrs Yes 33 34.27 839.14 4.05 4.47 No/DK 190 14.45 565.29 4.18 3.47 on shift Yes 27 39.49 839.14 4.17 4.57 No/DK 196 14.34 565.29 4.16 3.49 163 APPENDIX J WRITTEN INSTRUCTIONS FOR COLLECTING URINE 164