Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Long-term follow-up of neonates who have been immunized with hepatitis B vaccine Pastore, Marian Tomm 1993

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

Item Metadata


831-ubc_1993_fall_tomm_pastore_marian.pdf [ 6.59MB ]
JSON: 831-1.0086312.json
JSON-LD: 831-1.0086312-ld.json
RDF/XML (Pretty): 831-1.0086312-rdf.xml
RDF/JSON: 831-1.0086312-rdf.json
Turtle: 831-1.0086312-turtle.txt
N-Triples: 831-1.0086312-rdf-ntriples.txt
Original Record: 831-1.0086312-source.json
Full Text

Full Text

to the required standardE UNIVERSITY OF BRITISH COLUMBIAJune 1993© Marian Tomm Pastore, 1993LONG-TERM FOLLOW-UP OF NEONATES WHO HAVE BEENIMMUNIZED WITH HEPATITIS B VACCINEbyMARIAN TOMM PASTOREB.Sc., The University of Alberta, 1966A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTERS IN SCIENCE(Epidemiology)inTHE FACULTY OF GRADUATE STUDIESDepartment of Health Care and EpidemiologyWe accept this thesis as conformingIn presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature) Department of  Health Care and EpidemiologyThe University of British ColumbiaVancouver, CanadaDate^0.,t_crterf- /4, /99-3DE-6 (2/88)ABSTRACTAlthough the short-term effectiveness of hepatitis B vaccine hasbeen well established, few long-term studies of the effectivenessof this vaccine in neonates who are born to HBsAg positive mothershave been undertaken. It is also not known if supplementary dosesof vaccine are required to protect these children who arecontinuously exposed to the hepatitis B virus, and if so, how longafter primary immunization the additional dose should be given.To determine the hepatitis B vaccine effectiveness up to eightyears, a cohort of 770 (66%) of 1166 children who had beenimmunized with HBIG and hepatitis B vaccine at birth, between 1984and 1989 inclusive, were tested for the serological markers: anti-HBc, anti-HBs and HBsAg. (Anti-HBc indicates that viralreplication has occurred at some time (infection); anti-HBs is theantibody to HBsAg and may appear after natural infection orfollowing immunization; and HBsAg indicates presence of hepatitisB virus (HBV) infection.) The children in the cohort wereimmunized as part of the British Columbia Ministry of HealthProgram to prevent HBV infection in infants of carrier mothers.Blood samples were obtained by finger-prick and tested at theCanadian Red Cross Society, Vancouver Centre, Blood TransfusionService. The IMx assay kits were produced by Abbott Laboratories.Parents were interviewed for information on relevant variables.iiAssociations were determined using chi-square tests, Mantel-Haenszel trend tests, linear and logistic regression procedures.Best prediction models for the outcomes: anti-HBc, anti-HBs andHBsAg were determined in a stepwise fashion.Of the participants, 31% of the mothers were HBeAg positive aswell as HBsAg positive. The overall attack rate for ages 2 to 8years was 5.1%; the carrier rate was 2.3% and the seropositivityrate (>=10 mIU/ml) was 87.9%. The geometric mean titres of anti-HBs varied from a high of 272 mIU/ml at age two to a low of 23mIU/m1 at age seven. Vaccine efficacy for infection was 89%.The best predictors for infection were the mother's HBeAg status(P<0.0001), the age of the child when the first dose of vaccinewas given (P=0.0001), the number of years the mother spent in hercountry of birth (P=0.001) and inversely the mother's age(P=0.002). Anti-HBs titres were best predicted by the inverse ageof the child (P=0.0001), the number of doses of vaccine(P=0.0007), the age of the child when the first dose of vaccinewas given (P=0.01) and the number of months the child spent abroad(P=0.05).Since the infection rate was relatively stable between age groupsand the association between the child's age and anti-HBc was notsignificant (P=0.15), waning immunity is not suggested by thefindings of this study. This was in spite of the decline in anti-HBs titres with increasing age.iiiThe conclusions of this study are: 1) That a booster dose ofhepatitis B vaccine after the six month dose is not necessary atleast up to age eight; 2) Because a delay in the first dose ofvaccine resulted in increased infections, giving the first doseearly is critical; 3) It could not be concluded from the data thatHBIG was not important, therefore, continuation of HBIG at birthis recommended; 4) Since the timing of dose two later than twomonths after dose one was not found to be associated withincreased infections, this dose could be incorporated with theregular immunization schedule at two months of age ifadministratively more feasible; 5) Susceptibility appears to ariseearlier rather than later and may be due to non-response to thevaccine as a 14% non-response rate was found at 6 to 18 months ofage.ivTABLE OF CONTENTSABSTRACT^iiTABLE OF CONTENTSLIST OF TABLES^ixLIST OF FIGURES xACKNOWLEDGEMENTS^ xiDEDICATION^ xiiCHAPTER ONE^ 1INTRODUCTION AND GENERAL CONSIDERATIONS^ 11. DESCRIPTION OF THE PROVINCE-WIDE PROGRAM FOR THEPREVENTION OF HEPATITIS B VIRUS INFECTION IN INFANTSOF CARRIER MOTHERS^ 12. OBJECTIVES OF THE STUDY 23. JUSTIFICATION 34. RESEARCH QUESTIONS AND HYPOTHESES^ 4CHAPTER TWO^ 6LITERATURE REVIEW^ 61. BACKGROUND 62. SHORT-TERM EFFICACY OF HEPATITIS B VACCINE^73. LONG-TERM EFFICACY OF HEPATITIS B VACCINE 114. LONG-TERM DURATION OF ANTIBODIES^ 14CHAPTER THREE^ 16METHOD^ 161. SAMPLE SPECIFICATION^ 161.1. Included children 161.2. Excluded children 172. STUDY VARIABLES^ 192.1. Purposes for collection of variables^192.2. Definitions for dependent variables 202.3. Definitions for independent variables 213. DATA COLLECTION PROCEDURES^ 283.1. Study documents^ 284. PROCEDURES IN PREPARATION FOR DATA COLLECTION^314.1. Informing the community^ 314.2. Updating of records 324.3. Contacting the clients 33v5. CLINIC STAFFING AND DATA COLLECTION PROCEDURES^345.1. Orientation of staff to clinic procedures 355.2. The aims of the questionnaire-orientation session 355.3. Procedure for the finger-prick blood sample^366. DATA COMPILING PROCEDURES^ 377. COMPUTER SOFTWARE USED FOR THE ANALYSIS^37CHAPTER FOUR^ 3 9DEVELOPMENT OF THE QUESTIONNAIRE^ 391. PURPOSE OF THE QUESTIONNAIRE 392. CONTENT OF THE QUESTIONNAIRE 393. FORMAT OF THE QUESTIONNAIRE 414. PILOT STUDY OF THE QUESTIONNAIRE^ 424.1. Objectives of the pilot study 424.2. Results of the questionnaire pilot study^435. RELIABILITY OF THE QUESTIONNAIRE^ 446. QUESTIONNAIRE WEAKNESSES^ 467. VALIDITY OF THE QUESTIONNAIRE 477.1. Face validity^ 477.2. Content validity 47CHAPTER FIVE^ 48LABORATORY PROCEDURES FOR THE TESTING OF BLOOD SPECIMENS^481. IMMUNOASSAYS USED TO TEST FOR STUDY OUTCOMES^481.1. Introduction^ 481.2. Assessing the characteristics of the IMXimmunoassays 482. IMX CORE FOR THE ASSESSMENT OF ANTI-HBc OUTCOMES^492.1. Results of Anti-HBc positive retests for thisstudy^ 513. IMx AUSAB FOR THE ASSESSMENT OF ANTI-HBs OUTCOMES^513.1. Results of Anti-HBs negative retests for thisstudy^ 534. IMX HBsAG FOR THE ASSESSMENT OF HBsAG OUTCOMES^544.1. Results of HBsAg positive retests for this study 555. TECHNICAL PROBLEMS ENCOUNTERED IN TESTING THE BLOODSAMPLES^ 56CHAPTER SIX 57RESULTS^ 571. STATISTICAL ANALYSIS^ 572. BASELINE DISTRIBUTIONS AND COMPARISON OF PARTICIPANTSAND NONPARTICIPANTS 58vi3. VARIABLES COLLECTED TO ASSESS FOR BIAS IN THE SAMPLE 614. MAIN OUTCOME VARIABLES - DESCRIPTIVE INFORMATION^634.1. ANTI-HBc^ 634.2. ANTI-HBs 644.3. HBsAG 685. COMPARISON OF STUDY OUTCOMES WITH OUTCOMES AT 6 TO 18MONTHS OF AGE^ 696. ANTI-HBc OUTCOME - RESULTS OF ANALYSES^726.1. Association of covariates with Anti-HBc^726.2. Multiple logistic regression analysis for the Anti-HBc outcome^ 837. ANTI-HBs OUTCOME - RESULTS OF ANALYSES^897.1. Introduction 897.2. Associations of covariates with Anti-HBs titres 897.3. Multiple regression analysis for the Anti-HBsoutcome^ 918. HBsAG OUTCOME - RESULTS OF ANALYSES^ 958.1. Introduction 958.2. Associations of covariates with HBsAg^968.3. Multiple logistic regression analysis for HBsAgoutcomes - denominator all cases 1008.4. Multiple logistic regression analysis for the HBsAgoutcome - denominator restricted to anti-HBcpositive cases^ 103CHAPTER SEVEN^ 104DISCUSSION 1041. INTERNAL VALIDITY^ 1041.1. Considerations with respect to possible selectionbias 1041.2. Considerations with respect to information bias 1051.3. Considerations with respect to consistency ofresults^ 1062. EXTERNAL VALIDITY^ 1083. INFECTION AND CARRIER RATES^ 1094. THE OCCURRENCE OF INFECTION AND THE CHILD'S AGE^1115. PREDICTORS FOUND FOR ANTI-HBc AND HBsAG POSITIVEOUTCOMES^ 1116. SEROPOSITIVITY RATES^ 1177. PREDICTORS FOR ANTI-HBs OUTCOMES^ 1198. RELATIONSHIP BETWEEN ANTI-HBs TITRES AND INFECTIONS 1229. PROTECTIVE EFFICACY OF HB VACCINE 12510. SUMMARY^ 12611. RECOMMENDATIONS^ 12812. RECOMMENDATIONS FOR FURTHER STUDY^ 128viiBIBLIOGRAPHY^ 130APPENDIX 1 136STUDY DOCUMENTS^ 136APPENDIX 2^ 160RED CROSS REGISTRY FORM^ 161APPENDIX 3^ 162STUDENTIZED RESIDUAL PLOTS^ 163viiiLIST OF TABLESTableTableTableTable2. efficacy hepatitis B vaccineLong-term efficacy hepatitis B vaccineIncluded childrenExcluded Children8121718Table 5.1. Comparison of IMx Core and RIA assays for anti-HBc 50Table 5.2. Comparison of IMx AUSAB and RIA assays for anti-HBs 52Table 5.3. IMx HBsAg results compared to AUSZYME 55Table 6.1. Child's age at last birthday 59Table 6.2. Child's year of birth 59Table 6.3. Mother's HBeAg status 60Table 6.4. Baseline characteristics 62Table 6.5. Anti-HBs titres in mIU/ml 65Table 6.6. Results of 6 to 18 month tests 70Table 6.7. Comparisons between 6 to 18 month resultsand study results 70Table 6.8. Association of study outcomes with 6-18 monthoutcomes 72Table 6.9. Distribution of mothers' country of birth 73Table 6.10. Association of covariates with Anti-HBc 74Table 6.11. Number of doses of HB vaccine 77Table 6.12. Child's age at Dose one 78Table 6.13. Distribution of number of household contacts 82Table 6.14. Prediction models for the Anti-HBc outcomein children age 2 to 8 86Table 6.15. Association of covariates with anti-HBs titres 90Table 6.16. Final prediction model for Anti-HBs titresin children age 2 to 8 93Table 6.17. Association of covariates with HBsAg 97Table 6.18. Association of covariates with HBsAg(Anti-HBc positive cases only) 99Table 6.19. Prediction models for the HBsAg outcomein children age 2 to 8 102ixLIST OF FIGURES1. Infection and carrier rates according to age group 632. Infection and carrier rates according to year of birth 643. Seroconversion rates according to age groups 654. Anti-HBs levels 665. Geometric mean Anti-HBs levels - Both sexes 676. Geometric mean Anti-HBs levels - Both sexes 68>3 doses and late dose cases excluded7. Geometric mean Anti-HBs levels according to sex 698. Histogram of the log of Anti-HBs levels 92ACKNOWLEDGEMENTI would like to express my sincere gratitude to my supervisor, Dr.Richard Mathias, for his generous support, encouragement andguidance on this project. I would also like to thank the membersof my thesis committee: Dr. Stephen Marion for his advice onmethodology and for his patient and precise guidance with thestatistical analysis; Dr. David Pi for his support withfacilities, advice and encouragement; and Dr. John Farley for hisassistance with the proposal and the initiation of the project.Helpful suggestions or assistance on various phases of the projectwere appreciated from Dr. Joseph Tan, Dr. Alison Bell, Dr. RuthMilner, Ms. Sandra Wiggins and Ronnie Sizto.Appreciation is also extended to Margaret Elliot, her assistantsJudy Hrytzak and Glynda Stimson and all the technologists in theTransmissible Diseases Laboratory at the Red Cross for their workin keeping records on the children in the Program and for testingthe specimens. Also appreciated was the assistance of the HealthUnits and Health Departments with the updating of addresses andthe provision of space for clinics. A special thanks is expressedto the Richmond Health Department for their cooperation with thequestionnaire pretest, to Cheryl Lau and the nurses from TASC whocollected the blood samples and all the other project and U.B.C.staff members who assisted with the study.Finally, I would like to express my appreciation to the fundingagencies and organizations who provided the support which madethis research possible: The British Columbia Medical ServicesFoundation; SmithKline Beecham Pharmaceutical Inc.; The BritishColumbia Ministry of Health and the British Columbia HealthResearch Foundation.xiDEDICATIONI would like to dedicate this thesis to my husband, Giorgio, whowas 'my rock' and who unfailingly supported and encouraged methroughout the two years, as well as provided many hours oftechnical computer support for the project. And to my daughter,Carmen, who was 'my bright shining star' and who also encouragedme, volunteered at the clinics and licked hundreds of envelopes!xi iCHAPTER ONEINTRODUCTION AND GENERAL CONSIDERATIONSI. DESCRIPTION OF THE PROVINCE-WIDE PROGRAM FOR THE PREVENTIONOF HEPATITIS B VIRUS INFECTION IN INFANTS OF CARRIER MOTHERSIn July 1984, the British Columbia Ministry of Health extended itsProvince-Wide Program for the Prevention of Hepatitis B Virus (HBV)Infection in Infants of Carrier Mothers, (hereafter referred to asthe Program), to include the provision of hepatitis B (HB) vaccineat birth, one month of age and six months of age. Prior to this,infants of mothers who were found hepatitis B surface antigen(HBsAg) positive when screened by the Canadian Red Cross BloodTransfusion Service had been given only hepatitis B immune globulin(HBIG) as a prophylaxis against the HBV. The provision of HBIGbegan in 1977, with one dose at birth. In 1982 the Program wasexpanded to include a dose at three months of age. When hepatitisB vaccine was added to the protocol, the second dose of HBIG wasdropped. Initially, 10 ug of Heptavax, a plasma-derived vaccineproduced by Merck, Sharp and Dohme was provided. In March 1988,Heptavax was replaced by Engerix, a yeast-derived 20 ug vaccineproduced by SmithKline Beecham Biologicals. Both vaccines wereadministered intramuscularly. HyperHep, a preparation of 217international units per millilitre HBIG produced by Cutter was alsoadministered intramuscularly.The HBIG, since it was a blood product, was issued to the birthhospital by the Canadian Red Cross Society, Vancouver Centre, BloodTransfusion Service (hereafter referred to as the Red Cross). Withthe addition of the HB vaccine, the Red Cross continued toadminister the Program as well as keep a central registry of thedates of vaccine delivery and dose administration.The procedure of this on-going Program is as follows. Physiciansprovide a blood sample from pregnant women which is sent to the RedCross. Among other tests, this blood sample is screened for theHBV surface antigen. If it is determined that an expectant motheris positive for HBsAg, the physician is advised by the Red Crossfollowing which the HBIG and first dose of HB vaccine are sent tothe birthing hospital along with recording forms. A notificationletter is also sent to the family's area Health Department orHealth Unit with the second and third doses of vaccine andrecording forms. The respective hospital and Health Department orHealth Unit are expected to return the recording forms to the RedCross for recording of the dates of immunization on the centralregistry. If notification of immunization is not returned byapproximately one month after the due date, reminder notices aresent to the respective agencies by the Red Cross.2. OBJECTIVES OF THE STUDYIt was the purpose of this study to determine the effectiveness ofthe HB vaccine in the prevention of hepatitis B infection, asdelivered by this Program, between July 1984 through 1989. Thusthe study objectives were:1. To determine the long-term attack rate of HBV infection for thechildren immunized at birth as part of the Program between 19842and 1989 inclusive.2. To determine antibody titres (anti-HBs) as related to timesince birth in those children who were not infected.3. To determine what factors were related to attack rates andanti-HBs titres in this cohort of children.4. To evaluate the implications of infection rates and antibodystatus with regard to the (1) effectiveness of the HB vaccine,(2) evidence for waning immunity and (3) need for booster dosesof vaccine."Effectiveness", for the sake of this study will be defined bycomparing this Program's rate of HBV infection against the expectedrate of infection for this type of population without a HB vaccineimmunization program.3. JUSTIFICATIONEvaluation of the effectiveness of the HB vaccine in a mixed racialpopulation of neonates born to HBV carrier mothers of mixed HBeAgstatus is not available in the literature for children up to eightyears of age. It is not known if booster doses of HB vaccine arerequired for these children still at risk of HBV infection fromhousehold contacts. If booster doses are required, at whatinterval would they be most beneficial? This study attempts toanswer these questions by focusing on attack rates and anti-HBstitres at differing lengths of time since immunization. In as muchas certain population characteristics are similar and programdelivery practices are similar, the information from this study isexpected to be generalizable to other such programs.34. RESEARCH QUESTIONS AND HYPOTHESESWith respect to the neonates in this Program, who were born in 1984through to 1989 and who were immunized with HB vaccine immediatelyafter birth:1. What is the HBV attack rate over time?2. What are the anti-HBs titres and how do they change over time?3. Is there an association between attack rates and anti-HBstitres?4. What are the associations between the HBV attack rates andvarious demographic, program intervention, and exposurefactors?5. What are the associations between anti-HBs titres and variousdemographic, program intervention, and exposure factors?The null hypotheses corresponding to these research questions wereas follows:1. There is no association between infection rate and elapsed timesince immunization with HB vaccine.2. There is no association between infection rates and decreasinganti-HBs titres.3. There are no associations between the attack rates and variousdemographic, program intervention and exposure factors.4. There are no associations between the anti-HBs titres andvarious demographic, program intervention and exposure factors.The study design chosen to test these hypotheses and answer thesequestions was a cohort study of children born between 1984 and41989, inclusive, and tested cross-sectionally once in 1992. At thetime of the single follow-up the ages of the children ranged fromtwo to eight years.5CHAPTER TWOLITERATURE REVIEWI. BACKGROUNDHepatitis B (HB) is a viral infection of the liver.^It ischaracterized by three main serologic markers: HBsAg; which, ifstill present after 6 months, indicates the carrier state; anti-HBc; which indicates that viral replication (infection) hasoccurred at some time; and anti-HBs; which indicates immunity tohepatitis B virus (HBV). There is a greater likelihood oftransmission of infection where HBeAg is present in associationwith HBsAg than when HBeAg is not present. The HBsAg, anti-HBc andanti-HBs serologic markers are defined in more detail in the methodsection, Chapter Three, page 20.Hepatitis B virus is a major cause of acute and chronic hepatitisand cirrhosis. Over 300 million people throughout the world arepersistent carriers of HBV (Garrison and Baker, 1991). Theprevalence rate of current and past infection in many countries ofAsia, Africa and Oceania is in the range of 30-100% (West et al.1990). Although Canada is a country of low endemicity forhepatitis B infection, with a carrier rate of lower than 1% (West1990), many risk factors for acquiring HBV, such as illicit druguse, homosexual activity and multiple sexual partners, areprevalent in our society. Furthermore, increased immigration fromcountries where the disease is highly endemic has been thought tobe a factor in our rising incidence of hepatitis B (Canadian6Diseases Weekly 1987). Perinatal transmission from HBsAg positivemothers is one of the most common sources of HB infection. InTaiwan, approximately 40% of chronic carrier HBV infections are theresult of perinatal transmission (Beasley et al. 1983). Up to 90%of infants who are born to carrier mothers also become HBsAgpositive (Beasley et al. 1983). In British Columbia, prior to theinitiation of HB prophylaxis in 1977, 23% of infants born to HBsAgpositive mothers became carriers (Ballem et al. 1987).The most serious consequences of the HBV carrier state are thelong-term sequelae. The earlier in life that the infection fromHBV occurs the more likely it will progress to the chronic carrierstate, which in turn increases the risk of cirrhosis and cancer ofthe liver later in life (Polakoff and Vandervelde 1988). It isestimated that persons infected with the hepatitis B virus duringthe perinatal period have a 25% lifetime risk of death from livercirrhosis or primary hepatocellular carcinoma (West et al. 1990).2. SHORT-TERM EFFICACY OF HEPATITIS B VACCINEWhen the hepatitis B vaccine became available, neonatal programsbecame a priority in many countries, including Canada. Manystudies have established the short-term efficacy of HB vaccines innewborns. A summary of some of this research may be seen in Table2.1.7Table 2.1. Short-term efficacy hepatitis B vaccineAuthors AgeGroupMonths(age)1Number2AttackRate(%)HBsAg+2Anti-HBs+(%)Beasley (1983) neonates 9 159 6 100Esteban^(1985) neonates 12 52 4 88Stevens^(1985) neonates 18 113 14 96>50S/NStevens^(1987) neonates 12 122 7 100Lee^(1987) neonates 16 201 6 894Ballem (1987)(BritishColumbia)neonates 12 220 2 954Farmer (1987) neonates 12 39 18 77Hayashi (1987) preschool 24 203 0 74Polokoff (1988) neonates 12 102(79) 34 96Poovorawan (1989) neonates 13 55(46) 34 100The Gambian StudyGroup (1989)infants<lmonth12 710 1 98Waters^(1989)(Alberta)neonates 12 72(141)34 99Theppisai (1990) neonates 13 52 8 921 Denominator for attack and anti-HBs+ rates.2 Attack and anti-HBs+ rates are rounded.3 Denominator for anti-HBs+ rate.4 Not stated in the article if N excluded HBsAg+ children.In some of their earlier work, Beasley et al. (1983) studied acombination of HBIG and vaccine in different temporal combinations.They found that the combined attack rate at nine months, forinfants of HBeAg positive mothers, was 5.7%. For those that didnot become infected the seropositivity for anti-HBs was 100%.8Stevens et al. (1985) studied 113 infants of Asian-American motherswho were both HBsAg and HBeAg positive. They found that at 18months of age, 14.2% of infants were infected. Two differentvaccine schedules were employed, using HBIG at birth and plasma-derived Heptavax 20 ug. HB vaccine doses were given at either 0,1 and 6 or at 1, 2 and 6 months of age. The combinedseropositivity for anti-HBs of the two groups was 95.8%. Anotherstudy by Stevens et al. (1987) found similar results to Beasley(1983) when it compared two groups of neonates of HBeAg positivemothers: one received yeast-recombinant vaccine and the otherplasma-derived vaccine. The overall attack rate for the two groupswas 6.6% and the seropositivity rate was 100% for those that didnot become HBsAg positive. A study by Poovorawan et al. (1989) inThailand, found an even lower attack rate (3.6%) using 10 ug ofrecombinant DNA vaccine. At 13 months of age 100% of the childrenwho did not become positive for HBsAg had anti-HBs titres >10mIU/ml. The mothers were both HBsAg and HBeAg positive.Several studies using low doses of plasma-derived HB vaccine arereferred to in Table 2.1. Lee et al. (1987) reported an attackrate of 5.5% for 201 16-month-old infants who had either receivedthree 2.5 ug doses of HB vaccine or a schedule including only twodoses of 5 ug vaccine. Anti-HBs seropositivity over 10 mIU/m1 forthe two groups combined was 88.7%. Similarly, Theppisai et al.(1990) found an attack rate of 7.7% at 13 months for infants whohad received four doses of either 2 or 5 ug of vaccine. Theseropositivity rate for those not infected in the two groups was9100%. In contrast, in a study in New Zealand, a high attack rate(22%), was found in a group that received 5 ug of Hepavax vaccinealone. For the group that received vaccine combined with HBIG, theattack rate was 14.3% (Farmer et al. 1987). In the foregoingstudy, anti-HBs seropositivity was 72% for the vaccine-only groupand 81% for the vaccine plus HBIG group at one year of age.In 1989, The Alberta Neonatal HB Vaccination Program, which hadbeen in operation since 1985, reported a 4% (3/72) attack rate byone year of age (Waters, 1989). The seroconversion rate in thosetested was 98.6% (140/142). In an earlier report of the BritishColumbia Program, for the years 1983 to 1986, only 2% of 220infants were reported HBsAg positive and 95% were anti-HBs positiveafter 12 months of age (Ballem et al. 1987). The immunizationschedule for both the Alberta and British Columbia programsincluded HBIG given immediately after birth. In Alberta the HBvaccine was given at birth, two and six months of age, whereas, inBritish Columbia the HB vaccine was given at birth, one and sixmonths of age.The HBV infection attack rates for infants in the short-termstudies listed in Table 2.1. ranged between 0.28% for the GambianHepatitis Study Group (1989) and 18% for the study done by Farmeret al. (1987). The weighted-average attack rate was 3% for thesestudies which varied in length of follow-up from 9 to 24 months.The two studies with low attack rates by Hayashi et al. (1987) andthe Gambian Group (1989) did not necessarily involve HBsAg positive10mothers. It should also be noted that the attack rate statisticincluded only those children who were HBsAg positive. Children whowere anti-HBc positive were not included in the calculation of theattack rates since their antibodies could have been those acquiredin utero, and therefore would not necessarily have been induced bysubclinical infection. The weighted-average short-term anti-HBspositive rate for the studies in Table 2.1. was 94%. The range wasbetween 74 and 100%. The denominator for calculating theconversion rates did not include those children positive for HBsAg.(Possible exceptions are Lee et al. and Ballem et al., whereseroconversion rates were given with no indication of actualnumbers.)3. LONG-TERM EFFICACY OF HEPATITIS B VACCINEAfter birth, when the vertical risk of maternal transmission of HBVinfection is over, there is a continuing risk of horizontaltransmission in the household. According to Tong (1989), 40% ofcarriers acquire the virus as a result of perinatal transmission,and another 35-40% become infected during the preschool years. Sixto ten percent of infections are acquired after the fifth or sixthyear of life (Maynard et al. 1989). Franks et al. (1989) reportedthat nearly half the cases of HBV infection among the U.S.-bornchildren of refugees were not attributable to perinataltransmission from a mother with infectious disease. Thus continuedimmunoprophylaxis during the early years of life is important toprevent the chronic carrier state.11Most long-term trials on the efficacy of HB vaccine have focused onthe protection of adults, anti-HBs seropositivity rates and titresof antibodies A summary of attack rates and anti-HBsseropositivity rates for long-term studies may be seen in Table2.2.Table 2.2. Long-term efficacy hepatitis B vaccineAuthorsAgeGroupYearsafter1stdose1Number2AttackRateHBsAg&/oranti-HBc2Anti-HBs+>10 mIU/m1Wainwright (1984) all ages 5 1114 1 81 3Coursaget (1986) infants 4 37 3 98(any level)Coursaget (1986)Seven year studyinfants 7 135 7 -Jilg (1988) adults 6 39 - 66Horowitz (1988) adults 3 245 - 62Mannucci (1989) 1-54yrs 4 52 - 100Ip (1989) neonate 3 188 38 81Moyes (1990) neonate 4 70 0 781 Denominator for attack and anti-HBs+ rates.2 Attack and anti-HBs+ rates are rounded.3 Of those who had an initial response of 10 SRU's or greater.Coursaget et al. (1986) undertook two long-term HB vaccine efficacystudies with infants in a hyperendemic area. A series of threeinjections of 5 ug Pasteur vaccine at one month intervals as wellas a booster at 12 months were administered in both studies. Thethree-year study resulted in an attack rate of 2.7%, with one childof 37 tested, becoming positive for anti-HBc and none for HBsAg.After three years 97.5% of the children were still anti-HBspositive. The second study was of seven years duration followingadministration of the first dose of vaccine. Nine of the 13512children had HBsAg or anti-HBc events after this time demonstratingan attack rate of 6.7%. Another study of infants in an endemicarea was undertaken in New Zealand but in this case it wasdetermined that all mothers were non-carriers (Moyes et al. 1990).Three injections of 2 ug of plasma-derived vaccine wereadministered and 4 years later none of the 70 infants testedpositive for anti-HBc. Seventy-nine percent had anti-HBs titresgreater than 10 mIU/ml. Wainwright et al. (1989) also reported avery low attack rate in a study of all ages of an Eskimo populationin Alaska. An attack rate of 0.36% was found after five years and81% of 1114 individuals had anti-HBs titres of 10 SRU or greater.In contrast, Ip et al. (1989) reported a combined attack rate of38% in Hong Kong neonates of HBeAg-carrier mothers after threeyears. Three immunization schedules were used in this study,combining HBIG and HB vaccine. The average of the seropositivityrates (>10 mIU/ml) for the three groups after three years was80.5%.The attack rates for the long-term studies listed in Table 2.2.ranged between 0.36% and 38%. Although only the study by Ip et al.(1989) involves follow-up of neonates, the adult and "all age"studies were included in our calculation, for an average attackrate for long-term studies, in order to be conservative in ourestimate. Thus the weighted-average attack rate for the long-termstudies was 5.5%, somewhat higher than the short-term attack rateof 3%. Four of the above studies undertaken in endemic areasdemonstrated low attack rates even after seven years. However, the13attack rate of 38% in the perinatal study was in sharp contrast tothe rates in most of the other long-term studies. This raisesconcern that the protection afforded by vaccine against HBV is oflimited duration, particularly in neonatal populations that areexposed to a continued risk of infection.^This long-termeffectiveness study in a low prevalence population reassessesthat risk.4. LONG-TERM DURATION OF ANTIBODIESResults of long-term follow-up studies in adults showed evidence ofdecline in anti-HBs titres in these subjects. Jilg et al. (1988)reported that 34% of health-care workers had antibody titres thatdropped below 10 mIU/ml after six years. Horwitz et al. (1988)reported that 38% had antibody titres less than 10 mIU/ml afterthree years. These findings are not confirmed in a study of HIVnegative hemophiliacs (Mannucci et al. (1989). After four years,this group, with a mean age of nine years, had a 100%seropositivity rate for anti-HBs. Wainwright et al., 1989 reporteda rate of 81% for an Eskimo population of all ages after 5 years.The weighted-average seropositivity rate after three to seven yearsfor all the long-term studies listed in Table 2.2. is 79%. Forinfants and neonates alone, anti-HBs seropositivity rates afterthree to four years were in the range of 78.2 to 97.5%.There are studies that claim that protection against HE infectionclearly parallels the persistence of anti-HBs (Coursaget et al.,Seven-year study, 1986). Hadler et al. (1986) in a study of14homosexual men, found that the HBV infection was inversely relatedto the maximal antibody response to vaccine and that this riskincreased when antibody titres fell below 10 SRU. As a secondaryobjective, this study looked for an association between antibodytitres and HBV attack rates and explored the issue ofsusceptibility to HBV disease due to waning immunity.15CHAPTER THREEMETHOD1. SAMPLE SPECIFICATIONThe target population for this study was all neonates born tomothers who tested hepatitis B surface antigen positive as part ofthe Province-Wide Program for the Prevention of Hepatitis B VirusInfection in Infants of Carrier Mothers and who commenced the HBvaccine at birth during the years 1984 through 1989. There were atotal of 1398 records of children at the Red Cross who were bornduring these years. Of these, 1166 (84%) children qualified forinclusion, 224 (16%) were excluded and 8 (0.57%) records weremissed.1.1. Included childrenThe sample was comprised of all children from the target populationwho could be located in British Columbia and for whom parentalconsent was given to participate in the study. Of the 1166children who qualified for inclusion and were included in the database, 770 (66%) participated in that the parents or guardians wereinterviewed and the child provided a blood sample. The dispositionof included children can be seen in Table 3.1.Included in the 770 children were two children with evidence ofhaving commenced the HB vaccine series outside of British Columbia,but who had completed the vaccine series in British Columbia aspart of the Program. Also included were three children who wereborn in early 1984, before the Program officially started. For16these children, there was evidence in the Red Cross records of themother being HBsAg positive and there was a record of HB vaccinefor each child, although some of the doses were given late.Table 3.1. Included childrenDisposition of included children Number Percentof includedchildrenParticipants 770 66.0Unable to locate 231 19.8After inclusion determined to have movedout of B.C. 30 2.6After inclusion, determined to have died 1 .1Agreed to participate but did not showfor an appointment 16 1.4Agreed to participate but an appointmentcould not be arranged 32 2.7Refused to participate 86 7.4TOTAL INCLUDED 1166 1001.2. Excluded childrenThere were a total of 1398 records of children who were born in1984 through 1989 in the Red Cross registry for the Program.Categories and numbers of children excluded from the study can beseen in Table 3.2. Children excluded were those who had notreceived the HB vaccine, had moved out of British Columbia or hadmothers with no evidence of being HBsAg positive. One child wasexcluded because of death. Another child who had died wasinitially included in the data base since the Red Cross was unawareof this child's death.17Fifty seven (4%), of the 1398 children, were given the HB vaccineand had mothers who were anti-HBc positive but HBsAg negative. Afurther 22 (1.6%) were given the vaccine but there was noindication on the Red Cross record that the mother was either HBsAgor anti-HBc positive. These children were excluded since exposureto the HBV at birth was not verified. Children who lived in theYukon or had moved out of British Columbia were excluded because ofcost constraints.Table 3.2. Excluded ChildrenReasons for exclusion Number Percent oftotalrecordsMoved out of B.C.^(Red Cross records) 41 2.9Children given vaccine but Mother HBsAgnegative and Anti-HBc positive 57 4.0Children received vaccine but no recordof HB markers for the mother 22 1.6Mother converted to HBsAg negative beforechild's birth 7 0.5Mother HBsAg positive but no record of HBvaccine given to the child (85 born in1984 before the Program started in July1984)86 6.2Out of Province births in B.C.(Yukon Territories)7 0.5Parent ill with acute HB while child aninfant 2 0.14Adopted, adopting parents and addressunknown1 0.07Death as a newborn 1 0.07TOTAL EXCLUSIONS 224 16.0Two female children who were born in British Columbia hospitals,18one in 1986 and one in 1988, were excluded as no record ofimmunization with HB vaccine could be found at the Red Cross,Physician's office, Health Unit or in the Health Passport. Thesetwo children, for whom there was no record in the Red Crossregistry, were identified through participating siblings,therefore, the mothers were known HBV carriers.2. STUDY VARIABLES2.1. Purposes for collection of variablesBaseline information was collected for all 770 participatingchildren. The same information was collected on non-participantsto allow comparison for basic differences which might indicate biasin the sample obtained. These variables included the child's ageand sex and the mother's age and HBeAg status. Information onthree other variables was collected to allow assessment fordifferences between specific groups. Firstly, the language spokenin the home was asked of those who refused to join the study inorder to compared the languages of this group with the languagesspoken by participants. A difference between groups when languageis in some way associated with infection may indicate the resultsare biased. Secondly, information on another social variable,history of moving, was collected for all participants, to comparemoving history with infection status. If there was a difference ininfection rates between those who had moved (representing thosecases we could not locate) and those who had not moved, then theremay be bias in the results. If, for example, we were to find thatthose people who had moved had a higher infection rate, then our19results may give an underestimate of the true rate. Thirdly, toassess if there was a difference in the effort to contact thosechildren infected compared to those not infected, the number ofphone calls made to locate each participant was recorded.Information on other demographic, Program intervention and exposurevariables was collected to search for associations with outcomevariables and to build prediction models for these outcomes.2.2. Definitions for Dependent Variables1. Hepatitis B core antibody (anti-HBc) is a specific antibody tothe hepatitis B core antigen that develops during acuteinfection and persists for the lifetime of the previously-infected person (Evans ed. 1989). Presence of this antibodyin a child's serum was considered evidence of vaccine failure.This dependent variable was categorized as positive ornegative. Results of the serological test for anti-HBc weretaken from Red Cross records for the 6 to 18 month testing andfrom the 1992 testing of samples collected for the study.Specimens were not tested for anti-HBc IgM because oflimitations on the amount of serum available.2. Antibody to the surface antigen (anti-HBs) is the specificantibody to the hepatitis B surface antigen and appears afterexposure to this antigen either following naturally occurringinfection or through immunization (Evans ed. 1989). For thisstudy, and in the absence of anti-HBc, presence of thisserologic marker was considered evidence of response to the HBvaccine if titres were greater than or equal to 10 milli-20international units per millilitre (mIU/ml). Anti-HBs was acontinuous dependent variable. Anti-HBs titres taken from theRed Cross records for 6 to 18 month tests were measured in'ratio of sample counts per minute to mean counts per minuteof negative control' (S/N). Since this unit of measurement wasnot comparable to the mIU/m1 unit of measurement which was usedin this study, results were categorized as positive for 10S/N's or greater and negative for less than 10 S/N's.3. Hepatitis B surface antigen (HBsAg) is the glycoprotein coatof HBV and exists both associated with the complete virion andindependently. If found in the serum of an individual, HBsAgindicates acute or chronic HBV infection (Evans ed. 1989). Forthis study, a child was classified as a carrier if he or shehad a positive serologic test for HBsAg. We did not attemptto distinguish between chronic and acute carriers since it wasnot possible to obtain a second blood sample within the studyperiod to confirm the chronic carrier status. The variable,HBsAg, was categorized as positive or negative for both 6 to18 month results and study blood test results. The serologictest for HBsAg was done only if there was sufficient bloodsample after completing tests for anti-HBc and anti-HBs sinceHBsAg was not included in the primary hypotheses.2.3. Definitions for Independent Variables2.3.1. Program intervention variables1. Number of doses of HB vaccine received was a continuousvariable measured from 1 to 6. No distinction was made between21the types of HB vaccine administered since there was overlapbetween products. Heptavax (Merck Sharp & Dohme), a plasmaderived-product, was used at the program inception. Engerix(SmithKline Beecham) a recombinant vaccine, was initiated inMarch 1988, and was issued until after the study period.2. Age when Dose one was received was a continuous variableconsisting of the number of days of age the child was when thefirst dose of vaccine was received.3. Dose two timing was a dichotomous categorical variable. 'Yes',indicated the child had received the second dose of the HBvaccine within two months of dose one, and 'No' meant the childdid not receive the HB vaccine within two months of dose one.4. Dose three was a categorical variable, 'Yes/No', indicatingwhether or not the child had received the third dose of the HBvaccine series.5. Dose four and Dose five were categorical variables, 'Yes/No',indicating whether or not the child had received additionaldoses of HB vaccine beyond the protocol of the Program.6. Hepatitis B immune globulin or HBIG was a categorical variableclassified as 'Yes' when the child had received HBIG at birthand 'No' if HBIG was not received. For the multivariateanalyses, HBIG was classified as 'Yes/Unknown/No' and treatedas a continuous variable. HBIG, if administered, was givenwithin 48 hours of birth.2.3.2. Demographic variables1. The child's age was a continuous variable when used in the22multivariate analyses and categorical when used in univariateanalyses. As a categorical variable, age was grouped by yearof age at the last birthday (2 to 8 years). Age was calculatedfrom the day of birth to the day the blood sample was taken.For non-participating children the midpoint day (July 31, 1992)of the blood sample taking period was used as the sample dayin the calculation of age.2. Child's year of birth was a categorical variable groupedaccording to year of birth (1984 to 1989).3. Child's sex was a categorical variable, male/female.4. Mother's age was used as a continuous variable in themultivariate analyses and as a categorical variable in theunivariate analyses. The categorical variable was groupedaccording to quartiles with the age categories from 1 to 4being 15.7 to 26.7; 26.8 to 30.0; 30.1 to 33.1 and 33.2 to 42.1respectively.5. Mother's hepatitis B e antigen status was a measure of themother's infectivity and was a categorical variable measuredas positive or negative. Negative was the reference level.This information, when available, was taken from the Red Crossregistry for mothers who were tested before the birth of thechild.6. Mother's country of birth was a categorical variable and wasused only as a grouped variable. Countries were groupedaccording to the most common mode of transmission of the HBVin that country (vertical or horizontal). Verticaltransmission referred to perinatal transmission and horizontal23transmission referred to transmission by close association withother persons after birth. Classification was made based ongeographic patterns of frequency of mode of transmission of HBVinfection as given by Maynard, Kane and Hadler (1989) with theexception of Africa. The World Health Association report ofMaternal and Perinatal Infections, 1991, suggested that HBVinfections in children in Africa are associated with highhorizontal transmission rates in early childhood. Thecountries grouped according to vertical transmission were:China, Hong Kong, Viet Nam, Philippines, South East Asia,Taiwan, North and South Korea, and Macaw. Countries groupedaccording to horizontal transmission were Canada, India,Pakistan, the Middle East countries, East and North Africa,Sub-Sahara Africa, Eastern Europe, Japan, Western Europe, Southand Central America, United States of America, Australia andNew Zealand.7. Father's country of birth was categorized in the same way asmother's country of birth.8. Years mother spent in her country of birth was groupedaccording to quartiles and used in the univariate analyses asa categorical variable and as a continuous categorical variablein the multivariate analyses. Categories 1 to 4 were: 1 to 13years; 14 to 20 years; 21 to 25 years; and 26 to 40 yearsrespectively. So as to measure the effect of the number ofyears spent in a foreign birth country, this variableconsidered those mothers born in Canada as having zero yearsin her country of birth.249. Years father spent in his country of birth were coded andgrouped the same way as 'Mother's years in birth country' withquartiles categories 1 to 4 being: 1 to 12 years; 13 to 20years; 21 to 26 years; and 27 to 44 years respectively.10. Language was a categorical variable grouped by vertical orhorizontal transmission according to the countries representedby the languages spoken. Chinese, Korean, Vietnamese, SouthEast Asian and Philippine languages were included in thevertical transmission group. English, Japanese, European,African, Middle Eastern, and East Indian languages wereincluded in the horizontal transmission group. This variablewas also used to assess for selection bias in the sample asdiscussed in Chapter Three, page Illness and exposure variablesVariables 1 to 12 below were categorical variables classified as'Yes' or 'No' for both univariate and multivariate analyses.1. Hepatitis illness was considered 'positive' if the child hadbeen diagnosed by a physician or health care workers. Evidenceof laboratory confirmation was not requested.2. Jaundice was yellowing of the skin and eyes due to hepatitis.Physiological jaundice due to birth or breast feeding wereexcluded.3. History of breast feeding was defined as breast feeding for atleast a week in duration and at least one feeding a day.4. Ear piercing was defined as puncture of the ear for the purposeof wearing an ear ring.255. Acupuncture was defined as the Chinese practise of puncturingthe body with needles to cure disease or relieve pain.6. Tattooing was defined as puncturing the skin for the insertionof pigment to mark the skin.7. Surgical procedures were defined as procedures that involvedan incision or suturing as performed by a health professional.Dental procedures were excluded.8. Blood transfusions were defined as the child receiving bloodintravenously in any amount.9. Blood products were defined to include plasma, serum, thrombin,fibrinogen, packed red blood cells and cryoprecipitate. Theydid not include immune globulin, heat-treated plasm proteinfraction, albumin or fibrinolysin.10. Daycare was defined as any out-of-home, formal, paid-fordaycare in a group setting, for example, group or family daycare. It did not include in-home baby sitting, whether doneon a regular or occasional basis.11. Hospitalization was defined as overnight admission at leastonce. Visits to the Emergency Room not followed by admissionwere not included.12. Travel included only trips to overseas continents or to Mexico,Central or South America. Trips to the United States or withinCanada were not included.13. Minimum age at travel abroad was a categorical variableindicating the youngest age at which the child had travelledand was grouped according to the child's age at last birthday.In the multivariate analyses, minimum age at travel was treated26as a continuous variable.14. Total months abroad was a continuous variable representing thetotal time, in months, for each child, for all trips abroad.This variable was also grouped according to: 'no monthsabroad/less than three months abroad/three months or moreabroad'. The grouped variable was used in the multivariateanalyses and treated as a continuous variable.15. Contacts was a continuous variable indicating up to 10 thenumber of household contacts the child had had since birth.To qualify as a household contact, the person had to haveshared the same kitchen as the child and lived in the samehouse for at least one month. For the multivariate analyses,contacts were grouped as 1,2,3,4,5,6, and 7+ but treated as acontinuous variable.2.3.4. Variables collected to assess for selection bias1. Moved residences since the child was born was a categoricalvariable classified as 'yes' or 'no'. A change in permanentaddress after the child was first brought home from thehospital at birth constituted a move for that child.2. Phone calls was a continuous variable indicating the number ofcontact attempts made by phone for the purpose of locating aspecific case. Calls included those made to the family's home,place of business, physician and/or health department.3. Language was discussed above under demographic variables.273. DATA COLLECTION PROCEDURESThe main goals of the data collection procedures were:1. To collect blood samples from as many of the target populationas possible.2. To collect information on independent variables by conductinga questionnaire interview with parents and by recording datafrom children's records at the Red Cross.3.1. Study DocumentsThe documents prepared for the study were the following:1. Physician information letter2. Parent information letter3. Telephone introduction guide4. Consent for participation form5. Questionnaire and interviewer's guide6. Laboratory test report form and testing protocol7. Result letter forms for parents and physicians.Copies of the above documents may be seen in APPENDIX Physician information letterThe purposes of the physician letter were as follows:1. To inform the physician of the goals of the study and whatparticipation in the study would involve.2. To inform the physician that his/her patient would be contactedand may wish to consult with their physician before agreeingto participate.3. To advise the physician that they would receive a copy of theresults of the tests for their patient if the subject's parentsconsented.4. To request change of address or telephone information on their28patient to assist in locating of prospective subjects.5. To give the physician an opportunity to request that hispatient not be contacted by the study team in the case of deathof a child or other extenuating circumstances.3.1.2. Parent information letterThe purposes of the parent information letter were as follows:1. To advise parents of the study; its goals, who the researcherswere and what participation in the study would involve.2. To invite parents to consider allowing their child/ren toparticipate in the study and inform them that a member of thestudy team would contact them by telephone.3. To give parents a name and phone number to call for moreinformation.3.1.3. Telephone introduction guideThe purposes of the telephone introduction guide were as follows:1. To provide a format for phone contact with parents ofprospective subjects in order to invite permission for theirchild/ren to participate in the study and to arrange anappointment.2. To reinforce the importance of using the same format for allcalls in order to prevent selection bias.3. To provide instruction on how to respond to a parent who isuncertain about participation, has questions about the studyor refuses to participate.4. To provide guidance in the case of language barrier.293.1.4. Consent for participation formThe purposes of the consent for participation form were as follows:1. To review the purposes of the study with parents and give theman opportunity to ask questions.2. To describe what participation in the study would involve; thatis, providing a blood sample and completing a questionnaire.3. To advise parents of the confidential nature of study recordsand the voluntary nature of their participation.4. To advise parents that blood test results for their child wouldbe sent to them.There were three places for parent signatures on the consent form.As well as giving permission for their child to participate in thestudy, parents, if they wished, could give consent for a copy ofthe results of the blood tests to be sent to their physician and/orcould indicate willingness to be contacted for future studies onthe effectiveness of HB vaccine in their child.3.1.5. Interview questionnaireThe purpose of the questionnaire was to provide a tool for thecollection of data on demographic variables and exposure riskfactors. The questionnaire development and questionnaire guide arediscussed in Chapter Four.3.1.6. Laboratory test report form and protocolThe purposes of the laboratory test report form and protocol were:1. To provide a record of the study outcome results (anti-HBc,anti-HBs and HBsAg) for recording on the Red Cross registry30form and for entry into the study database.2. To provide instructions on the method for taking blood samplesand shipping procedures.3. To provide a protocol for testing to ensure sufficient bloodwas available for the primary outcome (anti-HBc) tests and toguide necessary retesting.3.1.7. Parent and physician result lettersThe purposes of the parent and physician result letters were asfollows:1. To provide parents and physicians with the results of thechild's blood test along with an interpretation of the results.2. To give recommendations if follow-up was required as a resultof blood test outcomes.3. To thank the parents and children for their participation inthe study.4. PROCEDURES IN PREPARATION FOR DATA COLLECTION4.1. Informing the communityBefore beginning the study, the goals and procedures of the studywere discussed with the Health Officers Council of BritishColumbia. Physicians in British Columbia were informed of thestudy in a protocol published in the British Columbia MedicalAssociation Journal two months before the study began. A letterwas sent by Dr. John Farley, Director, Division of CommunicableDisease Control, British Columbia Centre for Disease Control(BCCDC), to all Provincial Health Units and City Health Departments31to advise them of the goals of the study and their possibleinvolvement. Health Departments and Health Units were requested toassist by providing current addresses for the prospectiveparticipants and by providing space for clinics where possible.4.2. Updating of recordsStaff were hired to assist with the computer entry of 'locatingdata' for the 1166 children eligible for inclusion. These data,including the child's and mother's birth dates, the mother's name,address and phone number, and, when available, the father's andphysician's name and address, were taken from the Red Crossregistry form and entered onto a dBase database. A copy of the RedCross registry form may be seen in APPENDIX 2. Once these'locating data' were recorded in the computer, a printout for eachchild was made. This printout provided a form for the possiblecorrection of locating information, for the recording of contactattempts and for the disposition of the case.Since it was expected that many of the addresses on the Red Crossrecords would be out of date, updating of the children's 'locatinginformation' was done prior to sending out the physician and parentinformation letters. The Provincial Health Units and City HealthDepartments played a key role in this updating process. They didso in three ways: (1) by providing computer printout lists ofchildren who had been on the Program; (2) by providing access tocomputer or hard records to study staff; and (3) by providingcorrections to printouts of children's, parents' and physicians'32names, addresses and telephone numbers. In spite of the updatingprocess 167 (14%) of the parent letters and 32 (2.7%) of thephysician letters were none the less returned after the firstmailing. The response rate of subsequent mailings was enhancedwith the assistance of physicians' office records and telephonedirectories.4.3. Contacting the clientsOnce the 'locating information' in a particular area was updated,the physician information letter was sent out. After one week plusfive days delivery time, the parent information letters were mailedout. In the meantime, arrangements were made regarding locationsand times for clinics in Health Units, Health Departments andcommunity service rooms. (Home visits were made when the numbersin a community were too few to warrant a clinic or when clientswere without transportation.) Once clinic times were scheduled andapproximately one week had passed since the parent letters weremailed, attempts were made to contact the parents of the childrento request participation in the study and to set up appointmenttimes.Five research assistants, two of whom spoke Chinese, and twoVietnamese volunteers assisted with telephoning prospectivesubjects. Over 70% of the phone contacts were made by one researchassistant. Orientation to the contacting procedure included thefollowing:1. Review of the telephone introduction guide and format.332. Stressing the importance of using the format wording andfollowing guidelines consistently.3. Stressing the importance of not coercing the parents.4. Demonstration of the phoning procedure.5. Return demonstration of the phone procedure and feed back.5. CLINIC STAFFING AND DATA COLLECTION PROCEDURESClinic appointments were set up at approximately 30 minuteintervals. Appointments began with review and signing of theconsent form. Care was taken to make sure the parent could readthe face sheet. If they couldn't, the consent form was read tothem or explained to them. The three signature sections were alsoexplained, and questions about the study were answered. If theparent did not speak English, interpretation was provided.After parents or guardians signed the consent form, givingpermission for the child's participation, the questionnaire wascompleted by interview by the research staff. Finally, a fingerprick blood sample was obtained from the child.Blood samples were taken by three nurses experienced in takingfinger blood samples. These nurses were provided by TASC ResearchServices on a fee-for-service basis. The study coordinator and oneChinese-speaking research assistant were trained by the TASC nursesin the finger blood sampling technique as well. Research staff whotook blood samples were immunized against the HBV.34A second Chinese speaking research assistant interviewed parentsbut did not take blood samples. Volunteers assisted with thereception of clients and interpretation. In seven cases it wasnecessary to complete the interviews by telephone, due to languageproblems. In three other cases interviews were also completed bytelephone. A special clinic was held for Vietnamese-Canadians withVietnamese-language interpreters. Chinese interpretation wasprovided at most clinics.5.1. Orientation of staff to clinic procedures was as follows:1. A package of information including a list of clinic duties,study documents and background information on hepatitis Bdisease and vaccine was given to clinic staff.2. The study coordinator reviewed the documents with the staff,giving an opportunity for questions.3. The consent form completion and questionnaire interview weredemonstrated by the coordinator.4. A return demonstration was observed by the coordinator followedby feedback.5.2. The aims of the questionnaire-orientation session were:1. To ensure the interviewer understood the meaning of thequestions and the terms used. The questionnaire guide was usedto assist in clarification of terms.. 2. To ensure the interviewer understood how to complete thequestionnaire with the parent responses.3. To stress the importance of using the exact words of the35questions consistently to avoid bias. For example, rewordingthe question about hepatitis history in a leading manner suchas "Your child hasn't had hepatitis has he?" should not bedone.4. To stress the importance of completing all the questions,thereby avoiding missing data. (It was required that allquestionnaires be checked for completeness by the interviewerat the end of each interview.)5.3. Procedure for the finger-prick blood sampleBefore taking the blood sample the child's hand was warmed using aheating pad or warm water in order to increase the flow of blood tothe hand. All research staff wore gloves while taking bloodsamples. After cleansing the skin with an alcohol swab and dryingit, with sterile gauze, the side of the finger next to the nail bedwas pricked with a sterile single-use spring-loaded lancet. Theskin was pulled as this was done to assist in penetration. As theblood flow began, the hand was gently squeezed and released toallow blood flow to return. The nail bed was gently pumped aftereach drop. Too much pressure was avoided so as to preventhaemolysis of the blood sample. Nurses were requested to obtain atleast two microtainers of blood and more if possible. Once theblood sample was obtained, a bandaid was put on the child's fingerand the child was given a sticker as a reward. Blood samples werelabelled with I.D. numbers which were double checked against I.D.numbers on the questionnaire and laboratory form. Samples werekept cool until delivery to the Red Cross, which was, in most36cases, within 24 hours. (Please see Chapter Five for a discussionof the laboratory testing procedures.)In all, 770 blood samples were collected and 770 interviewquestionnaires were completed between the beginning of May and theend of October 1992. In order to do this, 57 clinics were held and47 home visits were made. Eight trips were made around theprovince of British Columbia and 31 communities were visited apartfrom Vancouver.Results of blood tests were merged with form letters appropriate tothe respective outcomes and sent to parents and physicians.Samples of these letters may be seen in APPENDIX 1.6. DATA COMPILING PROCEDURESA spreadsheet was constructed in Lotus 1-2-3 for the input of data.Blood test results, program intervention variables and demographicvariables were initially entered into a dBase IV database and thenimported into the Lotus spreadsheet. Questionnaire data wereentered directly into the Lotus spreadsheet. Finally, these datawere cleaned, transferred to the University of British Columbiamainframe computer and prepared for the analysis.7. COMPUTER SOFTWARE USED FOR THE ANALYSISThe SPSS statistical package was used for most of the analysisprocedures. Where Pearson exact statistics were required becausegreater than 20% of cells had expected frequencies less than five,37the SAS program was used. The Egret and Statxact testing packageswere used where the exact trend tests or exact confidence intervalswere required. Finally, where numbers could be taken frompreviously calculated tables in SPSS, Epi Info version 5, StatcalcMode, was used. In this instance the uncorrected chi-square orMantel-Haenszel statistic was used.38CHAPTER FOURDEVELOPMENT OF THE QUESTIONNAIRE1. PURPOSE OF THE QUESTIONNAIREThe purpose of the questionnaire was to gather information on thepresence of various risk factors that might be associated with HBVinfection in the study cohort. The information gathered would thenbe used to answer the third research question: "How are attackrates associated with various risk factors?" By including riskfactor variables in the model building process, the analysis coulddetermine if there was an association between these risk factorsand the HBV attack rate. In the development of the questionnaire,content and format were the main factors considered and will bediscussed in the following. (A sample of the questionnaire may beseen in APPENDIX 1.)2. CONTENT OF THE QUESTIONNAIREThe first step was to determine what risk factors for thetransmission of the hepatitis B virus (HBV) were known to beassociated with HBV infection and which of these risks wererelevant to the age-group of the study cohort. The literature wasreviewed and experts in Public Health and Communicable Diseaseswere consulted. There were no questionnaires regarding this agegroup available in the reviewed literature. Long-term studiespreviously undertaken indicated that data collected were limited todemographic or perinatal information (Coursaget et al., Seven-yearstudy, 1986; Ip et al. 1989).39There is evidence that HBV infections have been transmitted bypercutaneous and permucosal exposure to infected blood or bodyfluids (Benenson ed. 1990). Thus questions were designed toinclude assessment of opportunities for transmission viaintravenous, intramuscular, subcutaneous, intradermal, or mucousmembrane means. Opportunities for blood exposure include bloodtransfusion, receipt of blood products, surgical procedures andother procedures which might occur during hospitalization.Procedures in the community could include acupuncture, tattooingand earpiercing. Use of illicit intravenous drugs as a means oftransmission was excluded as this was not a likely means in thetwo-to-eight-year-old age group. Acupuncture and tattooing wereincluded on the recommendation of the British Columbia Centre forDisease Control.Evidence also suggests that the virus can be spread via nasalsecretions, saliva or skin ulcers or rashes that exude serum(Taylor et al. 1989). According to Davis et al. (1989)contamination of ulcers or abrasions with saliva or small amountsof blood may be a means of transmission. Thus questions about thepresence of close contacts (specifically household contacts), andthe use of day care for the child were included. Transmissionbetween children in day care in developed countries has beeninfrequent; however, occurrence has been documented (Shapiro et al.1989). Questions about history of sores and abrasions were notasked since almost all children experience these. Similarly,almost all children visit the dentist; thus, these exposures wouldnot be useful in discriminating between groups.40Prevalence rates in some overseas countries suggest that the mainmechanism of transmission between the ages of one to fourteen ishorizontal spread (Taylor et al. 1989). Furthermore, travel tosome of these countries may increase the risk of exposure. Forthis reason a question on foreign travel was included. Somecountries have higher prevalence rates for HBV infection and HBeAgcarriage; therefore, demographic information, such as country ofbirth, for the mother and father, was asked in the questionnaire.There is conflicting evidence on breast feeding as a mechanism oftransmission of HBV. Beasley et al. (1975) found no relationshipbetween breast feeding and the subsequent development ofantigenaemia in babies. Valjro et al. (1985) found that breast-feeding enhances the clearance of HBsAg in infants with HBVinfections. Because of this difference in results, it was decidedto include breast feeding as a variable, to examine its associationwith infection in this study cohort.Sexual contact, although a possible means of transmission in thisage group, would likely be rare, and therefore was excluded as aquestion item. The sensitivity of this type of question might alsoreduce compliance. Finally, since faecal-oral transmission of HBVhas not been demonstrated, questions related to eating practiceswere not included.3. FORMAT OF THE QUESTIONNAIREAll children enrolled in the study would have to be seen to obtaina blood sample. It was therefore decided that the questionnaire41would be conducted as a personal interview with the parents. Forthis reason, the questionnaire was designed to be easily used bythe interviewer. Exact phrases to be used by the interviewer whenasking parents the questions were included, as well as samplequestions and instructions to clarify how answers should berecorded. A questionnaire guide was also prepared which attemptedto anticipate problems or questions that might arise. (Pleaserefer to APPENDIX 1 for a sample of the questionnaire guide.)To facilitate recording, easy checking for completeness and quickercoding and data entry, adequate space for answers was placed on theright hand side of the page or in tables. Most answers simplyinvolved the circling of "Yes" or "No". A dichotomous scale wassuitable for most questions as the information sought was simplywhether a child had been exposed to the named risk or not.Question reduction was not a large issue as the first draft hadonly 12 questions. More time was spent on ensuring completeness.4. PILOT STUDY OF THE QUESTIONNAIRE4.1. Objectives of the pilot studyA pretest of the questionnaire was conducted with the followingobjectives:1. To identify ambiguous terms.2. To identify problems with question wording and understanding.3. To identify problems with question scaling.4. To identify problems with lay out.5. To identify problems in administering the questionnaire.6. To determine the length of time it would take to complete the42questionnaire.4.2. Results of the questionnaire pilot studyA sample of ten families was chosen to participate in the pilotingof the questionnaire. Of these, 6 agreed to be interviewed, 3could not be reached after 3 attempts and 1 refused.1. The terms 'hepatitis', 'jaundice' and 'breast feeding' werefound ambiguous. The wording was improved and clarificationswere included in the guide.2. The options given for relationship of 'contact' to the childwere not adequate. Therefore, it was decided to provide spacein a table in which the interviewer could list the relationshipof contacts and code numbers could be increased as foundnecessary.3. To make coding less confusing, length of breast feeding and thetime spent overseas would be measured in months and fractionsof months rather than months and weeks.4. It was found that the lay out could be improved with tables forthe questions on household contacts and trips abroad.5. Repetition was reduced by listing blood exposure risks in onequestion.6. There were no problems with refusal to answer or inability toanswer questions during the interviews. However, the followingconcerns were expressed:Will this be confidential?Will my child's name be made public?(The above questions were answered in the actual study by theparent information letter and the consent form.)43Can I give hepatitis to my children or other people?How is the virus spread?Why would my child get hepatitis if he/she had received thevaccine?Why do they use the vaccine if they are not sure it works?Are my children protected against the HBV?Do the eyes get yellow before the skin gets yellow?To answer some of the above questions and others that mayarise, a pamphlet titled "ADVICE FOR HEPATITIS B CARRIERS" wasobtained from the Canadian Liver Foundation and offered to eachparent at the clinics and home interviews.7. The length of time of appointments varied between 30-60 minutesand the actual questionnaire interview took 9 to 20 minuteswith an average of 12.5 minutes.5. RELIABILITY OF THE QUESTIONNAIRESince the subject of the questions represented factual informationand not opinions, problems of timing and subject were not expectedto be serious factors in reducing reliability of the questionnaire.For most of the information a criterion would be available or couldbe obtained; e.g., medical records for immunization,hospitalization and breast feeding history. Thus a true answerexists and "asking the same questions again and again ought toyield consistent results" or show high reliability (Oppenheim1992). Kelly et al. (1990) found in their study of hospitalpatients that reliability was good to excellent for demographicfactors such as birth place and medical information.44Following revisions to the questionnaire, two mothers wereinterviewed again. They gave the same answers to all thequestions, thus indicating consistency of response, at least fortwo cases. (Improvements had been made only in the format andwording prior to re-piloting.)Because of cost and time constraints, reliability studies usingtest-retest, split-half and inter-observer agreement studies werenot carried out. However, to improve questionnaire accuracy, andto reduce variation due to the interviewer, various activities wereplanned as follows:1. During training, interviewers were oriented to thequestionnaire using a questionnaire guide. They were thenexpected to observe an interview and finally to conduct oneunder supervision. Use of the exact wording was stressed.Ongoing feedback was given to ensure clarity and promotecompleteness.2. Parents were requested to bring certain information with them,such as Health Passports.3. If the parent knew that certain information existed but couldnot remember it, he/she was allowed to call back with theinformation. With respect to hepatitis B blood test resultsfor household contacts, parents often requested the study teamto contact the family physician.4. Terms such as 'hepatitis', 'breast feeding' and 'foreigntravel' were defined in the questionnaire and guide.5. Where language barriers might exist, parents were requested tobring along a family member who spoke English or an interpreter45was provided. Chinese was the most frequent language requiringinterpretation. Most of the interpretation was provided by aChinese-speaking interviewer research assistant.6. QUESTIONNAIRE WEAKNESSESRecall bias needs to be considered in questionnaires. Somequestions in this questionnaire may have been open to this bias.In particular, the questions pertaining to household contacts, eg.,year of birth, year of HB vaccine immunization and history ofhepatitis disease may have been most open to bias since parentswere reporting on events which happened to other people. Thenumber of exposures for the child such as hospitalizations, tripsabroad and months of breast feeding might also be biased, but to alesser degree, since the children were still quite young and theseevents, being important events, were more likely to be rememberedby the parents.Another possible weakness pertained to the history of hepatitis Bvaccine immunization in siblings who were born before 1984. Beforethe Hepatitis B Vaccine Program started for neonates, hepatitis Bimmune globulin (HBIG) was given without the vaccine. Althoughinterviewers attempted to clarify the difference, some parents mayhave interpreted the receipt of HBIG as being the hepatitis Bvaccine. The information on household contacts, such asrelationship to the child and history of hepatitis disease wassubsequently not used in the analyses. The total number of'contacts' was used as a surrogate measure of exposure for thechild to other household members with hepatitis B disease.467. VALIDITY OF THE QUESTIONNAIREThe validity of the questionnaire is dependent on face and contentvalidity. No 'gold standard' for measuring risk for HBV infectionis available to establish criterion validity. Therefore, knowntheory about HBV transmission was used in its construction.7.1. Face Validity"Face validity involves the common acceptance by all concerned thata particular measure indeed measures what it purports to measure"(Shortell and Richardson 1978). Five experts in the field ofPublic Health and Communicable Diseases reviewed the questionnaireand agreed it would measure what it was supposed to measure.7.2. Content ValidityFor content validity, the scale must have enough items andadequately cover the domain under investigation (Streiner andNorman 1992). The items in this questionnaire included all knownmeans of HBV transmission that are relevant to children and thatcan reliably be measured.47CHAPTER FIVELABORATORY PROCEDURES FOR THE TESTING OF BLOOD SPECIMENS1. IMMUNOASSAYS USED TO TEST FOR STUDY OUTCOMES1.1. IntroductionAll of the blood specimens collected by study personnel were testedin the Transmissible Diseases Laboratory of the Red Cross Society,Vancouver Centre, Blood Transfusion Service. The assays used todetect markers were from IMx kits produced by Abbott Laboratories:IMx Core for anti-HBc, IMx AUSAB for anti-HBs and IMx HBsAg forHBsAg.Hepatitis IMx assays are based on a microparticle enzymeimmunoassay (EIA) technology. The IMx are fully automated testsrequiring only the initial pipetting of the specimen into a samplewell and the use of one control per run of 23 patient samples (Ebleet al. 1991). Different hepatitis assays are performed in separateruns.1.2. Assessing the characteristics of the IMx immunoassaysFor the purpose of setting retest protocols for the study assayprocedures, the characteristics of each screening test werecalculated.^These characteristics were the sensitivity,specificity and posttest likelihoods. Likelihood ratios were usedfor calculation of the posttest likelihood values. The followingformulas were used:Sensitivity = a/a+c^Specificity = d/b+dWhere a equals true positives; b equals false positives; c equals48false negatives and d equals true negatives.Likelihood ratio(+): LR(+) = sensitivity/(1-specificity)Likelihood ratio(-): LR(-) = (1-sensitivity)/specificityPosttest likelihood of a positive outcome given a positiveresult:PTL+ = Prevalence * LR(+)/(1-Prevalence) + Prevalence * LR(+)Posttest likelihood of a positive outcome given a negativeresult:PTL- = Prevalence * LR(-)/(1-Prevalence) + Prevalence * LR(-)The prevalence values used in the formulas were the rates found inthis study: 5.1% for infection rates (anti-HBc); 87.9% for theseropositivity rate (>=10 mIU/ml anti-HBs); and 2.3% for the HBVcarrier rate (HBsAg). The numbers in the contingency tables arebased on previous research and compare IMx assay results to a "goldstandard".Where information was available, IMx assays were compared to theradioimmunoassay (RIA). For detecting HBsAg, the sensitivity ofthe RIA has been shown to be 99.2% and the specificity, 98.9%(Holland in Gerety ed. 1985). Although considered the mostaccurate test, the RIA is often impractical because of highexpense, complexity and the problem of disposal of radioactivereagents (McCready et al. 1991).2. IMx CORE FOR THE ASSESSMENT OF ANTI-HBc OUTCOMESThe sensitivity of IMx Core in picking up anti-HBc has been foundto be between 0.4 to 0.5 Paul Erhlich I (PEI) units per ml.compared to 0.6 to 0.7 for RIA (Spronk et al. 1991). Spronkreports results of five clinical sites testing 4,841 specimens in49171 47 2396IMxCore1752403parallel with IMx Core and Corzyme, a previously licensed EIAassay. Overall agreement between IMx Core and Corzyme was 99.1%(4794/4841). In the above research, 148 of 162 anti-HBc positivespecimens were correctly identified by the IMx Core giving asensitivity of 91.4%. The specificity of IMx Core could not bedetermined from Spronk's information as results for 3 of the 44discordants were not stated in the report. IMx Core comparedfavourably, however, to Corzyme in that the seven known discordantswere reactive only in IMx Core and these seven specimens werepositive also for other HB markers.In-house data were provided by Abbott Laboratories TechnicalService Division and can be seen in Table 5.1. These datarepresent assay results from 1989 on a mixed population of bloodbank, reference laboratories and hospital serum samples.Table 5.1. Comparison of IMx Core and RIA assays for anti-HBcCorab (RIA)178^2400^2578The sensitivity of the IMx assay according to data in Table 5.1.was 96.1% (95% confidence interval; 93.3%, 98.9%) and thespecificity was 99.8% (95% confidence interval 99.6%, 99.98%).Using the point estimate for sensitivity and specificity and astudy prevalence of 5.1%, the PTL+ was 96.3% and the PTL- was0.21%.50Since the PTL- was very low it was not necessary to retest negativeresults. To improve the specificity and increase the PTL+,positive assay results were repeated if sufficient specimen wasavailable.2.1. Results of Anti-HBc positive retests for this studyIn this study, 769 specimens were tested for anti-HBc. Of these,39 were found positive.^Sixteen of the 39 anti-HBc positivespecimens were positive for HBsAg as well. Of the remaining 23, 14specimens were retested and found positive while 9 were notretested because of insufficient specimen. Two borderline resultsnot included in the 39 anti-HBc positive results were retestedusing a second blood sample and found to be negative.3. IMx AUSAB FOR THE ASSESSMENT OF ANTI-HBs OUTCOMESThe assay, IMx AUSAB, was developed for the detection andquantitation of antibody against hepatitis B surface antibody."Anti-HBs concentrations in specimens are calculated automaticallyby comparison of the specimen rate to values determined from astored standard curve" (Ostrow et al. 1991). IMx AUSAB is able todetect anti-HBs in concentrations as low as 2 mIU/ml which isequivalent to a previously licensed assay AUSAB EIA (a non-automated enzyme immunoassay) and AUSAB RIA, a direct solid phaseradioimmunoassay. Both tests are used routinely to measure anti-HBs concentrations using a reference standard established inDecember 1977 (Barker et al. 1978;Courouce 1990). Ostrow (1991)has found that agreement of IMx AUSAB with AUSAB EIA and AUSAB RIAwas 97.8% (1265/1293) and 99.1% (1281/1293) respectively.51In the research done by Ostrow (1991) 1293 specimens taken randomlyfrom hospital sera and plasma, a sexually transmitted diseaseclinic, a volunteer blood bank and a plasmapheresis centre weretested in parallel by IMx AUSAB, AUSAB RIA and EIA. Sensitivityand specificity characteristics for IMx AUSAB were calculatedusing numbers given in the report by Ostrow and can be seen inTable 5.2. IMx AUSAB is compared to AUSAB RIA results in thistable and 10 mIU/ml is used as the cutoff point for positiveresults.Table 5.2. Comparison of IMx AUSAB and RIA assays for anti-HBsIMxAUSABAUSAB RIA169 12 1121171 112217011231293The sensitivity of the IMx AUSAB according to the numbers in table5.2. was 98.8% (95% confidence interval, 97.2%, 100%) and thespecificity was 99.9% (95% confidence interval, 99.7%, 100%).Using the estimated sensitivity and specificity and the studyprevalence of 87.9%, the PTL+ was 99.98% and PTL- was 8.03%.Ostrow et al. (1991) also report results of sera tested by IMxAUSAB, AUSAB RIA and EIA in 85 individuals previously vaccinatedwith HB vaccine. The vaccines used were Heptavax or Recombivax(n=74) (Merck Sharp and Dohme); Engerix B (n=6) (Smith Kline) orHevac B (n=5) (Pasteur). There was 100% agreement observed between52assays and 81 (95.3%) were reactive for anti-HBs. Although thisstudy indicated 100% sensitivity and specificity, the ages of theindividuals tested and duration from immunization were not stated.Retesting of positive results was not recommended for anti-HBs onthe basis of the high PTL+ for the IMx AUSAB assay, especially inimmunized individuals. However, it was laboratory policy to retestlow level reactive results (<6 mIU/ml) for reliability reasonsrelated to concerns about picking up false positive results. Sincethe 8% PTL- is a false negative concern, the ideal would have beento retest all samples with results below 10 mIU/ml since this wasthe cutoff point chosen for this study between positive andnegative results. The PTL- of 8% suggests that 8% of study resultsbelow 10 mIU/ml, or seven cases, were actually positive, andtherefore the seroprotection rate of 87.9% for the current studywas an underestimate. The 100% positive and negative predictivevalues for the immunized population described above would tend tosupport the 8% PTL- as being high. However, if the assumption weremade that there were in truth seven more positive cases, then theseropositivity rate for this study would have been 88.9% ratherthan the 87.9% found.3.1. Results of Anti-HBs negative retests for this studyOf the 756 children tested for anti-HBs, 64 (8.5%) were found tohave 0 or <6 mIU/ml. There was sufficient sample to retest 33 ofthese specimens.^Of the 33, the result sequence of 25 wasnegative/negative; four were positive/positive and four specimenswere discordant between the first and second run. One of two53specimens which were initially positive was positive on the thirdtest and one was negative. One specimen initially negative wasnegative on the third test. The fourth discordant sample could notbe resolved as there was insufficient specimen for a third test.The initial result for that specimen was 3.3 mIU/ml and the secondwas 0 mIU/ml. For this case, the second test result was the oneused. Of all the repeat assays on negative or low level results,none were 10 mIU/ml or greater on retest. Thus, there were nofalse negatives found by study samples being retested with the sametest.4. IMx HBsAG FOR THE ASSESSMENT OF HBsAG OUTCOMESThe IMx HBsAg assay has been shown to be able to detectconcentrations of HBsAg less than 0.5 ng/ml or 0.2 PEI units/ml,which is equivalent to the AUSZYME Monoclonal assay, a previouslylicensed standard method (Ebel et al. 1991). A total of 9,700specimens were assayed in clinical studies reviewed by Eble.Specimens tested included samples from blood donors, hospital anda public health patients, obstetric and gynaecology patients,dialysis patients, acute and chronic HBV patients, and patientswith diseases other than active hepatitis B. The overall agreementbetween IMx HBsAg and the licensed reference method was 99.77% inthis research. The results for which numbers were given in thereport by Eble (1991) can be seen in Table 5.3. The referencestandard, AUSZYME, was a previously licensed method with a reportedsensitivity of 97.8% and specificity of 97.9% when compared to RIA(McCready et al. 1991).54331 154 8467335^8482+IMxHBsAg -34684718817Table 5.3. IMx HBsAg results compared to AUSZYMEAUSZYME+The sensitivity calculated from Table 5.3. was 98.8% (95%confidence interval, 97.63%, 99.97%) and the specificity was 99.8%(95% confidence interval, 99.7%, 99.9%). Using these sensitivityand specificity estimates and a study prevalence of 2.3%, the PTL+was 92.1% and PTL- was .03%. Since the likelihood of a negativeHBsAg test result being false was so low this study's protocolrecommended a retest of only the positive results. Also, it wasfelt that the possible consequences of misdiagnosing a child as acarrier could be serious enough to justify confirmation.4.1. Results of HBsAg positive retests for this studyOf the 16 study specimens that were HBsAg positive, nine wereconfirmed positive by retest. There was insufficient specimen forretesting seven others. Three of the children for whom there wasinsufficient sample were known to be HBsAg positive at one year ofage.Five children who were positive for HBsAg were also positive forthe anti-HBs marker. Two of these children were HBsAg positive by12 months of age. Two of the remaining three children remainedpositive on retest for anti-HBs and one was found to be negative.555. TECHNICAL PROBLEMS ENCOUNTERED IN TESTING THE BLOOD SAMPLESThere were 76 specimens for which technical problems preventedobtaining test results: one anti-HBc; 14 anti-HBs; and 67 HBsAg.Technical problems were those due to inability to obtain asufficient volume of blood sample or failure of test runs.Occasionally a child or parent was unwilling for a second attemptto obtain a sample if the first prick clotted before the requiredamount was obtained. Test run failures were due to bubbles orgranules in the sample, machine failure or power surge. Haemolysisaffected four samples, only one of which was severe enough toprevent doing the third test.56CHAPTER SIXRESULTSI. STATISTICAL ANALYSISThe analysis was begun by producing descriptive statistics on allof the variables. Univariate chi-square analyses of all variables,separately, against the outcome variables were then performed. ThePearson chi-square statistic was used for 2 by 2 tables and fornot-ordered categorical variables. The Mantel-Haenszel trendstatistic was used for ordered categorical variables.Prediction models for the infection (anti-HBc) and carrier state(HBsAg) outcomes were developed using logistic regressionprocedures. Using logistic regression was necessary since theseoutcomes were dichotomous. The stepwise method, including forwardselection and backward elimination, was used to handle theconfounding between variables. The criterion for entry of avariable into the model was a significance level of 0.05 using thescore statistic. The likelihood-ratio test was used for removingterms from the model using a criteria of 0.10 significance level.Interaction terms were assessed in the final model.Linear regression procedures were used for the analysis of theanti-HBs outcome, since this variable was continuous. Interactionterms were assessed for the final model.To look for evidence that necessary assumptions were violated, bothstudentized residuals and standardized delta betas were calculated57for the final models. These variables were then plotted. A normalprobability plot of studentized residuals and a histogram of thestandardized residuals were produced for the anti-HBs outcome aswell.To avoid the effect of influential observations, some variableswere grouped and, where appropriate, included in the modellingprocedures as continuous variables.Statistical significance, unless otherwise qualified, refers to aP value of <= 0.05.2. BASELINE DISTRIBUTIONS AND COMPARISON OF PARTICIPANTS ANDNONPARTICIPANTSDescriptive statistics for some of the baseline demographicvariables are discussed in the following in order to compareparticipants with nonparticipants. A summary of these statisticsmay be found in Table 6.4.Child's sexOf the 1166 children who met the inclusion criteria, 44.8% weremale and 44.3% were female. The sex of 11% was unknown and thesewere all nonparticipants. Of 770 participants, 390 (50.6%) weremale, and 380 (49.4%) were female. Of the nonparticipants withknown sex group, 132 (49.3%) were male and 136 (50.7%) were female.The sex composition of the two groups was thus very similar and wasnot statistically different (P=0.69).58Child's ageFor the children who participated, ages were grouped in two ways,age at last birthday and year of birth. The distributions were asillustrated in Table 6.1. and 6.2.Table 6.1. Child's age at last birthday(on day blood sample taken)Age No. Percent2 75 9.73 163 21.24 147 19.15 133 17.36 125 16.27 114 14.88 13 1.7Total 770 100.0Table 6.2. Child's year of birthYear No. Percent1984 58 7.51985 121 15.71986 132 17.11987 144 18.71988 160 20.81989 155 20.1Total 770 100.0Because year of birth was highly correlated with age when the bloodsample was obtained, no attempt was made to separate the two.Results are generally presented for age at last birthday.The mean age for the total group, participants and nonparticipants,was 5.1 years with a range of 2.5 to 8.6 years. The mean ages andranges for participants and non-participants were 5.1 (2.5-8.6) and5.2 (2.6-8.3) respectively. The Mantel-Haenszel trend test fordifferences in ages between participants and non-participants59indicated that the groups were similar for this variable (P=0.46).Mother's ageMothers' ages were also comparable between participants andnonparticipants with an overall mean age of 29.8 years and a rangeof 15.7 to 42.1 years. The mean ages and ranges for participantsand nonparticipants were 29.9(15.7-42.1) and 29.7(15.8-39.8)respectively.^The Mantel-Haenszel trend test P value fordifferences^in mother's ages between participants andnonparticipants was 0.56.Mother's HBeAg statusResults of the mothers' HBeAg status were recorded on Redrecords for 883 eligible children.^Of the participants, 588mothers were tested for HBeAg.^The distribution of resultsparticipants and nonparticipants was as follows in TableTable 6.3. Mother's HBeAg statusCross(76%)between6.3.positive^negativeParticipants 182(31%) 406(69%) 588Non- 85(29%) 210(71%) 295Participants267^616^883Table entries are counts with row percentages.The difference between participants and non-participants was notsignificant (P=0.51; odds ratio, 1.11; exact 95% confidenceinterval, 0.81, 1.53).603. VARIABLES COLLECTED TO ASSESS FOR BIAS IN THE SAMPLELanguageInformation on which language was spoken in the home was asked ofthose people who participated as well as those people who refusedto participate. The languages (grouped by horizontal or verticaltransmission countries) represented by those who refused was thencompared with those language groups of the participants. Thedifference between the two groups was not significant (P=0.33).Moving residences after the birth of the child (Moving)Sixty-three percent of participants were found to have moved. Ofthose that did not participate, 66 percent were known to have movedor could not be located at their last known address.^Forparticipants who had moved, the infection rate was 0.06 while forthose who had not moved it was 0.04. This difference was notstatistically significant (P=0.26).Number of phone calls made contacting participants (Phone calls)The average number of phone calls made per participating child wasthree. The median was also three. The minimum was one and themaximum was 17. There was no significant difference in theattempts to contact those children infected as opposed to childrennot infected (P=0.69). The association was also absent for carrierchildren (P=0.33).61Table 6.4. Baseline characteristics^Variable Total^Partic's Non Partic's Partic's/group Non Partic's^Child sex M 44.8^M 50.6^M 33.3F 44.3^F 49.4^F 34.3Unknown 11.0^Unknown 32.3P value 0.69 1RR 1.01CI 0.94,1.09Child age mean^mean^mean^P value5.13 5.10 5.18 0.462range^range^range2.45-8.62^2.45-8.62 2.58-8.30Mother's mean^mean^mean^P valueage^29.81 29.88 29.67 0.562range^range^range15.67-^15.67-^15.82-42.06 42.06 39.78Mother's^ P valueHBeAg status 0.51 1RR 1.03CI 0.94, 1.14Anti-HBc^HBsAgPos/Neg Pos/NegPhone P value^P valuecalls^0.69 2^0.332Moved 0.26 1^0.341Language Partic's/Refusal P value0.33 1RR 1.33CI 0.74, 2.381 Uncorrected Pearson's chi-square test2 Mantel-Haenszel test for trend623^4^5^6^7^8CHILD'S AGE GROUP4. MAIN OUTCOME VARIABLES - DESCRIPTIVE INFORMATION4.1. ANTI-HBcOf 769 participants for whom anti-HBc results were available, 39(5.1%; 95% confidence interval, 3.5, 6.7) tested positive. For onechild who participated, results for the anti-HBc antibody markercould not be obtained. Infection rates for each age group areillustrated in Figure 6.1. and by year of birth in Figure 6.2.Figure 6.1INFECTION AND CARRIER RATESBY AGE0.24 -pL40• 02-• 0.18-00 0.16-0.14-F4 0.12-0J-OX8-GG0IN 0.08-0r4 0.04-0.02-2Legend^ Anti-HBc positiveHBsAg positive630.1 -0.090.08-r, 0.07 -m 0.06 -g0 0.05 -w8 O.04-0.03-1100 0.02-10.010^1984Figure 6.2INFECTION AND CARRIER RATESBY YEAR OF BIRTHLegendAnd-HBo positiveHBsAg positiveNNNNN1985^1986^1987^1988^1989CHILD'S YEAR OF BIRTH4.2. ANTI-HBsOf the 717 children for whom results of anti-HBs testing wereobtained and who were also anti-HBc negative, 87.9% (630) (95%confidence interval, 85.5, 90.2) showed evidence of seropositivity(>=10 mIU/ml).^Fifty-five (7.7%) of the 717 children had 0antibody titres. By age group cohort, seropositivity varied froma high of 99.4% for age three to a low of 70.4% for age seven.Figure 6.3. illustrates the seropositivity rates by age group andFigure 6.4. shows the distribution of anti-HBs titres for all ages.64Figure 6.3SEROPOSITIVITY RATES BY AGE GROUPws,98928884-80LegendANTI-HBet>-10m1U/m17672B8 1^i^12^3 4 5I81718CHILD'S AGE GROUPTable 6.5. also illustrates the distribution of anti-HBs titres butin larger groups.Table 6.5. Anti-HBs titres in mIU/mlNumber Percent<10 87 12.110-49 144 20.150-99 106 14.8100-499 264 36.8500-999 58 8.1>=1000 58 8.1w0gwa65Figure 6.4AIM - I-Uls LEVELSEXCLUDING ANTI-HBc POSITIVE CASESThe mean level of anti-HBs was 242 mIU/ml with a standard deviationof 303 mIU/ml. The minimum and maximum titres were zero and 1000mIU/ml. The geometric mean was 79.7 mIU/ml with a standarddeviation of 7.2 mIU/ml. The difference in geometric mean titresbetween age group cohorts is illustrated in Figure 6.5. A declinein titres between age groups, until age eight, occurs as ageincreases.66Figure 6 . 5GEOMETRIC MEAN LEVELS BOTH SEXES400350• 300• 2501-1 2001-1 150• 11-1 00500WITH CONFIDENCE BOUNDSLegendgeametdc meanlower cod. bound  upper cant. bound- - - -\\\\\\\\\'--2I3I4I^I5 6I^I7 8CHILD'S AGE GROUPThe difference is less than 40 mIU/ml (12%) between age groups twoand three followed by a sharp drop between the three and four yearold groups, from a geometric mean of 238 to 79 mIU/ml (67%). Thedifference is small again between age groups four to seven when thegeometric mean drops from a level of 79 to 23 mIU/ml (71%),however, a sharp rise occurs again in the age eight group. Thisincrease in titres for eight year old children from a geometricmean of 23 to 106 mIU/ml is partly due to additional doses and onelate third dose (1991). However, as can be seen in Figure 6.6.,673.-4 60--..H 30000-,-1 260tnart.' 200a150tr)MILI 100500there was still an increase in anti-HBs titres at age eight afterexcluding these cases. Males and females follow a similar patternin titres except for age two and age eight groups when males havehigher titres by approximately 100 mIU/ml. (Please see Figure 6.7.for geometric means by sex.)Figuie 6.6GEOMETRIC MEAN LEVELS BOTH SEXES>3 DOSES and LATE 3RD DOSE CASES EXCLUDEDLegend- -geometric mean- - lower cont. bound ^upper cont. bound----________\\\\\1 1 1^1 1^i2^3^4^5 6CHILD'S AGE GROUP4.3. HBsAGOf the 770 children seen, sufficient blood sample was obtained totest 703 (91.3%) children for the HBV surface antigen. It wasfound that 16 (2.3%; 95% confidence interval 1.2, 3.4) of thosetested were HBV surface antigen positive and considered carriers,68while 687 (97.7%) were negative and not considered to be carriers.All of the children who tested positive for HBsAg were also anti-HBc positive.LegendMALES---- FEMALES^\\\\\\\2 13 I4 I5 I^I^15 7 8CHILD'S AGE GROUP5. COMPARISON OF STUDY OUTCOMES WITH OUTCOMES AT 6 TO 18 MONTHSOF AGEChildren born between 1984 and 1987 were recalled by the Programfor HBV blood marker testing at around 12 months of age. Thosetested between 6 to 18 months of age were included in the analysis.The results had been recorded on Red Cross records as follows inTable 6.6.Figure 6.7Wm GEOMETRIC MEANS - ANTI - HBS BY SEX-1 300.....01-1 250zHco 200150coW 100HH50069Table 6.6. Results of 6 to 18 month testsNo. tested^Percent positiveANTI-HBc^169 33.7ANTI-HBs 183^85.8 (>=10 S/N's)HBsAg 196 3.0The comparisons between results of cases tested between 6 to 18months and study results for this group of children are illustratedin the two by two tables in Table 6.7. below.Table 6.7. Comparisons between 6 to 18 month resultsand study resultsAnti-HBc (study)+^-Anti -HBc^+ 7 51 586 -18mos 2 109 1119^160 169HBsAg (study)+^-HBsAg^+ 4 1 56-18 mos 1 168 1695^169 174Anti-HBs (study)+ -Anti -HBs +6-18 mos130 2010 1315023140 33 173Associations between each of the study outcomes against the 6 to 18month outcomes can be seen in Table 6.8. and are described in thefollowing.70Study anti-HBc results were significantly positively associatedwith anti-HBc results at 6-18 months (P=0.008). (Of those childrenanti-HBc positive in infancy a large proportion (88%) were nolonger positive at study testing, suggesting that some children mayhave been circulating maternal anti-HBc.) There was also a highlysignificant association between study anti-HBc results and HBsAgresults at 6 to 18 months (P<0.0001). The association betweenstudy anti-HBc and 6 to 18 month anti-HBs results was notsignificant (P=0.12).Study HBsAg results were positively associated with both anti-HBc(P=0.02) and HBsAg results (P<0.0001) at 6 to 18 months andnegatively associated with anti-HBs titres (P=0.008).Study anti-HBs results were strongly associated with infancy anti-HBs results (P<0.0001). However, there was no association betweenstudy anti-HBs results and anti-HBc (P=0.25) or HBsAg (P=0.36)results at 6 to 18 months.Of the 107 children who were known to be both anti-HBc negative andanti-HBs positive at 6 to 18 months, two were anti-HBc positive atstudy testing.71Table 6.8. Association of study outcomeswith 6-18 month outcomesVariable P value RR & CI Directionof effect1. Study Anti-HBc outcomes (N=169)Anti-HBc status at 6-18 months (pos/neg)0.0081 7.531.34,^75positive at6-18 monthsAnti-HBs status at 6-18 months (pos/neg)0.121 0.3030.06,^2.03negative at6-18 monthsHBsAg status at 6-18months (pos/neg)<0.00011 10538.22,^5230positive at6-18 months2. Study HBsAg outcomes (N=151)Anti-HBc status at 6-18 months (pos/neg)0.021 O.R.^infinite 31.19,^infinitepositive at6-18 monthsAnti-HBs status at 6-18 months (pos/neg)0.0081 0.0530.00,^0.62negative at6-18 monthsHBsAg status at 6-18months (pos/neg)<0.00011 672324,^22000positive at6-18 months3. Study Anti-HBs outcomes (N=159)Study anti-HBc positive cases excludedAnti-HBc status at 6-18 months (pos/neg)0^.25 2 highertitres ifnegative at6-18 monthsAnti-HBs status at 6-18 months (pos/neg)<0.00012 highertitres ifpositive at6-18 monthsHBsAg status at 6-18months (pos/neg)0.3621 Fisher's exact test2 Mantel-Haenszel test for trend3 Odds ratio and exact 95% confidence interval6. ANTI-HBc OUTCOME - RESULTS OF ANALYSES6.1. Association of covariates with Anti-HBcIn this section the associations of each of the variables with the72anti-HBc outcomes will be described.^A summary of theseassociations with anti-HBc outcomes can be seen in Table Baseline characteristics - associations with Anti-HBcNo significant associations were found between the anti-HBc outcomeand the child's sex (P=0.56), age (P=0.15) or year of birth(P=0.12).^However, there was a highly significant associationbetween occurrence of infection and the mother's age (P=0.0009) andthe mother's HBeAg status (P<0.0001; relative risk, 7.97; 95percent confidence interval, 3.51 to 18.08).6.1.2. Demographic variables - descriptive information andassociations with Anti-HBcMother's country of birthGrouped according to countries with primarily vertical orhorizontal transmission, the number and percent of mothers born ineach country was as follows in Table 6.9.Table 6.9. Distribution of mothers' country of birthVerticaltransmission No. PercentHorizontaltransmission No. PercentChina 212 27.5 Canada^74 9.6Hong Kong 140 18.2 India, Pakistan, Mid.E.Viet Nam 119 15.5 E. and N. Africa 21 2.7Philippines 65 8.4 Eastern Europe 20 2.6S.E. Asia 49 6.4 Japan 10 1.3Taiwan 15 1.9 Western Europe 8 1.0N. & S. Korea 15 1.9 Sub Sahara Africa 7 .9Macaw 6 .8 S. & Central America 6 .8U.S.A.^Aust.^N.Z. 3 .4Total 621 80.6 Total 149 19.473I Table 6.10. Association of covariates with Anti-HBcVariable P value RR & CI Directionof effectChild's Sex 0.561 1.20.65,2.22femalesChild year of age (2-8) 0.152 older agesChild birth year 0.122 earlierbirth yearsMother's age(quartiles)0.00092 youngermothersMother's birth country(vertical/horizontaltransmission)0.061 2.88.90,9.24verticalMother's years in birthcountry (quartiles)0.092 longer timeFather's birth country(vertical/horizontaltransmission)0.141 1.88.80,4.41verticalFather's years in birthcountry (quartiles)0.082 longer timeLanguage (by countries ofvertical/horizontaltransmission)0.031 2.361.06,5.28verticalMother's HBeAg status(positive/negative)<0.00011 7.973.5,18.08positiveNumber of doses HB vaccine 0.052 fewer dosesDose 1 (age received) 0.00052 >7 daysDose 2 timing (within 2months of dose 1, no/yes)0.523 1.444.03,10.0not within2 mosDose 3 (no/yes) 0.183 2.264.42,7.90no 3rd doseDose 4 (no/yes) 1.003 infinite40.10,infiniteDose 5 (no/yes) 1.003 infinite 40.01,infiniteHBIG (not given/given)(given/unknown/not given)0.0830.00125.224.52,^27not givenandunknown74Moved since child's birth(yes/no)0.261 1.47.74,2.91yes movedJaundice after first month(yes/no)1.003 040,46.06Breast feeding (yes/no) 0.331 0.74.40,1.36no breastfeedingDaycare (yes/no) 0.091 0.38.12,1.22no daycareEarpiercing (yes/no) 0.331 1.43.70,2.95earspiercedAcupuncture (yes/no) 1.003 0 40,^28.95Surgery (yes/no) 0.401 0.67.27,1.70no surgeryBlood transfusion (yes/no) 1.003 0 40,^20.90No. of household contacts(grouped 1,2,3,4,5,6,7+)0.0150.042># contacts7-10 cont'sHospitalization (yes/no) 0.291 0.58.21,1.61no hospi-talizationTravel (yes/no) 0.521 0.78.37,1.67no travelTotal months of travel 0.902Months abroad 0.532 no timeabroadMinimum age at travel 0.562 younger age1 Pearson's chi-square2 Mantel-Haenszel test for trend3^•Fisher's exact4 Odds ratio and exact confidence interval5 Exact test for trendThe difference between the two groups of countries in terms ofinfection outcomes just missed significance (P=0.06; relative risk,2.88; 95 percent confidence interval, 0.90, 9.24). The trendindicated that children of mothers born in primarily verticaltransmission countries were more likely to be infected.75Father's country of birthThe distribution for father's country of birth was similar to thatfor mothers with most fathers born in countries more commonlyassociated with vertical transmission of HBV (75%). There was nosignificant association between father's country of birth and anti-HBc outcomes (P=0.14).LanguageSixty-six percent of the participants spoke languages of countrieswhere vertical transmission of HBV is most common. Cantonese (43%)and English (29%) were the most common languages spoken. The chi-square test of association showed a significant relationshipbetween the language spoken and a positive outcome for anti-HBc(P=0.03). Children who spoke languages of countries where methodof transmission is more likely to be vertical were more likely tobecome infected.Mother's years in birth countryThe minimum number of years that the mother spent in her country ofbirth (other than Canada) was one year and the maximum was 40years. The average was 21.6 years with a standard deviation of 7.5years. When grouped in quartiles the mode was between 21 to 25years. There was a trend toward more infections with increasingnumber of years spent in the mother's country of birth but it didnot reach statistical significance (P=0.09).76Father's years in birth countryThe number of years fathers spent in their country of birth (otherthan Canada) averaged 19.2 with a range of 1 to 44 years. The modewas in the age group 21 to 25 years. As with the mother, there wasa trend towards increasing number of HBV infections with increasingnumber of years fathers spent in their country of birth but thistendency was not statistically significant (P=0.08).6.1.3. Program intervention variables - descriptive information andassociations with Anti-HBcDoses of hepatitis B vaccine (Number of Doses)The number of doses of vaccine received by the participatingchildren ranged from one to six with a distribution as illustratedin Table 6.11.Table 6.11. Number of doses of HB vaccine receivedNo. of Doses No. Children Percent1 9 1.22 20 2.63 732 95.14 7 .95 1 .16 1 .1There was an association between number of doses of hepatitis Bvaccine and anti-HBc outcome (P=0.05). Children with fewer doseswere more likely to have become infected.Age of child at first dose (Dose one)Most children received 'Dose one' on their first (452, 58.7%) or77second (268, 34.8%) day of life. Ninety-eight percent had receivedDose one by the end of their first 7 days. A few children had Doseone extremely late and to avoid dependence of statistics on thesefew (influential) observations, this variable was grouped asindicated in Table 6.12.Table 6.12. Child's age at Dose oneAge in days^No. children^1-3 743^4-7^128-61 962-307 6There was strong evidence to show that the age of the child when'Dose one' was administered is associated with the risk ofinfection (Mantel-Haenszel trend test, P=0.0005), with childrenhaving received Dose one after seven days of age more likely tohave become infected.HBIGOf the 770 children seen, 745 (97%) had documented evidence ofhaving received HBIG. Ten children had not received the injectionand for 15 it was unknown if HBIG had been administered. When theunknown cases were removed and this variable was classified as'given/not given', there was a tendency toward more infections inthose who did not receive the HBIG but it was not statisticallysignificant (P=0.08; odds ratio 5.22; exact 95 percent confidenceinterval, 0.52 to 27.47). However, when HBIG was classified as'given/unknown/not given', there was a statistically significanttrend test result for anti-HBc outcomes (P=0.001) in the same78direction.Timing of Dose two of hepatitis B vaccine (Dose two timing)Seven hundred forty two (96.4%) children received their second doseof hepatitis B vaccine within two months of Dose one. Fifteen(1.9%) received Dose two after this time. The date was unknown forfour of the remaining, while nine children did not receive a seconddose.^No association was found between Dose two timing andoccurrence of infection (P=0.52).Dose three of hepatitis B vaccine (Dose three)Seven hundred forty one (96.2%) children received three doses ofhepatitis B vaccine and 29 (3.8%) did not. The association betweenobtaining a third dose of vaccine and infection was not significant(P=0.18).Doses four to six of hepatitis B vaccineThere were only nine children who were given four or more doses ofvaccine and none of these children became anti-HBc positive, thusthe power was inadequate to determine whether there was an effector not.6.1.4. Association of anti-HBc with illness and exposure variablesHistory of illnessNo children were reported by the parents to have had any illnessesdue to hepatitis or to have had hospitalizations related tohepatitis or reactions to hepatitis B vaccine.^Deaths of twochildren who were born during the study period years were due to79causes other than those related to hepatitis.JaundiceAccording to the parents, three (0.4%) children had a history ofjaundice starting after the first month of life. There were noassociations found between this variable and any of the dependentvariables. None of the 39 infected children had a history ofjaundice reported.History of breast feeding (Breast feeding)Of the children seen, 395 (51.3%) were breast fed. No significantassociation was found between breast feeding and infection(P=0.33).History of ear piercing (Ear piercing)Of 770 children seen 133 (17.3%) were reported to have had theirears pierced. No association was found between this variable andthe occurrence of infection (P=0.33).AcupunctureOnly 4 (0.5%) of the children seen had experienced acupuncture andnone of these had become infected. Thus no associations were foundbetween acupuncture and outcome variables.SurgeryOf the 770 children seen, 138 (17.9%) had had surgery orexperienced suturing. There was no significant association foundwith infection (P=0.40).80Blood transfusion and blood products (Blood transfusions)Only 5 children in the study were reported to be recipients ofblood transfusions. One of these children had also received otherblood products. These children were not among the children who hadbecome infected and therefore no associations were found with theanti-HBc or HBsAg outcomes.HospitalizationOf the 126 (16.4%) children in the study who were hospitalized atleast once, 101 (80%) were admitted once and the remainder betweentwo to ten times. There was no significant association found withthe anti-HBc outcome (P=0.29).DaycareThe number of children in the study who had attended daycare was138 (17.9), while 632 (82.1%) were reported to have had no daycareexperience. The association between daycare and the anti-HBcoutcome (P=0.09) was not statistically significant; however, therewas a tendency for daycare attendance to be protective.Number of household contacts (Contacts)All children were reported to have had at least one householdcontact. The number of household contacts per child is illustratedin table 6.13.81Table 6.13. Distribution of number of household contactsNumbercontactsNumberchildrenNumbercontactsNumberchildren1 58 6 532 258 7 213 191 8 134 110 9 25 61 10+ 3Statistically significant association was found between the child'snumber of household contacts and the anti-HBc outcome (Mantel-Haenszel exact trend test P=0.01). The effect was for moreinfections when larger numbers of contacts had been present in thehousehold.Travel abroad (Travel)Of the 191 (25%) children reported to have travelled overseas, 73%had spent less than 3 months abroad and 27% more than 3 monthsabroad. The minimum time spent abroad was 2 days and the maximumlength of time was 32 months. The mode was one month. Eightchildren who had travelled overseas tested anti-HBc positive.However, no significant association was found between travellingabroad and HBV infection (P=0.52).Minimum age for travelForty-four percent of the children who had travelled were under twoyears of age when first taken abroad. A further 21 percent weretwo years of age and 35 percent were three years or older on theirfirst trip abroad. No significant association was found betweenearliest age of travel and the anti-HBc outcome (P=0.56).82Months abroadThe association between the amount of time spent abroad and theoccurrence of infection was not significant (P=0.53).6.2. Multiple logistic regression analysis for the Anti-HBc outcome6.2.1. Modelling procedure introductionIn the foregoing section, the associations of the study variableswith anti-HBc were described as analyzed singly. In this sectiongroups of variables are analyzed together for associations with theanti-HBc outcomes using multiple logistic regression procedures.As part of the logistic regression procedures, prediction modelsare built for the anti-HBc outcome. One component of the model isthe group of significantly associated variables which best explainsthe variation in the anti-HBc outcomes. The other component is thecoefficient. The variables in the model are chosen using a forwardstepwise process, and what was found to occur in this stepwiseprocess will be described. The modelling procedure, which involvessimultaneous adjustment of all included variables, producescoefficients for each variable selected to be in the model.Finally, the variables and coefficients are combined in thelogistic regression formula which is used to predict theprobability of infection.For ease of reading, variable names will be placed in quotationmarks and in some cases will be abbreviated as indicated above inthe results of associations with anti-HBc.Two models were built for the anti-HBc outcome. Because of 19083missing values for the variable 'Mother's HBeAg status', the firstmodel was built without including 'Mother's HBeAg status' thusretaining as much information as possible. 'Mother's HBeAg status'was included in the second modelling procedure. The variablesexcluded from both procedures because of negligible associationswere the following: 'Dose two timing', 'Dose three, 'Dose four','Dose five', 'Moving', 'Jaundice', 'Acupuncture', 'Bloodtransfusions', 'Months abroad', 'Minimum age of travel', 'Father'scountry of birth' and 'Father's years in birth country'.6.2.2. Model one Anti-HBc (without 'Mother's HBeAg status')The variables considered in the first model were: 'Child's age','Child's sex', 'Mother's age', 'Mother's country of birth','Mother's years in birth country', 'Language', 'Number of Doses' ofHB vaccine, age of the child when 'Dose one' was received, receiptof 'HBIG', 'Contacts', histories of 'Moving', 'Breast feeding','Ear piercing', 'Surgery', 'Daycare', 'Hospitalization' and'Travel'. The final variables selected, by the forward stepwiseprocess, to remain in the first model were as follows: 'Dose one'(P=0.01); 'Mother's age' (P=0.0001); and 'Mother's years in birthcountry' (P=0.003).The stepwise process revealed that 'Dose one' was highly confoundedwith 'HBIG' and also explained some of the variation due to 'Numberof doses' while being negatively confounded with 'Mother's age'.'Mother's age' was positively confounded with 'Number of doses',but negatively confounded with 'Mother's country of birth' and'Mother's years in birth country'. 'Mother's years in birth84country', being more highly statistically significant, acted as asurrogate for 'Mother's country of birth' and 'Language' as well asexplaining some variation due to 'Daycare'. 'Daycare' was the onlyremaining variable with a significance level less than 0.10 butentry into the model was denied because the criterion for entry was0.05. 'Child's age' reached a significance level of only Model two Anti-HBc (with 'Mother's HBeAg status')When 'Mother's HBeAg status' was included in the procedure formodel two, positive confounding was apparent between 'Mother'sHBeAg status' and 'Mother's age', 'Language', 'Number of contacts'and 'Mother's country of birth'. 'Child's age' was negativelyconfounded with the 'Mother's HBeAg status'. This resulted in ascore test P value of 0.04 for 'Child's age' after 'Mother's HBeAgstatus' was the first variable to be chosen for the model. Butstatistical significance was lost for 'Child's age', after 'Doseone', 'Mother's years in birth country' and 'Mother's age' wereselected for entry into the model. There remained a tendency for'Daycare' (P=0.09) to be associated with anti-HBc outcomes. Arelative risk of 0.27 (95 percent confidence interval, 0.05 to1.35) indicated that this trend toward association with HBVinfection was protective. Thus, the final model with statisticallysignificant variables which best explained the variation in theanti-HBc outcomes were 'Mother's HBeAg status' (P<0.0001), 'Doseone' (P=0.0001), 'Mother's years in birth country' (P=0.001) and'Mother's age' (P=0.002). The final significance level reached by'Child's age' was only 0.24 (score test).85R2 , the proportion of variation in outcome accounted for in thefinal model was (247.63-188.37)/ 247.63= 0.24. The interactionbetween 'Mother's HBeAg status' and each of the other variables inthe model was assessed. None of the interaction terms reachedstatistical significance. A summary of statistics for the anti-HBcmodels may be seen in Table 6.14.Table 6.14. Prediction models for the Anti-HBc outcomein children age 2 to 8P value Coef.N=7580.01 1 0.750.0 001 1 -0.170.003 1 0.56-0.53N=580<0.0001 1 2.370.0001 1 0.340.001 1 0.750.002 1 -0.15-2.37Variables in model 1Age at Dose oneMother's ageMother's years inbirth countryConstantVariables in model 2Mother's HBeAg statusAge at Dose one squaredMother's years inbirth countryMother's ageConstantRR^C I^2. ^1.26, 3.53^0.85^0.78, 0.921.75^1.19, 2.5710.70^3.98, 28.781.40^1.20,^1.622.13^1.29, 3.470.86^0.77, 0.95Interaction terms (not in the model)Mother HBeAg*Dose one sq.Mother HBeAg*Mother's ageMother HBeAg*Mother's yearsin birth country0.89 20.68 20.25 21 -2 Log likelihood Ratio2 Score testBecause 'Dose one' was so highly confounded with HBIG in theforegoing procedures, it was decided to re-run the procedures formodel one and model two, forcing HBIG into the model. In modelone, once 'HBIG' was forced in, 'Dose one' was not significant86(P=0.15, score test). The P value for 'HBIG' when in the model was0.02, (relative risk, 2.95; 95 percent confidence interval, 1.35,6.48.). Forcing 'HBIG' into the model when 'Mother's HBeAg status'was also in the model did not change the final model. 'Dose one'remained significant (P=0.04, score test) and when included,rendered 'HBIG' non-significant.6.2.4. Prediction formula for HBV infection (positive Anti-HBc)The prediction of a positive anti-HBc outcome in this study samplemay be modelled as follows:Log odds Anti-HBc positive=(-2.37 + 2.37 * MOM HBeAg + 0.34 * DOSE ONE 2 +0.75 * MOM YEARS IN BIRTH COUNTRY - 0.15 * MOMS AGE)If, for example, a child had a mother who was HBeAg positive (1) [0if the mother was negative], the child's age at time of Dose onewas two days (category 1, squared), the child's mother spent 5years in her country of birth (category 1) and the mother's age was25 years, the odds of this child being infected would be:log odds anti-HBc positive=(-2.37 + 2.37 * 1 + 0.34 * 1 + 0.75 * 1 - 0.15 * 25)odds anti-HBc positive= exp (-2.66) = 0.07If the mother were HBeAg negative the odds of infection for thechild would be Anti-HBc outcome diagnostic proceduresResidualsMore large positive studentized residuals were found in the finalprediction model for anti-HBc outcomes than would be expected by87chance. However, only one was outside the 3 standard deviationrange. For the variable 'Mother's age' the studentized residualswere spread evenly and with no particular pattern. The largeresiduals were also evenly dispersed between the four quartiles fornumber of years in mother's country of birth. However, for 'Doseone', the nine large residuals fell in the first category which wasone to three days. It appears, then, that the risk of infection isgreater for those children who have their first dose of HB vaccinewithin three days of birth than the model would predict. Thesechildren were evenly spread between positive and negative statusfor 'Mother's HBeAg'. (The plot of residuals against 'Dose one'may be seen in APPENDIX 3, page 161.)In an attempt to improve the fit of the model, the variable 'Doseone' was squared such that category one was 1 2 , category two was 2 2and so forth. The model was re-run with both 'Dose one' and 'Doseone squared' in the model as well as the other final modelvariables which were 'Mother's HBeAg status', 'Mother's age' and'Mother years in birth country'. With 'Dose one squared' in themodel, 'Dose one' was no longer significant (P=0.26, likelihoodratio test). Thus, the model with significant variables that wasbest able to explain the variation in anti-HBc outcomes was'Mother's HBeAg status' (P<0.0001); 'Mother's age' (P=0.002);'Mother's years in birth country' (P=0.001); and 'Dose one squared'(P=0.0001).Influential observationsThe changes in beta coefficients (delta betas) when cases were88removed one at a time from the analysis were calculated andplotted. It was found that the delta betas for 'Mother's age','Mother's years in birth country', 'Dose one squared' and 'Mother'sHBeAg status' were all within the standard error for thatcoefficient. Thus, there were no influential cases.7. ANTI-HBs OUTCOME - RESULTS OF ANALYSES7.1. IntroductionBefore looking for associations with covariates or building aprediction model for the anti-HBs outcome, cases which were anti-HBc positive were excluded from the analysis. This was done sinceanti-HBs titres in children who are anti-HBc positive may have beenboosted by natural infection rather than be due to immunizationwith the hepatitis B vaccine.7.2. Associations of covariates with Anti-HBs titresAssociations of study variables with the anti-HBs outcome may beseen in Table 6.15. An inverse association was found between anti-HBs titres and the 'Child age' (P<0.0001). The opposite was truefor the direction of effect with 'Mother's age', but theassociation was not statistically significant (P=0.06). The'Number of doses' of hepatitis vaccine was significantly associatedwith anti-HBs outcomes (P=0.02), with more than two doses resultingin higher titres. 'Dose one' was also associated with anti-HBstitres (P=0.01), the effect being that titres were higher if 'Doseone' was given later. 'HBIG', classified as given/notgiven(P=0.28), was not associated with anti-HBs titres, nor was it whenclassified as given/unknown/not given (P=0.93).89A significant association was found between anti-HBs titres and'Minimum age of travel' (P=0.001). This was not the case, however,for the variable 'Months abroad' grouped as: no time/<3 months/>=3months (P=0.70). There was also no significant association betweenanti-HBs titres and the 'Number of contacts' (P=0.22) or 'Breastfeeding' (P=0.85).Table 6.15. Association of covariates with anti-HBs titresvariable p value direction ofeffectChild's sex 0.251 more males have titres500+ mIU/mlChild's age in years(2 to 8 years)<0.00012 older ages have lowertitresMother's age in quartiles 0.062 older moms have childrenwith higher titresNumber doses HB vaccine 0.022 >2 higher titresDose one (age rec'd grpd1-3,^4-7,^8-61,^62+ days)0.013 higher titres for laterdose oneDose 2^(received within2 months of dose 1, Y/N)0.972 no differenceHBIG (given/unk/notgiven) 0.932No.^of householdcontacts (grouped1,2,3,4,5,6,7+)0.222Breast fed (Y/N) 0.852Months abroad (grouped-no time/ <3mos/ >3mos)0.702Minimum age at travel 0.0012 younger ages had highertitres2 Mantel-Haenszel trend test3 Mantel-Haenszel exact test1 Pearson's chi-square907.3. Multiple regression analysis for the Anti-HBs outcome7.3.1. IntroductionThe natural logarithm of the anti-HBs titres resulted in adistribution which closely resembled a normal distribution apartfrom a cluster of 55 cases at the low end (log of 0 + 0.5) and 58cases at the high end (log of 1000 + 0.5). The testing proceduresdid not allow identification of actual values above 1000 mIU/ml.It was calculated that there should be 18 cases in the right handtail beyond 2 standard deviations and there were actually 17. Thusthe distribution of the logarithm of the anti-HBs titres resembledthe normal distribution. A histogram of the log anti-HBs titresmay be seen in Figure Model building procedureThe model for the prediction of anti-HBs titres was built bybackward elimination. In the first analysis, all of the studyvariables which could have some biologically plausible relationshipto anti-HBs titres were included. These were 'Child's sex','Child's age', 'Number of doses', 'Dose one', 'Dose two timing','HBIG', 'Breast feeding', 'Number of contacts' and 'Months abroad'.After the first procedure those variables which had missing valuesand did not reach a significance level of 0.05, and other variableswhich did not reach a significance level of 0.05 were removed. Theprocedure was then repeated for the final model. The final modelwith statistically significant variables that best predicted anti-HBs titres was 'Child's age' (P<0.0001), 'Number of doses'(P=0.001), 'Dose one' (P=0.01) and 'Months abroad' (P=0.05). There911Figure 6.8HISTOGRAM OF THE LOG OF ANTI-HBs LEVELSD,..qt-q. 7rlf- 4-Wnr. f•A42=42 mler, lnAmAwas positive confounding between 'Dose one' and 'Dose two timing'while 'Child's age' and 'Months abroad' were highly negativelyconfounded. In the univariate analysis there was no associationbetween anti-HBs and 'Months abroad' (P=0.70), but after adjustingfor 'Child's age' in the multivariate model it was apparent thatlonger time abroad was significantly associated with higher titresof anti-HBs. None of the interaction terms tested, ('Child's age'with 'Number of doses' (P=0.51), 'Child's age' with 'Dose one'(P=0.11) and 'Child's age' with 'Months abroad' (P=0.20)), reached92statistical significance when entered into the final model.R squared for the final anti-HBs model was 0.20.^Summarystatistics for the anti-HBs model may be seen in Table 6.16.Table 6.16.^Final prediction model for Anti-HBs titresin children age 2 to 8N=707^ P value Coef. Ratio CIVariables in the modelChild's age^<0 .0 001' -0.53 0.59 0.54, 0.64Number doses vaccine^0.0007 1 0.75 2.12 1.38, 3.26Dose one(age rec'd) 0.01 1 0.53 1.71 1.14, 2.56Months abroad 0.05 1 0.22 1.25 1.002, 1.56Constant 4.22Interaction terms (not in the model)Child's age*Number dosesChild's age*Dose oneChild's age*Months abroad0.51 10.11 10.20 11 t test7.3.3. Prediction formula for Anti-Ms titresThe exponential of the coefficients obtained from the final modelprovided the ratios for calculating the prediction of anti-HBstitres. For every one year increase in the 'Child's age', anti-HBstitres are reduced by a factor of 0.59. Each additional dose ofvaccine results in a 2.12 times increase in anti-HBs titres.Giving 'Dose one', one category of time later results in a 1.71times increase over the previous level and a one category increasefor 'Months abroad' increases anti-HBs titres by a factor of 1.25.For the study sample, the prediction formula derived from thismodel can be depicted as follows:93Log anti-HBs + 0.5=(4.22 - 0.53 * CHILD'S AGE + 0.75 * NUMBER DOSES +0.53 * DOSE ONE category + 0.22 * MONTHS ABROAD category)Where 'Dose one' categories are: 1= 1-3 days; 2= 4-7 days; 3= 8-61days and 4= >=62 days and'Months abroad' categories are: 0= no time abroad; 1= <3 monthsabroad; 2= >=3 months abroad.Thus, a five year old child who had 3 doses of hepatitis B vaccine,had the first dose within 3 days of birth and travelled at leastthree months abroad can be expected to have an anti-HBs level of:Log Anti-HBs + 0.5=exp(4.22 - 0.53 * 5 + 0.75 * 3 + 0.53 * 1 + 0.22 *2)Anti-HBs = exp(4.79) - 0.5= 120.3 - 0.5= 119.8 = 120 mIU/mlIf the same child had not travelled abroad, the predicted anti-HBstitres would be 77 mIU/ml.7.3.4. Diagnostic procedures for the Anti-HBs modelResidualsThe histogram of studentized residuals resembled the normal curvewith a small amount of increased concentration of cases around the-1 to -2 standard deviation area and a small cluster beyond +2standard deviations. The normal probability plot approximated astraight line and was fairly linear. Studentized residual plotsshowed that 42 out of 707 cases plotted fell outside the 2 standarddeviation range. For the variable 'Child's age' there did notappear to be any pattern to these cases. However, the residuals94for 'Number of doses' were concentrated with the third dose below-2 standard deviations indicating the model over predicted thelevel of antibodies for some children who had three doses ofvaccine. The first category of 'Dose one' (1-3 days) contained allof the residuals below 2 standard deviations. The model predicteda higher level of anti-HBs for these 42 children than was actuallythe case. The large residuals for 'Months abroad' were spreadbetween the three categories but the category of 'no months abroad'had 34 residuals indicating an over prediction of anti-HBs titresfor some children who did not travel. Residual plots for 'Numberof doses' and 'Dose one' may be seen in APPENDIX 3.Influential observationsThe studentized beta coefficients (delta betas) were within 0.2standard deviations for 'Child's age'; within 0.6 standarddeviations for both 'Number of doses' and 'Dose one', and within0.3 standard deviations for 'Months abroad'. This indicates therewere no influential observations.8. HBsAG OUTCOME - RESULTS OF ANALYSES8.1. IntroductionAlthough it was not in the primary hypotheses to look for factorsassociated with the development of the hepatitis B virus carrierstate as opposed to infection, it was decided to examine the studyvariables in this respect for hypothesis generation purposes.Univariate associations with HBsAg were first explored followed bymultivariate analyses, which was done in two ways. Firstly,logistic regression procedures were run with the denominator being95'all cases for which results for the HBsAg tests were obtained'(N=703). Missing values reduced this number to 525 when mother'sHBeAg status was added. Secondly, the procedures were run when thedenominator was restricted to 'those cases which were anti-HBcpositive' (N=39). There were only 32 cases when the procedure wasrun including the variable 'Mother's HBeAg status'.8.2. Associations of covariates with HBsAgStatistics for the associations of covariates with the HBsAgoutcome are summarized in Table 6.17. There were no significantassociations found between HBsAg outcomes and the followingvariables: 'Child's sex', 'Child's age', 'Mother's country ofbirth', 'Mother's years in birth country', 'Father's country ofbirth', 'Father years in birth country', 'Number of doses', 'Dosetwo timing', 'Language', history of 'Moving' since the child'sbirth, 'Jaundice', 'Breast feeding', 'Ear piercing', 'Acupuncture','Surgery', 'Blood transfusions', 'Daycare' or 'Hospitalization'.Significant associations were found between the HBsAg outcome andthe following variables: 'Mother's age' (inverse), 'Dose one','HBIG', 'Mother's HBeAg status', 'Number of contacts', 'Travel' and'Months abroad' (inverse).When the analysis was restricted to anti-HBc positive cases only,there were significant associations only with the followingvariables: 'Number of contacts', 'Travel' (inverse), 'Total monthsabroad' (inverse), and 'Months abroad' grouped (inverse). Asummary of associations between variables and the HBsAg outcome,excluding anti-HBc negative cases, may be seen in Table 6.18.96Table 6.17. Association of covariates with HBsAgVariable P value RR & CI Directionof effectChild's sex 0.591 1.300.49,^3.46femalesChild's age in years(2 to 8)0.232 older agesChild's birth year 0.212 earlieryearsMother's age(quartiles)0.0022 youngermothersMother's birth country(vertical/horizontaltransmission)0.753 1.6840.38,15.40verticalMother's years in birthcountry (quartiles)0.782 longertimeFather's birth country(vertical/horizontaltransmission)0.383 2.3640.53,^21.59verticalFather's years in birthcountry (quartiles)0.862Language (by country ofvertical/horizontaltransmission0.071 3.560.82,^15.54verticalMother's HBeAg status(positive/negative)<0.00013 30.3644.46,1293.71positiveNumber of doses HBvaccine0.192 fewerdosesDose 1 (age received,4 groups)0.042 > 7 daysDose 2 (received outside2 months of dose 1)1.003 0 40,^14.84within 2monthsDose 3^(no/yes) 0.443 1.840.04,^12.96no 3rddoseDose 4 (no/yes) 1.003 infinite4.04,infiniteDose 5 (no/yes) 1.003 infinite40, infiniteHBIG (not given/given)(given/unknown/not given)0.1830.0325.9540.13,^49.58not given97Table 6.17. Association of covariates with HBsAgVariable P value RR & CI Directionof effectMoved since child's birth 0.341 1.72 moved(yes/no) 0.56,^5.27Jaundice after first 1.003 04month (yes/no) 0,^108.58Breast fed (yes/no) 0.551 0.75 no breast0.28,^1.98 feedingEars pierced (yes/no) 0.503 1.594 ears0.37,^5.36 piercedAcupuncture (yes/no) 1.003 0 40,^108.58Surgery (yes/no) 0.333 0.324 no surgery0.01,^2.11Blood transfusion 1.003 04(yes/no) 0,^68.77Daycare (yes/no) 0.093 0 4 no daycare0,^1.16Number household contacts > number(ungrouped) <0.00012 contacts(grouped 1,2,3,4,5,6,7+) 0.00032 7+Hospitalization (yes/no) 0.493 0.354 no hospit-0.01,^2.30 alizationTravel (yes/no) 0.023 04 no travel0,^0.76Months abroad (no months/<3months/ >3months)0.032 no timeabroad2 Mantel-Haenszel test for trend3 Fisher's exact test4 Odds ratio and exact confidence interval5 Exact test for trend1 Pearson's chi-square98Table 6.18. Association of covariates with HBsAg(Anti-BSc positive cases only)Variable PvalueRR & CI Directionof effectChild's sex 0.701 1.160.54,^2.46femalesChild's age in yrs (2-8) 0.912Child birth year (84-89) 0.972Mother's age in quartiles 0.182 oldermothersMother's birth country(vert/horiz transm)0.563 0.3340.01,^7.18Mother's years in birthcountry0.212Father's birth country(vert/horiz transm)1.003 1.1140.11,14.86Father's years in birthcountry0.172Number of doses(ungrouped)0.942Dose one (age rec'd) 0.982Dose two (rec'd outside 2months of dose 1, yes/no)1.003 O.R.^0 40,^58.50Dose 3 (Yes/No) 1.003 1.5040.07,94.31HBIG (not given/given)(given/unk/not given)1.0030.9721.364.02,112.06Mother's HBeAg status(positive/negative)0.103 6.504.61,321.49if motherpositiveLanguage (for countriesof vert/horiz transm.1.003 1.5640.19,19.33Moved since child's birth(yes/no)1.003 1.1340.21,^6.68Breast-fed (yes/no) 0.862 1.070.51,^2.26Earpiercing (yes/no) 1.003 1.1340.18,^6.55Surgery or stitches(yes/no)0.373 0.4440.07,^2.64nosurgery99Table 6.18. Association of covariates with HBsAg(Anti-HBc positive cases only)Variable PvalueRR & CI Directionof effectDaycare (yes/no) 0.503 O.R.^0 40,^7.32nodaycareNumber of contacts 0.032 > numberHospitalization (yes/no) 0.623 0.4240.01,^6.01no hosp'nTravel (yes/no) 0.013 O.R.^0 40,^0.80no travelTotal months of travel 0.052 0 monthsMonths abroad (grouped) 0.022 0 months1 Pearson's chi-square2 Mantel-Haenszel test for trend3 Fisher's exact test4 Odds ratio and exact 95% confidence intervalExact test for trend8.3. Multiple logistic regression analysis for HBsAg outcomes -denominator all casesModel oneModels were determined in the same fashion as for the anti-HBcoutcome. Variables included in the analysis to determine the firstmodel (without 'Mother's HBeAg status' included) were: 'Child'sage', 'Child's sex' 'Mother's age', 'Mother's country of birth','Mother's years in birth country', 'Language', 'Number of doses','HBIG', 'Number of contacts', histories of 'Moving', 'Breastfeeding', 'Ear piercing', 'Surgery', 'Day care', 'Hospitalizations'and 'Travel'. The statistically significant variables selected bythe stepwise procedure which best explained the variation in HBsAgoutcomes were: 'Number of contacts' (P=0.001), 'Mother's age'100(P=0.002), 'Language' (P=0.02) and 'Travel' (P=0.006). In theprocess of building the model it was revealed that there wasconfounding between 'Number of contacts', the strongest predictor,and 'Mother's age', 'Language', 'Daycare' and 'Travel'. However,the variable 'Number of contacts' could not explain all of thevariation due to 'Mother's age' resulting in 'Mother's age' beingentered into the model as well. Negative confounding between'Mother's age' and 'Language' resulted in the increasedsignificance of 'Language', causing it to be selected into themodel. 'HBIG' was significant initially (P=0.02) but graduallylost significance as other variables were added, ending up atP=0.07. 'Travel', on the other hand, could not be explainedfurther by 'Mother's age' or 'Language' and thus retained itsstatistical significance.Model twoModel two was run with the same variables as model one with theaddition of 'Mother's HBeAg status' and 'Dose one'. Only twovariables qualified for selection for the final model. These were'Mother's HBeAg status' (P<0.001) and 'Dose one' (P=0.03).Positive confounding was apparent between 'Mother's HBeAg status'and the variables 'Language', 'Mother's age', 'Number of contacts'and 'Travel'. Variation due to 'HBIG' and the 'Number of doses'was largely explained by the inclusion of 'Dose one'. Betacoefficients were determined for 'Mother's age' (P=0.07), 'Numberof contacts' (P=0.06) and 'Travel' (P=0.07), as these variablestended toward significance but the evidence was not sufficient toqualify them for entry into the final model. When these variables101Table 6.19. Prediction models for the HBeAg outcomein children age 2 to 8Denominator-all casesP value Coef.^RR^CIVariables in model 1 (Mother's HBeAg status not included)N=703Number contacts 0.0011 0.48 1.62 1.21, 2.18Mother's age 0.0021 -.18 0.83 0.74, 0.94Language 0.021 1.66 5.24 1.04, 26.49Travel 0.0061 -7.36 0.0006 1.3E-19, 3.2E+12Constant -1.54Variables in model 2 (Mother's HBeAg status included)N=525Mother's HBeAg status <0.0001 1 3.62 37.33 4.46, 312.33Dose one (age rec'd) 0.031 1.18 3.25 1.29, 8.16Constant -7.39Variables not in model 2Number of contacts 0.098 3 0.26 1.29 0.95, 1.76Mother's age 0.09 3 -0.11 0.90 0.79, 1.02Travel abroad 0.73 3 -7.05 0.0009 1.6E-21, 4.6E+14Interaction term (not in the model)Mother HBeAg*Dose one^0.42Denominator Anti-HBc positive cases only N=32Variables in model 1 (Mother's HBeAg status included)Travel abroad^0.011^-9.28^.0001^1.1E-42,7.9E+33Constant 0.07Variables in model 2 (Mother's HBeAg status forced in)Mother's HBeAg status^0.061^1.87^6.50^0.68,^62.09Constant^ -1.791 -2 Log likelihood Ratio2 Score test3 Wald test102were forced into the model, the instability of the variable'Travel', was revealed by a P value of 0.73 (relative risk, 0.0009)and a very wide 95 percent confidence interval, (1.6E+ 21 , 4.6E+14 ).A summary of the statistics for this model may be seen in table6.19.The proportion of variation in the null model accounted for by thefinal prediction model ('Mother's HBeAg status' and 'Dose one')was: R2 = 129.105 - 100.324/129.105 = Multiple logistic regression analysis for the HBsAg outcome -denominator restricted to anti-H8c positive casesBy limiting the analysis to only those cases that were anti-HBcpositive, it was hoped to determine more specifically the studyvariables which may be associated with the progression of infectionto the carrier state. The only variable that turned out to besignificant in the first procedure (model one) was 'Travel'(P=0.01). When 'Mother's HBeAg status' was forced into the modelfor model two, no variables reached statistical significance. Pvalues were as follows: 'Travel' (P=0.07, score test); 'Mother'syears in birth country' (P=0.06, score test); 'Mother's HBeAgstatus' (P=0.06; relative risk 6.50; 95 percent confidenceinterval, 0.68 to 62.09; N=32). Summary statistics related to thismodel may be seen in Table 6.19. above. None of the children whobecame HBsAg positive had a history of travel.103CHAPTER SEVENDISCUSSION1. INTERNAL VALIDITYBecause of the design of this study, two of the main concerns withrespect to its internal validity were the possibilities ofselection and information bias. Bias due to confounding wasreduced by adjustment in the multiple regression analyses and isdiscussed later.1.1. Considerations with respect to possible selection biasIn order to retain sufficient power for analysis with the smallnumber of events expected, an attempt was made to include theentire population as the sample. The 67% response rate opened thepossibility for systematic error which could have resulted indifferences between the group that participated and the group thatdid not participate in the study. Error created in selecting anonrepresentative sample could lead to the results correlating withparticipation rather than with the study variables. In this study,analysis of the information collected on variables which werecommon to both participants and nonparticipants suggested that thesample was representative of the population being investigated.There were no significant differences found in the children's agesand sex or in the mothers' ages and HBeAg status. The similarityin the mothers' ages and HBeAg status between the two groups wasparticularly important as it was found that the variables mother'sage and HBeAg status were predictors of infection.104Information on language and a history of moving was collected toprovide a surrogate measure of confounding due to socioeconomicfactors. It was thought this surrogate measure might differentiatebetween those infected and those not infected. No significantdifference was found in language or history of moving. Nor wasthere a difference in effort to contact children who were found tobe infected, versus not infected, as measured by the number ofphone calls made in an effort to enrol participants. Similarity inknown factors which may be confounders for the outcomes does notrule out selection bias but it would support the likelihood ofsimilarity in unknown factors which may influence study outcomes.1.2. Considerations with respect to information biasThe three information biases that could affect study outcomes wereinterviewer bias, recall bias and other sources ofmisclassification. Since interviewers did not known the results ofblood tests before completing questionnaires, prior knowledge ofoutcomes could not have led to systematic bias. Furthermore, therewere no results for which interviewer bias appeared to be anexplanation. The only unexplainable result, the association oftime abroad with anti-HBs titres, was also not readily explained byinterviewer bias. As described in Chapter Four on thequestionnaire, a concerted effort was made to prevent bias due toimproper interview technique.Recall bias could have resulted in misclassification of exposures,but the factual nature of the questions may have reduced thisthreat to validity. Misclassification of timing and number of HB105vaccine doses was possible, but using written records rather thanrelying on parents' recall likely minimized this source of error.The grouping of some variables into broad categories likely reducederror due to inability to recall lengths of time, (e.g., in thevariable 'Months abroad'). However, error due to confusion withrespect to which child travelled, or was hospitalized or was breastfed in the case of larger families could still have occurred.Misclassification may have occurred in the HBsAg and anti-HBcoutcomes. An over estimate was possible, but unlikely, in theHBsAg outcomes since confirmation of the carrier state was notcarried out within the study period. An over estimate in the anti-HBc outcome was possible due to circulating maternal antibodies inthe two year old group. Since maternal anti-HBc usually disappearsby two years of age this source of error is likely to be minimal.Some misclassification due to laboratory error in outcome resultswas also possible, but this was likely limited because of the goodposttest likelihoods of the blood sample assays. Ifmisclassification was the case it would have led to an underestimation of the magnitude of associations between outcomes andprediction variables making our results conservative.1.3. Considerations with respect to consistency of resultsThere were several factors which support internal validity in thisstudy. Firstly, there was consistency in the anti-HBc and HBsAgoutcome results, between those obtained when 'Mother's HBeAgstatus' was excluded and those obtained when it was included in themodelling procedure. The model for the anti-HBc outcome was the106same for both model building procedures except for the addition ofmother's HBeAg status when this variable was included. In theHBsAg outcome model building procedure it was determined that themother's HBeAg status was a confounder for the variables thatappeared in the model without mother's HBeAg status, and thereforethese variables were simply surrogate measures for the effect ofmother's HBeAg status.Secondly, the appearance of the same variables, 'Mother's HBeAgstatus' and age at 'Dose one', in both the anti-HBc and HBsAgoutcome prediction models, would tend to support the validity ofthe results.Thirdly, evidence for internal validity was found by applying theprediction formulas for the anti-HBc and anti-HBs. The parameterestimates provided predictions of risk for infection which aresupported by similar neonatal studies in the literature; 0.7% and7% for mothers with negative and positive HBeAg statusrespectively. The estimate of geometric mean anti-HBs levelderived from the parameter estimates was close to that found in thecurrent study (77 versus 79.7 mIU/ml).Ideally, one would want to include measurements on all possibleconfounders which may affect the outcomes. For this study, allvariables that were thought to have plausible associations with theoutcomes and that could feasibly be measured were included in themodel-building procedures. Nevertheless, all the variation due tothe outcomes was not accounted for by the prediction models.107Therefore, the variables in the final models cannot be consideredthe only explanations for the outcome variation. Only 24% of thevariation for anti-HBc outcome and 22% for the HBsAg outcome wasexplained by the variables in the prediction models. The majorityof variation was due to unknown factors, among which could bevirulence and dose of the HB virus as well as the geneticpredisposition of the child.Thus far the internal validity of this study has been discussed.Internal validity is important in terms of the conclusions drawnfrom the results of the present study and forms the basis forexternal validity which is discussed next.2. EXTERNAL VALIDITYThe advantage of this study of hepatitis B vaccine, within thecontext of a Program, is that it will be useful for generalizationto similar programs. However, to apply the results of this studyto other programs, those factors that were not measured by thecurrent study, but that may account for variation in outcomesshould be similar to this Program. Variables measured in thisstudy need not be the same as in the comparison population. Forthe variables that were measured and found to be significant,adjustment can be made.^The adjustment is made by using thecoefficient for that variable in the prediction formula. For thosevariables measured but not found significant, adjustment is notnecessary. For example, given similarity in unmeasured factorswhich may account for variation in outcomes, these results could beapplied to a program that gave the second dose of HB vaccine at or108close to two months of age, since the coefficient for "Dose twotiming' was zero. In addition, using the prediction formulas toextrapolate at the extremes of continuous variables in theprediction formula (relative to what occurred in these study data),may not be valid.3. INFECTION AND CARRIER RATESThe infection rate found in this study was 5.1% in children rangingfrom two to eight years of age. This is close to the long-termweighted average of 5.5% estimated from the literature reviewedprior to this research. Both higher and lower rates have beenfound in long-term studies of immunized neonates born to HBsAgpositive mothers. Ip et al. (1989) found a 38% rate of infectionafter three years in infants born to HBeAg positive mothers. Thesechildren had received four doses of 3 ug plasma derived HB vaccinewith or without HBIG. This 38% rate is not consistent with otherfindings. An 11.9% rate of HBV events was found by Tajiri et al.(1989) after 4 years in infants who had a successful response to apreventive schedule including two doses of HBIG and three doses of10 ug vaccine. Lo et al. (1988) in Taiwan, found an even lowerrate in infants who initially responded with >=10 mIU/ml anti-HBs.No children were anti-HBc positive at three years of age and nonebecame HBsAg positive by five years of age. All of the motherswere HBeAg positive as well as HBsAg positive and a four doseschedule of 5 or 2.5 ug plasma-derived vaccine with or without HBIGwas used.The carrier rate (HBsAg positive) in this study was 2.3%. This is109similar to a rate of 2.8% by five years of age found by Resti etal. (1990) in Italy. These infants were given HBIG at birth and athree dose series of plasma-derived vaccine with doses at 20 days,two months later and in the twelfth month of life. Zhu et al.(1992) found a 13.7% carrier rate after 6 years using 3 differentschedules of vaccination, two with vaccine only. The HBeAg statusof the mothers in this study was not stated, but the research wascarried out in Shanghai, an area of high e antigen prevalence.Attack rates after immunization have also been studied in otherhigh prevalence populations and have generally been shown to belower than in populations of neonatally immunized infants whosemothers are carriers. After five years of follow up only four(0.36%) persons became anti-HBc positive in a follow-up of 1114Yupik Eskimos of all ages in Southwest Alaska (Wainwright et al.1989). After three years, a two percent attack rate and a 0.7%carrier rate has been reported in the Gambian hepatitisintervention study (Chotard et al. 1992). HB vaccine in thisprogram was given to all infants, not just infants of carriermothers. Coursaget (1986) noted a 1.8% attack rate three to fouryears after a booster dose given 12 months after the first dose.This rate increased to 8% after five to six years. This increaseover time may imply waning immunity or time to exposure to aninfective dose of virus. Because an 8% vaccine failure rate issimilar to other neonatal studies, it is likely the latterexplanation.1104. THE OCCURRENCE OF INFECTION AND THE CHILD'S AGEIn this present study no significant association was found betweenthe overall attack rate and the child's age from two years to eightafter adjustment for covariates. Also, it was found, usinglogistic regression, that the higher attack rate in childrenimmunized in the earlier years of the Program could be accountedfor by 'Dose one' timing. The change of attack rates with timesince immunization reported in the literature for similar neonatalresearch are consistent with this result. Lo et al. (1988) andResti et al. (1990) do not show an increase over five years.The infection rates in this present study varied between 2.6 and 7%for ages less than eight with no consistent trend. At age eight,however, the rate increased to 23% (3/13). Of the anti-HBcpositive eight year old children, one child had only one dose of HBvaccine at 126 days of age, the second child had three doses ofvaccine but the first one was at 158 days of age and the thirdchild had three doses of vaccine but no HBIG at birth. Also, forthis third child the age at the first dose was not known. Sincethese children did not receive vaccines according to Programguidelines the high rate of infection at age eight is not evidenceof lack of vaccine efficacy or waning immunity.5. PREDICTORS FOUND FOR ANTI-HBc AND HBsAG POSITIVE OUTCOMESAccording to the logistic regression analysis, the best group ofindependent predictors of HB infection consisted of the followingfactors: the mother's HBeAg status, the child's age at the firstdose of HB vaccine, the number of years the mother spent in her111country of birth and the mother's age. The carrier state was bestpredicted by only two variables: the mother's HBeAg status and theage of the child when the first dose of vaccine was administered.The statistical models used to determined which predictors weremost strongly related with the outcomes were association models,and therefore do not necessarily imply causation. Rothman (1986)states that:"To an extent, every variable measured in an epidemiologicstudy can be considered only a surrogate variable for somemore appropriate measure of the underlying phenomena."A causal explanation may be true in some cases but this judgementwould depend on other factors such as biological plausibility andconsistency of results. There are four other possible reasons forcertain variables to be chosen in a prediction model. These are:reverse causation, confounding, bias or chance. How one might viewthe variables that were selected to be in the prediction models isdiscussed in the following.Mother's HBeAg statusThat the 'Mother's HBeAg status' is a predictor of infection andcarrier status has previously been shown by many other researchers(Beasley et al. 1977; Theppisai et al. 1990; Stevens et al. 1979;Assateerawatt et al. 1991). Also, biological plausibility issupported by evidence of viral replication in mothers who are HBeAgpositive. In this study there was a strong association between'Mother's HBeAg status', 'Mother's country of birth' and'Language', and the association of infection with 'Mother's country112of birth' and 'Language' was accounted for by 'Mother's HBeAgstatus'.Age at Dose oneThe importance, as seen in this study, of the age of the child whenthe first dose of HB vaccine is administered is not so clearlystated in the literature. Several researchers have done studieswith a delayed first dose of vaccine, on the theory ofphysiological immunosuppression of newborns producing a lesseffective response to vaccination (Resti et al. 1990). After onedose of HBIG at birth, Resti et al. (1990) delayed the first doseof HB vaccine to 20 days and found a 2.8% (2/72) HBsAg positiverate after five years. Only 3.9% of the mothers were HBeAgpositive and most were anti-HBe positive. Monna et al. (1988)observed after one year, an 8% (5/65) rate of HBsAg positivity whenthe first vaccine dose was given at one, two or three months ofage. All of the mothers were HBeAg positive and HBIG was given attwo months of age as well as at birth. In a randomized study,Schalm et al. (1989) found a four percent (4/90) rate of infectionafter a four dose schedule which included vaccine at birth but onlya one percent (1/90) rate after a four dose series starting atthree months of age. The group receiving the first dose of vaccineat three months also received a second dose of HBIG at three monthsof age. The difference in rates between early and late vaccinestart groups was not statistically significant. The results ofthese studies imply that the rate of infection is not necessarilygreater when the first dose of HB vaccine is delayed. However, ourfindings indicated that a delay in the first dose of vaccine was113more likely to result in infection. Furthermore, the early timingof the first dose of HB vaccine was more important than HBIG inpreventing infections.During the analysis, it was apparent from the stepwise regressionthat 'Dose one' was highly confounded with 'HBIG'. In one modelbuilding procedure which did not include the 'Mother's HBeAgstatus', 'HBIG' was selected over 'Dose one'. However, when'mother's HBeAg status' was included, 'Dose one' was selected over'HBIG' even when HBIG was forced into the model. Thus it wouldappear, from our data, that getting 'Dose one' on time was a moreimportant factor than getting the HBIG. However, the importance ofHBIG cannot be ruled out by this finding since three of the fourchildren who had a late 'Dose one', and became infected (4 of 39anti-HBc positive cases), also did not have HBIG at birth or it wasunknown if HBIG was given at birth.Furthermore, how soon after birth 'Dose one' must be given toprevent infection is not clear from these data. The associationbetween delay in 'Dose one' and infection, in the current study,arose because of the high rate of infection in children whoreceived 'Dose one' after seven days. Because there were only 12children who received 'Dose one' between four and seven days of age(none of whom became infected), it is not possible to say whetherchildren who receive the vaccine between four and seven days of ageare at higher risk of infection than those who receive it beforefour days of age.114The use of HB vaccine alone at birth, without concomitantadministration of HBIG, has also been studied by other researchersbut with varied results. Poovarawan et al. (1990) found noinfections when both treatments were given (0/48), but a 3.4%(2/58) carriage rate when only vaccine was administered. In alater randomized controlled study by Poovarawan et al. (1992) nostatistical difference was observed in long-term protectiveefficacy (carrier) between vaccine alone (3/60) or vaccine and HBIGat birth (1/61). Sehgal et al. (1992) found a trend toward greaterefficacy when HBIG was not given but the power of the study wasinadequate.Mother's years in birth countryThe variable 'Mother's years in birth country' would appear to bea surrogate measure for other risk factors. These factors may becultural or social practices that put children at greater risk ofinfection and thus may be a marker for horizontal transmission.One such practice may be the care of children in the home and byextended family members who are also infected. This would besupported by the stepwise regression results when the variable'Daycare' appeared, on the basis of change in score statistics, tobe confounded with 'Mother's years in birth country'. Mothers whospent more years in their country of birth were less likely to use'Daycare' (P=0.0002) and there was a univariate trend for 'Daycare'attendance to be protective of infection (P=0.09). Use of daycaremay also be a marker for socio-economic conditions such as lesscrowding at home.115Mother's age'Mother's age' was a predictor of infection, resulting in a higherrate of infection for younger mothers even after adjusting for'Mother's HBeAg status'. This variable may be a surrogate forother factors related to vertical transmission which are not beingmeasured directly, such as the mother's anti-HBe status.Information on anti-HBe status for the mothers was not collected.Apart from being more infectious with respect to verticaltransmission, younger mothers may also have been more likely toinfect their children by horizontal transmission. This variable,therefore, also raises the possibility of cultural or socialfactors. Younger mothers may adhere to practices which may throughthemselves or through other contacts increase the child's risk forinfection. It could also be, that the mothers of children whobecame infected (who were more likely to be older), were youngerjust by chance.Breast feeding in this study was not associated with risk ofinfection. This is compatible with the findings of Beasley (1975)and Woo (1974).Of the variables studied, the 'Mother's HBeAg status' and thechild's age at 'Dose one' were the only variables significantlyassociated with becoming a carrier, after adjustment in themultiple logistic regression model. These are infection relatedvariables and they accounted for only 22% of the variation in HBeAgoutcomes. Of more importance to the persistence of viral infectionmay be host related factors, some of which were not measured in116this study. Some authors have suggested that the age of initialinfection, race and other genetic factors may influence the abilityto suppress viral replication (Evans ed. 1989).6. SEROPOSITIVITY RATESThe overall seropositivity rate in this study was 87.9% (>=10mIU/ml). As this was a one-time follow-up study, this overall ratemay not be comparable to other long term study rates where the samechildren are tracked over time. In this study each age groupconstitutes a separate cohort of individuals. Thus each rate, ascalculated, applies to only its age cohort. The seropositivityrates by age group were as follows: two years 98.6%; three years,99.4%; four years, 90%; five years, 88%; six years, 80%; sevenyears, 70.4%; and eight years, 90%. These rates are compared toother study results in the following discussion.After two years, Poovorawan et al. (1992) found a seropositivityrate of 94.9% (37/39) using 10 ug vaccine. This compares to aseropositivity rate of 98.6% for the age two group for the currentstudy using a 20 ug vaccine. In other respects the immunizationschedules were the same.The current study's age three seropositivity rate was 99.4%compared to 100% (47/47) at three years found by Poovorawan et al.(1992) (using a four dose schedule of 10 ug vaccine and HBIG). Incontrast, Ip et al. (1989) in Hong Kong, reported a 73%seropositivity rate after three years. The treatment schedule forIp's study used four doses of 3 ug plasma-derived vaccine whereas117the current study used three doses of 20 ug yeast-derived vaccinefor this age cohort. Ninety-five percent of infants immunized inThe Gambia Program using a four dose series of 10 ug vaccine hadprotective titres of antibody after a period of three years(Chotard et al. 1992).The age four seropositivity rate for the current study was 90%while the rate for Poovorawan et al. (1992) after four years was100% (44/44) in children immunized with four doses of 10 ug vaccineand no HBIG.While the current study found a seropositivity rate of 88% for theage five group, Resti et al. (1991) reported a seropositive rate of95.8% after five years for infants immunized with plasma-derivedvaccine: the first dose at 20 days, the second two months later andthe third in the twelfth month of life. Infants in the YupikEskimo population study had a more rapid loss of antibody. Only 11(61%) of 18 had anti-HBs titres of 10 SRU's (protective level) orgreater after five years (Wainwright et al. 1989). They hadreceived three 10 ug doses of plasma-derived vaccine in the firstyear of life. In a subsequent follow-up of this Yupik Eskimopopulation, Wainwright et al. (1991) found that in the age group 19years and under, 87% had anti-H8s titres of 10 or more mIU/ml afterseven years (rate not stated for infants alone). The median anti-HBs level at seven years for infants immunized in their first yearof life was only 21 mIU/ml. In the current study 70.4% in the ageseven group had anti-HBs titres above 10 miU/m1 and the mediananti-HBs level for the age seven group was 27.1 mIU/ml.118Seropositivity rates appear to remain higher in children whoinitially respond with at least 10 mIU/ml of anti-HBs. Lo et al.(1988) found rates of 94% (60/64), 91% (31/34) and 97% (98/101) forages 3, 4 and 5, respectively, in a cohort of 199 infants. Thesechildren had an initial response of >=10mIU/m1 anti-HBs followingHBIG at birth and a four dose series of 5 or 2.5 ug of plasma-derived vaccine. Using a three dose series and 20 ug, Hwang et al.(1990) found, in infants who initially responded, that 95% hadpersistence of anti-HBs after five years. These rates at ages 4and 5 were slightly higher than those for the current study (at 90%and 88% respectively).7. PREDICTORS FOR ANTI-HBs OUTCOMESWe have seen that seropositivity rates vary considerably betweenstudies. This may be related to the population immunized, vaccinetype, dosage and schedule and other possible factors. There is,however, generally a decline in rate over time. In this study itwas found that the 'Child's age' was the strongest predictor foranti-HBs titres. 'Number of doses', age at 'Dose one' and 'Monthsabroad' were the other significant predictors.Child's age and Anti-HBs titresThe anti-HBs geometric mean titre at three years of age, which inthis study was 238 mIU/ml, can be compared to two other long-termstudies which also gave estimates at three years of age. Ip et al.(1989) found a geometric mean of 30 mIU/ml using a 3 ug dose ofhepatitis B vaccine. Schalm et al. (1989) found between 400-800mIU/ml using a four dose schedule and 10 ug of vaccine.119In the present study, the geometric mean anti-HBs titres differedmost dramatically between age groups three and four: 238 mIU/mlcompared to 79 mIU/ml, respectively. The difference in titres,between age groups, then dropped more slowly to 23 mIU/ml for theage seven group. Other authors have found a decline of anti-HBstitres associated with age but the greatest decline is generallydescribed to occur within the first two years after immunization(Hadler et al. 1988; Jilg et al. 1988). We did not have anti-HBsdata in mIU/ml units for children under two years of age.Wainwright et al. (1989) also found age to be associated with anti-HBs titres, but they found that the initial antibody level is astronger predictor of the persistence of protective antibodytitres. This finding is supported by other authors (Chotard et al.1992; Jilg et al. 1988) and by the current research as well sincestudy anti-HBs titres were highly associated with titres at orabove 10 mIU/ml at 6 to 18 months of age (P<0.0002).Age at Dose one and Anti-HBs titresAn interesting phenomena occurred in this study in that there wasa geometric mean titre increase of 83 mIU/ml anti-HBs in the ageeight group over the age seven group level. This increase to 106mIU/ml at age eight is explained by late initiation of the vaccineseries in some children around the time the vaccine was first madeavailable and thus may be attributed to a program effect. Four outof the ten eight year old children had the first dose of HB vaccineafter five days of age and these children also had anti-HBs titreswell over 100 mIU/ml. Resti et al. (1990) found after a delayed120first dose at 20 days that only 3 of 72 children at five years ofage had less than 10 mIU/m1 anti-HBs: four percent as compared to12% in this study. Schalm et al. (1989) found statisticallysignificantly higher titres of anti-HBs at 11 and 24 months inchildren who received a delayed vaccine series compared to vaccinestarting at birth.Several authors (Chotard et al. 1992; Coursaget et al. 1986; Lo etal. 1988) have noted boosting of anti-HBs titres in the absence ofimmunization. An explanation suggested for this 'boosting' wasviral exposure without cell invasion and replication. This is nota likely explanation for the increase in titres for eight year oldchildren in this study. If it were, one would have expected tohave seen this effect in the seven year old group as well.Number of doses and Anti-HBs titresEvidence from these data supports the effect of 'Number of doses'on antibody titres. Children with fewer than three doses were morelikely to be non-responders or have less than 100 mIU/ml of anti-HBs and those with three or more doses were more likely to haveover 100 mIU/ml of anti-HBs. Chotard et al. (1992) also found theproportion of children with high antibody titres increased with thenumber of doses received, though only a slight increase occurredbetween three and four doses.Months abroad and Anti-HBs titresChildren who spent time abroad had higher anti-HBs titres but theinterpretation of this is not clear and the association requires121further study.8. RELATIONSHIP BETWEEN ANTI-HBs TITRES AND INFECTIONSThe effect of anti-HBs titres on subsequent infection rates cannotbe determined directly in this study.^Because the two weremeasured simultaneously the effect of anti-HBs on infection riskcannot be separated from the effect of infection on anti-HBstitres.^For 183 (24%) children, 6 to 18 month results wereavailable but this number is small and the analyses based on thissample lacks power.That infections may not be related to anti-HBs titres was suggestedin this study in several indirect ways. First of all, anti-HBcoutcomes were not statistically significantly associated with 6 to18 month anti-HBs outcomes (P=0.12; relative risk 0.30; 95%confidence interval, 0.06, 2.03). This suggests that infectionsoccurred even in children who had responded to the initial seriesof HB vaccine. As well, initial non-responders did not necessarilybecome infected. Since the power of this test was low and therelative risk of 0.30 suggests protection (seropositivity) theargument that infections are not related to anti-HBs titres on thisbasis is weak.Study HBsAg positive outcomes were associated with early anti-HBsoutcomes (P=0.008; relative risk, 0.05; 95% confidence interval, 0,0.62). The direction indicated that chronic carriers were morelikely to be those who did not initially respond to the HB vaccine.These results would also suggest that if infection does occur after122an initial response, progression to the carrier state is lesslikely.Secondly, this study demonstrated little increase in cumulativeinfection rates from ages two to seven years. This suggests thatif infections occur, they are more likely to occur early. For thesub-sample of children tested twice, 7 of 9 (78%) infections foundon study results had already occurred by 6 to 18 months (P=0.008).However, 51 of the 58 children who were anti-HBc positive ininfancy converted to being negative by study testing. Thus, onemight conclude that the 6 to 18 month anti-HBc result does notaccurately reflect infection. This persistence of what are mostlikely maternal anti-HBc has been noted by other authors as well.Chotard et al. (1992) found that of 31 anti-HBc positive childrenat one year, 27 (87%) were negative at two years. Lo et al. (1988)found that 4 of 137 (2.9%) infants retained the anti-HBc at twoyears but were negative at three years. The association withinfancy HBsAg positivity is stronger for HBsAg study results. Offive HbsAg positive study results four were already HBsAg positiveat 6 to 18 months (P<0.0001).The results of Coursaget et al. (1986) do not agree with thefinding that most infections occur early. They found that nochildren became infected until the third to fourth year period,following which the rate for HBV events jumped from 1.8% (for the3 to 4 year period) to 8% at the fifth to sixth year period and thecarrier rate rose correspondingly from 0% to 4%. On the otherhand, some researchers have noted that children who became chronic123carriers are the ones who were already identified as being HBsAgpositive at birth or within the first year of life (Resti et al.1990; Poovarawan et al. 1992). Ip et al. (1989) found both earlyand later chronic carrier cases, the later cases occurringfollowing loss of anti-HBs titres to below 10 mIU/ml. Anti-HBcpositivity also arises in children with greater than 10 mIU/ml ofanti-HBs but infections with greater than 100 mIU/ml are rare(Hwang et al. 1990; Schalm et al. 1989; Coursaget et al. 1986;Chotard et al. 1992).Since cases were not tracked over time in this study, it is notknown when the infections occurred and what the prior anti-HBstitres were. However, as the infection rate was fairly consistentover age groups, this would suggest that infection was not due towaning immunity. Therefore it would appear that susceptibility wasmore likely to arise early rather than later in infected children.This possibility is supported by the 14% of inadequately protectedchildren at the 6 to 18 month testing and the fact that some ofthese non-responders became infected. Later susceptibilityfollowing early protection cannot be ruled out because two of thechildren tested at 6 to 18 months who were anti-HBc negative andanti-HBs positive initially, later became infected. But perhaps,as Ip et al. (1989) suggest, what appears to be early response isactually evidence of passive antibodies which have not yet wanedand the lack of an active response becomes apparent with latersusceptibility. This is when other factors such as those which arerelated to the 'Mother's age', 'Mother's years in birth country'and 'Number of contacts' may possibly take on more importance.1249. PROTECTIVE EFFICACY OF HB VACCINEUsing the formula: (expected rate - observed rate)/expected rate,the overall protective efficacy l of HB vaccine in this BritishColumbia Neonatal Program in preventing infections was 89.1%, andin preventing the carrier state was 94.2%. These rates are similarto those found by other studies after 3 to 5 years except for Ip etal. (1989) who found a vaccine efficacy rate of 80% for the carrierstate after 3 years (for the group on a similar vaccine schedule)and Coursaget et al. (1986) who found a vaccine efficacy rate of67.2% after 5 to 6 years. For the current study expected estimateswere taken from control groups cited in the literature and were asfollows:1. For infants of HBsAg and HBeAg positive mothers: 96% forinfection and 85% for the carrier state (Stevens et al., 1979).2. For infants of HBsAg positive and HBeAg negative mothers dataare less adequate because of low numbers but the rates usedwere 25% for infection (Zanetti et al. 1982) and 19% for thecarrier state (Theppisai et al. 1989).Calculations were based on the proportions of children with motherswho were HBeAg positive or HBeAg negative as follows:Infection vaccine efficacy rate =(0.96 * 182(HBeAg+) + 0.25 * 406(HBeAg-))/588 = 0.47(0.47 - 0.051)/0.47 = 0.891'The term vaccine 'protective efficacy' is commonly used inthe literature to compare the outcomes of disease between use ofvaccine and no use of vaccine. Although the term 'protectiveeffectiveness' may be more appropriate for this effectivenessstudy, the conventional term is used.125Carrier vaccine efficacy rate =(0.85*182(HBeAg+) + 0.19*406(HBeAg-))/588 = 0.394(0.394 - 0.023)/0.394 = 0.942A sensitivity analysis using lower expected rates indicated thatthe vaccine efficacy remained high. Using values of 0.65 for HBeAgpositive mothers and 0.10 for HBeAg negative/HBsAg positive mothersthe vaccine efficacy would be 81%.10. SUMMARYThe questions this research attempted to answer were:1. What is the HBV attack rate over time?2. What are the anti-HBs titres and how do they change over time?3. Is there an association between attack rates and anti-HBstitres?4. How are the HBV attack rates and anti-HBs titres affected byvarious demographic, program intervention, and exposurefactors?The attack rate for HBV infections was 5.1% and the carrier ratewas 2.3% over seven years. These rates were not associated withage and were similar to those found in other long-term studies.Anti-HBs seropositivity overall was 87.9% The geometric mean anti-HBs level was 79.7 mIU/ml and declined from 272 mIU/ml to 23 mIU/mlin two year old versus seven year old children. An increase to 106mIU/ml in eight year old children was seen which can be explainedby late initiation of the HB vaccine series and additional doses.No association was found between the study infection results andinfant anti-HBs titres but this was on the basis of only 183children. The relatively stable infection rate does not suggestwaning immunity but rather early susceptibility which may be due tolack of active response to the HB vaccine or some other reason.126Lack of early protection is consistent with the 14% non-responserate in infancy.The best group of predictors for infection were the mother's HBeAgstatus, the timing of dose one, the number of years the motherspent in her birth country and the mother's age. The last twovariables are probably related to cultural or social factors or maybe indirectly measuring a factor related to the mother's anti-HBestatus. The mother's HBeAg status and the timing of dose one werethe best predictors for the HBsAg carrier state. The age of thechild, number of doses of HB vaccine, timing of the first dose andthe number of months the child spent abroad were the factors whichwere significantly associated with anti-HBs titres.According to the multivariate results children were more likely tobe infected if their mothers were HBeAg positive, the first dose ofHB vaccine was delayed, the mother spent a longer time in her birthcountry and/or if the mother was younger. Children were morelikely to become a carrier if their mother was HBeAg positiveand/or if they had a delayed first dose of vaccine. Children whowere younger, had more doses of vaccine, had a delayed first doseand/or had travelled abroad were more likely to have higher titresof anti-HBs.The protective efficacy of the HB vaccine found in this study was89% for infection with the HBV and 94% for the carrier state.12711. RECOMMENDATIONS1. There is no evidence from these data of the need for additionaldoses of vaccine up to age eight. Therefore a booster dose forthe children studied, after the six month dose, is notrecommended at this time.2. A delay in the first dose resulted in increased infections andit is therefore recommended that efforts are made to assurethis dose is given within the first three days after birth.3. Although the timing of dose one was found to be critical, itcould not be concluded from these data that HBIG was notimportant and therefore it is recommended that HBIG at birthbe continued.4. The timing of dose two of HB vaccine later than two monthsafter dose one was not found to be associated with increasedinfections; therefore, this dose could be incorporated with theregular immunization schedule at two months of age ifadministratively more feasible.5. Identification of non-responders to the HB vaccine within oneor two months after the booster dose at six months followed bya supplemental dose is advisable because of the 14% non-response rate found at 6 to 18 months. This may, however, notbe a cost effective endeavour.12. RECOMMENDATIONS FOR FURTHER STUDYIt is recommended that a similar study be undertaken in three tofive years from the last collection of blood specimens to look forevidence of infection and to determine whether a booster dose maybe needed. Although this study found decreasing anti-HBs titres128over time there was no evidence that this increased susceptibility.However, we cannot be sure that waning immunity will not occur overa longer term and therefore for the protection of these children,who are continuously exposed to the HBV by contact with the mother,further follow-up should be considered. Since infections are insome way related to younger mothers who may be new immigrants,consideration should be made to include in future studies variablessuch as the mother's anti-HBe status which may help to explain thisassociation.129BIBLIOGRAPHYAnonymous, Symposium on Hepatitis B. Can Dis Wkly Rep Nov. 21,1987;13(46):207-10.Assateerawatt A. Tanphaichitr VS. Suvatte V. In-ngarm L.Immunogenicity and Protective Efficacy of Low Dose RecombinantDNA Hepatitis B Vaccine in Normal and High-Risk Neonates. AsianPac J Allergy Immunol 1991;9(2):89-93.Assateerawatt A. Suvatte V. Tanphaichitr VS. Long Term Efficacy ofHepatitis B Immunoprophyaxis in Neonates at Risk: Using DifferentVaccine and Schedule. J Med Assoc Thai 1992;75(6):328-36.Ballem PJ. Stout TD. Elliot M. Buskard NA. Review of a Province-Wide Program for the Prevention of HBV Infection in Infants ofCarrier Mothers (British Columbia). Abstract from The AmericanSociety of Hematology Annual General Meeting, August 1987.Barker LF. Gerety RJ. Lorenz DE. Rastogi SC. and Seligman EB.Biological Standardization in Viral hepatitis. In:G. Vyas, S.Cohen and R. Schmid [Eds], Viral Hepatitis:...Proceedings of the Second Symposium on Viral Hepatitis, University of California, San Francisco Philadelphia: Franklin Institute Press, 1978:581-7.Beasley RP. Hwang LY. Lee GCY. et al. Prevention of PerinatallyTransmitted Hepatitis B Virus Infections with Hepatitis B ImmuneGlobulin and Hepatitis B Vaccine. The Lancet 1983;ii(8359):1099-1102.Beasley RP. Trepo C. Stevens CE. Szmuness W. The e Antigen andVertical Transmission of Hepatitis B Surface Antigen. Am JEpidemiol 1977;105(2):94-8.Beasley RP. Shiao IS. Stevens CE. Meng HC. Evidence Against Breast-feeding as a Mechanism for Vertical Transmission of Hepatitis B.The Lancet 1975;2:740-1.Benenson AS. [Ed.] Control of Communicable Diseases in Man [15thEdition]. Washington, DC:American Public Health Association,1990, 202.Chotard J. Inskip HM. Hall AJ. et al. The Gambia HepatitisIntervention Study: Follow-Up of a Cohort of Children Vaccinatedagainst Hepatitis B. J Infect Dis 1992;166:764-8.Courouce AM. Quantification of post-vaccination humoral immunity,In:P. Coursaget and MJ. Tong [Eds], Progress in Hepatitis BImmunization. London: John Libbey & Company, 1990, 67-72.Coursaget P. Yvonnet B. Chiron JP. Diop-Mar I. Long term efficacyof hepatitis B vaccine in infants from an endemic area. AnnVirol (Institute Pasteur) 1986;137E:221-4.130Coursaget P. Yvonnet B. Chotard J. et al. Seven-Year Study ofHepatitis B Vaccine Efficacy in Infants from an Endemic Area(Senegal). The Lancet 1986;ii(8515):1143-5.Davis LG. Weber DJ. Lemon SM. Horizontal Transmission of HepatitisB Virus. The Lancet, 1989;1(8643):889-93Eble K. Clemens J. Krenc C. et al. Differential Diagnosis of AcuteViral Hepatitis Using Rapid, Fully Automated Immunoassays. J MedVirol 1991;33:139-150.Esteban JI. Genesca J. Esteban R. et al. Immunoprophylaxis ofPerinatal Transmission of the Hepatitis B Virus: Efficacy ofHepatitis B Immune Globulin and Hepatitis B Vaccine in a Low-Prevalence Area. J Med Virol 1986;18(4):381-91.Evans AS. [ed.] Viral Infections of Humans Epidemiology and Control [3rd Ed.]. New York and London:Plenum Medical Book Company,1989, 362, 366, 368.Farmer K. Gunn T. Woodfield DG. A combination of hepatitis Bvaccine and immunoglobulin does not protect all infants born tohepatitis B e antigen positive mothers. N Z Med J 1987;100:412-4.Franks AL. Berg CJ. Kane MA. et al. Hepatitis B Virus InfectionAmong Children Born in the United States to Southeast AsianRefugees. N Engl J Med 1989;321(19):1301-5.The Gambian Study Group. Hepatitis B vaccine in the expandedprogramme of immunisation: The Gambian experience. The Lancet1989;1(8646):1057-9.Garrison MW. Baker DE. Therapeutic advances in the prevention ofhepatitis B: Yeast-derived recombinant hepatitis B vaccines. DICPAnn Pharmacother 1991;25:617-27.Ghendon Y. WHO strategy for the global elimination of new cases ofhepatitis B. Vaccine 1990;8(Suppl):S129-33.Hadler SC. Francis DP. Maynard JE. et  al. Long-Term Immunogenicityand Efficacy of Hepatitis B Vaccine in Homosexual Men. N Engl JMed 1986;315(4):209-14.Hayashi J. Kashiwagi S. Nomura H. et al. The control of hepatitisB virus infection with vaccine in Japanese nursery schools. AmJ Epidemiol 1987;126(3):474-9.Holland PV. Hepatitis B Surface Antigen and Antibody (HBsAg/Anti-HBs) In: RJ. Gerety [Eds], Hepatitis B. Toronto: Academic Press,1985, 10.Horowitz MM. Ershler WB. McKinney WP. Battiola RJ. Duration ofImmunity After Hepatitis B Vaccination: Efficacy of Low-DoseBooster Vaccine. Ann Intern Med 1988;108(2):185-9.131Hwang LY. Lee CY. Beasley RP. Five-Year Follow-up of HBVVaccination with Plasma-derived Vaccine in Neonates: Evaluationof Immunogenicity and Efficacy against Perinatal Transmission.In Hollinger FB. Lemon SM. Margolis HS. [Eds]:Viral Hepatitis andLiver Disease. Baltimore: Williams & Wilkins, 1990, 760.Ip HMH. Lelie PN. Wong VW. et al. Prevention of Hepatitis B VirusCarrier State in Infants According to Maternal Serum Levels ofHBV DNA. The Lancet 1989;i(8635):406-10.Jilg W. Schmidt M. Deinhardt F. Persistence of specific antibodiesafter hepatitis B vaccination. J Hepatol 1988;6:201-7.Kelly JP. Rosenberg L. Kaufman DW. and Shapiro S. Reliability ofPersonal Interview Data in a Hospital-Based Case-Control Study.Am J Epidemiol 1990;131(1):79-90.Lee KS. Lee H. Moon SJ. et al. Hepatitis B Vaccination of NewbornInfants: Clinical Study of New Vaccine Formulation and DoseRegimen. Hepatology 1987;7(5):941-5.Lo KJ. Lee SD. Tsai YT. et al. Long-Term Immunogenicity andEfficacy of Hepatitis B Vaccine in Infants Born to HBeAg-PositiveHBsAg-Carrier Mothers. Hepatology 1988;8(6):1647-50.Mannucci PM. Zanetti AR. Gringeri A. et al. Long-termImmunogenicity of a Plasma-Derived Hepatitis B Vaccine in HIVSeropositive and HIV Seronegative Hemophiliacs. Arch Intern Med1989;149:1333-7.Maynard E. Kane M.A. Hadler S.C. Global Control of Hepatitis BThrough Vaccination: Role of Hepatitis B Vaccine in the ExpandedProgramme on Immunization. Clin Infect Dis May-June 1989;11(Suppl3):S574-8.McCready JA. Morens D. Fields HA. et al. Evaluation of enzymeimmunoassay (EIA) as a screening method for hepatitis B markersin an open population. Epidemiol. Infect. 1991;107:673-684.Monna T. Kuroki T. Oka H. et al. Prevention of VerticalTransmission of HBV by Administration of Hepatitis B VaccineCombined with HBIG and Long-term Follow-up of HBsAb Titer. OsakaCity Med J 1988;34(1):9-17.Moyes CD. Milne A. Waldon J. Very Low Dose Hepatitis B Vaccinationin the Newborn: Anamnestic Response to Booster at Four Years. Jof Med Virol 1990;30:216-8.Oppenheim AN. Questionnaire Design, Interviewing and AttitudeMeasurement [New Edition] . New York:Pinter Publishers, 1992, 145.Ostrow DH. Edwards B. Kimes D. et al. Quantitation of hepatitis Bsurface antibody by an automated microparticle enzymeimmunoassay. J Virol Methods 1991;32(2,3):265-76.132Polakoff S. Vandervelde EM. Immunisation of neonates at high riskof hepatitis B in England and Wales: national surveillance. BMJJuly 1988;297(6643):249-52.Poovorawan Y. Sanpavat S. Pongpunglert W. et al. Long-term efficacyof hepatitis B vaccine in infants born to hepatitis B e antigen-positive mothers. Pediatr Infect Dis J. 1992;11(10):816-21.Poovorawan Y. Sanpavat S. Pongpunlert W. et al. Comparison of arecombinant DNA hepatitis B vaccine alone or in combination withhepatitis B immune globulin for the prevention of perinatalacquisition of hepatitis B carriage. Vaccine 1990;8(Suppl):S56-9;discussion S60-2.Poovorawan Y. Sanpavat S. Pongpunlert W. et al. Protective Efficacyof a Recombinant DNA Hepatitis B Vaccine in Neonates of HBeAntigen-Positive Mothers. JAMA 1989;261(22):3278-81.Resti M. Azzari C. Rossi ME. et al. Five-year follow-up ofvaccination against hepatitis B virus in newborns vaccinated witha reduced number of doses. Vaccine 1991;9:15-18.Rothman KJ. Modern Epidemiology. Boston and Toronto: Little Brownand Company, 1986, 109.Schalm SW. Mazel JA. de Gast GC. et al. Prevention of Hepatitis BInfection in Newborns Through Mass Screening and DelayedVaccination of All Infants of Mothers With Hepatitis B SurfaceAntigen. Pediatrics 1989;83(6):1041-7.Sehgal A. Gupta I. Sehgal R. Ganguly NK. Hepatitis B Vaccine Aloneor in Combination with Anti-HBs Immunoglobulin in the PerinatalProphylaxis of Babies Born to HBsAg Carrier Mothers. Acta Virol. 1992;36:359-366.Shapiro CN. McCaig LF. Gensheimer KF. et al. Hepatitis B virustransmission between children in day care. Pediatr Infect Dis J1989;8(12):870-5.Shortell SM. Richardson WC. Health Program Evaluation, St. Louis:The C.V. Mosby Co., 1978, 78.Spronk AM. Schmidt L. Krenc C. et al. Improvements in Detection ofAntibody to Hepatitis B Core Antigen by Treating Specimens withReducing Agent in an Automated Microparticle Enzyme Immunoassay.J Clin Microbiol 1991;29(3):611-6.Stevens CE. Taylor PE. Tong MJ. et al. Yeast-Recombinant HepatitisB Vaccine Efficacy with Hepatitis B Immune Globulin in Preventionof Perinatal Hepatitis B Virus Transmission. JAMA1987;257(19):2612-6.Stevens CE. Toy PT. Tong MJ. et al. Perinatal Hepatitis B VirusTransmission in the United States. JAMA 1985;253(12):1740-5.133Stevens CE. Neurath RA. Beasley RP. Szmuness W. HBeAg and Anti-HBeDetection by Radioimmunoassay; Correlation with VerticalTransmission of Hepatitis B Virus in Taiwan, J Med Virol 1979;3:237-41.Streiner DL. Norman GR. Health Measurement Scales. New York: OxfordUniversity Press, 1992, 15.Tajiri H. Nose 0. Shimizu K. et al. Prevention of Neonatal HBVInfection with the Combination of HBIG and HBV Vaccine and ItsLong-Term Efficacy in Infants Born to HBeAg Positive HBV CarrierMothers. Acta Paediatr Jpn 1989;31:663-8.Taylor R. Montaville B. Levy S. et al. Hepatitis B Infection inVanuatu: Age of Acquisition of Infection and Possible Routes ofTransmission. Asia Pac J Public Health 1989;3(3):205-12.Theppisai U. Thanuntaseth C. Chiewsilp P. Siripoonya P. Long-TermImmunoprophylaxis of Hepatitis B Surface Antigen Carrier inInfants Born to Hepatitis B Surface Antigen Positive MothersUsing Plasma Derived Vaccine. Asia Oceania J Obstet Gynaecol 1989;15(2):111-115.Theppisai U. Thanuntaseth C. Chiewsilp P. Siripoonya P. Preventionof hepatitis B infection in infants born to hepatitis B carriermothers: low dosage vaccination. Int J Gynecol Obstet1990;32(4):353-7.Tong MJ. Hepatitis B Vaccination of Neonates and Children. Am J Med1989;87(Suppl 3A):33S-35S.Vajro P. Fontanella A. Avellino N. Greco L. Auricchio S. Breastfeeding enhances the clearance of HBsAg in infants with hepatitisB virus infection [Abstract]. Gastroenterology 1985;88(5):1702.Wainwright RB. McMahon BJ. Bulkow LR. et al. Duration ofImmunogenicity and Efficacy of Hepatitis B Vaccine in a YupikEskimo Population. JAMA 1989;261(16):2362-6.Wainwright RB. McMahon BJ. Bulkow LR. et al. Protection Provided bythe Hepatitis B Vaccine in a Yupik Eskimo Population Seven-YearResults. Arch Intern Med 1991;151:1634-6.Waters JR. Universal Prenatal Screening For Hepatitis B, Alberta,1985-1988. Can Dis Wkly Rep Feb. 1989; 15(6):29-32.West DJ. Calandra GB. Ellis RW. Vaccination of Infants and ChildrenAgainst Hepatitis B. Pediatr Clin North Am June 1990; 37(3):585-97.Woo D. Cummins M. Davies PA. et al. Vertical transmission ofhepatitis B surface antigen in carrier mothers in two west Londonhospitals. Arch Dis Child 1979;54:670-5.134World Health Organization, Report of WHO Consultation on Maternal and Perinatal Infections, Division of Family Health,WHO/MCH/91.10, Distr.: General(e).Zanetti AR. Ferroni P. Magliano EM. et al. Perinatal Transmissionof the Hepatitis B Virus and of the HBV-Associated Delta AgentFrom Mothers to Offspring in Northern Italy. J Med Virol 1982;9:139-48.Zhu Q. Duan S. Xu H. A Six-Year Survey of Immunogenicity andEfficacy of Hepatitis B Vaccine in Infants Born to HBsAgCarriers. Chinese Med J 1992;105(3):194-8.135APPENDIX 1STUDY DOCUMENTSPhysician information letter 137Parent information letter 138Telephone introduction guide 139Consent for participation 140-1Questionnaire 142-5Questionnaire guide for interviewers 146-8Laboratory test report form 149Results of tests - form lettersfor parents and physicians 150-59136Dear Physician:Some time ago, your patient born toreceived hepatitis B vaccine after birth as part of the Ministry of Health, NeonatalHepatitis B Vaccine Program. We will soon be sending a letter to his/her parent(s)requesting participation in a study on the effectiveness of hepatitis B vaccine given toneonates.The goals of the study are to measure the HBV infection rate in children who receivedhepatitis B vaccine during 1984 through 1989 and to find out how many of these childrenstill have antibodies to HBV. Since very few studies have been done on the long-termeffectiveness of hepatitis B vaccine in infants this information will be valuable in terms ofassessing the need for boosters.We will be arranging for a finger-prick blood sample to be taken from those children whoseparents agree to their participation in the study. The blood samples will be tested forpresence of anti-HBc, anti-HBs and HBsAg. Parents will also be asked to answer a fewquestions concerning risk factors their child may have for hepatitis B infection. •As well as advising parents of the blood test results, we will request consent from theparents to inform the child's physician of the results so that the necessary follow-up maybe provided.Enclosed is a copy of the letter we are sending to your patient's parent(s). We wouldappreciate hearing from you in the event of an address and/or telephone number change forthis child. Also, if circumstances exist with this family which would suggest that contactwith these parents is not advisable, we would appreciate hearing from you. If we don'thear from you within a week of your receipt of this letter, we will assume you have noobjection to us contacting these parents.If you would like to speak with us or have any questions about this study, you may call Dr.Richard Mathias, Principal Investigator, at 822-2772, Marian Tomm-Pastore, StudyCoordinator, at 879-7551 local 307 or myself at 660-6063. Please refer any address ortelephone number changes to Marian Tomm-Pastore as soon as possible.erely.;Dr. John FarleyEpidemiologist,B.C. Centre for Dise^Control,Ministry of Heal Preventive Health Services.Long Term Follow-up of Neonates Immunized with Hepatitis B VaccineOr. R.O. kleirdorTHE UNNERSITY OR BRITISH COLUMBIAFacoay d Ma/ideaDsoommani d Hoorn, Cam i EpabaeretlogyJamas Kuhr Building5104 F411141M AvonuoVarecvom. B.C. VC 123TEL: (00411122-2772FAX: (1104) (122:4994Dr. 1 PaleyTHE PFIOVINCE OF BRITISH COLUMBIAMammy a HealthBraun Columbia Comm Mr Moms Control11211 Woo 10th AvenueVan:ovum. B.C. VIZ 1111TEL. (:104) 060.4063Dr. D. PITHE CANAOAN RED CROSS SOCIETYVarcouvar ContraMow Trarartioion Sento4750 011k S77•01Varcouvor. B.C. V6H asTEL: (104) il79.7551Ma. Marian Toirem-PlunowSlip* CoordaiMarTEL. (804)1179-75610/ (SOO 177.7803137Dear Parent:You may recall that after being born your child received several doses of hepatitis Bvaccine to protect him or her from hepatitis B infection.The Ministry of Health, Centre for Disease Control, the Canadian Red Cross Society(Vancouver Blood Centre), and the Department of Health Care and Epidemiology at theUniversity of British Columbia are conducting a study to obtain information on how wellthis vaccine has worked in children who have received it. This information will beimportant in helping make decisions about whether booster doses will be required in thefuture.We would like to enrol your child born^in the study. This research willbe conducted by Dr. R. Mathias (a Communicable Disease Specialist at U.B.C.), Dr. DavidPi (Deputy Medical Director, Canadian Red Cross Society). and Dr. I. Farley(Epidemiologist, Ministry of Health).Children taking part in the study will have a finger-prick blood sample taken. The parentswill be asked to answer a few questions that will help us to better understand hepatitis Binfections in children.A member of the study team will contact you in the near future to offer you moreinformation. Please understand that your participation in this study is entirely voluntary.If you would like to contact us, vou may call Dr. Richard Mathias at S22-2772 or MarianTomm-Pastore, Study Coordinator, at 879-7551 local 307. We will gladly answer yourquestions. If .ou have not h• om us in a week, we would appreciate a call from you at879-7551 Iq al 307 as w^v not have been able to locate a phone number for vou./311 rce7 v,13r. John FarleyEpidemiologist,B.C. Centre for^ase Control,Ministry o^ealth.Long Term Follow-up of Neonates Immunized with Hepatitis B VaccineDr. R.O. 5541NsaTHE uNivERsirY Of Barr COLUMINAorar, d 1.144PaneC445,4v,40 414443h Cre I E.R6wricacqyJars 4.40+44 lv14IgSad Falnwe Arer444Van:awe'. S.C. VSTTEL moo 622 -2772FAX (ON) 1274004Dr. 1 FerryTHE PROVINCE OF BRRISN CC4.U1ARIAMnouv d 14•44+114443 Coort. Colt, RR Noes. Como(QS Wen 1 AvenuaVarcouvw. B.C. VSZTEL 1004) 305-0333Dr. D. PITHE CANADIAN RED CROSS SOCIETYV4recrwer CentroBbd Tranevsce SoNce4750 0.4 StromVarcouver. B.C. V1H 250mod) s71-7561lArtrian Tenon-PasterStudy Co-onlinffircrTEL 0:104) 11711.7351or (MAI 177. 7300138TELEPHONE INTRODUCTIONLONG-TERM FOLLOW-UP OF NEONATES WHO HAVE BEEN IMMUNIZED WITHHEPATITIS VACCINE"Hello, my name is ^Is this the ^ residence?I am calling about the Hepatitis B vaccine study.Did you receive a letter in the mail inviting your child ^to be part of this study?Do you have any questions about the study?We would very much like to include ^ in the study.It will help us to understand how well the hepatitis B vaccine hasworked in children who were immunized at birth.Many short term studies have shown that the vaccine has workedvery well over a short term. We would like to find out if this istrue for periods three years or longer.Would you be willing to allow^ to be a part of thisstudy? It would involve answering a few questions about riskfactors for getting hepatitis B and taking a finger blood sample.Would you be willing to come to   with you child forthis interview?(make appointment if agree to participate)Thank you for agreeing to participate in the study."Options: clinic appointmenthome visit appointmentArrange for interpreter to telephone if language is a problem.Also advise Marian if it is expected there will be difficulty withcommunication at the appointment.If the parent is hesitant about participating, ask: "Would youlike to talk to your doctor about the study and then we'll callyou back?"Give a name and phone number to contact in case of questions.The Red Cross number for the study is 879-7551 local 307.If the parent refuses to participate, please ask: "What is theusual language spoken in your home?" Record the language on theprintout and forward to Marian (Study Coordinator). 139NEONATAL HEPATITIS B VACCINE STUDY CONSENTSTUDY TITLE: LONG-TERM FOLLOW-UP OF NEONATES WHO HAVEBEEN IMMUNIZED WITH HEPATITIS B VACCINEThe purpose of this study is to establish whether the hepatitis B vaccine which your child received atbirth has protected him/her from infection. This can be assessed by testing for the presence of certainmarkers in the blood. Your child does not need to have had symptoms to have experienced an infection.The test will also determine if antibodies to hepatitis, stimulated by the vaccine, are still circulatingin your child's blood.We are requesting your permission to obtain a finger blood sample from your child. Pain from a finger-prick blood sample is mild and short lived. The blood sample will be tested only for evidence of past orpresent hepatitis B infection and for the presence of hepatitis B antibodies. The specific markersinclude anti-HBs, anti-HBc and HBsAg.Results of the tests on your child's blood will be sent to you with an explanation of their meaning. Somefollow-up by your physician may be recommended if your child's blood shows evidence of hepatitis Binfection. Information on blood test results will be released to your physician with your consent only.All information from the questionnaire and on blood test results will be STRICTLY CONFIDENTIAL.The study records will be available only to members of the study team. Your child will not be identifiedin any public report of the study.Your participation in this study is COMPLETELY VOLUNTARY. Your refusal will in no way affectyour or your child's future medical care.With your consent, we would also like to contact you in the future for similar follow up.Investigators:Dr. R. Mathias, Associate Professor, Department of Health Care and Epidemiology, U.B.C.Dr. D. Pi, Deputy Medical Director, Canadian Red Cross, Vancouver Blood Center, and ClinicalInstructor, Department of Pathology, U.B.C.Dr. J. Farley, Epidemiologist, B.C. Centre for Disease Control, Clinical Assistant Professor, Departmentof Health Care and Epidemiology, U.B.C.Marian Tomm Pastore, M.Sc. student, Department of Health Care and Epidemiology, U.B.C.If you have any questions about the study, please contact Dr. Mathias, Principal Investigator, at 822-2772, or Marian Tomm Pastore, Study Coordinator, at 879-7551 Loc. 307.Page 1 of 2140I CONSENT FOR MY CHILD TO PARTICIPATE IN THE HEPAIH1S B VACCINESTUDY DESCRIBED ABOVE. THE STUDY HAS BEEN CLEARLY EXPLAINED TOME AND I UNDERSTAND THE INFORMATION PROVIDED. I ACKNOWLEDGETHAT I HAVE RECEIVED A COPY OF THE CONSENT FORM.Child's family name:^ givennarre^Parent or Guardian Signature: ^Please print full name^Relationship to child: Date^Witness Signature:Print full name^Date^I CONSENT TO BEING CONTACTED IN THE FUTURE FOR SIMILAR FOLLOW-UP ON MY CHILD.IN THE EVENT OF A CHANGED ADDRESS, MY CHILD'S PERSONAL HEALTH NUMBER MAY BEUSED FOR LOCATING PURPOSES ONLY.Child's family name:^ given name:^Patent or Guardian print nave^Relationship to child:Date^Witness Signature:^Print full name:Date:^I CONSENT TO THE RELEASE OF MY CHILD'S HEPATITIS B BLOOD TEST RESULTS TO MYCHILD'S PHYSICIAN, DR^ , FOR THE PURPOSE OF FOLLOW-UP.Child's family name: given name^Parent or Guardian Signature^Please print name:^Relationship to child: Date^Witness Signature:^Print full name:Date:^Page 2 of 2141YY MM DDYY MM DDI.D. NUMBER:LONG-TERM FOLLOW-UP OF NEONATES WHO HAVE BEENIMMUNIZED WITH HEPATITIS B VACCINEINTERVIEW QUESTIONNAIRE FOR PARENTS OR GUARDIANSSECTION AInterviewer Instructions:Complete Date of interview and identifying information in SECTION A.PLEASE PRINT CLEARLYDate of interview: YY MM DDChild's given name^  family name^Male ^ Female ^ Date of birth I^ YY MM DDMother's given name ^last^  D.O.B.(if different from child's)Father's given name ^last^  D.O.B.(if different from child's)Address^ Ph. No.^No.^Street^city^ postal codeCountry of birth for: [round off years to the nearest year]Mother ^ Length of residence in country of birth:^yearsFather  Length of residence in country of birth:^yearsUsual language at home ^ Has child moved since birth? Yes NoPhysician's name^  Ph. No.^City^COMPLETE THIS SECTION AT THE END OF THE INTERVIEWCan you give us the name of an alternate contact who could assist us inlocating you in the event that:a) you move before we can get the results of this study to you orb) we would like to reach you for a follow-up to this study in a few years?Name: last^  first^Phone number:  Relationship: Or, may we have your child's Personal Health Number? ^Your PHN will be kept confidential and will not be used for any otherpurposes.192I.D. NUMBERSECTION BInterviewer instructions:Complete SECTION B by asking the parent or guardian the questions as stated,except, using the child's name for "this child". Answer by circling thecorrect response as indicated in the sample question unless otherwisedirected in the question. Interviewer instructions are in square brackets.Refer to the questionnaire guide for more complete instructions.Sample question only:Has "this child" had chickenpox?1(a) To your knowledge, has "this child" ever been illwith hepatitis?^ Yes NoIf answered "Yes" for 1(a):1(b) What type was it? [circle the number next tothe correct response]^1^A23^NonA NonB4^Don't know2. Has "this child" ever had jaundice starting afterthe first month of life?3(a) Was "this child" breast fed?[Must be at least 1 week in durationand at least one feeding per day]Yes NoYes NoIf answered "Yes" to 3(a):3(b) What was the duration of breast feeding in months? ^[consider 1 week to equal .25 months]4. Has "this child" experienced any of the following?4(a) Ear piercing^ Yes Nocountry if other than Canada^4(b) Acupuncture Yes Nocountry if other than Canada^4(c) Tattooing^ Yes Nocountry if other than Canada^4(d) Surgical procedure^ Yes Nocountry if other than Canada^4(e) Blood transfusions Yes Nocountry if other than Canada^4(f) Receipt of blood products Yes Nocountry if other than Canada^4(g) Day care^ Yes Nocountry if other than Canada^143Y N Y N ^Y NY NY NY NY NY NY NY NY NY NY N^Y NY N^Y NM FM FM FM FM FM FM FM FY NY NY NY NY NY NY NY N5(a) Other than the mother, does "this child"live with anyone else in the home?^ Yes No[Include only those household members who share the same kitchenas the child and have lived in the home for at least one month.][If answered "Yes" to 5(a), then complete the table below with respect toquestions 5(b) to 5(j) as illustrated.]Sample question only:5(b)^5(c)^5(d)^5(e)^5(f)^5(g)^5(h)^5(i)^5(j)Relationship sex ,^ year hepatitis type year^lah HA^Yearof birth confirmed vaccine rct+k er ^OF  196o^N^_ail.^N^N 111315(b) What is their relationship to "this child"?5(c) What is their sex? [Circle "M" for male and "F" for female]5(d) What is their year of birth? [If uncertain, approximate the year]5(e) Have they had hepatitis? [Circle "Y" for yes and "N" for no]5(f) If so, what kind? [Indicate if A, B, NonA NonB]5(g) In what year? [If uncertain, approximate the year of diagnosis]5(h) Was it laboratory confirmed? [Circle "Y" for yes and "N" for no oruncertain]5(i) Have they had hepatitis B vaccine? [Circle "Y" for yes and "N" for no]5(j) What year was the hepatitis B vaccine received? [Approximate the year ifuncertain)5(b)^5(c)^5(d)^5(e)^5(f)^5(g)^5(h)^5 ( 1 )^5(j)^Relationship sex^Year hepatitis type year^lab HB^Yearof birth confirmed vaccine1446(a) Has "this child" ever been hospitalized?^Yes NoIf answered "Yes" to 6(a):6(b) How many times has "this child" been hospitalized?^6(c) What were the reason(s) for hospitalization?7(a) Has "this child" ever travelled overseas?^Yes No[If answered "Yes" to 7(a) complete the table below withrespect to questions 7(b) to 7(d):]7(b) What countries did "this child" visit on each trip?7(c) How many months was the child abroad?7(d) What was "the child's" age at the time of leaving on the trip?7(b)^ 7(c)^7(d) visited # of months(one week=.25months)Child's age attime of visitin years12345SECTION CTHIS COMPLETES THE QUESTIONNAIRE. THANK YOU VERY MUCH FOR ANSWERING THESEQUESTIONS. WE APPRECIATE VERY MUCH YOUR ASSISTANCE IN THIS STUDY.Respondent: [check] Mother ^ Father ^ Other ^What is their relationship to the child if "other"? Interviewer: Name:[print] Phone #:^Comments:REMEMBER TO ASK THE RESPONDENT FOR AN ALTERNATE CONTACT (PAGE 1)145QUESTIONNAIRE GUIDE FOR INTERVIEWERSLONG-TERM FOLLOW-UP OF NEONATES WHO HAVE BEEN IMMUNIZED WITHHEPATITIS B VACCINEThe purpose of this study is to determine how effective thehepatitis B vaccine has been in preventing infections in thosechildren who were immunized with the hepatitis B vaccine at birth.The purpose of this questionnaire is to determine if there arefactors which explain the occurrence of infection in somechildren. Children who have had an infection may not have had anysigns of illness. A positive blood test for anti-HBc or HBsAgmarkers may be the only indication that an infection has takenplace.To make sure the results of the questionnaire are as accurate aspossible it is important to follow these guidelines:1. Use the same wording as in the questionnaire.2. Use simple clarifications if necessary. (Refer to the guide.)3. Arrange for an interpreter to be available if parents havedifficulty with English.4. Complete every question.SECTION ABe sure to complete the identifying information. This is essentialto allow us to match records and to assist in locating familiesfor future follow-up.In the case of Asian children's names, some children are called bythe family name first, followed by given names. We would considerthe first name the family name as long as it is common to othersin the family, and the rest of the names which are different asthe given names.Parent's length of residence in country of birth should be roundedoff to the nearest number of years. For under one year indicateone year.With regard to physician, it is sufficient to indicated the cityname only unless the physician has recently moved his/her office.Parents may have concerns about giving alternate contactinformation. They may be reassured that this information will bekept confidential and giving this information does not constitutegiving consent to participate in future studies. Parents would beasked to give consent at the time of contact. The purpose is onlyto make it easier to locate them if another study is undertaken.If parents are reluctant to give this information do not pressthem.PLEASE PRINT CLEARLY IN BLOCK LETTERS146SECTION BQuestion 1 "Yes" implies the hepatitis disease was diagnosed by a physicianor health care worker.Question 2 Indicate "Yes" only if the child had jaundice after one month ofage. Do not include jaundice or yellow skin caused by breastfeeding or physiological jaundice due to birth.Ouestion 3 3(a) Indicate "Yes" only if the mother continued breast feedingafter the first week of the child's life.3(b) Duration of breast feeding should include the length of timeat least one feeding per day of breast milk was fed, even ifformula was fed for all other feedings.Question 4 4(a), (b), (c) (d) (e) (f) (g) For ear piercing, tattooing,acupuncture, surgery, receipt of blood products or bloodtransfusions and day care, indicate the country or place wherethis procedure took place if other than Canada, for example, HongKong or Taiwan.4(d) Surgical procedure refers to any procedure that has involvedan incision or suturing.4(f) Blood products include plasma, serum, thrombin, fibrinogen,packed red blood cells and cryoprecipitate. They do not includeimmune globulin, heat-treated plasma protein fraction, albumin orfibrinolysin.4(g) Day care includes any out-of-home, formal, paid-for day carein a group setting, for example, group or family day care. It doesnot include in home baby sitting whether done on a regular oroccasional basis.Question 5 5(b) Include all household contacts who share the same kitchen asthe child and have lived in the same house for at least a month.Contacts should include relatives such as the father,grandparents, cousins, aunts, uncles etc. as well as nanny's,live-in baby sitters and boarders. If more than one grandparentlives in the home, indicate if maternal or paternal.5(d) If the exact date of birth of the contact is not known,indicate the approximate age.5(f) If the exact year of hepatitis is not known, indicate theapproximate year of diagnosis.5(g) If the hepatitis diagnosis was confirmed by laboratory test,circle "Y" for yes and circle "N" for no or uncertain. If thecontact's blood was not tested, and diagnosis was made on thebasis of clinical symptoms or contact with someone else who wasdiagnosed with hepatitis B circle "N".Add more lines or use the back of the page if there are more than8 household contacts.147Ouestion 6 6(a) "Hospitalization" refers to actual admission for overnightstay, not visits to the Emergency Room which are not followed byadmission.6(b) refers to the number of separate admissions, not the lengthof stay.6(c) indicate the reason/s for each separate admission. Add extralines if necessary or use the back of the page.Question 7 "Overseas" implies travel to another continent such as Asia orAfrica. Include Mexico, Central America and South America but donot include the U.S.A. If the child visited three differentcountries, for example, Taiwan, Thailand and Japan on one trip,indicate these countries on the same line. If one country wasrevisited on another trip, include that country again on aseparate line so that each line id trip specific. Include anyextended periods of time the child may have lived abroad. If indoubt whether a country or place qualifies, include it. Length oftime should be recorded in months using .25 (one week), .5 and .75as portions of a month. Record the age of the child on the lastbirth date at the time of leaving on the trip. Please use the backof the page if there were more than 5 trips.Under comments, include any special circumstances such asdifficulty communicating or concerns about the responses obtained.Also indicate if a satisfactory blood sample was not obtained andwhy. At least 2 microtainer tubes should be filled.If you have any questions or concerns about the questionnaire orprocedures please call Marian Tomm Pastore at ph.^ or Dr.Rick Mathias at ph. ^Thank you very much for your assistance with this study.148^11^1LAB TEST REPORT FORMLONG-TERM FOLLOW-UP OF NEONATES WHO HAVE BEEN IMMUNIZED WITHHEPATITIS B VACCINESECTION A^ I.D. NUMBER:Child's given name^Child's D.O.B. YY MM DDMother's given name^Mother's D.O.B. YY MM DDfamily namefamily name ^Date blood sample take by ^YY MM DDLocation of sample taking ^ ph#^Section BYY MM DDDate blood sample received at CRCSTestReading C/ODateDD/MM/YYTech.InitialResetReadingResetC/ODateDD/MM/YYTech.InitialPoshNeg1.Anti-HBc(IMX) s/co2.Anti-HBs(IMX) mIU/ml3.HBsAg (IMX)S/NIs the sample haemolysed?^circle)^Y^NTechnologist comments:Note to person taking blood sample: The minimum amount of blood required is 1.2 mls. To get this amount, fill 2microtainers to the top. If blood is flowing freely, fill a third microtainerin case retesting is needed. Write I.D. number on the microtainer, checkingwith the questionnaire I.D. number. Complete part A of the LAB TEST REPORTFORM. Place form and microtainer into the baggie and then into a deliverycontainer. Deliver to the Red Cross Blood Transfusion Centre by 9 am the nextmorning or as soon as possible. Areas outside the Lower Mainland will requirespecial arrangements for delivery. Blood samples should be kept at about 4degrees C until delivered. Do not freeze or allow them to get too hot.149Long Term Follow-up of Neonates Immunized with Hepatitis B VaccineOr. an WatilasrrIE tAWYEASITY OF WWI esw COLUSUSIA*DAT of WOW*Corairnoni of W.. C. a Egsdenvoicq,Jam.* MawrlirqSIM Fanny* ArroyoVorfcconor. E.C.VC l23TEL. (354) s2t-E7T2FAX. (004) 4224904Or. J. FmkriTfE PROVIPCS OF IIIRMISO. octutaguMowry er NoahBolan Cohort. Coon for Duras* Contra421 Wort 'OM ArrowsV erceanw. B.C. VS2TEL: )00N)Or.1).14TIE CANADtm RED CAMS 9CCETYVonerrovor ComoWoos irsofkrofoo SowsOSO Oak StmaV arcauret, B C VtIN 2toTEL (11047$711.7561Sis Avian T4 IVI.P14101%study Co-enkuraTEL 004)117,7561Or (IDS) S77.7633ADear Ms.We would like to thank you for your child's participationin the hepatitis B vaccine study. Your child, ^born ^ has the following blood test results:Anti-HBc: NegativeAnti-HBs: PositiveHBsAg: NegativeThese results indicate that your child has an adequateprotective response to the hepatitis B vaccine.A. copy of the results will be sent to your physician.Please keep this record of results in your child'simmunization booklet. It may be necessary for referencewhen your child is offered the hepatitis B vaccinethrough the grade six program.Yours sincerely,Dr. Richard MathiasPrincipal Investigator150Long Term Follow-up of Neonates immunized with Hepatitis B VaccineOr. R.O. WeiseriE LINNIRSiTe of BArreu4 col-MONANewly ot MeekOnoOkenrent d teak Caro 6 EneterrieogyJanes Mather fluiblevaNOG Fanner AvenueVIOCCanet. I.C. WIT 1Z,TEL. (804) 022.Z772FAX. (80') K2 -amOr. .I. Forte!114 PROVIRCE^COLUIIIILAWebby of MeathVito Columba Centre ter Nowa Cork%WI West ttkt AvenueVarsouver, I.C. VSZ ILSTELMOO RIORCOSOr.O. PITkE CANADIAN RED CAMS 30CIETYVs/cower Centro'lbw Trio.4 ,14.0. Sevres4752 Oak StrewVonore. I.0 VON VIITEL (804)1171.1561M . Marten TenviektelsoeauT COevellnrorTEL. Mk 111711.7$61ar (8 4) v7.7IesBDear Ms.We would like to thank you for your child'sparticipation in the hepatitis B vaccine study. Yourchild,  , born  has the following bloodtest results:Anti-HBc: PositiveAnti-HBs: PositiveHEsAg: NegativeThese results indicate that your child has been infectedwith the hepatitis B virus in the past, but is not acarrier and now has an adequate protective response.Further immunization is not indicated.A copy of the results will be sent to your physician.Please keep this record of results in your child'simmunization booklet. It may be necessary for referencewhen your child is offered the hepatitis B vaccinethrough the grade six program.Yours sincerely,Dr. Richard MathiasPrincipal Investigator151Long Term Follow-up of Neonates immunized with Hepatitis B VaccineDr. OLD. lisitiasTwE uttAtEnrry Cr BATT Dm COLUMBIAracuity 0 IA.:flareCmaaratera at l4•021 Cart 4 EsmarmtegyJarmo W0.0 *lingMN 'whim ArenasV arroamat. B.C. VET lbTEL (604)FAX: wwca.00dO►. 1 PansyniE rosovrck Oc awns++ COLUMBIAMinotry d IMAMBMW C.:APO. Celan 101 Camas. Coaxallb Bic lOP Amax.Vara:mem. B.C, XSZ ILATEL(010001041010Or. D. PIIME CIV4ADAN RED CROSS SOCIETYVara:ewer Coveetas Trtralusan Stows4750 00 3~VOlCCINOt, S.C. VSFITEL. Ma, 1179.1361Y. Marlan Twarn•PatilereCoettlinaKeTEL. (604) 1171.7SSItx (1104) 677.711103CDear Ms.We would like to thank you for your child's partici-pation in the hepatitis B vaccine study. Your child,^ , born ^ has the following bloodtest results:Anti-HBc: NegativeAnti-HBs: NegativeHBsAg: NegativeThese results indicate that your child's response to thehepatitis B vaccine is not considered adequate atpresent. Please see your physician concerning theseresults.A copy of the results will be sent to your physician.Please keep this record of results in your child'simmunization booklet. It may be necessary for referencewhen your child is offered the hepatitis B vaccinethrough the grade six program.Yours sincerely,Dr. Richard MathiasPrincipal Investigator152Long Term Follow-up of Neonates Immunized with Hepatitis B VaccineOr. R.CI. Maras114 uNivERsfTY Of IR(T ^001.1.114111AFOICIAr or IJoelanoDopoolYort of HooN4 Cr. 6 Esdorrodopy.111117•M Witto• MA %+pSOCA FonAro Avow..Vareouvor. S.C. VC ILITEL. 16041 622.2772FAX. (604) 572 490441)Cs. J. FarleyTHE pooyiNc€ OF BRITISH COI-061UOdinsfy of Hootlitiorar Comma Come for Oa•ra Cana826 Woo ,06 Avow.VoiceA.m. B.C. VSZ ILITEL(634)6W-02641Or. O. PIToE CANADIAN RED CROSS SOCIETYVorcc._ ComoStood Tranglyg.0”4750 Oat StromVarccAmor, I C VON V*TEL: mocio+nvI. II•rion Temns.Pairargluey CoerellnamfTEL .101) 11757$111a (IN) 177•7903DDear Ms.We would like to thank you for your child's partici-ation in the hepatitis B vaccine study. Your child,^ , born ^ has the following bloodtest results:Anti-HBc : NegativeAnti-HBs : Insufficient blood to testHBsAg : Insufficient blood to testThese results indicate that your child has not beeninfected with the hepatitis B virus. However, sufficientblood was not obtained to measure response to thevaccine.A copy of the results will be sent to your physician.Please keep this record of results in your child'simmunization booklet. It may be necessary for referencewhen your child is offered the hepatitis B vaccinethrough the grade six program.Yours sincerely,Dr. Richard MathiasPrincipal Investigator153CO. 1 Foley11.6 PAOVII,OE Ot sarrss COLOSIMOAlrootro of OkoolhWOO Goode Gore far Noose Control12111 West OM, AvowsVonomoot. S.C. VS2 ILITEL: (90')Long Term Follow-up of Neonates Immunized with Hepatitis B VaccineOt. s.n IlittiooJDear Ms.We would like to thank you for your child's partici-pation in the hepatitis B vaccine study. Your child,^ , born ^ has the following bloodtest results:Anti-HBc: PositiveAnti-HBs: NegativeHBsAg: PositiveThe blood tests indicate that your child is infected withthe hepatitis B virus. We suggest that you arrange to seeyour physician about these results. Further immunizationwith hepatitis B vaccine is not indicated.A copy of the results will be sent to your physician.Please keep this record of results in your child'simmunization booklet. It may be necessary for referencewhen your child is offered the hepatitis B vaccinethrough the grade six program.Yours sincerely,Dr. Richard MathiasPrincipal InvestigatorTHE uNryEAsnY OF lArTem COLUMBIAFoo.ey of liodidnoOecoAmors O Moat, Go 6 EgoOl000logyMoos WM./ &4d.SAW Form., Av.,w.Vorccuow. S.C. VC 123TEL (104)FAA: pot)ar.0.111TWE CANAOUV4 AEO cnoss scourVoncovolo GooOboe Torefuoce &roc,47500.). StroutVorcouvor. S.C. vw 214TEL: 11041117167661Y. Mortar Timmo.Pool000Study C000llootoTEL; 61011) 6764561es (101) 627-7600154Long Term Follow-up of Neonates Immunized with Hepatitis B VaccineINE UNNER3MY OF fIRMON COLUMBIAfealty eV tNeon,DmArenurt d FI..M Caro 4 EveAr•alogyNolo Mohr BudAny5404 Farvwrs AINPANVancouver. B.C.VC 125TEL/No=147nPAcmurm-4994Or. Z. FoleyT14 PROVINCE OF BRIT COLLAISIAYour of Near&San Courrta Cann rce Dooase Um"FM Woe •Co Ammo,VareAror. B.C. VSZrELJOUIPW00WTHE CANADIAN RED CROSS Socieryvmx,... cow.(mop 1,.7sh.p.on SpNYre4750 046 SowVarcovver.11C VON 2N111TEL: 104; 679. 756IAA 14.64,A Terren-PsalentSPAN DomTEL. 1104) 471F 7341ar (054) 177- MOSEDear Dr.Your patient, ^ , born ^ to ^has participated in the study on long-term effectivenessof hepatitis B vaccine.The results of this child's blood tests are asfollows:Anti-HBc: NegativeAnti-HBs: 100 mIU/mlHBsAg: NegativeWe consider this child to have an adequate protectiveresponse to the hepatitis B vaccine.The parents of this child have also been advised of theblood test results. Please feel free to discuss theseresults with me at 822-2772 if you have any questions.Yours truly,Dr. Richard MathiasPrincipal Investigator155Long Term Follow-up of Neonates Immunized with Hepatitis B VaccineLINNEPtirTY CP BRITS.' C01.04111APearty cl MedWsOpparo-ori d MooRIN Coe Eptl•rnacqylamas WANNe.11p4erre5404 Faeorprr Avow.V yew.% S.C.VC IDTEL: ta04) 122•2772FAX: IOC tZ2.4994Or. 1 ForelTILE PROVINCE Or BRITISPI COLUMBIAklinalry a NomaMtn earn:. Gaup ter Disoas• CarpRS Pleat 'Col knew,Vancouver. B. C. VSZ ILITEL:MAIM-80COr.O.MT1iE CANADIAN RED CROSS SCCIEryVarcowv CoresBlood Trarelvac. Seem*4750 O. SuerVarocuvor.^2N11TEL: (vot) 57+7561W. Madan Tortm.ParrameSA*, CoonflrwiceTEL. 1!04) 571.75{1et (Goo 1177.71103Dear Dr.Your patient, ^ , born ^ to ^has participated in the study on long-term effectivenessof hepatitis B vaccine.The results of this child's blood tests are asfollows:Anti-HBc: PositiveAnti-HBs: 50 m1U/m1HBsAg: NegativeWe consider that this child has had a past infectionwith the hepatitis B virus, but is not a carrier and nowhas an adequate antibody response. Revaccination is notrecommended.The parents of this child have also been advised of theblood test results.Please feel free to discuss these results with me at822-2772 if you have any questions.Yours sincerely,Dr. Richard MathiasPrincipal Investigator156Long Term Follow-up of Neonates Immunized with Hepatitis B VaccineOr. OA. MoihlosnE upwvarry OF 1111773114 COLUMBIAFaculty al ModkolneOftwoonsor A 14•01 Coro 4 Eoclofr000lmlorom IWO.. boldingSOX iarNw AvenueV revive t. B.C. VC laTEL (104)1222172FAR: (034) Su45Q4Or. A FreonTHE PROVINCE OF BRITISH COLUMBIAhIrsoiry of YloolTBONO Cowl:or Cantor fer Doom. Cool*11211 West .00 AmmoVoncoymr. B.C. KA NJin:mx)G0041010-Or.^PI714 CANAO.AN RED CROSS SOCIETYYancovor CentreOboe T. wolumon Sway1710 NO Sow*Vancouver. B.C. YIN 3IV•TEL: (1104) 179.755 INM. Morton Yorron-PooromSlue, Co-ordinakeTEL. (504) SAM,et (SO4) 1774103GDear Dr.Your patient, ^ , born ^ to ^has participated in the study on long-term effectivenessof hepatitis B vaccine.The results of this child's blood tests are asfollows:Anti-HBc: NegativeAnti-HBs: 6.5 mIU/m1HBsAg: NegativeThe results indicate that this child has not beeninfected with the hepatitis B virus. However, weconsider this child to have below protective levels ofanti-HBs and would recommend a booster dose of hepatitisB vaccine.The parents of this child have also been advised of theblood test results.If you wish your patient to receive the booster dose,the vaccine can be obtained by leaving a message at879-7551 local 307. It will be sent to your office byMarian Tomm Pastore for you to administer.Please feel free to discuss these results with me at822-2772 if you have any questions.Yours sincerely,Dr. Richard MathiasPrincipal Investigator157Long Term Follow-up of Neonates Immunized with Hepatitis B VaccineOr. R.O. lArlriasTHE LINIVEFLSITV or aRrrem COLD,astaFserrtyRanoC4R4vvr444 at Hiram Carr I fpRomanyJanos rAgror Budding5104 Farh•rr Aver..Varcarorr. I.C. VC lbTEL. 0304l 8224772FAX. (504) 522.4004R. 1 Fehr,THE PROVT•CE OF BRITISH COLUMBIAMowry al SaabStarr C.:rums Came kr PeAare Carrel11211 Mel , C01 knowsVerecarro. S.C. VSZ ILITEL: (604)A. O.THE CANADIAN RED CROSS SCCIETVVarrx4,44. GrasBtfloe TrarrirtAren Sorra4750 044 StromVIVIOXPAir.S C V6N 2NOTEL 0100 579-755?Marla Terros.P.41earswycor.swmTEL (604)wmaorn.minDear Dr.Your patient, ^, born ^ to ^has-participated in the study on long-term effectivenessof hepatitis B vaccine.The results of this child's blood tests are asfollows:Anti-HBc: NegativeAnti-HBs: Insufficient blood to testHBsAg: Insufficient blood to testWe consider that this child has not been infected withthe hepatitis B virus. However, sufficient blood was notobtained to assess the child's response to the hepatitisB vaccine.The parents of this child have also been advised of thisblood test result.If you have any questions, please feel free to discussthe outcome of this child's participation in the studywith me at 822-2772.Yours sincerely,Dr. Richard MathiasPrincipal Investigator158Long Term Follow-up of Neonates Immunized with Hepatitis B VaccineOr. KM WiseopevEasrry OF WV T eM CXXUMIll■Forzoly d Medan.Dordirtrndi of Head, Caro E.DesemalogyJana Umber gofersFar.4.. Avon.*Voreowfw. LC. VET 123TEL: MN) 1122.2772FAX: (104)1224994R. 1. PrioryTHE PROVNCE Or SADISM fXXUMBIAUnary of Noahaftesh CoRdrea Canes for Osoas. CartedVI Wed To AvenueVareouvw. t.C. V5ZTEL MOO 550-5053a.0.11154E eArfALAAN RED CROSS SOCIETYVarcoive, Gemsgb:ed Tr anshwon Swot4750 ORA StyeVorcovwf, re. Verl IreTEL. (104) 171.7551IAA Morton Tondo-PostonShier Oa-o:MaceTEL: (104)1171.7551te MI) 177.71103IDear Dr.Your patient, ^ , born ^ to ^has -participated in the study on long-term effectivenessof hepatitis B vaccine.The results of this child's blood tests are as follows:Anti-HBc: PositiveAnti-HBs: 0 mIU/mlHBsAg: PositiveThe results indicate that this child is infected withthe hepatitis B virus and is considered a carrier atthis time.The parents of this child have also been advised of theblood test results.Please feel free to discuss these results with me at822-2772 if you have any questions.Yours sincerely,Dr. Richard MathiasPrincipal Investigator159APPENDIX 2RED CROSS REGISTRY FORM^161HBIG FOLLOW-UP OF NEONATES FROM HBV INFECTIOUS MOTHERSNAME OFMOTHER:D.O.B: E.D.C: D.O.B:(BABy)ADDRESS: NAME OFFATHER:HBIG HEPATITIS B VACCINELOT NO: (1) LOT NO: (2) LOT NO: (3) LOT NO:DATE GIVEN: DATE GIVEN: DATE GIVEN: DATE GIVEN:SAMPLEFROM:TESTDATEVANCOUVER TEST RESULTS NRL TEST RESULTSHBsAg(EIA)HBsAg(IMX)Anti-HBsAnti-HBcHBsAgCONFHBsAg(EIA)HBsAg(RIA)Anti -HBcMOTHERMOTHERCOMMENTS:PHYSICIAN(S) :^ 'HOSPITAL: 'HEALTH UNIT:APPENDIX 3STUDENTIZED RESIDUAL PLOTSDose one for the Anti-HBc model^163Number of doses for the Anti-HBs model 164Dose one for the Anti-HBs model^165Frequencies and symbols used (not applicable for control or overlay plots)1 -^1 11 -^Et 21 -^L 31 -^V2 -^2 12 -^C 22 - M 32 - W3 -^3 13 -^D 23 - N 33 -^X4 -^4 14 -^E 24 - 0 34 -^Y5 -^5 15 - F 25 -^P 35 -^Z6 -^6 16 -^G 26 - 0 36 -^•7^- 7 17 -^H 27 -^R8 -^8 18 -^I 28 -^S9^- 9 19 -^J 29 -^T10^- A 20 - K 30 -^U580 cases plottedStudentized Residuals for the Anti-HBc Model163NUMBER OF DOSES OF HS VACCINE707 cases Plotted.Studentized Residuals for the Anti-ns Model164707 cases plotted.Studentized Residuals for the Anti-HBs Model165


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items