Occupational and Environmental Medicine 1997;54:629-635

Risk of congenital anomalies in children of parentsoccupationally exposed to low level ionisingradiation

Lois M Green, Linda Dodds, Anthony B Miller, Darrell J Tomkins, Jiehui Li,Michael Escobar

Health ServicesDepartment, OntarioHydro, 700 UniversityAvenue, Toronto,Ontario, CanadaMSG 1X6L M Green

Reproductive CareProgram, IWA GraceHealth Centre, 5980University Avenue,Halifax, Nova Scotia,Canada B3H 4N1L Dodds

Department ofPreventive Medicineand Biostatistics,University ofToronto,Toronto, Ontario,Canada MSS 1A8A B MillerJLi

M Escobar

Department ofPediatrics, McMasterUniversity, 1200 MainStreet West, Hamilton,Ontario, Canada,L8N 3Z5D J Tomkins

Correspondence to:Dr L Green, Health ServicesDepartment, Ontario Hydro,700 University Avenue,Toronto, Ontario, CanadaM5G 1X6.

Accepted 12 March 1997

AbstractObjectives-To evaluate the risk of havinga child with a congenital anomaly in rela-tion to occupational exposure to low levelionising radiation in the pre-conceptionperiod.Methods-A case-control study based onthe Canadian congenital anomalies regis-try used record linkage techniques toidentify congenital anomalies among maleand female workers in Canada's largestelectric company. Cases were defined asparents of a child with a congenitalanomaly born between April 1979 andDecember 1986 who had a congenitalanomaly diagnosed within the first year oflife. Controls were an individuallymatched sample of parents of a livebornchild without an anomaly. Risk ofcongeni-tal anomaly was assessed in relation toparental exposure to ionising radiationacquired through work within a nucleargenerating station of an electric powercompany. Exposure was assessed accord-ing to employment, whether or not theworker was monitored for radiation expo-sure, and quantitative estimates of radia-tion dose.Results-Employment within the electricpower industry was not associated with anincreased risk of congenital anomalies inthe offspring of mothers or fathers. Riskestimates for workers monitored (thosewho are likely to be exposed to ionisingradiation) were 1.75 (95% confidenceinterval (95% CI) 0.86 to 3.55) for mothersand 0.84 (95% CI 0.68 to 1.05) for fathers.Exposure for fathers before conception,defined cumulatively and for six monthsbefore conception, was not associated withincreased risk of anomalies in their off-spring. There were no significant in-creases in risk found between type ofanomaly and any measure of exposure,although the statistical power in thesegroups was limited. The study had insuffi-cient numbers to evaluate the effects ofionising radiation in mothers as only threemothers had recorded doses >0 mSv.Conclusions-Overall, workers in a nu-clear power industry, and specificallythose exposed before conception to lowlevels of ionising radiation, do not appearto be at an increased risk of having a live-born child with a congenital anomaly.

(Occup Environ Med 1997;54:629-635)

Keywords: congenital anomalies; ionising radiation;parental occupation; preconceptional exposure

The cause of most congenital anomalies isunknown and their pathogenesis is not wellunderstood.' Those factors which have beenestablished as causing congenital anomaliesinclude: single gene disorders, chromosomalaberrations, multifactorial causes, and discreteenvironmental teratogens.' Occupational orenvironmental exposures might be implicatedin any of these causal roles.The possibility that hazardous exposures

which parents may receive at work could causeadverse reproductive outcomes is a topic ofgrowing interest. Increasingly there are de-mands on the part of both the public andworkers to know whether reproductive risksexist and to assess their impact. Concertedefforts are being made in many countries toensure that occupational health policies forworkers consider possible risks for theiroffspring.3Although earlier research was primarily lim-

ited to prenatal exposures in the mother, therehas been increasing focus in the epidemiologi-cal literature on exposures received in theperiod before conception, and especially tothose of the father.4' Evidence from laboratorystudies indicates that defects in offspring canarise from mutagenic exposure of thefather.'1" Paternal contributions to birthdefects can occur as a result of direct genemutation and chromosomal mutation, orthrough alterations in the seminal fluid trans-mitted to the mother.'2 It is known thatradiation has the potential to adversely affectcells, particularly the DNA.'3 Damage to theDNA is more likely to occur during periods ofmeiotic division than during periods of inactiv-ity. It has therefore been suggested that menmay be at higher risk than women for germ celldamage from radiation as spermatogoniaundergo multiple mitotic divisions from thetime of sexual maturity.'4

Studies evaluating the effects of exposures toionising radiation and reproductive outcomesin humans, specifically congenital anomalies,are very limited. Congenital anomalies consid-ered together are an indication of underlyinggenetic defects. About half of congenitalanomalies in a population based registry werefound to have a genetic aetiology."5 For the past

629

on Septem

ber 27, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.54.9.629 on 1 Septem

ber 1997. Dow

nloaded from

Green, Dodds, Miller, Tomkins, Li, Escobar

50 years, considerable effort has been made toidentify genetic effects of the atomic bombs bystudying chromosomal abnormalities, germi-nal mutations altering protein electrophoresis,and more recently germinal mutations de-tected at the DNA level.'6 17 These studies havefailed to show an increase in germinal muta-tions in atomic bomb survivors.'8However, it has been reported that children

born to atomic bomb survivors, conceived onaverage about five years after the bombings,had an increased, but not significant, risk of amajor congenital malformation, stillbirth, andneonatal death with increasing levels of paren-tal exposure to ionising radiation.'

Recently an increased germinal mutationrate was found at the DNA level in familiesexposed to caesium-137 surface contaminationafter the Chernobyl nuclear power stationaccident.20 Although differences between thetypes of radiation exposure from the atomicbomb and Chernobyl accidents and the appro-priateness of the selected group for comparisonmay account for the conflicting results, suchdata suggest a reconsideration of the geneticeffects of ionising radiation.2On the other hand, little information is avail-

able for chronic low level exposures, which aremore representative of contemporary expo-sures. Few studies have used direct quantitativemeasures of exposure to radiation. Sever et a12'examined the risk of congenital malformationsand exposure to radiation before conceptionamong employees in a plutonium and electricalenergy production plant. Although no associ-ation was found between maternal exposurebefore conception and the different birthdefects, paternal exposure before conceptionwas significantly related to risk of neural tubedefects.An increased prevalence at birth of central

nervous system defects and Down syndromewas found to be associated with the highestconcentrations of airborne tritium released inthe vicinity of one nuclear plant in southernOntario.23 Inconsistencies related to the expo-sure data, testing for multiple hypotheses, andthe ecological design of the study were cited bythe authors as reasons for cautiousinterpretation of the findings.23 The possibilityof an association between chromosomal non-disjunction and maternal radiation before con-ception has been studied for decades with overa dozen studies showing a positive associationbut most failing to reach significance.24The present case-control study was carried

out to evaluate the risk of congenital anomaliesin the offspring of parents employed within anelectric power company (Ontario Hydro). Thiscompany has been responsible for nuclearpower generation for the province of Ontariosince 1962. Risk of congenital anomalies wasassessed from radiation exposure derived fromdosimetric measurements and surrogates ofthis exposure. The case-control design wasused as this enabled the same information to beassessed as in a cohort design, but with greatereconomy of effort.

MethodsCases and controls were identified as part of apreviously completed study by Dodds et al.25The figure shows the data sources for thisstudy. Cases were ascertained from the Cana-dian congenital anomalies surveillance system.This is a population based registry of allcongenital anomalies diagnosed during the firstyear of life. In Ontario, the congenital anomalydiagnoses were based on discharge reportsfrom hospital admissions, reports of birth, anddeath certificates for children under one year ofage. The selection of case parents was accord-ing to the criteria: child born in Ontariobetween 1 April 1979 and 31 December 1986,mother resident in Ontario at time of birth ofthe child, child was liveborn, child had at leastone diagnosis with international classificationof diseases (ICD-9) congenital anomaliesrubrics 740.0-759.9.26 A maximum of 15separate diagnoses of congenital anomalies arerecorded for each child. One control parent percase was randomly selected from Ontario birthregistrations and matched according to the cri-teria: year of birth of the child, birth order ofthe child (first, second to third, and fourth orgreater), exact maternal age (according to birthyear), marital status of mother (single versusnot single), and birthplace ofeach parent (bornin Ontario versus not born in Ontario). Parentswho were resident outside of Ontario at thetime of the index birth and adoptive parentswere excluded. To avoid the problem of corre-lated exposure information among the sub-jects, parents were only included once even ifthey had more than one child during the studyperiod. Information from the first birth whichoccurred during the study period was used,regardless of whether a case or control.A total of 45 200 case-control sets were

identified for the mothers and 41 158 case-control sets were identified for the fathers.There are about 9% fewer fathers than mothersbecause fathers are not always named on thebirth certificates-for example, in the case of asingle mother.The case and control parents were then

linked with a stepwise set of computer linkagesbased on a probabilistic linking model27 28 witha cohort of 32 975 men and 9507 women whoworked for Ontario Hydro. This cohort repre-sented employees with at least 12 consecutivemonths of employment at Ontario Hydro dur-ing the period 1978-86 inclusive. The inclu-sion criterion of 12 consecutive months ofemployment was imposed because the com-pany did not require complete retention ofrecords for employees who worked less than 12months. As the accuracy and completeness ofthe record linkage process depended onpersonal identifiers being present, the groupwhere these identifiers might be incompletewas excluded. To resolve any uncertainties onpossible matches between the case-control setsand the employee cohort, additional personalidentifiers were obtained from companyrecords.For men, the period of aetiological interest is

before conception. The date of conception wasdetermined by subtracting gestational age from

630

on Septem

ber 27, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.54.9.629 on 1 Septem

ber 1997. Dow

nloaded from

Risk of congenital anomalies in children ofparents occupationally exposed to low level ionising radiation

the date of birth for the child of the case orcontrol parent. A comparison was madebetween date of hire with the company anddate of conception. If the date of conceptionpreceded the date of employment, theseparents were excluded. For women, the periodsof interest included employment before con-ception as well as employment during preg-nancy.

In accordance with legislative requirementsfor regulatory control,29 all workers whoseexposures to ionising radiation may exceed thedose limits for the members of the generalpublic are monitored and dose records areretained. Radiation doses acquired beforeemployment with Ontario Hydro are collectedand are added to those doses received duringemployment with Ontario Hydro.'0 Doses wereestimated by methods and measuring devicesused for personal monitoring calibrated inaccordance with the specifications of theAtomic Energy Control Board of Canada.

Detailed dose information is kept for exter-nal whole body dose (which includes any neu-tron dose), external skin dose, and internaldose (primarily from uptakes of tritium) whichare the most common exposures associatedwith work in the nuclear power generation sec-tor. For the years 1962-76, radiation dosesfrom x rays, y rays, and fi rays were monitoredwith a film badge. From 1976 to the present,doses of x rays, y rays, and fi rays weremonitored with a thermoluminescent dosim-eter. Internal dose, reflecting the internaluptake of radionuclides, mainly tritium, wasmeasured with urinary bioassay.The risk of congenital anomalies was deter-

mined for three main types of exposure:employment with Ontario Hydro, monitoringstatus (those employees monitored in accord-ance with regulatory requirements), and quan-

titative estimates for radiation dose. For thequantitative estimates, three periods of aetio-logical interest were evaluated: total wholebody dose before conception, whole body dosesix months before conception, and tritium dose60 days before conception. The six monthperiod allowed comparability with the study ofGardner et aPl' in which an excess risk of child-hood cancer was found among offspring ofparents occupationally exposed to ionisingradiation and which was more pronounced inthe period six months before conception. The60 day period corresponded with the period ofhuman spermatogenesis."

Congenital anomalies may be multiple inexpression and may have diverse aetiologies. Asmany as 15 anomalies per child coded toICD-9 were recorded. A child with more thanone anomaly was counted in more than oneICD group in the analyses according to ICDcategory.To categorise anomalies according to their

possible aetiology, a hierarchical system devel-oped by Baird et atP5 was used to establishaetiological groupings. The assignment ofaetiological groupings based on ICD codes wascarried out by a geneticist (DMT) who was notaware of the employment or exposure status ofthe parent. Each child was assigned to one ofthe aetiological groups in the following hierar-chy: single gene disorders (autosomal domi-nant, autosomal recessive, X linked), chromo-somal (autosomal and X chromosomal),multifactorial, genetic (unspecified), and un-known. This hierarchical system ensured thatcases were not counted more than once andalso permitted comparisons with otherwork.'5 25

Because the mechanisms by which ionisingradiation might affect the risk of congenitalanomalies is different for offspring of men and

Identification of children in Ontario with a congenitalanomaly diagnosed before age 1 year, from Canadian congenital

anomalies surveillance system

Computerised record linkage between anomalies registry and Ontariobirth certificates

Cases: parents of children with acongenital anomaly identified from child's

birth certificateI

Controls: parents of children without acongenital anomaly selected from

birth certificate file

Additional identifying data on parents abstracted fromoriginal birth certificates and added to case control file

Computerised record linkage betweencase-control file and Ontario Hydro cohort

Case and control parents with Ontario Hydro employment

Occupational exposure to ionising radiation determined fromOntario Hydro radiation dose information system

Sources of data.

r-

I~~ -.

631

on Septem

ber 27, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.54.9.629 on 1 Septem

ber 1997. Dow

nloaded from

Green, Dodds, Miler, Tomkins, Li, Escobar

Table 1 Exposure informationforfaes* according toperiod before conception

Exposure Cases Controls

Employed but not monitored for radiation exposure 189 247Monitored for radiation exposure, recorded dose = Ot 63 53Monitored for radiation exposure, recorded dose > 0 (mSv)t:

Cumulative whole body dose before conception:Number of employees 89 126Mean (SD) dose 39.9 (64.4) 38.7 (58.9)Dose range 0.04-262.1 0.10-307.0

Whole body dose six months before conception:Number of employees 65 73Mean (SD) dose 3.6 (4.5) 3.6 (4.1)Dose range 0.05-18.2 0.01-17.4

Cumulative tritium dose before conception:Number of employees 74 97Mean (SD) dose 16.9 (27.2) 14.4 (23.9)Dose range 0.02-107.2 0.03-137.9

Tritium dose received 60 days before conception:Number of employees 51 53Mean (SD) dose 0.4 (0.5) 0.5 (0.6)Dose range 0.01-2.8 0.02-2.8

*Only three mothers had a non-zero dose.tDose refers to that received at Ontario Hydro from date of hire to date of conception including any dose reported at the time ofhire by Ontario Hydro.

women, fathers and mothers were analysedseparately. The closeness of the matchingtogether with the type of records (birthregistrations) from which the data for this studywere gathered limited the number of variablesavailable for evaluation as possible risk factorsor confounders. Other than the matching vari-ables, history of stillbirth" and paternal age'4were the only available variables which mightbe predictive of risk of congenital anomaly. Asthe data were collected with a matchedcase-control design, the standard McNemartest" and conditional logistic regression" wereused for the analysis. Maximum likelihoodestimates for odds ratios (ORs) were calculatedwith SAS for matched pairs that were discord-ant for the exposure of interest in the parent.The method ofMiettinen was used to calculatethe 95% confidence interval (95% CI).`7

ResultsA total of 763 pairs of case-control fathers and165 pairs of case-control mothers were identi-fied in which at least one parent had employ-ment at Ontario Hydro. For fathers, four pairswere concordant for employment with OntarioHydro-for example, both case and controlhad been employed with Ontario Hydro.Analysis of all anomalies combined showed noevidence of increased risk in relation toemployment of the parents at Ontario Hydro(for men, OR 0.79, 95% CI 0.69 to 0.92; forwomen, OR 0.94, 95% CI 0.69 to 1.28).As a surrogate of exposure to ionising radia-

tion, monitoring status was examined inrelation to congenital anomalies in the off-spring. A monitored worker may be expected tohave exposure to ionising radiation greater thanthat which would normally be experienced bymembers of the general public. A monitoredworker is not necessarily synonymous with anexposed worker. There were 149 pairs wherethe case father was monitored for exposure toradiation and the control father was not and177 pairs where the case father was not moni-tored and the control father was. The analysisof these discordant pairs gave an OR of 0.84(0.68 to 1.05). For mothers, the correspondingOR was 1.75 (0.85 to 3.55) which was based

on substantially fewer discordant pairs, 21pairs where the case mother was monitoredand the control mother was not and 12 pairswhere the case mother was not monitored andthe control mother was.For fathers, more detailed analyses were car-

ried out according to radiation dose and periodbefore conception. Table 1 shows exposureinformation relating to dose and period beforeconception. Of the 331 workers who weremonitored for exposure to radiation, 35% hada recorded dose of zero mSv. The categories ofzero dose and no exposure opportunity werecombined to form the control exposure vari-able in the conditional logistic regression mod-els. Regulations in effect during the studyperiod were such that few women worked inareas where radiation exposure would beexpected. It was not possible to conduct thislevel of analysis for the mothers because onlythree mothers (two cases, one control) hadrecorded radiation exposure >0 mSv.

Table 2 shows adjusted ORs, according toaetiological subgroups, and father's wholebody dose, and tritium dose for specified peri-ods before conception. With the exception ofthe risk estimate for anomalies classified ashaving a multifactorial aetiology, which didattain significance for cumulative exposureonly (OR 0.61, 95% CI 0.42 to 0.90), the ORsfor specific aetiological groups were not signifi-cantly different from unity. Risk estimates forchromosomal anomalies were above unity, butwide CIs indicate the instability of theestimates which are based on few discordantpairs.

Adjusted ORs, ICD groupings, father'swhole body dose, and tritium dose (table 3),were also not significantly increased or re-duced, with the exception of anomalies of thecirculatory system which showed a significantlyreduced OR for cumulative exposure only(0.51, 95% CI 0.27 to 0.95) and musculo-skeletal anomalies which showed a reduced ORof borderline significance (0.62, 95% CI 0.38to 1.01). Odds ratios above unity (not signifi-cant) were found for chromosomal anomaliesfor fathers who had a recorded dose greaterthan zero either six months or 60 days before

632

on Septem

ber 27, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.54.9.629 on 1 Septem

ber 1997. Dow

nloaded from

Risk of congenital anomalies in children ofparents occupationally exposed to low level ionising radiation

Table 2 Adjusted ORs (95% CIs) for aetiological groups of congenital anomalies according to dose* and radiation exposure offathers during the periodbefore conception

Cumulative whole body dose beforeconception Whole body dose 6 months before conception Tritium dose 60 days before conception

Aetiological group Discordant pairst OR (95% CI) Discordant pairs OR (95% CI) Discordant pairs OR (95% CI)

Single gene disorders 0/3 - 0/0 - 0/0Chromosomal disorders 4/4 0.97 (0.24 to 3.91) 3/2 1.46 (0.24 to 8.80) 3/2 1.46 (0.24 to 8.80)Multifactorial§ 43/70 0.61 (0.42 to 0.90) 31/37 0.92 (0.56 to 1.50) 28/28 1.13 (0.66 to 1.94)Genetic, unspecified 13/14 0.93 (0.44 to 1.98) 8/10 0.83 (0.33 to 2.12) 6/8 0.80 (0.27 to 2.32)Unknown§ 28/34 0.80 (0.48 to 1.32) 22/23 0.91 (0.50 to 1.65) 14/15 0.84 (0.40 to 1.76)Total 88/125 0.72 (0.55 to 0.95) 64/72 0.90 (0.60 to 1.27) 51/53 0.99 (0.67 to 1.47)

* Dose >0 v dose = 0 or not exposed.tCase exposed, control not exposed, or case not exposed, control exposed.jAdjusted for father's age only.§Adjusted for father's age and history of stillbirths.-Could not be calculated.

conception, but the estimates were based on

small numbers with wide 95% CIs (table 3).There was a tendency in several ICD groups ofanomalies for the ORs to increase for theperiod closest to the date of conception but allwere accompanied by wide 95% CIs, of whichnone were significant at the 5% level.

DiscussionWe found no evidence of a significantlyincreased risk for fathers having a livebornchild with a congenital anomaly after occupa-

tional exposure to radiation in the periodbefore conception. Congenital anomalies donot all share the same aetiological factors andcombining all anomalies together may dilutethe risk that ionising radiation may have on

some specific anomalies. However, analysesaccording to subgroups, defined by possibleaetiology or by anatomical ICD rubrics, alsoshowed no evidence that fathers with exposureto ionising radiation were more likely to have a

child with a birth defect. There is not sufficientevidence from this study to suggest whetherthis relation is causal or not. Further studieswith larger numbers are required to confirm or

rule out this relation.For mothers, the opportunities for exposure

during the study period were minimal as

dictated by regulatory requirements, andhence, the data for mothers were limited. Theavailable data indicated, as was shown forfathers, no significantly increased risks of hav-ing a child with a congenital anomaly relative toemployment or monitoring status. The non-

significant increase in risk for mothers whowere monitored might be explained by chancealone. In this study we did not have the powerto fully explore the relation between exposure

to ionising radiation in mothers and congenitalanomalies in their offspring, but the suggestionof an increased risk among those who were

monitored may warrant further study.Calculations carried out at the end of the

work38 indicated a statistical power of over 80%to detect a 1.5 increase in risk for all anomaliescombined among fathers relative to their expo-sure before conception. However, the statisticalpower was greatly reduced in the analyses formothers and by aetiological subgroups for off-spring of fathers and hence a true associationmight have been missed because of this.The design of this study provided distinct

strengths in the evaluation of risks of congeni-tal anomaly potentially associated with expo-sure to ionising radiation. There is no knownreporting bias in the population based registryfrom which the cases were ascertained. The

acquisition of dosimetric measurements ofexposure to radiation with record linkage tech-niques and no direct contact with studyparticipants, obviated any possibility of recallbias, which can compromise the validity ofstudies examining a health outcome of rela-tively high emotion and an earnest desire tounderstand the cause.

The measurement of exposure to ionisingradiation was blind to case-control status andthereby avoided any possible bias in exposureassessment. Quantitative dose estimates were

Table 3 Adjusted ORs (95% CIs) for ICD groupings of congenital anomalies according to dosed and radiation exposure offathers during the periodbefore conception

Cumulative whole body dose beforeconception Whole body dose 6 months before conception Tritium dose 60 days before conception

ICD group Discordant pairs OR (95% CI) Discordant pairs OR (95% CI) Discordant pairs OR (95% CI)

Nervous system 6/7 0.99 (0.32 to 3.08) 5/5 1.23 (0.33 to 4.63) 5/4 1.65 (0.39 to 6.94)Facialregion§ 4/3 1.43 (0.29 to 7.04) 3/2 1.74 (0.24 to 12.65) 3/2 1.74 (0.24 to 12.65)Circulatory system* 15/30 0.51 (0.27 to 0.95) 11/12 0.86 (0.38 to 1.98) 11/9 1.14 (0.47 to 2.81)Respiratory system 6/5 1.38 (0.37 to 5.07) 4/3 1.20 (0.26 to 5.59) 3/1 2.71 (0.27 to 27.05)Cleft palate or lip* 4/3 1.17 (0.25 to 5.57) 2/2 0.93 (0.13 to 6.82) 1/1 0.96 (0.06 to 15.42)Digestive system 12/16 0.78 (0.36 to 1.67) 10/3 0.80 (0.34 to 1.88) 6/9 0.74 (0.26 to 2.17)Genitourinaryt 16/17 0.91 (0.38 to 2.20) 12/11 0.95 (0.34 to 2.64) 10/9 1.04 (0.32 to 3.36)Musculoskeletal§ 28/46 0.62 (0.38 to 1.01) 21/24 0.88 (0.48 to 1.61) 16/18 0.90 (0.45 to 1.80)Integumentt 8/12 0.65 (0.26 to 1.61) 5/11 0.46 (0.16 to 1.35) 4/8 0.52 (0.16 to 1.75)Chromosomal anomaliest 4/4 0.97 (0.24 to 3.91) 3/2 1.46 (0.24 to 8.80) 3/2 1.46 (0.24 to 8.80)

*Dose >0 v dose = 0 or not exposed.tCase exposed, control not exposed, or case not exposed, control exposed.4Adjusted for father's age.§Adjusted for father's age and history of stillbirths.

633

on Septem

ber 27, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.54.9.629 on 1 Septem

ber 1997. Dow

nloaded from

Green, Dodds, Miller, Tomkins, Li, Escobar

assigned in advance of the outcome occur-rence, having been previously documented inaccordance with independently determinedregulatory guidelines and controls which estab-lish standards of accuracy.

In this study, it was expected that availabilityof personal identifiers from employmentrecords would contribute to the accuracy ofcase and control ascertainment through thepopulation based records as record linkagetechniques were used. Whether this accuracyhas indeed been optimised might be ques-tioned in the light of several estimates of riskwhich were below unity. There was, at the startof the study, no reason why exposure to ionis-ing radiation should be associated with areduced risk of anomalies, unless this exposureselectively decreased fertility (reduced abilityto conceive among parents who would havebeen more likely to have a fetus with acongenital anomaly or increased spontaneousabortions or stillbirths among fetuses with ananomaly). Therefore, it is possible that thestringent criteria used for linkages might havecontributed to an underascertainment of a fewcases although it would not likely be differentialto case-control status.

Limitations of the study relate to incompletecontrol for possible confounders. However, thecloseness of the matching in the study designdid control for several possible confounders.Information relating to medical history of theparents was not available and therefore, wewere unable to account for the possible role ofmaternal illnesses before and during preg-nancy, relevant family medical history, repro-ductive history, or socioeconomic status. Theanalyses pertaining to aetiological subgroupswere further limited by the availability of onlyfour digit ICD-9 codes making it difficult toassign aetiology with certainty. For this reason,many anomalies had unknown aetiology orunspecified genetic aetiology. However, instudies of congenital anomalies, it is importantto group together anomalies with commonaetiology. Seldom will studies be large enoughto have the power to analyse any singleanomaly.This study is confined to those congenital

anomalies in liveborn infants resulting from"successful conceptions" and pregnancies andtherefore, is restricted to a subset of fertilemothers and fathers. It was not possible for thisstudy to consider if ionising radiation compro-mises fertility in any way. The lack of an effectin liveborn children may be the result ofincreased effects of other adverse reproductiveoutcomes. However, because the selection ofcontrols took parity into account, and by doingso provided some control for fertility, it isunlikely that this factor constitutes a majorlimitation.The epidemiological evidence to which these

findings could be related is sparse. The findingsreported here do not correspond with thosereported by Sever etal,22 one of the few studiesthat used quantitative individual measure-ments of exposure to radiation. They found anincrease in neural tube defects related to expo-sure of fathers to ionising radiation before con-

ception. The ecological design of the study byJohnson and Rouleau," who found an increasein Down's syndrome relative to tritium releasesfrom nuclear power generating stations causesdifficulties in interpretation of the results at theindividual level, and in relating the findings tothose from studies such as this one.The present study is unique in its examin-

ation of the possible role of exposure toionising radiation before conception as we usedexisting individual dosimetric exposure recordsand a population based birth defects registry.Although workers involved in the generation ofnuclear power receive higher exposures of ion-ising radiation than environmental exposureswhich might be received by members of thepublic, the absence of a significantly increasedrisk of having a child with a congenital anomalyin fathers exposed to radiation within an occu-pational setting suggests that a public healthrisk is unlikely.

We acknowledge Claus Wall who performed data managementactivities and record linkage and Warren Christiani whoprovided the exposure information related to ionising radiation.This study was supported by Grant 04450 Ontario, Ministry ofHealth, and by Ontario Hydro.

1 Kalter J, Warkany J. Congenital malformations: etologicfactors and their role in prevention. N Eng; J? Med1983;308:424-31.

2 Baird PA. Reproductive hazards and the workplace [edito-rial). Can MedAssoc7 1992;147:157-60.

3 Stijkel A, van Dijk FJH. Developments in reproductive riskmanagement. Occup Environ Med 1995;52:294-303.

4 Taskinen HK. Effects ofparental occupational exposures onspontaneous abortions and congenital malformation.Scandj Wb* Environ Health 1990;16:297-314.

5 Roman E, Doyle P, Ansell P, Bull D, Beral V. Health of chil-dren born to medical radiographers. Occup Environ Med1996;53:73-9.

6 Brender JD, Suarez L. Paternal occupation and anen-cephaly. AmJEpidemiol 1990;131:517-21

7 McDonald AD, McDonald JC, Armstrong B, Cherry NM,Nolin AD, Robert D. Fathers' occupation and pregnancyoutcome. BrJ IndMed 1989;46:329-33.

8 Olshan A, Teschke K, Baird PA. Paternal occupation andcongenital anomalies in offspring. Am J Ind Med 1991;20:447-75.

9 Olshan AF, Baird PA, Teschke K. Paternal occupationalexposures and the risk of Down's syndrome. Am Y HumGenet 1989;44:646-51.

10 Nomura T. Parental exposure to x-rays and chemicalsinduces heritable tumours and anomalies in mice. Nature1982;2%:755-7.

11 Kirk KM, Lyon MF. Induction of congenital malformationsin the offspring of male mice treated with x-rays atpre-meiotc and post-meiotic stages. Mutat Res 1984;125:75-85.

12 Olshan AF, Faustman E. Male-mediated developmentaltoxicity. Annu Rev Public Health 1993;14:159-81.

13 Parker RG. Principles of radiation oncology. In: HaskellCM, ed. Cancer treatment. Philadelphia: WB Saunders,1985, 14-20.

14 Connor JM, Ferguson-Smith MA. Essential medical genetics.Oxford: Blackwell, 1987.

15 Baird PA, Anderson TW, Newcombe HB, Lowry RB.Genetic disorders in children and young adults: apopulation study. Am J Hum Genet 1988;42:677-93.

16 Kodaira M, Satoh C, Hiyama K, Toyama K Lack of effectsof atomic bomb radiation on genetic instability of tandem-repetitive elements in human germ cells. Am J Hum Genet1995;57: 1275-83.

17 Neel NV. New approaches to evaluating the genetic effects ofthe atomic bombs. AmJ Hum Genet 1995;57:1263-6.

18 Sankaranarayanan K. Ionizing radiation, genetic riskestimation and molecular biology: impact and inferences.Trends Genet 1993;9:79-84.

19 Otake M, Schull WJ, Neel NV. Congenital malformations,stillbirths, and early mortality among the children ofatomic bomb survivors: a reanalysis. Rad Res 1990;122:1-11.

20 Dubrova YE, Nesterov VN, Krouchinsky NG, OstapenkiVA, Neumann R, Nell DI, Jeffreys AJ. Human minsatellitemutation rate after the Chernobyl accident. Nature1996;380:683-6.

21 Satoh C, Kodaira M. Effects ofradiation on children. Nature1996;383:226.

22 Sever LE, Gilbert ES, Hessol NA, McIntyre JM. Acase-control study of congenital malformations and occu-pational exposure to low-level ionizing radiation.AmJ Epi-demiol 1988;127:226-42.

634

on Septem

ber 27, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.54.9.629 on 1 Septem

ber 1997. Dow

nloaded from

Risk of congenital anomalies in children ofparents occupationally exposed to low level ionising radiation

23 Johnson KC, Rouleau J. Tritium releases from the Pickeringnucleargenerating station and birth defects and infant mortalityin nearby communities, 1971-88. Ottawa, Canada: AtomicEnergy Control Board, 1991. (AECB Project No 7156.1.)

24 Uchida IA. Radiation-induced non-disjunction. EnvironHealth Perspect 1979;31:13-7.

25 Dodds I, Marrett LD, Tomkins DJ, Green B, Sherman G.Case-control study of congenital anomalies in children ofcancer patients. BMJ 1993;307:64-168.

26 World Health Organization. The international classification ofdiseases, 9th revision. Vol 1. Geneva: World HealthOrganisation.

27 Roos LI, Wajda A, Nicol JP, The art and science of therecord linkage: methods that work with few identifiers.Comput Biol Med 1986;16:45-57.

28 Roos LI., Sharp SM, Nichol JR Software for health careanalysts; a modular approach. JMed Syst 1987;11:445-64.

29 Atomic Energy Control Act. Atomic Energy Control BoardRegudations June 1974. Ottawa, Canada: Minister of Supplyand Services, 1974. (Registration SOR/74-334.)

30 Nuclear Generation Division. Radioactivity Management andEnvironmental Protection Department. Toronto, Canada:Ontario Hydro, 1980.

31 Gardner MJ, Snee MP, Hall CA, Downes S, Terrell JD.Results of a case-control study ofleukaemia and lymphomaamong young people near Sellafield nuclear plant in WestCumbria. BMY 1990;300:423-9.

32 Courot M, Hochereau-de Reviers MT, Ortavant R.Spermatogenesis: In: Johnson AD, Gomes WR, VandemarkNI, eds. The testis. New York: Academic Press, 1970.

33 Lie RT, Wilcox AJ, Skjaerven R. A population-based studyof the risk of recurrence of birth defects. N Engl Y Med1994;331:1-4.

34 McIntosh GC, Olshan AF, Baird PA. Paternal age and therisk ofbirth defects in offspring. Epidemiology 1995;6:282-8.

35 Fleiss J. Statistical methodsfor rates andproportions. New York:Wiley, 1981.

36 Hosmer DW, Lemeshow S. Applied logistic regression. NewYork. Wiley, 1989.

37 Miettinen OS. Estimation of relative risk from individuallymatched sets. Biometrics 1970;26:75-86.

38 Schlesselman Jj. Case control studies: design, conduct andanalysis. Oxford: Oxford University Press, 1982.

Vancouver style

All manuscripts submitted to Occup EnvironMed should conform to the uniform require-ments for manuscripts submitted to biomedi-cal journals (known as the Vancouver style.)

Occup Environ Med, together with manyother international biomedical journals, hasagreed to accept articles prepared in accord-ance with the Vancouver style. The style(described in full in the AMA[l1]) is intendedto standardize requirements for authors, and isthe same as in this issue.

References should be numbered consecu-tively in the order in which they are first men-tioned in the text by Arabic numerals on theline in square brackets on each occasionthe reference is cited (Manson[l] confirmedother reports[2] [3] [4] [5]). In future ref-erences to papers submitted to Occup EnvironMed should include: the names of all

authors if there are seven or less or, if there aremore, the first six followed by et al; the title ofjournal articles or book chapters; the titles ofjournals abbreviated according to the style ofIndex Medicus; and the first and final pagenumbers of the article or chapter. Titles not inIndex Medwcus should be given in full.Examples of common forms of references

are:

1 International Committee of Medical Journal Editors.Uniform requirements for manuscripts submitted tobiomed journals. JAMA 1993;269:2282-6.

2 Soter NA, Wasserman SI, Austen KF. Cold urticaria:release into the circulation of histmaine and eosinophilchemotactic factor of anaphylaxis during cold challenge.N EnglJ Med 1976;294:687-90.

3 Weinstein I, Swarm MN. Pathogenic properties of invad-ing micro-organisms. In: Sodeman WA Jr, Sodeman WA,eds. Pathologic physiology, mechanisms of disease. Philadel-phia:W B Saunders, 1974:457-72.

635

on Septem

ber 27, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.54.9.629 on 1 Septem

ber 1997. Dow

nloaded from