Pesticide exposure in male farmers and survival of in vitro fertilized pregnancies

N.H. Hjollund1,6, J.P. Bonde1, E. Ernst2, S. Lindenberg3, A.N. Andersen4 and J. Olsen5

1 Department of Occupational Medicine, Aarhus University Hospital, DK-8000, Aarhus, 2 Department of Gynaecology and Obstetrics, Aarhus University Hospital, DK-8200, Aarhus, 3 The Fertility Clinic, Copenhagen University Hospital, Herlev, DK-2730, Herlev, 4 The Fertility Clinic, Copenhagen University Hospital, Rigshospitalet, DK-2100, Copenhagen and 5 Danish Epidemiology Science Centre, Aarhus University, DK-8000, Aarhus, Denmark

6 To whom correspondence should be addressed at Åbyhojvej 17, DK-8210 Aarhus V, Denmark. e-mail: henrik.hjollund{at}dadlnet.dk


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgments
 Appendix 1
 Appendix 2
 References
 
BACKGROUND. Male-mediated spontaneous abortion is well described among animals, but less well documented in humans. Studies that include early pregnancy experience are expensive to conduct, but IVF provides new opportunities. MATERIALS AND METHODS. The Danish IVF register covers all IVF treatments performed after 1993. We mailed a questionnaire to 5879 women (response rate 68.2%). A subgroup of exposed male partners received a questionnaire on specific exposure to pesticides and growth retardants (n = 128 men, response rate 81.3%). Information on outcome was collected from national health registers. Survival of the first HCG-positive pregnancy was analysed using Cox regression. RESULTS. The proportion of pregnancies terminated by spontaneous abortion before 28 gestational weeks was 19.7% (n = 66 pregnancies), 19.7% (n = 61), 21.3% (n = 47) and 22.2% (n = 18) in pregnancies with paternal exposure to herbicides, fungicides, pesticides and growth retardants. respectively. In the reference group of 2925 pregnancies, the abortion rate was 28.4%. The differences in survival were not statistically significant. CONCLUSIONS. We found no increased risk of spontaneous abortion in IVF-treated women attributable to paternal agricultural application of pesticides and growth retardants. Exposure to potentially harmful pesticides in Denmark is relatively low, and the findings are restricted to countries with similar standards of protection.

Key words: abortion/IVF/male mediated/occupation/pesticides


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgments
 Appendix 1
 Appendix 2
 References
 
Male-mediated spontaneous abortion is a well-established phenomenon in some animal experimental studies (Olshan and Faustman, 1993Go; Marchetti et al., 1999Go). The so-called dominant lethal assay detects genetic defects occurring in the male germ cell that allows fertilization, but results in early embryonic death in the female. Several chemicals, including pesticides such as dibromchloropropane and captan, produce harmful reproductive effects in this model (Olshan and Faustman, 1993Go; Wyrobek, 1993Go; Ratcliffe, 1994Go; Friedler, 1996Go). Paternal effects of pesticides on spontaneous abortion have been examined in humans with conflicting results, and none of these studies could detect abortions in early pregnancy (Savitz et al., 1994Go; Arbuckle and Sever, 1998Go;).

It is now well documented that in humans some 30% of implanted embryos are lost, and around two-thirds of these are lost before the woman knows that she is pregnant (Rothman and Greenland, 1998Go). Few studies have included data on pre-clinical abortion (Wilcox et al., 1988Go; Zinaman et al., 1996Go; Bonde et al., 1998Go; Hjollund et al., 2000Go; Cho et al., 2002Go), but such studies are time consuming and expensive. IVF patients provide opportunities to follow a large number of pregnancies from the start, making analyses of even rare exposures possible. The male partner is not medically treated and is less likely to alter occupational activities before semen donation. Consequently, IVF pregnancies constitute an interesting model for research into paternal risk factors for spontaneous abortion.

The aim of this study of pregnancies identified in the Danish In Vitro Fertilization (DIVF) Registry was to investigate the survival of early pregnancies following paternal exposure to pesticides.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgments
 Appendix 1
 Appendix 2
 References
 
Information on IVF treatment
The DIVF register was established in January 1994 and involves compulsory reporting from both public and private clinics of all treatment cycles involving standard IVF, ICSI, frozen embryo replacement and egg donation (Danish National Board of Health, 1998Go). The coverage is believed to be close to 100%, although with missing information on some variables (Andersen et al., 1999Go). The unit of registration is the initiation of a treatment cycle. The woman is identified by her unique 10-digit personal registration number (central population register; CPR). Data are reported to the Danish National Board of Health when the cycle is terminated with a negative result or a clinical pregnancy is established, normally documented by ultrasonography at approximately the 7th gestational week, corresponding to 4 weeks after embryonal transfer (ET) (Andersen et al., 1999Go).

For each treatment cycle, the following information is available: indication for IVF treatment, origin of egg and sperm, fertilization method, date of initiation, aspiration and ET, information on the HCG analysis, and whether or not a clinical pregnancy was detected. In the year 2000, there were six public and 11 private fertility clinics. From 1994 to 2000, there were 46 614 treatments given to 15 013 women (Danish National Board of Health, 2002Go). Information from the DIVF register previously has been linked to register information on birth and spontaneous abortion for descriptive purposes (Danish National Board of Health, 1998Go, 2002).

Information on pregnancy outcome
Information of pregnancy survival until clinical detection was collected from the DIVF register, while data on pregnancy outcome were collected from other registers (the Danish hospital register and the Danish birth register) by linkage with the woman’s CPR number. Diagnoses coded as spontaneous abortion, induced abortion, ectopic pregnancy and mola (ICD-10 codes DO00–DO06) were classified as termination of a pregnancy (see Appendix 1). If the difference between gestational age calculated from the treatment cycle and the gestational age registered in the health register was ≤2 weeks, the event was assigned to this treatment cycle and the pregnancy was assigned the calculated gestational age as the date of termination of the pregnancy: if no such record was found, the pregnancy was classified as a spontaneous abortion and was assigned a gestational age of 7 weeks. Home deliveries are registered in the Birth Register, but a few births may escape registration, e.g. if the woman gives birth outside the country. To account for this, women with a clinical pregnancy but no information on either abortion or birth were linked to their CPR. In 362 of these apparently early subclinical abortions, the woman had a child with a birth date within the expected time of delivery and this child had Danish citizenship at birth. These cycles were deleted since it could not be confirmed if the children were outcomes of the cycles in question or from treatments abroad, or a child adopted later by the woman (see Appendix 2).

Treatment with donated eggs, fertilization with donor sperm or thawed eggs were excluded. We also excluded treatment cycles ending with ectopic pregnancies, mola or induced abortion (1.9% of pregnancies). Treatments after 31 December 2000 were excluded to insure complete follow-up in the health registers. The criteria for selected cycles are listed in Table I. A total of 5879 women had at least one pregnancy fulfilling the criteria.


View this table:
[in this window]
[in a new window]
 
Table I. Stepwise selection of pregnancies for the study
 
Selection of couples
In October 2001, we received data from the DIVF register to identify couples who were to receive questionnaires, and on January 10, 2003, an updated version was received. A proportion, totalling 37.9%, of the women never had an HCG-positive cycle and were initially excluded.

For the remaining women, the following information was retrieved from the Danish Central Population Register: postal address, actual marital status, CPR numbers of any current and any previous spouses and the date of changes in marital status, and CPR number of any man living at the same address as the IVF-treated woman on the day the file was processed (October 2001) as well as at the time of her first HCG-positive treatment cycle. The identity of the father is not recorded in the DIVF register, but Danish legislation demands that couples treated with IVF should be cohabiting at the time of treatment. If the woman was married at the date of the first treatment, the spouse was considered to be the father (70.6%). If the woman was unmarried but cohabiting with one man at the time of the first treatment, the cohabiting man was considered to be the father (29.4%). Otherwise the woman was excluded. Four men appeared as partner to more than one woman, and these couples were excluded. For ethical reasons, only couples who were still married or cohabiting in 2001 were selected to receive the questionnaire. A few couples had declined contact from researchers (1.2%). A total of 5879 couples who had received their first treatment after January 1, 1996 were included (Table I).

Basic questionnaire information
The basic questionnaire addressed the situation in October 2001 and was divided into a section for the mother and a section for the father. The questionnaire had 40 items on education, job, trade, specific occupational exposures as well as smoking, drinking and other lifestyle issues. The participants were asked to indicate if their actual job or trade and lifestyle had changed during the last 5 years. The women were also asked questions on reproductive history prior to their first IVF treatment. The questionnaires were mailed on October 29, 2001, and 4007 useable questionnaires (response rate 68.2%) were received after one reminder. The characteristics of the couples are shown in Table II.


View this table:
[in this window]
[in a new window]
 
Table II. Register information of first pregnancy by response status
 
Exposure-specific questionnaire to farmers
Based on type of company, trade and job function, each father was assigned a Danish ISCO code. Fathers with ISCO codes 1311 or 6130 (farmers, agricultural workers) were selected together with fathers who, in a list of job functions, had marked ‘cultivation of plants’ or ‘work at machine pool’. On April 22, 2002, these 128 men were mailed a second 20-item questionnaire. We defined the exposed group as men who personally sprayed pesticides or growth retardants. Fathers were asked to indicate use of herbicides, fungicides, insecticides and growth retardants together with several work characteristics (Table III). For each type of chemical, the father was asked if he sprayed it 5 years ago. Furthermore, an exposure index developed in previous studies was calculated to quantify dermal exposure (van Hemmen, 1993Go; Larsen et al., 1998bGo). The model allocates 0 to 2 points to each man: spraying of >20 ha per year (1 point), cultivation of potatoes or beet (1 point) and use of manual sprayer (1 point). Employees at machine pools were allocated 2 points. A total of 104 questionnaires were received after one reminder (response rate 81.3%).


View this table:
[in this window]
[in a new window]
 
Table III. Characteristics of work with pesticides and growth retardants among non-organic farmers and those employed at machine pools (n = 66)
 
Statistical methods
For each couple, only the first pregnancy was selected for analysis (n = 4007 pregnancies). The unexposed group were respondents who had answered ‘no’ to all specific occupational exposures in the basic questionnaire (n = 2925). Another reference group comprised respondents of three other trade-specific questionnaires, who had thus been subjected to the same selection mechanisms as the farmers (metal workers, greenhouse workers and workers in the reinforced plastic industry, n = 362 pregnancies). Survival curves were constructed by the Kaplan–Meier method (Allison, 1995Go). Pregnancies were right-censored at completion of the 28th week. A positive HCG test with no clinical recognized pregnancy was encountered in 14.7% of pregnancies and was assigned a survival time of 29 days. Another 5.5% became clinically pregnant but had no registered abortion or birth. A survival time of 50 days was assigned to these abortions. Abortion rate ratios were calculated by Cox regression (SAS PROC PHREG) (Allison, 1995Go) after taking these a priori selected potential confounding factors into consideration: male and female age, smoking, coffee consumption, alcohol consumption and number of embryos transferred (see Table IV). Differences between clinics, e.g. cut-off level in HCG analyses and timing of ultrasonography, were adjusted for by including clinic ID number in the STRATA statement. The proportional risk assumption was evaluated by including time-dependent covariates in the model (Allison, 1995Go).


View this table:
[in this window]
[in a new window]
 
Table IV. Risk rate ratios for spontaneous abortion with reference to exposure to pesticides and growth retardants compared with unexposed pregnancies
 
The study was approved by the Research Ethics Committee and the Danish Data Protection Agency.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgments
 Appendix 1
 Appendix 2
 References
 
Characteristics of exposure among the fathers of the 66 pregnancies are shown in Table III. All fathers were spraying herbicides, and the majority also sprayed insecticides and fungicides in 2001. In 1996, 78.8% of these men had the same job and a further 11.5% worked in another job in the agricultural trade.

The proportion of miscarriage before the 28th gestational week was 19.7, 19.7, 21.3 and 22.2% in the non-mutually exclusive groups of herbicide-, fungicide-, insecticide- and growth retardant-exposed groups, respectively. In the unexposed reference group, 28.4% of pregnancies ended in a spontaneous abortion. The survival curve of pregnancies by application of herbicides, fungicides, insecticides and growth retardants is shown in Figure 1, together with the two reference groups. The survival was consistently higher among farmers compared with the reference groups.



View larger version (18K):
[in this window]
[in a new window]
 
Figure 1. Survival of IVF pregnancies by paternal exposure to pesticides and growth retardants compared with two groups of reference pregnancies.

 
The abortion rate was at its lowest in the exposed groups, although this was not statistically significant (Table IV). After adjustment for potential confounders, the risk estimates were closer to unity.

The outcome of pregnancy was also categorized according to whether or not the father sprayed pesticides or growth retardants in the month when he supplied the semen: 1 month before; 2 months before; and 3 months before. The largest departure from unity [risk rate ratio 0.5) was found with a 2 month lag time [95% confidence interval (CI) 0.2–1.4].

Also, the risk of a hospitalized abortion was at its lowest among those exposed, although this was not statistically significant (data not shown). Similar results were found in analyses of clinically detected pregnancies only (n = 2515, 6.7% abortions).

In supplementary analyses, the following pregnancies were excluded, but with limited effect on the risk estimates: a reported IVF treatment before the first entry in the DIVF registry (n = 143), a reported spontaneous abortion before the woman’s first IVF treatment (n =696), centres with <200 recorded pregnancies (n = 327), private IVF centres (n = 848), indications other than ‘tubal factor’ (n = 1814), no indications of male factor (n = 1619), only one, or more than two embryos transferred (n = 531) and pregnancies resulting from ICSI (n = 1002). Likewise, unchanged estimates were found in the following presumably highly exposed pregnancies (see Table III ): if the father (i) sprayed more >40 h in the last 12 months; (ii) did not use gloves when mixing chemicals; (iii) used a tractor without an air filter; (iv) used a manually controlled sprayer; and (v) had the same job characteristics in both 1996 and 2001 (data not shown).


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgments
 Appendix 1
 Appendix 2
 References
 
We found no increased risk of spontaneous abortion in IVF-treated women impregnated by partners exposed to agricultural application of pesticides or growth retardants. Previous studies of reproductive function in Danish farmers found no effect of pesticides on semen quality or time to pregnancy (Larsen et al., 1998aGo; Thonneau et al., 1999Go). A Dutch study found a decreased fertilization rate, but an increased implantation rate in two partly overlapping groups of IVF patients was reported (Tielemans et al., 1999Go, 2000). The authors suggested that the ability for fertilization and successful implantation may share mechanisms. In that case, an exposure (e.g. paternal exposure to pesticides) that compromises the first step (fertilization) will result in a population of surviving embryos which may be more viable than average. The tendency in our data towards increased pregnancy survival in exposed pregnancies could theoretically be explained by such a biological mechanism, which may be augmented by the active selection of the (normally two) most promising embryos for transfer by the laboratory technician. Information on fertilization is not recorded in the DIVF register.

Validity of the IVF model
The validity of our results depends on the comparability of IVF pregnancies and normal pregnancies with respect to mechanisms leading to spontaneous abortion. Spontaneous abortion occurred in ~30% of reference pregnancies. This corresponds well to the proportion of spontaneous abortion in other IVF studies (Liu and Rosenwaks, 1991Go) and in normal pregnancies, when early pregnancy losses are included (Rothman and Greenland, 1998Go). On the other hand, in IVF pregnancies, normally two or more embryos are transferred (Danish National Board of Health, 2002Go). If more than one embryo is implanted successfully and results in an HCG-positive pregnancy, a spontaneous abortion is only registered in the case of death of all fetuses. Thus, we overestimate the survival of the individual fetus and, although we obtained similar findings when stratified on the number of transferred embryos, it should be acknowledged that in >80% of cases, more than one embryo was transferred. Little is known about the physiological processes involved in human gestation in the period from implantation of the blastocyst until a pregnancy is detectable, and the initial adhesion of the blastocyst has never been observed (Robertson et al., 1985Go; Norwitz et al., 2001Go). Moreover, mechanisms responsible for early embryonal loss may be different in IVF pregnancies and natural pregnancies, and findings based upon IVF data perhaps cannot be generalized to pregnant women in general.

Sources of bias
Information on exposure was collected retrospectively in relation to the outcome, and the scope of the study was known to the participants. Information bias as well as selection bias may therefore play a role. The pregnancy survival was highest among respondents to the basic as well as to the specific questionnaires. At least with respect to the basic questionnaire, this does indicate selection. The same mechanisms should, however, be operating in the three other groups who received a specific questionnaire, and analyses with these groups as reference revealed results similar to the main analyses. Active selection of eggs, sperm and embryos for transfer may represent the most important limiting factor for the external validity of the findings.

In conclusion, we found no indication that paternal exposure to pesticides and growth retardants confers a risk to pregnancy survival in IVF pregnancies. Exposure to pesticides does occur among Danish tractor drivers, mainly through dermal exposure (Larsen et al., 1998bGo), but bioactive exposure to potentially harmful pesticides in Denmark is believed to be low. The findings are restricted to countries with similar standards of protection.


    Acknowledgments
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgments
 Appendix 1
 Appendix 2
 References
 
Jane Boilesen is acknowledged for skilful secretarial assistance during all parts of the study, Lone Mortensen, Danish National Board of Health, for competent cooperation and fruitful discussions, and Violetta Rossen MD for valuable comments in the planning phase. The study was supported by the Danish Medical Research Council (Jnr 9902790) and The Danish Ministry of Health, Research Centre for Environmental Health (Jnr 383-16-2000). The activities of the Danish Epidemiology Science Centre are financed by a grant from the Danish National Research Foundation.


    Appendix 1
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgments
 Appendix 1
 Appendix 2
 References
 
Contradictory register data on outcome were found in some cases. In 1.4% of treatment cycles, a negative clinical pregnancy was recorded in the DIVF register despite a subsequent birth or abortion. In another 0.015% of cycles, a negative HCG, but a positive clinical pregnancy was recorded. These discrepancies were also found in the original paper forms and are considered to be coding errors. In 0.06% of treatment cycles, a hospitalization with an ICD-10 diagnosis DO00–DO06 was found followed by a code for a birth within the expected time of delivery. In these cases, the first hospitalization was regarded as a suspected event (e.g. a threatening abortion).


    Appendix 2
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgments
 Appendix 1
 Appendix 2
 References
 
The procedure to identify subclinical abortions was evaluated in treatment cycles terminated by an early spontaneous abortion. In seven cases, the woman also had a registered child with a birth date within the expected time of delivery. In all seven cases, the child in question had a registered previous citizenship, indicating that the child was adopted from abroad. Adoption is not registered in CPR. Adoption is common in this population, but only very few Danish-born children are available for adoption. The 100% match indicates that few births have escaped identification.


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgments
 Appendix 1
 Appendix 2
 References
 
Allison PD (1995) Survival Analysis Using the SAS System: A Practical Guide. SAS Institute, Cary, NC.

Andersen AN, Westergaard HB and Olsen J (1999) The Danish in vitro fertilisation (IVF) register. Dan Med Bull 46,357–360.[ISI][Medline]

Arbuckle TE and Sever LE (1998) Pesticide exposures and fetal death: a review of the epidemiologic literature. Crit Rev Toxicol 28,229–270.[Medline]

Bonde JP, Hjollund NH, Jensen TK, Ernst E, Kolstad HA, Henriksen TB, Giwercman A, Skakkebæk NE Andersen A and Olsen J (1998) A follow-up study of environmental and biologic determinants of fertility among 430 Danish first pregnancy planners: design and methods. Reprod Toxicol 12,19–27.[CrossRef][ISI][Medline]

Cho SI, Goldman MB, Ryan LM, Chen C, Damokosh AI, Christiani DC, Lasley BL, O’Connor JF, Wilcox AJ and Xu X (2002) Reliability of serial urine HCG as a biomarker to detect early pregnancy loss. Hum Reprod 17,1060–1066.[Abstract/Free Full Text]

DanishNationalBoardofHealth(1998) [In vitro befrugtning. Behandlinger og resultatet heraf for private og offentlige fertilitetsklinikker i Danmark i 1994og 1995] English summary. National Board of Health, Copenhagen.

DanishNationalBoardofHealth(2002) [IVF-behandlinger i Danmark i perioden 1994–2000. Del 1: Fertlitetsbehandling] Nye tal fra Sundhedsstyrelsen, 6.

Friedler G (1996) Paternal exposures: impact on reproductive and developmental outcome. An overview. Pharmacol Biochem Behav 55,691–700.[CrossRef][ISI][Medline]

Hjollund NH, Bonde JP, Jensen TK, Henriksen TB, Andersson AM, Kolstad HA, Ernst E, Giwercman A, Skakkebaek NE and Olsen J (2000) Male-mediated spontaneous abortion among spouses of stainless steel welders. Scand J Work Environ Health 26,187–192.[ISI][Medline]

Larsen SB, Giwercman A, Spano M and Bonde JP (1998a) A longitudinal study of semen quality in pesticide spraying Danish farmers. The ASCLEPIOS Study Group. Reprod Toxicol 12,581–589.[CrossRef][ISI][Medline]

Larsen SB, Joffe M and Bonde JP (1998b) Time to pregnancy and exposure to pesticides in Danish farmers. ASCLEPIOS Study Group. Occup Environ Med 55,278–283.[Abstract]

Liu HC and Rosenwaks Z (1991) Early pregnancy wastage in IVF (in vitro fertilization) patients. J In Vitro Fertil Embryo Transf, 8,65–72.[ISI][Medline]

Marchetti F, Lowe X, Bishop J and Wyrobek AJ (1999) Absence of selection against aneuploid mouse sperm at fertilization. Biol Reprod 61,948–954.[Abstract/Free Full Text]

Norwitz ER, Schust DJ and Fisher SJ (2001) Implantation and the survival of early pregnancy. N Engl J Med 345,1400–1408.[Free Full Text]

Olshan AF and Faustman EM (1993) Male-mediated developmental toxicity. Reprod Toxicol 7,191–202.[CrossRef]

Ratcliffe JM (1994) Paternal exposures and embryonic or fetal loss: the toxicologic and epidemiologic evidence. In Mattison DR and Olshan AF (eds), Male-mediated Developmental Toxicity. Plenum Press, New York, pp. 185–195.

Robertson WB, Brosens I and Landells WN (1985) Abnormal placentation. Obstet Gynecol Annu 14,411–426.[Medline]

Rothman KJ and Greenland S (1998) Modern Epidemiology. Lippincott-Raven, Philadelphia, pp. 311–326.

Savitz DA, Sonnenfeld NL and Olshan AF (1994) Review of epidemiologic studies of paternal occupational exposure and spontaneous abortion. Am J Ind Med 25,361–383.[ISI][Medline]

Thonneau P, Abell A, Larsen SB, Bonde JP, Joffe M, Clavert A, Ducot B, Multigner L and Danscher G (1999) Effects of pesticide exposure on time to pregnancy: results of a multicenter study in France and Denmark. ASCLEPIOS Study Group. Am J Epidemiol 150,157–163.[Abstract]

Tielemans E., van Kooij R, te Velde ER, Burdorf A and Heederik D (1999) Pesticide exposure and decreased fertilisation rates in vitro. Lancet 354,484–485.[CrossRef][ISI][Medline]

Tielemans E, van Kooij R, Looman C, Burdorf A, te Velde E and Heederik D (2000) Paternal occupational exposures and embryo implantation rates after IVF. Fertil Steril 74,690–695.[CrossRef][ISI][Medline]

vanHemmen JJ (1993) Predictive exposure modelling for pesticide registration purposes 2. Ann Occup Hyg 37,541–564.[ISI][Medline]

Wilcox AJ, Weinberg CR, O’Connor JF, Baird DD, Schlatterer JP, Canfield RE, Armstrong EG and Nisula BC (1988) Incidence of early loss of pregnancy. N Engl J Med 319,189–194.[Abstract]

Wyrobek AJ (1993) Methods and concepts in detecting abnormal reproductive outcomes of paternal origin. Reprod Toxicol, 7 Suppl 1,3–16.[CrossRef]

Zinaman MJ, Clegg ED, Brown CC, O’Connor J and Selevan SG (1996) Estimates of human fertility and pregnancy loss. Fertil Steril 65,503–509.[ISI][Medline]

Submitted on November 18, 2003; accepted on January 6, 2004.