Effects of Pesticide Exposure on Time to Pregnancy - Oxford Journals

80 downloads 129 Views 641KB Size Report
groups were included: conventional farmers and green- house workers. ... Alsace health insurance plan distributed 1,326 question- naires between the start of ...
American Journal of Epidemiology Copyright © 1999 by The Johns Hopkins University School of Hygiene and Public Health All rights reserved

Vol. 150, No. 2 Printed in U.S.A.

Effects of Pesticide Exposure on Time to Pregnancy Results of a Multicenter Study in France and Denmark

P. Thonneau,1 A. Abell,2 S. B. Larsen,2 J. P. Bonde,2 M. Joffe,3 A. Clavert,4 B. Ducot,1 L. Multigner,1 and G. Danscher2 for the ASCLEPIOS* Study Group The aim of this study was to determine whether there was a relation between male exposure to pesticides and the amount of time needed to conceive (time to pregnancy) for farmers and agricultural workers in France and Denmark. The authors used retrospective studies to compare the time to pregnancy of couples in which the man was exposed to pesticides during the year before the birth of their youngest child with that of couples in which the man was not exposed. In 1995 and 1996, the authors studied 362 French rural workers (142 exposed to pesticides and 220 not exposed), 449 Danish farmers (326 conventional farmers exposed to pesticides and 123 nonexposed organic farmers), and 121 Danish greenhouse workers exposed to pesticides. The fecundability ratio for exposure to pesticides (Cox model, before and after adjustment for confounding factors) did not differ from 1 in any of the three populations. In France, the adjusted fecundability ratio was 1.17 (95% confidence interval (Cl) 0.89-1.55) for exposed and nonexposed agricultural workers. In Denmark, it was 1.09 (95% Cl 0.82-1.43) for exposed and nonexposed farmers and 0.83 (95% Cl 0.69-1.18) for greenhouse workers and nonexposed farmers. Thus, this study found no relation between fertility (time to pregnancy) and male exposure to pesticides. Am J Epidemiol 1999; 150:157-63. agrochemicals; fertility; men; occupational exposure; pesticide residues; pesticides; pregnancy

Several recent international studies have indicated a reduction in sperm counts and an increase in the numbers of testicular cancers and male genital malformations (cryptorchidism, hypospadia) during the past 20 years (1-6). It has been suggested that these phenomena may be caused by endocrine disrupters competing directly with androgens or affecting various androgendependent processes. Some of the most important known endocrine disrupters are pesticides (7). Few epidemiologic studies have investigated the effects of pesticides on human fertility. In 1979,

Whorton et al. (8) demonstrated lower fertility levels and sterility in agricultural workers exposed to dibromochloropropane on the banana plantations of Costa Rica. The spermograms performed at the time demonstrated clear changes in sperm characteristics (8). Ratcliffe et al. (9) demonstrated changes in the sperm characteristics of agricultural workers exposed to ethylene dibromide (a pesticide used in papaya production). In the Netherlands, de Cock et al. (10) showed that there was a correlation between the pesticide spraying season and the amount of time needed by couples to conceive a child. Fertility, defined as successful conception, is poorly estimated by the number of children born in industrially advanced societies. The time span from the point at which a couple decides to have a child to the clinical detection of a pregnancy is a useful alternative. The amount of time taken to conceive has been put forward by Baird et al. (11) as a suitable outcome variable for epidemiologic research concerning environmental agents that may affect reproduction. This indicator, time to pregnancy, expressed as number of menstrual cycles, is directly correlated with the fecundability of each couple (time to pregnancy = 1/fecundability). The fecundability of a couple is therefore the probability per menstrual cycle of a detectable pregnancy. Fecundability depends on many biologic events,

Received for publication May 18,1998, and accepted for publication November 20, 1998. Abbreviations: ASCLEPIOS, Occupational Hazards to Male Reproductive Capacity: A European Concerted Action; Cl, confidence interval. 1 Institut National de la Sante et de la Recherche Medicate (INSERM), Unite 292, Bicetre Hospital, Le Kremlin-Bicetre, France. 2 Steno Institute of Public Health, Department of Occupational Medicine, Aarhus University Hospital, Norrebrogade 44, Aarhus C, Denmark. 3 Imperial College School of Medicine, Norfolk Place, London W2 1PG, United Kingdom. 4 Centre d'Etude et de Conservation du Sperme, Hopital Civil de Strasbourg, 67091 Strasbourg Cedex, France. •ASCLEPIOS [Occupational Hazards to Male Reproductive Capacity: A European Concerted Action] is coordinated by the Steno Institute of Public Health, University of Aarhus (Aarhus, Denmark). Reprint requests to Dr. P. F. Thonneau, Groupe de Recherche sur la Fertilite Humaine, Service d'Urologie-Andrologie, Hopital La Grave, Place Lange, 31052 Toulouse Cedex, France.

157

158

Thonneau et al.

including gametogenesis, transport of germ cells, fertilization, transport of the embryo, and implantation. Environmental and toxic substances may interfere with these processes. Time to pregnancy therefore appears to be an effective, easy-to-determine indicator (obtained by a simple questionnaire) for evaluating any deleterious environmental effects on fecundability. In 1992, Rowland et al. (12) analyzed mailed questionnaires completed by female dental assistants and showed that women exposed to high levels of nitrous oxide in the workplace were significantly less fertile than women who had not been exposed to such an atmosphere. In a Danish population-based survey of lifestyle factors and subfecundity, Olsen (13) showed that smoking and consumption of hot caffeinated beverages were significantly associated with waiting times of up to 1 year. Delayed conception associated with caffeine consumption was also reported in 1993 by Hatch and Bracken (14) in a cross-sectional study of a population of US married women. Alderete et al. (15) investigated time to conception, cigarette smoking, and coffee consumption in primigravid women participating in the Child Health and Development Studies in the San Francisco Bay Area between 1959 and 1966. They found that cigarette smoking was associated with low levels of fertility (expressed as time to pregnancy). Bolumar et al. (16) carried out a retrospective study of 4,000 couples from 10 European countries. They also demonstrated an association between female smoking and time to pregnancy. We investigated the relation between male pesticide exposure and time to pregnancy by conducting a retrospective epidemiologic study. This study was carried out in two European countries as part of a European research project (ASCLEPIOS). In 1995, we studied conventional farmers, organic farmers, and greenhouse workers in Denmark. In 1996, we studied vineyard workers and rural workers in France. MATERIALS AND METHODS Target populations and data collection

In France, all male agricultural workers (vineyard workers and other rural workers) of the Alsace region were asked at their annual medical visit to participate in the study. For men who agreed to participate, the first part of the questionnaire (concerning male health) was completed with the help of the occupational health physician during the medical examination. The part of the questionnaire concerning female health was then given to the man, completed at home by the couple, and returned by mail to the occupational health physician. The exposed group consisted of workers who had used pesticides during the calendar year before the

birth of their youngest child. The control group consisted of rural workers from Alsace who had not been exposed to pesticides during the year before the birth of their youngest child. In Denmark, we selected two samples from an official Ministry of Agriculture list: all organic fanners in Jutland and a random sample of conventional farmers in Jutland. We also selected male greenhouse workers (all exposed to pesticides) from a list of members of the Danish Gardeners' Union. Thus, two exposed Danish groups were included: conventional farmers and greenhouse workers. The Danish control group consisted of organic and conventional farmers who had not been exposed to pesticides during the calendar year before the birth of their youngest child. Telephone interviews were conducted by trained female interviewers. For both countries, some couples were excluded from the analysis: those in which the man had never tried to make his partner pregnant (married or not) or in which the wife was less than 6 months pregnant (because this study was focused on the youngest child and we could not make assumptions about the outcome of a pregnancy of less than 6 months—e.g., abortion or miscarriage) and those in which the man had not fathered a child. Cases in which the last pregnancy was due to failure of birth control were excluded. (In these cases, the couples systematically gave a time to pregnancy of 1. However, since we did not know when the period of "exposure to pregnancy" began, we could not calculate an exact time to pregnancy.) We also excluded couples in which the man had not worked in agriculture or in greenhouses during the year before the birth of their youngest child. Questionnaire

The questionnaire was developed by an expert working group of the ASCLEPIOS European network. The English version was translated into Danish and French. Before it was used in the study, the questionnaire was validated by testing it in a sample of about 100 workers. We also made use of the experience and results of a European network that had previously dealt with the same topics using similar methodology (infertility and environmental risk factors). The most important question concerned the amount of time needed by the couple to conceive their youngest child (time to pregnancy): "How many months did it take your wife to get pregnant?" We also obtained information on age, parity, smoking, and contraceptive method used (if any) during the year before the birth of the youngest child. Information on the job title, type of work, and pesticide exposure of the man during the year before the birth of the couple's youngest child was also requested. In cases of pestiAm J Epidemiol

Vol. 150, No. 2, 1999

Pesticide Exposure and Time to Pregnancy

cide exposure, the men were asked to supply a list of the pesticides (by chemical family) they had used.

TABLE 1. Selection of the study population in a study of male pesticide exposure and time to pregnancy, ASCLEPIOS,* 1995-1996 France (vineyard workers and controls)

Statistical analysis

The various study populations were analyzed separately. With SAS statistical software (SAS Institute, Cary, North Carolina), we analyzed time to pregnancy as a function of pesticide exposure using a discrete Cox model to calculate crude fecundability ratios— i.e., the ratio between the probability of conception in exposed couples and the probability in nonexposed couples. (We calculated not a "true" fecundability ratio but a conditional fecundability ratio, because we only included couples who conceived a child.) We censored times to pregnancy of more than 13 months to avoid medical intervention bias (e.g., infertility treatment for the man or the woman) and because 95 percent of couples conceived within 1 year of unprotected intercourse. We then incorporated all of the potentially confounding variables (female age, parity, male and female smoking, and recent use of oral contraceptives) into the Cox model to obtain adjusted fecundability ratios. RESULTS

In France, the occupational health physicians of the Alsace health insurance plan distributed 1,326 questionnaires between the start of February 1996 and the end of April 1996. Of these, 976 were returned by mail (74 percent). We excluded 164 couples who had never tried to have a child, one couple in which the wife was less than 6 months pregnant at the time of the study, 386 couples whose youngest child was born before 1986, five couples in which pregnancy was due to failure of birth control, and 58 couples for whom time to pregnancy was unknown. Of the remaining 362 questionnaires, 142 were from men exposed to pesticides during the year before the birth of their youngest child and 220 were from men not exposed to pesticides (table 1). In Denmark (1995), 73 of the 1,146 targeted conventional and organic farmers who were sent a letter about the study were ineligible for inclusion; 169 of the 1,073 eligible couples in the study population refused to participate (response rate = 84 percent). Of the remaining 904 couples, we excluded 133 because they had never tried to have a child or because the wife was less than 6 months pregnant at the time of the study, 137 whose last child was born before 1986, 89 whose most recent pregnancy was due to failure of birth control, and 30 whose time to pregnancy was unknown. We also excluded 66 men who had not worked in agriculture during the year before the birth of their youngest child. Of the 449 completed quesAm J Epidemiol Vol. 150, No. 2, 1999

159

Target population Ineligible Study population Nonresponders

Denmark 1 (conven- Denmark (greentional and house organic farmers) ers 1,146

73

899 312

1,326 350

1,073 169

587

Participants Wife