Parental Concentration of Dichlorodiphenyl
Dichloroethene and Polychlorinated Biphenyls
in Michigan Fish Eaters and Sex Ratio
in Offspring

Journal of Occupational Medicine v.44, n.1, Jan02

Wilfried Karmaus, Dr med, MPH
Suiying Huang, MS
Lorraine Cameron, PhD

Contamination of fish in the Great Lakes generated three surveys assessing polychlorinated biphenyls (PCB) serum concentration in Michigan anglers: 1973 to 1974, 1979 to 1982, and 1989 to 1991. This cohort provided 1177 individuals with PCB determinations. In 2000, we conducted telephone interviews with parents on their children's birth characteristics. We estimated the sex odds ratio for parental PCB and dichlorodiphenyl dichloroethene concentrations using generalized estimation equations. We identified 208 offspring, within 101 families, born after 1963, with paternal measurements of both PCB and dichlorodiphenyl dichloroethene. When controlling for maternal exposure and parental dichlorodiphenyl dichloroethene, the sex odds ratio was increased if paternal PCB concentrations exceeded 8.1 µg/L (sex ratio, 2.29; 95 % confidence interval, 1.11 to 4.74). Thus, paternal exposure was linked to a higher proportion of male offspring. These findings are opposite those reported for the Seveso study and are in accordance with those for dioxin exposure in the American veterans study. (J Occup Environ Med. 2002;44:8-13)

From the Department of Epidemiology, Michigan State University (Dr Karmaus, Dr Huang), and Division of Environmental Epidemiology, Michigan Department of Community Health, Lansing (Dr Cameron).

Address correspondence to: Dr Wilfried Karmaus, Department of Epidemiology, Michigan State University, 4660 S. Hagadorn Road, Suite 600, East Lansing, MI 48823; karmaus@pilot.msu.edu.

The primary sex ratio is defined as the number of boys to girls before or at conception; the secondary sex ratio is ascertained at birth.(1,2) Several studies have linked changes in the secondary sex ratio to exposure to organochlorine compounds (OC).(3-7) There is disagreement whether the effect may be contributed to maternal or paternal exposure or both.(6,8)

Previous investigations have shown that anglers and fish eaters in Michigan have higher serum concentrations of dichlorodiphenyl dichloroethene (DDE) and polychlorinated biphenyls (PCBs).(9-11) DDE is the organic metabolite of the pesticide dichlorodiphenyl trichloroethane, otherwise known as DDT. DDT had widespread use as an insecticide from 1940 to 1960. The use of DDT was banned in the United States during the early 1970s. DDT is a lipophilic organic compound that is known to bioaccumulate. The long biological half-life of the metabolite p,p'-DDE (7 years) accounts for the ubiquitous exposure in the general population.(12)  PCBs are a class of lipophilic compounds. From the

1920s, until they were banned in the United States in 1975, PCBs had widespread industrial use in capacitors and dielectric fluids. Reports also indicate that the half-lives of various PCB congeners can range up to 70 years.(13,14)  DDE and PCB are hypothesized as endocrine disruptors, and a variety of studies have demonstrated their estrogenic, antiestrogenic, and androgen competing properties. (15-16)

The Great Lakes have been polluted with DDE and PCB, and high concentrations of these substances were found in fish samples from the Great Lakes. Contamination of fish in the Great Lakes had generated three surveys to assess the PCB burden in Michigan anglers in 1973 to 1974, 1979 to 1982, and 1989 to 1991. This cohort provided a total of 1177 individuals with PCB and DDE determinations." We investigated whether the secondary sex ratio in the offspring of anglers was associated with maternal or paternal exposure to DDE and PCB. According to the definition of the sex ratio, we investigated the odds of offspring being of male sex for maternal and paternal OC exposures.

In three surveys conducted from 1973 to 1991, fish eaters and their spouses were recruited by field visits to sites of fishing activities (eg, docks, marinas, and bait shops) in 11 Lake Michigan shoreline counties. Direct contacts and referrals were used in the field to recruit participants. In the first survey, 1973 to 1974, 156 anglers were recruited. The second survey, 1979 to 1982, included 115 members from the first survey plus 1140 new participants. In the third survey, 1989 to 1991, 717 of the sample of the second survey participated again plus 11 additional individuals.(11,17)  This cohort provided a total of 1177 individuals with PCB, polybromated biphenyls, and DDE determinations.

In 2000, we approached this cohort again. We excluded deceased individuals based on the Michigan Vital Records data up to summer 1999 (n = 119). The Michigan Department of Community Health sent out 886 letters to separate addresses, families, and/or individuals representing 621 original families. If the couple became separated/divorced, a letter was sent to each individual. Approximately 2 weeks later, trained interviewers called each participant and reexplained the purpose of the study. After verbal consent was given, an interview was conducted regarding their offspring's birth characteristics. Of this cohort, an additional group of 27 individuals were deceased, 88 could not be located, and 47 were ineligible (had no children). Participants constituted 412 families with 647 individual parents. The new study was approved by the review board on human subjects of Michigan State University and the Michigan Department of Community Health.

At the time of each of the past three surveys, participants were encouraged to provide a nonfasting blood sample for analyses of serum levels of PCB and other contaminants. All serum samples were analyzed at the Health Risk Assessment Laboratory of the Michigan Department of Community Health in Lansing. Serum PCB and DDE levels were determined, using a modification of the Association of the Official Analytical Chemists, approved Webb-McCall packed column gas chromatography technique. The laboratory analytical methods used are described in detail elsewhere.(18-20)  The procedure used methanol-ether/ hexane extraction, microflorisil column cleanup, and silicon gel column separation technique before injection into a programmed electron detection chromatograph. In the 1973 to 1974 and 1979 to 1982 surveys, contamination was based on Aroclor 1254 and 1260 standards. In the 1989 to 1991 survey, Aroclor 1016 and 1260 standards were used. For the sake of comparison of the exposures, we used PCB measurements based on the Aroclor 1260 standard, which were available for all samples. Laboratory values reported as less than the detectable limit for Aroclor 1260 (3 µg/kg) were assigned as 1.5 µg/kg. Less than 5% of the samples were below the threshold in all three surveys.

We asked the parents for birth date, sex, birth weight, gestational age, and breast/formula-fed information for all children. In addition, for each child, we inquired whether the parents were biological. To investigate the reliability of the information, we repeated the survey in a subsample.

The sex ratio is defined as an odds, male to female. We investigated the odds of offspring being of male sex for maternal and paternal exposures. The unit of the analyses is the biological child. We linked each child with his or her maternal and paternal DDE and PCB serum concentrations. We used the parental exposure measurement that was closest to the birth of the child. Parental exposure values included up to three measurements for PCB (1973 to 1974, 1979 to 1982, and 1989 to 1991) and two measurements for DDE (1979 to 1982, and 1989 to 1991). We categorized the PCB and DDE measurements into two groups so that approximately half of the children were in the lower and higher categories for their maternal and paternal exposures. For PCB, we used >8.1 µg/L versus <8.1 µg/L; for DDE, we used >11 µg/L versus <11 µg/L All values below the detection limit were in the less-exposed group (reference).

We estimated the sex odds ratio (SOR) and the 95% confidence intervals (CIs) using Generalized Estimation Equations.(21,22)  This statistical model took into account that offspring data were correlated if coming from the same family. In all models, we controlled for the calendar period of child's birth (1963 to 1972, 1973 to 1982, after 1983), age of the mother at birth (15 to 24, 25 to 29, 30 years and older), and whether there were older brothers in the family. The associations of maternal and paternal DDE and PCB exposure were estimated simultaneously. For all children born after 1963, we estimated the SOR 10 years before the first PCB determination in their parents.

For quality control purposes, we randomly chose 30 parents and reinterviewed them. To assess reliability of the information, we calculated kappa and the intraclass correlation coefficient (23)  In addition, we investigated the stability of the different halogenated organic compounds measurements by means of correlation analyses.

TABLE 1 Timing, Number, and Results of PCB and DDE Measurements in Parents*

                                                Year of Determination                  
                               1973-1975        1979-1982          1989-1991      Total
PCB
  Maternal values (n)             30          237               178               245
    Median (5%-95% values, µg/L)  5.0 (1-22)  9.1 (1.5-34.6)      8.0 (0-28.6)
  Paternal values (n)             34          281               214               354
    Median (5%-95% values, µg/L)  12 (3-38)    14.9 (64.9)       13.4 (3.8-65.8)
  No. of families included        42          357               277               366
DDE
  Maternal values (n)                         234               177               239
    Median (5%-95% values, µ/L)    -           12.5 (3.5-36.0)    7.3 (1.6-26.3)
  Paternal values (n)                         279               212               286
    Median (5%-95% values, µg/L)   -           18.5 (4.3-140.2)  10.8 (2.0-53.1)
  No. of families Included         -          357               276               361

* PCB, polychlorinated biphenyls; DDE, dichlorodiphenyl dichloroethene.

We received consent from 398 families to participate in our new study (proportion of participation, 64.1 %). We excluded 22 adult offspring from the parental group when the adult offspring had participated in the original study, but we kept them as offspring. The sample of the parents constituted 361 families (Table 1).

We repeated the interview with 30 parents. Kappa for the number of children was 0.91 (89 vs 91 children, including one stillbirth in 1952; the other live birth was born in 1949, lower 95% CI for kappa 0.8). We had total agreement for the sex of the child among those 89 reported.

The total number of all children was 1050. We then restricted the offspring sample to those born after 1963 and for whom we had both biological parental measurements (n = 208). Among them, 57.7% were born between 1963 and 1972, 33.7% between 1973 and 1982, and 8.7% were born after 1983.

Median PCB values for the mothers, including repeated measurements in the three surveys, were lower than for the fathers (Table 1). For offspring born after 1963 with both parental exposure measurements (n = 208), the median maternal exposure was 6.9 µg/L for PCB and 9.4 µg/L for DDE, The median paternal exposure was 9.9 µg/L for PCB and 12.2 µ/L for DDE. Maternal and paternal PCB exposures were moderately correlated (Spearman's rank correlation coefficient, 0.59; P < 0.0001), as were maternal and paternal DDE exposures (Spearman's rank correlation coefficient, 0.56; P < 0.0001). Also, maternal PCB and DDE values (Spearman's rank correlation coefficient, 0.65; P < 0.0001) and paternal PCB and DDE measurements were correlated (Spearman's rank correlation coefficient, 0.79; P < 0.0001).

In the assessment of risk factors for the SOR, we took into account that observations were correlated if coming from the same family (208 offspring from 101 families, generalized estimation equations). _ The results show that for paternal PCB concentrations exceeding 8.1 µ/L, the SOR of fathering a boy was increased (SOR, 2.29; 95% CI, 1.11 to 4.74). None of the other risk factors (maternal DDE and PCB, paternal DDE) and none of the confounders (maternal age, calendar year of birth, and preceding sisters) showed statistically significant associations, with the exception of preceding brothers (Table 2). The SOR of having a boy was reduced when there were previous brothers (SOR, 0.56; 95% CI, 0.32 to 0.96).

Anglers and fish eaters in Michigan have been exposed to a mixture of OC present in fish from the Great Lakes. We used data on PCB and DDE determined between 1973 and 1991 by the Michigan Department of Community Health. Our study suggested that fathers with a PCB concentration >8.1 µ/L were more likely to have boys then those with a serum PCB concentration <8.1 ppb.

Our study group included only parents who were still alive by the year 2000. Nevertheless, this approach should not be a major drawback, because the deceased individuals most likely had had their children before 1963, and our analyses included only children. born after 1963. Of the original cohort, 64.1% of the families participated in our study (n = 398). This proportion is a conservative estimate because we could exclude deceased parents on the basis of Michigan vital statistics only and not on statistics from other states when the participants had moved. Of the 886 letters we mailed, we were unable to locate 88 participants.

In the questionnaire, we asked the parents for birth date, sex, birth weight, gestational age, and breast-/ formula-fed information for all children, We observed minor disagreement for the number of children that were reported when we reinterviewed the randomly chosen 30 parents. Nevertheless, the kappa was acceptable (91 %). The agreement for the sex of the children was 100%.

TABLE 2
Proportion of Boys in Maternal and Paternal PCB and DDE Exposure Groups and SOR Controlled for Year of Birth, Previous Sisters and Brothers, and Maternal Age* (n = 208 of 101 Families, Generalized Estimating Equations)

                                             Adjusted
     Exposure (µg/L)          n    Boys (%)    SOR     95% CI  
Maternal PCB concentration                     0.71   0.35-1.45
    <8.1                     119    54.6
    >8.1                      89    49.4
Paternal PCB concentration                     2.29   1.11-4.74
    <8.1                      82    45.1
    >8.1                     126    57.1
Maternal DDE concentration                     0.71   0.35-1.45
    <11                      115    53.9
    >11                       93    50.5
Paternal DDE concentration                     0.82   0.39-1.74
    <11                       94    51.1
    >11                      114    53.5

* PCB, polychlorinated biphenyls; DDE, dichlorodiphenyl dichloroethene;
  SOR, sex odds ratio; CI, confidence interval.

For parental OC measurements, we used PCB and DDE determinations conducted between 1973 and 1981 using the Webb-McCall method. These measurements did not provide PCB congeners. Therefore, we have only total PCB values, compared with the Aroclor 1260 standard. Some PCB congeners may exhibit estrogenic behavior, some antiestrogenic or androgenic.(15,16) Because we were unable to distinguish PCB congeners in this study, we could not separate the potentially different congener behavior. We used the parental PCB concentrations that were determined closest to the child's birth date. We did not estimate the concentration of PCB at times other than at serum collection.(24,25)  Models to estimate year-to-year concentrations were developed for situations characterized by one major short-term exposure, such as the accident in Seveso or the polybromated biphenyls accident in Michigan. However, fish eating anglers have more continuous exposure. Our data showed that the three measurements taken in the 1970s, 1980s, and 1990s were quite stable. Among consumers of sport-caught fish in the Great Lakes, serum PCB levels did not significantly decrease, probably because of continued exposure and the long half-life of PCB (17)

Our results are contradictory to findings in two other studies. One Canadian study showed an altered proportion of offspring gender .related to wood preservatives encompassing a range of toxic substances contaminating male sawmill workers. (26) Among 19,675 offspring, 48.6% were boys and 51.4% were girls. James concluded from these findings that men exposed to dioxin-contaminated wood preservatives might father an increased number of daughters. (4)  In 168 couples from Seveso, Mocarelli et al found a decreased number of boys (113 boys vs 137 girls) related to the paternal exposure to 2,3,7,8-tetracWorodibenzop-dioxin (TCDD).(6) Maternal exposure was included, but the association with sex of offspring was not significant. Among all original families, 130 men and 154 women with measured TCDD concentration were excluded because they were married to partners outside the stud) region. Among the excluded 15z mothers, however, the number o boys was increased (127 boys vs 106 girls).

In Taiwan, thousands of people were exposed to PCB and polychlorinated dibenzofurans from cooking oil (Yu-chen study). A follow-up study of women registered with the health department showed . that ?4 women reported 137 live births occurring between June 1978 and spring 1985. The study showed that the sex ratio was not altered (68 boys to 69 girls).(27)

Our results, however, are in accordance with three other studies. Contrary to the abovementioned study from Taiwan, parents exposed to PCB and polychlorinated dibenzofurans from rice oil in Japan (Yusho accident) gave birth to a higher number of boys (49 boys to 36 girls; SOR, 1.36). (7) However, the deviation of the sex ratio was not statistically significant. In the Veterans of Operation Ranch Hand study, Michalek et al found that a reduced number of daughters was associated with TCDD exposures. (5) With a restriction to 483 children born no later than 1 year after the father's service, 394 had a paternal exposure of < 10 ppt of TCDD and 88 had a paternal exposure of > 10 ppt. These data showed an SOR of 1.52 (95% CI, 0.96 to 2.43). In this study, dioxin was not measured in wives and partners of the veterans, and their exposures to dioxin or other chemicals were unknown. For mixed exposures of polychlorinated dibenzodioxins and dibenzofurans and heavy metals in residents of a development site in Germany, Fertmann et al showed a significant increase in male children in parents who themselves were exposed to these contaminants at a younger age. (28)

Of the above-reviewed sex ratio studies regarding exposure to OC, two showed a lower proportion of boys, three showed a higher proportion of boys, and one did not find an altered sex ratio. Our study, and three other reports, found an increase in boys in exposed fathers. Different exposure measurements do not seem to explain these different results. The Michigan fish eaters study, along with the Seveso, the American veteran, and the Yu-chen and Yusho studies, used actual measured OC concentration. Individual OC measurements were not used in the Canadian wood-preservative study and the German waste-site investigation.

Biggar et al suggested that previous sex ratio studies have overlooked the numbers and sex of the older children as predictors .(29) Using national Danish data, they suggested that a declining sex ratio is mainly due to declining family size, biological heterogeneity, and child sex preference. In our study, we also controlled for the effect of older children. We found that having previous brother(s) reduced the chance of having more boys (SOR, 0.56; 95 % CI, 0.32 to 0.96).

The question whether these suggested effects were due to paternal or maternal exposure is still not clear. In the Taiwan sex ratio study on cooking oil, only female exposure was addressed. In the cooking oil study from Japan, at least one parent was included (mother or father), whereas Michalek et al investigated male exposures in veterans, and no exposure values were available for the veterans' wives, who were not exposed. Mocarelli et al included both maternal and paternal exposures in the explanatory model. Our analysis included only children with both parental measurements. Not including the partners' OC concentration in the model could produce confounded effects if both parents were exposed and their levels were associated.

Nevertheless, a correlation of maternal and paternal OC values may introduce multi-colinearity. Our data did not indicate that the correlation of maternal and paternal DDE and PCB produced numerical problems (tolerance >0.61 for all predictors). Maternal and paternal DDE serum concentrations did not gain statistical significance. However, it confounded the PCB effect and could not be removed from the model.

The biological mechanisms involved are not yet understood. One endocrinological explanation suggested that hormone levels of both parents around the time of the conception partially control the sexes of mammals .(30) Exposure of men to OC seems to be associated with low testosterone or high gonadotropin levels.(31)

Toxicological explanations focus on a differential survival of female and male fetuses. An equal number of X- and Y-chromosomes is reported for human sperm .(2, 32)  Kukharenko, however, claimed for normal conditions that the sex ratio is increased at the beginning of the pregnancy and then declines. At the end of the second month, he observed 151 male to 100 female fetuses, at the end of the third month, 132 male to 100 female fetuses.' Because male exposure to halogenated hydrocarbons may be associated with fetal loss,(33-35) an increased proportion of female infant miscarriages is another pathway that may increase the number of male offspring.

Our results suggest that paternal PCB serum concentrations in fish eaters and anglers, but not maternal concentrations or DDE concentrations, are associated with increased odds of male offspring. The contradictory results from two previous studies, the Seveso investigation (fewer boys) and the Operation Ranch Hand study (fewer girls), and poorly understood biological mechanisms point to the need for more research to better protect public health from potentially endocrine-disrupting chemicals.

The authors thank the Michigan Department of Community Health, Environmental Epidemiology Division, for providing access to the historic cohort of anglers and the Analytical Chemistry Section for providing the data on PCB and DDE exposure. The authors appreciate suggestions from Elizabeth Marshall on a prior manuscript. This work was supported by a grant from the Agency for Toxic Substances and Disease Registry (H75/ATH582536-04-2).

1. Kukharenko VI. The primary sex ratio in man (analysis of 1014 embryos). Genetika. 1970;6:142-149.

2. Lobel SM, Pomponio RJ, Mutter GL. The sex ratio of normal and manipulated human sperm quantitated by the polymerase chain reaction. Fertil Steril. 1993;59: 387-392.

3. Mocarelli P, Brambilla P, Gerthoux PM, Patterson DGJ, Needham LL. Change in sex ratio with exposure to dioxin. Lancet. 1996;348:409.

4. James WH. The sex ratio of offspring sired by men exposed to wood preservatives contaminated by dioxin. Scand J Work Environ Health. 1998;24:416.

5. Michalek JE, Rahe AJ, Boyle CA. Paternal dioxin and the sex of children fathered by veterans of Operation Ranch Hand. Epidemiology. 1998;9:474-475.

6. Mocarelli P, Gerthoux PM, Ferrari E, et al. Paternal concentrations of dioxin and sex ratio of offspring. Lancet. 2000;355: 1858-1863.

7. Yoshimura T, Kaneko S, Hayabuchi H. Sex ratio in offspring of those affected by dioxin and dioxin-like compounds: the Yusho, Seveso, and Yucheng incidents. Occup Environ Med. 2001:58:540541.

8. James WH. Maternal control of offspring sex ratio. Hum Reprod. 1996;1 1:1139.

9. Hovinga ME, Sowers M, Humphrey HE. Historical changes in serum PCB and DDT levels in an environmentally-exposed cohort. Arch Environ Contam Toxicol. 1992;22:362-366.

10. Hovinga ME, Sowers M, Humphrey HE. Environmental exposure and lifestyle predictors of lead, cadmium, PCB, and DDT levels in Great Lakes fish eaters. Arch Environ Health. 1993;48:98104.

11. Humphrey HE, Budd ML. Michigan's fisheater cohorts: a prospective history of exposure. Toxicol Ind Health. 1996:12: 499-505.

12. Longnecker MP, Rogan W1, Lucier G. The human health effects of DDT (dichlorodiphenyltrichloroethane) and PCBS (polychlorinated biphenyls) and an overview of organochlorines in public health. Annu Rev Public Health. 1997;18211-244.

13. Matthews HB, Dedrick RL. Pharmacokinetics of PCBs. Annu Rev Pharmacol Toxicol. 1984;24:85-103.

14. Masuda Y, Haraguchi K, Kuroki H, Ryan JJ. [Change of PCDF, and PCB concentrations in the blood of Yucheng and Yusho patients for 25 years). Fukuoka Igaku Zasshi. Fukuoka Acta Medica. 1995;86:178-183.

15. Danzo BJ. Environmental xenobiotics may disrupt normal endocrine function by interfering with the binding of physiological ligands to steroid receptors and binding proteins. Environ Health Perspect. 1997;105:294-301.

16. Fielden MR, Chen 1, Chittim B, Safe SH, Zacharewski TR. Examination of the estrogenicity of 2,4,6,2',6'-pentachlorobiphenyl (PCB 104), its hydroxylated metabolite 2,4,6,2',6'pentachloro-4-biphenylol (HO-PCB 104), and a further chlorinated derivative, 2,4,6,2',4',6'hexachlorobiphenyl (PCB 155). Environ Health Perspect. 1997;105:1238-1248.

17. He JP, Stein AD, Humphrey HEB, Paneth N, Courval JM. Time trends in sport-caught Great Lakes fish consumption and serum polychlorinated biphenyl levels among Michigan anglers, 1973-1993. Environ Sci Technol 2001; 34:435-440.

18. Needham LL, Burse VW, Price HA. Temperature-programmed gas chromatographic determination of polychlorinated and polybromated biphenyls in serum. J Assoc Off Anal Chem. 1981;64:11311137.

19. Humphrey HEB. Evaluation of Humans Exposed to Water Borne Chemicals in the Great Lakes. Final report to the Environmental Protection Agency for Cooperative Agreement. Washington, DC: Environmental Protection Agency; 1983. Report No. CR-807192.

20. Price HA, Welch RI, Scheel RH, Warren LA. Modified multiresidue method for chlordane, toxaphene, and polychlorinated biphenyls in fish. Bull Environ Contam Toxicol. 1986;37:1-9.

21. Stokes ME, Davis CS, Koch GG. Categorical Data Analysis Using the SAS System. Cary, NC: SAS Institute; 1995.

22. Statistical Analysis Systems. SAS/STAT Software. Cary, NC: SAS Institute; 1999.

23. Armstrong B, White E, Saracci R, Principles of Exposure Measurement in Epidemiology. 1st ed. Oxford: Oxford University Press; 1992:78-114.

24. Blanck HM, Marcus M, Hertzberg V, et al. Determinants of polybromated biphenyl serum decay among women in the Michigan PBB cohort. Environ Health Perspect. 2000;108:147-152.

25. Kreuzer PE, Csanady GA, Baur C, et al. 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TODD) and congeners in infants. A toxicokinetic model of human lifetime body burden by TCDD with special emphasis on its uptake by nutrition. Arch Toxicol. 1997;71:383-400.

26. Dimich-Ward H, Hertzman C, Teschke K, et al. Reproductive effects of paternal exposure to chlorophenate wood preservatives in the sawmill industry. Scand J Work Environ Health. 1996;22:267273.

27. Rogan WJ, Gladen BC, Guo YL, Hsu CC. Sex ratio after exposure to dioxin-like chemicals in Taiwan. Lancet. 1999; 353:206-207.

28. Fertmann R, Schumann M, Karmaus W, Schmid-Hopfner S, Kuppers-Chinnow M. Sex-ratio variation in the Bille settlement. Scand J Work Environ Health. 1997;23:308-310.

29. Biggar RJ, Wohlfahrt, Westergaard T, Melbye M. Sex ratios, family size, and birth order. Am J Epidemiol. 1999;150: 957-962.

30. James WH. Evidence that mammalian sex ratios at birth are partially controlled by parental hormone levels at the tune of conception. J Theor Biol. 1996;180:271286.

31. James WH. Sex ratio of offspring of residents of a highly polluted housing site. Scand J Work Environ Health. 1998; 24:74-75.

32. Graffelman J, Fugger EF, Keyvanfar K, Schulman JD. Human live birth and sperm-sex ratios compared. Hum Reprod. 1999;14:2917-2920,

33. Morgan RW, Kheifets L, Obrinsky DL, Whorton MD, Foliart DE. Fetal loss and work in a waste water treatment plant. Am J Public Health. 1984;74: 499-501.

34. Savitz DA, Brett KM, Dole N, Tse CK. Male and female occupation in relation to miscarriage and preterm delivery in central North Carolina. Ann Epidemiol. 1997;7:509-516.

35. Stellman SD, Stellman JM, Sommer JFJ. Health and reproductive outcomes among American Legionnaires in relation to combat and herbicide exposure in Vietnam. Environ Res. 1988;47:150-174.