Polychlorinated biphenyls (PCBs) and 2,2-bis(p-chlorophenyl)-1,1,1-trichloroethane (DDT) are organochlorine compounds that were widely used after World War II but are now restricted from use because of their persistence in the environment and effects on wildlife. Laboratory studies reported an estrogenic effect of DDT and some PCBs 1 and an androgen-blocking effect of 1,1-dichloro-2,2-bis(p-chlorophenyl) ethylene (p,p ′-DDE), the persistent metabolite of DDT. 2 Hormonal activity and limited experimental evidence 3 suggest the potential for effects on reproductive function or menstrual characteristics, but these outcomes have not been extensively examined in epidemiologic studies of organochlorine exposure. 4–6
Age at natural menopause may be a marker for damage to the follicular pool 7 or disruption of hormonal feedback during the perimenopausal transition. 8 Smoking is associated with a 1- to 2-year decrease in age at natural menopause, with the strongest and most consistent relations seen for smoking around the time of menopause. 9,10 However, little is known about the effect of other potential toxicants, including organochlorines, on the timing of ovarian senescence. We examined the association between plasma organochlorine levels and age at natural menopause among 1,407 participants in a population-based, case-control study of breast cancer.
The Carolina Breast Cancer Study is a population-based, case-control study of breast cancer conducted in 24 counties of central and eastern North Carolina. 11,12 The study protocol was approved by the Institutional Review Board of the University of North Carolina School of Medicine, and informed consent was obtained from study participants. Cases 21 to 74 years of age with a first diagnosis of histologically confirmed invasive breast cancer were identified using a rapid case ascertainment system within the North Carolina Central Cancer Registry. 13 Controls were selected using Division of Motor Vehicles and Medicare beneficiary lists. Randomized recruitment 14 was used to identify approximately equal numbers of African-American and white women and equal numbers of women younger than age 50 and age 50 years or over.
Participation was higher among breast cancer cases than among controls (Table 1), with 72% of eligible cases and 50% of eligible controls interviewed. Patterns of response are described in detail elsewhere. 15 Between May 1993 and December 1996, 861 cases and 790 controls completed the full interview. Participants who were interviewed during this period were asked to provide a blood sample.
Plasma levels of organochlorines were used as an estimate of the total body burden of these compounds. 16–19 As described in detail elsewhere, 12 organochlorine analyses were conducted at Research Triangle Institute (Research Triangle Park, NC). Thirty-five individual PCB congeners were quantitated (International Union of Pure and Applied Chemistry [IUPAC] numbers 74, 99, 101, 105, 114, 118, 137, 138, 141, 146, 149, 153, 156, 157, 158, 167, 170, 171, 172, 174, 177, 178, 180, 182, 183, 185, 187, 190,194, 195, 196, 197, 200, 201 and 203) using calibration solutions prepared from certified standard solutions (AccuStandard, New Haven, CT).
Quantitation limits (defined as two times the detection limit) were 0.125 ng per gram for DDT, o,p ′-DDE, and p,p ′-DDE, and 0.025 ng per gram for individual PCB congeners. The interassay coefficient of variation was 12% for total PCBs, below 10% for the individual congeners, and 16% for p,p ′-DDE based on the repeated analyses of the standard between batches. Imputation was conducted to facilitate log transformation and was performed by setting zero values and values below the detection limit (0.625 for o,p ′-DDE and p,p ′-DDE and 0.0125 for individual PCB congeners) to the detection limit divided by the square root of 2. Total PCBs were determined by summing the individual PCB congeners (raw or imputed) for each person. All participants had detectable levels of one or more PCB congeners.
Only 1% of participants had detectable levels of o,p ′-DDE and 40% had detectable levels of p,p ′-DDT, whereas 99.7% had detectable levels of p,p ′-DDE. Results are therefore presented for p,p ′-DDE only (hereafter referred to as DDE).
Plasma lipid profiles were determined at the Core Laboratory for Clinical Studies at Washington University School of Medicine (St. Louis, MO). The interassay coefficient of variation was 3.4% for net triglycerides and 2.1% for total cholesterol. Lipid adjustment was performed by dividing DDE or total PCB levels by total lipids 20 (see Eq 2 in Ref 20) to yield micrograms per gram lipid. Organochlorine and lipid measurements were available for 748 of the cases (87%) and 659 of the controls (83%) who completed the study interview, with some losses because of insufficient plasma or interference in the sample.
Participants were asked whether they had ceased to have menstrual cycles and, if so, the reason and age at which this had occurred. Age at time of the last menstrual period was used to define age at menopause among women who reported that their menstrual periods had stopped by themselves (natural menopause). Thirty-six women who reported a natural menopause within 1 year of their current age were classified as premenopausal because it was not clear whether they had met the commonly used criterion of at least 12 months since last menstrual period. We excluded 28 women with unknown menopausal status (25 because of hormone use that began before cessation of periods and 3 for missing data).
We considered DDE and PCBs in categories, with cutpoints at the 50th (the reference category), 75th and 90th percentiles of the distribution for controls. We also analyzed log-transformed continuous measures of organochlorines. Time to natural menopause was examined using proportional-hazards modeling of the hazard or rate of onset of natural menopause as a function of age. Women who were classified as premenopausal were censored at their age at interview; women who were classified as menopausal as a result of surgery were censored at age at surgery and those who were menopausal for other medical reasons were censored at age at last menses. The models included age at the time of interview, race, education, parity and lactation, physical activity, thyroid condition, body mass index and smoking. Smoking status was considered for ages 18–34 among women whose age was less than 35 years and for ages 35–44 among women age 35 years or older. These ages were used to reflect smoking before menopause rather than smoking at the time of the interview. We also examined the effect of adjustment for age at interview squared, but the results were similar to those adjusted for a continuous measure of age, and the latter results are presented here.
We first conducted analyses separately for breast cancer cases and controls, but because we found the patterns of association in the two groups to be quite similar, we combined them to enhance precision. Because African-American women had higher levels of DDE and (to a lesser extent) PCBs in our study, 12 we conducted analyses stratified by race (classified by self-report). Additional analyses were conducted dividing PCBs into high-chlorinated congeners (IUPAC numbers 194, 195, 196, 197, 200, 201 and 203) and low- to medium-chlorinated congeners (the remaining). There was little difference in the results, so we present the total PCB analysis.
The median plasma DDE level, unadjusted for lipids, was 3.09 ng/ml (range = 0.04–93.84 ng/ml) and median PCBs was 1.82 ng/ml (range = 0.26–26.08 ng/ml). Median lipid-adjusted values were 0.60 μg per gram for DDE and 0.36 μg per gram for total PCBs. More than half of the women were postmenopausal, but fewer than half of the women who had reached menopause did so naturally (Table 1). Age at natural menopause was distributed similarly among cases and controls. Among women with a natural menopause, the median time between age at menopause and age at interview was 14 years, and for 25% the difference was 20 or more years.
Women with higher plasma levels of DDE had a higher rate of menopause (Table 2). Although the hazard ratios (HRs) were not large, we observed a fairly monotonic gradient, good precision and consistency between the categorical and continuous analyses. The effect estimate in the highest category of DDE exposure (HR = 1.4; 95% confidence interval [CI] = 0.9–2.1 for levels at or above the 90th percentile) was similar to the association between smoking and menopause (HR = 1.4; CI = 1.1–1.9) and corresponds to an approximately 1-year difference in the median age at natural menopause comparing women in the highest DDE category with women in the lowest DDE category. There was little evidence for modification of hazard ratios for DDE by race, and there was little evidence of an association between PCB exposure and menopause.
One possible explanation for these observations is that the loss of menstrual blood and tissue might be a means of excreting organochlorines, leading to lower blood levels among women who menstruated for a longer period. However, when we used a similar proportional-hazards model, age at surgical menopause, which would also cause menstrual bleeding to cease, was not associated with organochlorine measures. The HR per log unit change was 1.0 (CI = 0.9–1.1) for DDE and 1.0 (CI = 0.8–1.3) for PCBs.
Our analysis of plasma DDE and PCB levels was based on women age 21–74 years in North Carolina between 1993 and 1996. The DDE measures were somewhat lower than in most other studies 21 in which the median lipid-adjusted values ranged from 0.41 to 1.67 μg per gram, compared with our median of 0.60 μg per gram. The median PCB values in other studies ranged from 0.53 to 0.67 μg per gram compared with our median of 0.36 μg per gram. Differences in laboratory methods, demographic composition and temporal trends toward declining levels over time would affect the distribution of these compounds.
Our data suggest that higher body burdens of DDE may be associated with earlier onset of natural menopause but provide little evidence for such an influence from PCBs. Only one other study has considered the role of organochlorines in relation to age at menopause 6; in this study of high levels of PCBs in Taiwan from contaminated cooking oil, no difference was seen in menopausal status when comparing women within and outside the affected geographic area.
Associations between estrogenic exposures, such as age at menarche, oral contraceptive use and age at natural menopause have not been consistently or strongly seen in previous studies, 22 and estrogenic effects are not necessarily part of the mechanisms through which exposures affect timing of menopause. Tobacco smoke may affect age at menopause by a mechanism that involves increased apoptosis and oocyte death through activation of Bax gene transcription. 23 Binding of polycyclic aromatic hydrocarbons to the aromatic hydrocarbon receptor in the ovary is needed for this process.
DDE is only weakly estrogenic, but has been shown to have strong androgen-blocking effects. 2 Relatively little is known about androgen-related effects on follicular development and atresia, and available studies indicate that effects vary by species. Of particular relevance, however, is a recent experimental study in primates (rhesus monkeys). 24 Administration of testosterone or dihydrotestosterone resulted in increased growth of preantral and small antral follicles and a decrease in apoptotic granulosa cells. These data suggest that an androgen blocker, such as DDE, could lead to decreased follicular development (which may be particularly important if it occurs during the menopausal transition period) and to increased apoptosis or atresia. Either of these effects could influence timing of menopause.
One limitation of our analysis is the reliance on cross-sectional self-reported data on menopausal status and age at menopause. However, previous studies have reported fairly high levels of accuracy or reliability for data on age at natural menopause based on interview or questionnaire. 25–27 The mean age at last menstrual period is younger than expected (mean or median 50–51 years), 28,29 but this may reflect the cross-sectional nature of the design in which women with later menopause are relatively underrepresented in the age distribution.
Another limitation is that organochlorine measures were taken at the time of study enrollment. For 25% of postmenopausal women, this was 20 or more years after their reported age at natural menopause. Biological measures of organochlorines do reflect long-term history, but the predictive value of current measurements for the remote past may be limited. 30,31 Although breast cancer or its treatment may affect organochlorine measurements, 32 we saw little difference in effect estimates in separate analyses for cases and controls. Prospective studies, in which organochlorine measurements are taken before the onset of the menopausal transition, would provide a means to reduce any nondifferential or differential misclassification bias that could have affected our results.
The proportion of all potentially eligible subjects who participated in the study and provided a usable blood sample was 51%, so selection bias must also be considered. It does not seem likely, however, that participation would be differentially associated with the exposure and outcome under study because participants were not aware of their organochlorine levels or of the interest in timing of menopause. Finally, for any given hypothesized biologic pathway, such as estrogenicity or antiandrogenicity, the measures of DDE and PCBs are incomplete, given the variable biological activities and varying half-lives of the contributing compounds.
Our data provide some evidence to suggest an effect of DDE on timing of natural menopause. Early natural menopause has been associated with an increased risk of all-cause mortality 33,34 and cardiovascular mortality. 35,36 Additional research on the ovarian effects of DDE, and on the mechanism through which an androgen-blocking exposure could affect ovarian failure, would be warranted if the association we observed is replicated.
We thank Beth Gladen, Matthew Longnecker and Mary Wolff for helpful comments on the manuscript.
1. Shekhar PVM, Werdell J, Basrur VS. Environmental estrogen stimulation of growth and estrogen receptor function in preneoplastic and cancerous human breast cell lines. J Natl Cancer Inst 1997; 89: 1774–1782.
2. Kelce WR, Stone CR, Laws SC, et al
. Persistent DDT
is a potent androgen receptor antagonist. Nature 1995; 375: 581–585.
3. Barsotti DA, Marlar RJ, Allen JR. Reproductive dysfunction in rhesus monkeys exposed to low levels of polychlorinated biphenyls (Aoroclor 1248). Food Cosmet Toxicol 1976; 14: 99–103.
4. Longnecker M, Rogan WJ, Lucier G. The human health effects of DDT
(dichlorodiphenyl-trichloroethane) and PCBs (polychlorinated biphenyls) and an overview of organochlorines in public health. Annu Rev Public Health 1997; 18: 211–244.
5. Mendola P, Buck GM, Sever LE, et al
. Consumption of PCB
-contaminated freshwater fish and shortened menstrual cycle length. Am J Epidemiol 1997; 146: 955–960.
6. Yu ML, Guo YLL, Hsu CC, Rogan WJ. Menstruation and reproduction in women with polychlorinated biphenyl (PCB
) poisoning: long-term follow-up interviews of the women from the Taiwan Yucheng cohort. Int J Epidemiol 2000; 29: 672–677.
7. Mattison DR. Clinical manifestations of ovarian toxicity. In: Dixon RL, ed. Reproductive Toxicology. New York: Raven Press, 1985; 109–130.
8. Richardson SJ, Senikas V, Nelson JF. Follicular depletion during the menopausal transition: evidence for accelerated loss and ultimate exhaustion. J Clin Endocrinol Metab 1987; 65: 1231–1237.
9. Midgette AS, Baron JA. Cigarette smoking and the risk of natural menopause
. Epidemiology 1990; 1: 474–479.
10. Cooper GS, Sandler DP, Bohlig M. Active and passive smoking and age at natural menopause
. Epidemiology 1999; 10: 771–773.
11. Newman B, Moorman PG, Millikan R, et al
. The Carolina Breast Cancer Study: integrating population-based epidemiology and molecular biology. Breast Cancer Res Treat 1995; 35: 51–60.
12. Millikan R, DeVoto E, Duell EJ, et al
. Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in North Carolina. Cancer Epidemiol Biomarkers Prev 2000; 9: 1233–1240.
13. Aldrich TE, Vann D, Moorman PG, Newman B. Rapid reporting of cancer incidence in a population-based study of breast cancer: one constructive use of a central cancer registry. Breast Cancer Res Treat 1995; 35: 61–64.
14. Weinberg CR, Sandler DP. Randomized recruitment in case-control studies. Am J Epidemiol 1991; 134: 421–432.
15. Moorman PG, Newman B, Millikan R, Tse CKJ, Sandler DP. Participation rates in a case-control study: the impact of age, race and race of interviewer. Ann Epidemiol 1999; 9: 188–195.
16. Wolff MS, Toniolo PG, Lee EW, Rivera M, Dubin N. Blood levels of organochlorine residues and risk of breast cancer. J Natl Cancer Inst 1993; 85: 648–652.
17. DeWailly E, Ayotte P, Brisson J, Dodin S. Breast cancer and organochlorines (Letter). Lancet 1994; 344: 1707–1708.
18. Blackwood A, Wolff M, Rundle A, et al.
O rganochlorine compounds (DDE
) in plasma and breast cyst fluid of women with benign breast disease. Cancer Epidemiol Biomarkers Prev 1998; 7: 579–583.
19. Stellman SD, Djordjevic MV, Muscat JE, et al.
Relative abundance of organochlorine pesticides and polychlorinated biphenyls in adipose tissue and serum of women in Long Island, New York. Cancer Epidemiol Biomarkers Prev 1998; 7: 489–496.
20. Phillips DL, Pirkle JL, Burse VW, Bernert JT, Henderson LO, Needham LL. Chlorinated hydrocarbon levels in human serum: effects of fasting and feeding. Arch Environ Contam Toxicol 1989; 18: 495–500.
21. Laden F, Collman G, Iwamoto K, et al.
-chlorophenyl)ethylene and polychlorinated biphenyls and breast cancer: combined analysis of five U.S. studies. J Natl Cancer Inst 2001; 93: 768–776.
22. Cooper GS, Baird DD, Darden FR. Measures of menopausal status in relation to demographic, reproductive, and behavioral characteristics in a population-based study of women ages 35 to 49 years. Am J Epidemiol 2001; 153: 1159–1165.
23. Matikainen T, Perez GI, Jurisicova A, et al.
A romatic hydrocarbon receptor-driven BAX
gene expression is required for premature ovarian failure caused by biohazardous environmental chemicals. Nat Genet 2001; 28: 355–360.
24. Vendola KA, Zhou J, Adesanya OO, Weil SJ, Bondy CA. Androgens stimulate early stages of follicular growth in the primate ovary. J Clin Invest 1998; 101: 2622–2629.
25. Bean JA, Leeper JD, Wallace RB, Sherman BM, Jagger H. Variations in the reporting of menstrual histories. Am J Epidemiol 1979; 109: 181–185.
26. Colditz GA, Stampfer MJ, Willett WC, et al.
Reproducibility and validity of self-reported menopausal status in a prospective cohort study. Am J Epidemiol 1987; 126: 319–325.
27. Horwitz RI, Yu EC. Problems and proposals for interview data in epidemiological research. Int J Epidemiol 1985; 14: 463–467.
28. McKinlay SM, Bifano NL, McKinlay JB. Smoking and age at menopause
in women. Ann Intern Med 1985; 103: 350–356.
29. Cramer DW, Xu HJ. Predicting age at menopause
. Maturitas 1996; 23: 319–326.
30. Schildkraut JM, Demark-Wahnefried W, DeVoto E, Hughes C, Laseter JL, Newman B. Environmental contaminants and body fat distribution. Cancer Epidemiol Biomarkers Prev 1999; 8: 179–183.
31. Wolff MS, Anderson HA, et al
. Correspondence re: J. M. Schildkraut, Environmental contaminants and body fat distribution (letter). Cancer Epidemiol Biomarkers Prev 1999; 8: 951–952.
32. Gammon MD, Wolff MS, Neugut AI, et al
. Treatment for breast cancer and blood levels of chlorinated hydrocarbons. Cancer Epidemiol Biomarkers Prev 1996; 5: 467–471.
33. Snowdon DA, Kane RL, Beeson WL, et al
. Is early natural menopause
a biologic marker of health and aging? Am J Public Health 1989; 79: 709–714.
34. Cooper GS, Sandler DP. Age at natural menopause
and mortality. Ann Epidemiol 1998; 8: 229–235.
35. Jacobsen BK, Knutsen SF, Fraser GE. Age at natural menopause
and total mortality and mortality from ischemic heart disease: the Adventist health study. J Clin Epidemiol 1999; 52: 303–307.
36. Van der Schouw YT, van der Graaf Y, Steyerberg EW, Eijkemans MJC, Banga JD. Age at menopause
as a risk factor for cardiovascular mortality. Lancet 1996; 347: 714–718.