Influence of exposure assessment methods on risk estimates in an epidemiologic study of total trihalomethane exposure and spontaneous abortion

Trihalomethanes are common contaminants of chlorinated drinking water. Studies of their health effects have been hampered by exposure misclassification, due in part to limitations inherent in using utility sampling records. We used two exposure assessment methods, one based on utility-wide sampling averages, and one based on measurements from the utility sampling site closest to the subject's residence, to reestimate total trihalomethane (TTHM) exposure for 4212 participants in a preexisting study of risk factors for spontaneous abortion (SAB). For both approaches we performed unweighted, weighted, and subset analyses. The weighted and subset analyses were intended to reduce exposure misclassification, and were based on within-utility variance in TTHM measurements for the utility-wide average approach, and the distance between the subject's residence and sampling site for the closest-site approach. In general, the utility-wide average methods produced odds ratios equivalent to or slightly higher than the closest-site methods. Odds ratios obtained using the utility-wide average, but not the closest-site, approach also became progressively stronger in the weighted and subset analyses. A dose-response was seen between SAB and an exposure metric incorporating both TTHM concentration (utility-wide average approach) and personal ingestion, with SAB rates ranging from 8.3% to 13.7% (unweighted), 7.9% to 16.6% (variance weighted), and 6.6% to 23.1% (low-variance subset). Utility-wide average TTHM exposure assessment methods together with variance-based weights and subsets are relatively simple exposure assessment techniques, which may increase the epidemiologic usefulness of utility sampling records.

Intrauterine exposure to diethylstilbestrol: long-term effects in humans

DES is the most carefully scrutinized EDC and its history provides valuable insights into the current evaluation of less well-studied EDCs. This review summarizes the health effects of prenatal exposure to diethylstilbestrol (DES) and emphasizes the role of DES as the first endocrine disrupting chemical (EDC). Vaginal clear cell adenocarcinoma (CCAC), the most severe consequence of prenatal exposure to DES, affected only 0.1% of exposed females, while the far more prevalent teratogenic and reproductive effects of DES were only discovered when DES daughter were screened for CCAC. Initial studies, conducted before most DES daughters had tried to conceive, examined vaginal cancer and vaginal, cervical and uterine abnormalities. Subsequently, several controlled studies demonstrated the increased risk of adverse reproductive outcomes in DES daughters. While most DES daughters can eventually experience a live birth, this is less likely in women with genital tract abnormalities, in whom there is a two-thirds chance that each pregnancy will be unsuccessful. In DES sons, who have been far less studied, results suggest male reproductive toxicity, but are less consistent. The importance of dose and gestational age at initial exposure are discussed, and the implications of DES findings for the evaluation of risks from current EDCs emphasized.

The question of declining sperm density revisited: an analysis of 101 studies published 1934-1996

In 1992 Carlsen et al. reported a significant global decline in sperm density between 1938 and 1990 [Evidence for Decreasing Quality of Semen during Last 50 Years. Br Med J 305:609-613 (1992)]. We subsequently published a reanalysis of the studies included by Carlsen et al. [Swan et al. Have Sperm Densities Declined? A Reanalysis of Global Trend Data. Environ Health Perspect 105:1228-1232 (1997)]. In that analysis we found significant declines in sperm density in the United States and Europe/Australia after controlling for abstinence time, age, percent of men with proven fertility, and specimen collection method. The declines in sperm density in the United States (approximately 1.5%/year) and Europe/Australia (approximately 3%/year) were somewhat greater than the average decline reported by Carlsen et al. (approximately 1%/year). However, we found no decline in sperm density in non-Western countries, for which data were very limited. In the current study, we used similar methods to analyze an expanded set of studies. We added 47 English language studies published in 1934-1996 to those we had analyzed previously. The average decline in sperm count was virtually unchanged from that reported previously by Carlsen et al. (slope = -0.94 vs. -0.93). The slopes in the three geographic groupings were also similar to those we reported earlier. In North America, the slope was somewhat less than the slope we had found for the United States (slope = -0.80; 95% confidence interval (CI), -1.37--0.24). Similarly, the decline in Europe (slope = -2.35; CI, -3.66--1.05) was somewhat less than reported previously. As before, studies from other countries showed no trend (slope = -0.21; CI, -2.30-1.88). These results are consistent with those of Carlsen et al. and our previous results, suggesting that the reported trends are not dependent on the particular studies included by Carlsen et al. and that the observed trends previously reported for 1938-1990 are also seen in data from 1934-1996.

Prenatal active or passive tobacco smoke exposure and the risk of preterm delivery or low birth weight

We examined the association of exposure to environmental tobacco smoke with birth weight and gestational age in a large, prospective study. We also compared these endpoints between infants of active maternal smokers and those of non-smoking, non-ETS exposed women. Pregnant women were interviewed by telephone during the first trimester, and pregnancy outcome was determined for 99%. Among the 4,454 singleton live births that could be linked to their birth certificate, we confirmed increased risks of low birth weight and small for gestational age with heavier maternal smoking (> 10 cigarettes/day), as well as noting an increased risk for "very preterm" birth (< 35 weeks). These associations were generally stronger among infants of older (> or = 30 years) than those of younger mothers, as well as among non-whites. High environmental tobacco smoke exposure (> or = 7 hours/day in non-smokers) was moderately associated with low birth weight (adjusted odds ratio (AOR) 1.8, 95% confidence limits (95% CL) = 0.82, 4.1), preterm birth (AOR 1.6, 95% CL = 0.87, 2.9), and most strongly with very preterm birth (AOR 2.4, 95% CL = 1.0, 5.3). These associations were generally greater among non-whites than whites. The data support earlier studies suggesting that prenatal environmental tobacco smoke exposure, in addition to maternal smoking, affects infant health.

Cosmetic and postmastectomy breast implants: Finnish women's experiences

The purpose of this study was to compare women's satisfaction with and short-term problems of silicone breast implants after cosmetic breast augmentation and after mastectomy. Women (n = 224) were recruited through advertising in mass media, and 91% responded to a questionnaire asking for their experiences, both positive and negative, with silicone breast implants. Approximately equal numbers of women received their implants for cosmetic reasons (augmentation group) and postmastectomy (113 and 111, respectively). Mean time from first implantation was 9 years (SD 7.3) in the cosmetic group and 8 years (SD 4.9) in the postmastectomy group. Women in the postmastectomy group received their implants at an older age than women in the cosmetic group (percent of women 45 and older, 59% and 3%, respectively). Women's overall preoperative knowledge of and postoperative satisfaction with their implants were similar in the two groups; 58% of women said that they had insufficient knowledge of breast implants preoperatively, 26% of women said they would not choose the implants again, and 44% of women expressed no dissatisfaction with their breasts. However, women in the cosmetic group were better informed about possible physical problems. One third of the women in the postmastectomy group had one or more reoperations, most frequently because of implant slippage (30%), encapsulation (26%), or implant size and shape (23%). Because insertion of breast implants is a lifelong decision, in-depth counseling about complication rates and possible risks should be given to women before implantation, and nonsurgical alternatives should be discussed, particularly for cosmetic implantation.

Declining semen quality: can the past inform the present?

By using instrumentation initially designed for counting white blood cells, sperm counts have been utilized by clinicians since 1929, particularly to evaluate cases of suspected infertility. Although this basic biological parameter might be assumed to be stable over time, several studies over the past 20 years have suggested a decline in sperm count or density. The most controversial of these analyses was published in 1992. A flood of criticism followed this analysis of 61 studies that found a 50% decline in sperm density between 1938 and 1990. Critics suggested that historical methods (of counting sperm or conducting studies) were variable and unreliable, differing from modern methods both qualitatively and quantitatively. To address this issue we analyzed these studies for trends in counting methods or their variability. We found neither. Alternative analyses produced some differences in trend estimates, but statistical factors alone could not account for the total decline in sperm density. We reviewed study populations to identify trends in population characteristics, such as abstinence time, that might explain the decline. However, controlling analytically for such factors only increased the rate of decline. We conclude that historical data on sperm density, despite large random error, are surprisingly reliable. Nonetheless, understanding causes of temporal and geographic differences in sperm density must await contemporary data.

Caffeine consumption and menstrual function

The relation between caffeine intake and menstrual function was examined in 403 healthy premenopausal women who belonged to Kaiser Permanente Medical Care Program in 1990-1991. A telephone interview collected information about caffeinated beverage intake as well as other lifestyle, demographic, occupational, and environmental factors. Subjects collected daily urine samples and completed a daily diary for an average of five menstrual cycles. Metabolites of estrogen and progesterone were measured in the urine, each cycle was characterized as anovulatory or ovulatory, and a probable day of ovulation was selected when appropriate. Logistic regression and repeated measures analyses were performed on menstrual parameters. Women whose caffeine consumption was heavy (>300 mg of caffeine per day) had less than a third of the risk for long menses (> or =8 days) compared with women who did not consume caffeine (adjusted odds ratio = 0.30, 95% confidence interval 0.14-0.66). Those whose caffeine consumption was heavy also had a doubled risk for short cycle length (< or =24 days) (adjusted odds ratio = 2.00, 95% confidence interval 0.98-4.06); this association was also evident in those whose caffeine consumption was heavy who did not smoke (adjusted odds ratio = 2.11, 95% confidence interval 1.03-4.33). Caffeine intake was not strongly related to an increased risk for anovulation, short luteal phase (< or =10 days), long follicular phase (> or =24 days), long cycle (> or =36 days), or measures of within-woman cycle variability.

Cigarette smoking and effects on menstrual function

OBJECTIVE

To examine the relationship between smoking and menstrual function, using biologic measures rather than self-report of menstrual cycle characteristics.

METHODS

In a prospective study, 408 women collected urine daily for one to seven menstrual segments (cycles), maintained daily diaries, and completed detailed interviews. Smoking data from the diaries were averaged over each segment and verified by cotinine assay. Urine samples were analyzed for metabolites of steroid hormones to define the day of ovulation and various menstrual characteristics, including: 1) segment, follicular, luteal phase, and menses length, 2) variability, and 3) anovulation.

RESULTS

Heavy smoking (at least 20 cigarettes per day) was associated with nearly four times the risk of short segment (less than 25 days) as was nonsmoking (adjusted odds ratio 3.8, 95% confidence limits 1.1, 12.7). Mean segment length was on average 2.6 days shorter with heavy versus no smoking (95% confidence limits 0.14, 5.0), due almost entirely to shortening of the follicular phase. Women who smoked an average of ten or more cigarettes per day had significantly more variable segment and menses lengths than nonsmokers. Based on small numbers, the data suggested that with greater smoking, there was a possible increased risk of anovulation and short luteal phase. Segments of exsmokers with ten or more pack-years of exposure were more likely to be short and have shorter luteal phases than those of never smokers.

CONCLUSION

The effects found in this study of smoking on the menstrual cycle might explain in part associations of smoking with other reproductive endpoints, such as subfecundity and early menopause.

Use of urine biomarkers to evaluate menstrual function in healthy premenopausal women

A total of 403 healthy, premenopausal women, residing near Santa Clara, California, were recruited from a large health care plan in California for a study of menstrual function. After a telephone interview, participants collected daily urine samples and recorded bleeding and other information in diaries. Data were collected during 1990-1991. Urine samples were analyzed for creatinine and for estradiol and progesterone metabolites by enzyme-linked immunoassay. Computer algorithms were developed to derive menstrual segment length, ovulatory status, day of ovulation, and other parameters from the urine and diary data. (We use "segment" rather than "cycle" to avoid implying that normal cycling occurred.) The average length of participation was 141 (standard deviation, 45) days. The mean segment length was 28.8 (standard deviation, 4.4) days; follicular phase length, 16.0 (standard deviation, 4.4) days; and luteal phase length, 12.9 (standard deviation, 1.7) days; 19 (4.7%) women experienced anovulatory episodes. In exploratory multivariate analyses, important associations included the following: age of > or = 35 years with decreased segment and follicular phase lengths; heavier weight (upper quartile) with anovulation and increased follicular phase and decreased luteal phase lengths; Hispanic ethnicity with anovulation and increased segment length; and past difficulty in achieving pregnancy with anovulation and increased length and variability of segments and follicular phases. Urine biomarkers can be used successfully to evaluate menstrual function in epidemiologic studies.

A prospective study of spontaneous abortion: relation to amount and source of drinking water consumed in early pregnancy

In 1992, we published four retrospective studies, conducted primarily within a single California county, which found higher spontaneous abortion rates among women who drank more tapwater than bottled water in early pregnancy. The current prospective study extends that investigation to other water systems. Pregnant women from three regions in California were interviewed during their first trimester. Multivariate analyses modeled the amount and type of water consumed at 8 weeks' gestation in each region in relation to spontaneous abortion rate. In Region I, which was within the previous study area, the adjusted odds ratio (OR) comparing high (> or = 6 glasses per day) consumption of cold tapwater with none was 2.17 [95% confidence interval (CI) = 1.22-3.87]. Furthermore, when women with high cold tapwater and no bottled water consumption were compared with those with high bottled water and no cold tapwater consumption, the adjusted odds ratio was 4.58 (95% CI = 1.97-10.64). Conversely, women with high bottled water consumption and no tapwater had a reduced rate of spontaneous abortion compared with those drinking tapwater and no bottled water (adjusted OR = 0.22; 95% CI = 0.09-0.51). Neither tap nor bottled water consumption altered the risk of spontaneous abortion in Regions II and III. Although controlling for age, prior spontaneous abortion, race, gestational age at interview, and weight somewhat strengthened the association in Region I, the distribution of these confounders did not vary appreciably across regions. This study confirms the association between cold tapwater and spontaneous abortion first seen in this county in 1980. If causal, the agent(s) is not ubiquitous but is likely to have been present in Region I for some time.

Trihalomethanes in drinking water and spontaneous abortion

Trihalomethanes (chloroform, bromoform, bromodichloromethane, and chlorodibromomethane) are common contaminants of chlorinated drinking water. Although animal data indicate that these compounds may be reproductive toxicants, little information exists on their relation to spontaneous abortion in humans. We examined exposure to trihalomethanes and spontaneous abortion in a prospective study of 5,144 pregnant women in a prepaid health plan. Seventy-eight drinking water utilities provided concurrent trihalomethane sampling data. We calculated total trihalomethane levels by averaging all measurements taken by the subject's utility during her first trimester. We calculated exposures to individual trihalomethanes in an analogous manner. Women who drank > or = 5 glasses per day of cold tapwater containing > or = 75 micrograms per liter total trihalomethanes had an adjusted odds ratio (OR) of 1.8 for spontaneous abortion [95% confidence interval (CI) = 1.1-3.0]. Of the four individual trihalomethanes, only high bromodichloromethane exposure (consumption of > or = 5 glasses per day of cold tapwater containing > or = 18 micrograms per liter bromodichloromethane) was associated with spontaneous abortion both alone (adjusted OR = 2.0; 95% CI = 1.2-3.5) and after adjustment for the other trihalomethanes (adjusted OR = 3.0; 95% CI = 1.4-6.6).

Have sperm densities declined? A reanalysis of global trend data

In 1992 a worldwide decline in sperm density was reported; this was quickly followed by numerous critiques and editorials. Because of the public health importance of this finding, a detailed reanalysis of data from 61 studies was warranted to resolve these issues. Multiple linear regression models (controlling for abstinence time, age, percent proven fertility, specimen collection method, study goal and location) were used to examine regional differences and the interaction between region (United States, Europe, and non-Western countries) and year. Nonlinear models and residual confounding were also examined in these data. Using a linear model (adjusted R2 = 0. 80), means and slopes differed significantly across regions (p = 0. 02). Mean sperm densities were highest in Europe and lowest in non-Western countries. A decline in sperm density was seen in the United States (studies from 1938-1988; slope = -1.50; 95% confidence interval (CI), -1.90--1.10) and Europe (1971-1990; slope = -3.13; CI, -4.96- -1.30), but not in non-Western countries (1978-1989; slope = 1.56; CI, -1.00-4.12). Results from nonlinear models (quadratic and spline) were similar. Thus, further analysis of these studies supports a significant decline in sperm density in the United States and Europe. Confounding and selection bias are unlikely to account for these results. However, some intraregional differences were as large as mean decline in sperm density between 1938 and 1990, and recent reports from Europe and the United States further support large interarea differences in sperm density. Identifying the cause(s) of these regional and temporal differences, whether environmental or other, is clearly warranted.

Caffeinated beverages, decaffeinated coffee, and spontaneous abortion

We examined the relations between spontaneous abortion and the consumption of caffeine, individual caffeine-containing beverages (coffee, tea, and soda), and decaffeinated coffee in a prospective study of 5,144 pregnant women. We collected information about potential risk factors for spontaneous abortion, including consumption of caffeinated beverages and decaffeinated coffee before and during pregnancy, by interview in the first trimester. Neither total estimated caffeine nor individual caffeinated beverage consumption during the first trimester was associated with an appreciable increase in risk for spontaneous abortion. The adjusted odds ratio for consumption of greater than 300 mg per day of caffeine was 1.3 [95% confidence interval (CI) = 0.8-2.1] after adjustment for maternal age, pregnancy history, cigarette and alcohol consumption, employment, race, gestational age at interview, and marital and socioeconomic status. The adjusted odds ratio for spontaneous abortion related to consumption of three or more cups of decaffeinated coffee during the first trimester was 2.4 (95% CI = 1.3-4.7) in the same model. Although we could not demonstrate this with available data, we suspect that this association was due to bias resulting from the relations among fetal viability, symptoms of pregnancy such as nausea, and consumption patterns during pregnancy.

Moderate maternal alcohol consumption and risk of spontaneous abortion

Maternal alcoholism is known to have adverse effects on reproduction and fetal development, but the effects of moderate consumption remain controversial. In a previous case-control study, we found a doubled risk of spontaneous abortion with an average consumption of seven or more drinks per week during the first trimester. To confirm this finding while avoiding potential biases from the case-control design, we examined moderate alcohol consumption in a prospective cohort study of over 5,000 pregnant women. An interview in the first trimester asked about alcohol consumption during the week before interview ("during the first trimester") and before pregnancy. We found an increased risk of spontaneous abortion in women who drank more than three drinks per week during the first trimester, with an adjusted odds ratio (OR) of 2.3 [95% confidence interval (CI) = 1.1-4.5]. The increased risk associated with this moderate alcohol consumption may be higher in first than in second trimester abortions, and it is even higher in the first 10 weeks (OR = 3.8; 95% CI = 1.7-8.7), based on small numbers. Consumption of alcohol before pregnancy was not strongly associated with spontaneous abortion.

Effects of psychological stress on human semen quality

We investigated the relationship between psychological stress and sperm concentration, motility, and morphometry in a prospective study of 157 volunteers who were enrolled in a prepaid health plan. We measured psychological job stress and life-event stress by telephone interview. Sperm-kinematic and nuclear-morphometric variables were measured using computer-assisted image analyses. Sperm concentration, percent motility, and semen volume were determined by objective visual methods. We performed multiple linear regression for each semen variable to examine its relationship to stress, controlling for potential confounders. Stress at work and total number of life events were not related to differences in semen quality. However, the recent death of a close family member was associated with a reduction in straight-line velocity (P = 0.002) and percent of progressively motile sperm (P = 0.02); it was also marginally associated with an increase in the fraction of sperm with larger and more tapered nuclei. These findings suggest that the fecundity of men experiencing the stress of a family member's death might be temporarily diminished.