You are here:
SHORTER MENSTRUAL CYCLES ASSOCIATED WITH CHLORINATION BY-PRODUCTS IN DRINKING WATER
Citation:
Windham, G., K. Waller, M. Anderson, L. Fenster, P Mendola, AND S. Swan. SHORTER MENSTRUAL CYCLES ASSOCIATED WITH CHLORINATION BY-PRODUCTS IN DRINKING WATER. Presented at International Society for Environmental Epidemiology, Vancouver, British Columbia, Canada, August 11-15, 2002.
Description:
Shorter Menstrual Cycles Associated with Chlorination by-Products in Drinking Water.
Gayle Windham, Kirsten Waller, Meredith Anderson, Laura Fenster, Pauline Mendola, Shanna Swan. California Department of Health Services.
In previous studies of tap water consumption we reported increased risks of spontaneous abortion with increasing tap water consumption in one region of California, as well as with high consumption of tap water containing high levels of total trihalomethanes (TTHM) in general. To determine whether trihalomethanes (THMs) are associated with other reproductive endpoints and to identify possible mechanisms for the association with spontaneous abortion, we conducted a prospective study of menstrual cycle function and early pregnancy loss. Women of reproductive age (n=403) collected urine samples daily for up to 6 months (average 4.4 menstrual cycles) for measurement of steroid metabolite levels. These were used to calculate menstrual parameters such as cycle length, phase length, menses length and ovulatory status. At baseline, women were asked about consumption of various types of water, as well as other habits and demographics. A THM exposure database was assembled using quarterly measurement data from water utilities in the area. We initially examined exposure at one point in time around the start of urine collection, but we have conducted further analyses estimating an exposure level for each cycle. This measurement represents the utility-wide average based on subject residence, as measured during a 90-day time period starting 60 days before the individual cycle start date. We analyzed menstrual parameters as continuous and categorical variables (e.g. mean cycle length or risk of short cycle) in relation to various categorical exposure measures, including water consumption, cycle-specific TTHM level and individual THM levels, in models that account for repeated measures. Adjusting for age, race, income, body mass index, pregnancy history, caffeine, alcohol and cigarette intake, we found a dose-related trend towards decreasing cycle length with increasing TTHM level. At >60 ug/L (about the top quartile) the adjusted decrement was 1.1 day (95% confidence interval (CI) ?1.8, -0.40), compared to <40 ug/L. Results were similar for >80 ug/L, the current USEPA maximum contaminant level. This finding was also reflected as a reduced follicular phase length (difference ?0.92; CI ?1.6, -0.23, at >60 ug/L). Examining the individual THM's by quartile, the greatest risk was with chlorodibromomethane and the lowest with chloroform. We summed the brominated THM's and again found a pattern of decreasing cycle and follicular phase length with increasing exposure, so that at the highest quartile, cycle length was decreased 1.2 days (CI ?2.0, -0.40). Calculating a "dose" incorporating the amount of tap water usually consumed, showed a similar pattern of reduced cycle length with increasing TTHM exposure. Considering the amount of time spent showering did not reveal greater risks. In conclusion, we find a consistent pattern of shorter mean cycle length with several different measures of chlorination by-product levels. As this is the first study to examine menstrual function it will be important to replicate the findings, but they add to the growing literature indicating effects of chlorination by-products on the reproductive system.
This is an abstract of a proposed presentation and does not necessarily reflect EPA policy.