Chlorination disinfection by-products and adverse birth outcomes: a retrospective cohort study in three regions in England: ISEE-183

Toledano, M.B., Nieuwenhuijsen, M.J., Best, N.G., Whitaker, H., de Hoogh, C., Cobley, N., Fawell, J. and Elliott, P. (2003) Chlorination disinfection by-products and adverse birth outcomes: a retrospective cohort study in three regions in England: ISEE-183. Epidemiology, 14 (5 Suppl), S39.

Abstract

Recent epidemiological studies of the relationship between disinfection by-products (DBPs), from chlorination, in public drinking water supply and adverse birth outcomes, have reported inconsistent and inconclusive findings. We report here on the largest study to date to examine the relationship of total trihalomethanes (TTHMs) and the individual trihalomethanes (THMs) to birth weight and still birth prevalence, between 1993 and 1998, for regions covered by three water companies in England. Initial results from this ongoing study using modelled annual TTHM estimates were found to differ between each of the three water supply areas. There was some evidence of confounding by social deprivation and ethnicity, which was further explored by analysis of information on the lifestyle characteristics of all women of reproductive age in the study regions from National survey data. Here, we report on the next stage of the study using modelled quarterly (3 monthly) estimates of the individual THMs in water zones. We modelled THM measurements using a Bayesian hierarchical mixture model, taking into account heterogeneity in THM levels between water originating from different source types (e.g. ground, lowland surface or upland surface), quarterly variation in THM levels and uncertainty in the true value of undetected and rounded measurements. These modelled estimates were linked using Geographical Information Systems to routine birth and stillbirth records based on location of maternal residence at the time of birth to obtain a weighted third trimester exposure estimate for each birth. Exposure categories were constructed as follows: for chloroform (low (< 20 mg/l). medium (20-40 mg/l) and high (>40 mg/l), bromodichloromethane (low (< 6 mg/l), medium (6-12 mg/l) and high (>12 mg/l), dibromoehloromethane (low (< 5 mg/l) and high (>5 mg/l) and total brominated compounds low (< 6 mg/l), medium (6-12 mg/l) and high (>12 mg/l) respectively. After exclusion of multiple births and births from water zones without valid THM data, we studied a total of 934,843 live and stillbirths. Exposure-response relationships were explored using multiple logistic regression analysis with weighted third trimester individual THM estimates and adjustment for e.g. gender, maternal age and deprivation. Analysis of the individual THMs, as opposed to TTHMs. is essential to characterise more precisely the effects of complex mixtures of DBPs on the developing foctus, and refine our understanding of this important area of research.

This record has no associated files available for download.

Contributors

Download statistics