Disinfection question: how should we measure exposure?

Chlorination of public water supplies began in the early 1900s and quickly led to a massive reduction in waterborne diseases. At the same time, chlorination and other disinfection methods introduced a subtler threat to public health: disinfection by-products (DBPs), chemicals formed when disinfectants react with organic or inorganic matter in the water. Several DBPs cause carcinogenic, mutagenic, reproductive, or developmental effects, and many DBP studies have focused on effects of exposure in utero. Traditionally, total trihalomethane (THM) exposure is used as a surrogate in epidemiological studies to estimate maternal DBP exposure during pregnancy, partly because of data availability (THMs have been measured since the late 1970s) and partly because THMs are so prevalent. However, research led by J. Michael Wright of the U.S. Environmental Protection Agency's National Center for Environmental Assessment suggests that estimating DBP exposure from total THM concentrations alone provides insufficient information about risks to public health [EHP 112:920-925].

Wright and his colleagues focused on potential DBP-associated developmental effects, using specific end points including birth weight among full-term newborns (born at 37-45 weeks of pregnancy), birth before 37 weeks (preterm delivery), and pregnancy duration (gestational age). Birth weight and gestational age were further used to identify small-for-gestational-age (SGA) infants, defined as being in the 10th percentile for birth weight given gestational age. Preterm delivery and low birth weight are strongly associated with significant infant health problems--U.S, infant mortality is largely attributable to preterm delivery.

Birth certificates from Massachusetts towns with 10,000 or more residents provided Wright's group with information on 196,000 infants born during 1995-1998 and their mothers. This information was matched with exposure data for three types of DBPs: THMs, haloacetic acids, and MX. The team estimated the mother's average third-trimester exposure to THMs and haloacetic acids based on each town's water supply monitoring data. Data for MX came from earlier studies by Wright and colleagues, in which they had measured drinking water concentrations of MX and mutagenic activity (as measured by the Ames test). For each DBP, the researchers categorized exposure as low, intermediate, or high.

Of the 196,000 births, 6% were preterm, and 9% of full-term infants were SGA. Intermediate to high total THM exposure was linked to a small but statistically significant 12- to 18-gram reduction in birth weight. High exposure to MX and to mutagenic activity was also related to birth weight reduction, with mutagenic activity associated with the largest reduction (27 grams). Intermediate to high exposure to either THMs or mutagenic activity was associated with an increased risk of being SGA.

The clinical significance of the small changes observed is unknown, but it could potentially have a large public health impact, given the extent of DBP exposure. The researchers note that the associations seen between mutagenic activity and mean birth weight are unlikely to be due to THMs or haloacetic acids, because these chemicals are not considered strong mutagens. Even MX, a potent mutagen, did not appear to substantially affect the association between mutagenic activity and mean birth weight. The researchers suggest that mutagenic activity may eventually prove to be a better indicator of exposure to the complex mixtures of DBPs than measurement of traditional DBP surrogates, but more research is needed to determine the relationship between mutagenicity and individual DBPs.