Pregnancy fears on drinking water byproducts may be overstated

Aug 1 2005

Fears that chemical byproducts resulting from purifying drinking water with chlorine boost the chances that pregnant women will miscarry were not supported by the results of a major new study. If such threats exist at all, which is uncertain, they likely are modest, it concludes.

The national study, directed by University of North Carolina at Chapel Hill scientists, contrasts with earlier, less detailed work done in Northern California and published in 1998. That research suggested an association between byproducts known as trihalomethanes and loss of pregnancy.

"We think our new work should be an important contribution to policy studies," said principal investigator Dr. David A. Savitz of the UNC School of Public Health. "While it is not the final answer, what we found is largely reassuring relative to what had come before.

"The vast majority of the U.S. population is living with these exposures to drinking water byproducts," Savitz said. "If they clearly increased women's miscarriage rates, that would be a very big, very expensive problem to solve. Reducing exposure would be quite expensive for the water utilities and ultimately their customers."

The methods used in the new work make it the most ambitious and sophisticated study ever done on this issue, he said.

Savitz, Cary C. Boshamer distinguished professor and chair of epidemiology, and colleagues released a report on the investigation today (July 29). The American Water Works Association Research Foundation, a non-partisan, non-profit research group, sponsored it.

Others involved include Dr. Philip C. Singer, professor of environmental sciences and engineering and co-principal investigator, and Drs. Katherine E. Hartmann of obstetrics and gynecology and epidemiology, Amy H. Herring of biostatistics and Howard S. Weinberg of environmental sciences and engineering.

The scientists selected and repeatedly tested three properly functioning water purification facilities, including one with moderate levels of chlorinated disinfection byproducts, one with moderate levels of brominated disinfection byproducts and one with low levels of all disinfection byproducts. None of the utilities was selected because of water quality or health problems.

Between December 2000 and April 2004, the team recruited 3,132 women in the three areas who were planning a pregnancy or who had been pregnant for fewer than 12 weeks. They collected extensive information on the volunteers, including their patterns of water use such as drinking and bathing and what later happened with their pregnancies and births. Among tests the women underwent were ultrasound examinations to collect more detailed information about their pregnancies.

Researchers then, after adjusting for such factors as age, education, alcohol use and prior pregnancy histories, analyzed the data to see what effects the byproduct compounds might have had on pregnancy loss (miscarriage).

The study uncovered no clear-cut evidence that trihalomethanes harmed women or their developing infants and as a result differed significantly from the Northern California research, Savitz said. The new work did suggest that brominated compounds and total organic halides at normal-range levels in purified drinking water might modestly increase the risk of miscarriages among pregnant women.

"These latter findings suggest that we or others should take a closer look at individual and groups of chemicals that might have a negative effect on pregnancy," he said. "I don't want to downplay those findings and say they were perfectly reassuring because they were not. But overall -- on balance -- I'd say this work moves the evidence in a reassuring direction and should serve to lessen concerns."

Because of the overall negative results, he and his colleagues do not plan to make policy recommendations to water managers and municipalities, he said. If possible, future comparable studies ought to include methods for determining actual exposures to purification byproducts more accurately.

"We also looked at exposure to these chemicals in relation to preterm births and did not find an association," Savitz said. "If anything, we found the opposite -- that higher-exposure populations had slightly lower risks. There was some indication, however, that reduced fetal growth might be associated with higher exposure levels."

Particularly helpful to the study were the women who participated and the participating water work utilities, he said.

"We are grateful to all these people since we could not have done this useful study without their active cooperation," Savitz said.

A number of earlier investigations have suggested that long-term exposure to water purification byproducts might be related to bladder cancer, an issue not addressed in the present study.