Chlorinated water and birth defects

“Chlorine in tap water 'nearly doubles the risk of birth defects'” read the headline in the Mail On Sunday on June 1 2008. A study has found that women who consume water that is “heavily disinfected with chlorine” have nearly twice the risk of having babies with “heart problems, a cleft palate or major brain defects”, the newspaper said. The Mail added that these findings appear to contradict a major study carried out in 2007 by Imperial College, London, which found “little evidence” of an association between THM levels – a group of chemical by-products in chlorinated water – and birth defects in Britain.

The newspaper story is based on a Taiwanese study, which has a number of limitations, including the fact that women’s THM exposure was not measured directly, but was estimated based on where she lived. It also did not take into account several factors known to be related to risk of birth defects, such as smoking and alcohol consumption. This study does not provide robust enough results to conclude that THMs affect the risk of any type of birth defect. Pregnant women should not be concerned about drinking tap water and should stick to their doctor’s advice on the foods and drink that should be avoided during pregnancy.

Where did the story come from?

Dr Bing-Fang Hwang and colleagues from Universities in Taiwan and the University of Birmingham carried out this research. The study was funded by the National Science Council. It was published in the peer-reviewed scientific journal Environmental Health .

What kind of scientific study was this?

This was a cross-sectional study looking at birth defects and water quality in Taiwan. The researchers also carried out a meta-analysis (pooling) of data from other studies.

In the cross-sectional part of their study, the researchers identified all births in Taiwan between 2001 and 2003, using records from the Department of Health. Looking at only five areas of the country, where there was data available on water quality for the same period from the Waterworks registry, the researchers included 396,049 births. The researchers used the birth records to identify all cases of the 11 most common birth defects, including various brain and heart defects, cleft palates and lips, kidney and urinary tract defects, and chromosome defects. These records include any defects diagnosed between 20 weeks of pregnancy and seven days after birth, but do not include details of any abortions due to birth defects. The records also provided data about the sex of the baby, whether it was a multiple birth (e.g. twins), the mother’s age and whether the mother had certain medical conditions such as heart, lung, or kidney disease, diabetes, genital herpes, high blood pressure, too much or too little fluid in the amniotic sac or uterine bleeding.

The researchers also obtained records of water quality between 2001 and 2003 from water treatment plants in the five areas of interest that used chlorination to disinfect their water. They specifically looked at the total concentration of one group of chemicals (the trihalomethanes – THMs) that are formed as by-products of the chlorination process (as well as other water disinfection processes). The water treatment plant measured and recorded levels of THMs at least four times a year. Based on where a mother lived, the researchers estimated her exposure to THMs in the water supply during her pregnancy. Estimated THM exposure was classified as negligible, low, medium or high. The researchers compared the risk of birth defects in women with low to high THM exposure with women with negligible exposure. They adjusted these analyses for factors that would affect results, such as the mother’s age, whether it was a multiple birth and the population density in the area where the mother lived.

The researchers also searched a database of scientific literature (Medline) to look for other studies published between 1966 and 2007 that assessed the effect of the by-products of chlorination on birth defects. They also looked in relevant journals and at the reference lists of relevant scientific papers to identify further relevant studies. They included cross-sectional, cohort and case control studies. They then pooled the results of their study with those of the studies they identified.

What were the results of the study?

Overall, only five in every 1,000 babies had any of the 11 birth defects. When compared with mothers with negligible THM exposure, there was an increase in the odds of any birth defect in women with low THM exposures, but not those with medium or high exposures. When looking at each of the 11 birth defects individually, the odds of having some of the defects (including defects of the wall separating the lower chambers of the heart) was increased in the foetuses with a high exposure to THMs in the water, but that this increase was not large enough to be statistically significant. There was also a 56% increase in the risk of cleft palate in foetuses with high exposure, but this increase only just reached statistical significance (odds ratio 1.56, 95% confidence intervals 1.00 to 2.41).

In their literature search, the researchers identified three cross-sectional studies and two case control studies that looked at the effects of by-products of chlorination on birth defects in different countries (Sweden, Norway, USA and England and Wales). When they pooled the results from these studies, they found that high exposure to THMs increased the odds of the baby having defects of the wall separating the lower chambers of the heart, but there was no significant increase in the risk of the other birth defects.

What interpretations did the researchers draw from these results?

The researchers concluded that exposure to the by-products of water disinfection increased the risk of anencephalus (a fatal condition in which most of the upper part of the brain and covering skull has not developed), cleft palate and defects of the wall separating the lower chambers of the heart.

What does the NHS Knowledge Service make of this study?

This study has a number of limitations and there are a number of points to consider when interpreting the results:

• This study did not follow women during pregnancy and assess the water they drank; instead it estimated women’s exposure to THMs depending on where they lived. They had no information about what each woman drank during her pregnancy, or about other possible exposures to chlorination by-products, for example, by swimming or bathing. Therefore, the estimates of THM exposure may not be reliable. Reports in the Mail that women can put themselves at risk by “drinking the water, taking a bath or shower, or even standing close to a boiling kettle” are assumptions that are not founded in this research.
• The researchers relied on birth records to identify birth defects; some defects may have been missed and some diagnoses may have been incorrectly recorded, which may have affected results.
• This study carried out a large number of statistical tests. The more statistical tests a study carries out, the more likely it is to find a significant result purely by chance. Most of the increases in risk did not reach statistical significance, which means that it is not possible to say whether THM exposure has any effect on risk.
• Although the study tried to adjust for factors that affected results (such as maternal age) there are many factors that it did not take into account that will affect the risk of birth defects, including the mother’s nutritional state during the pregnancy, smoking, alcohol consumption and other environmental factors, as well as genetic factors. Therefore, it is not possible to be certain that THMs, rather than any of these other factors, were responsible for any of the increases seen.
• It is unclear to what extent the levels of disinfection by-products in Taiwan represent the situation in other countries such as the UK. Therefore it is uncertain whether the results are applicable to other countries. 
• Birth defects are very rare and the numbers of cases for each type of birth defect in each exposure category were very small (for example, there were only four cases of defects of the wall separating the lower chambers of the heart in women who had high exposure to THMs). Analysing such small numbers can also lead to finding significant results by chance.

Overall, the authors themselves state that “our results showed no consistent association between exposure and risk of birth defects in general”, but go on to suggest that it is better to look at specific birth defects. However, this study does not provide results robust enough to conclude that THMs affect the risk of any type of birth defect, and should not cause women to be concerned about drinking water. Pregnant women should stick to their doctor’s advice on the foods and drink that should be avoided in pregnancy.

Sir Muir Gray adds...

We need to see more studies in different countries before stopping adding chlorine to water.