Chen, A and WJ Rogan. 2003. Nonmalarial infant deaths and DDT use for malaria control. Emerging Infectious Diseases 9(8):960-964.

A growing body of scientific evidence indicates that DDT may have a substantial impact on infant mortality, by increasing the risk of pre-term birth and by decreasing the duration of breast-feeding after birth. In this paper, Chen and Rogan conclude that DDT may cause a comparable increase in infant mortality through these mechanisms as the decrease in infant mortality it causes by killing mosquitoes and thus reducing malaria cases.

They conclude that debates over the value of using DDT for malaria control should incorporate consideration of the unintended consequences of exposure.

The authors acknowledge uncertainty in their calculations because the available research does not yet prove conclusively that DDT causes the effects on pre-term birth and breast-feeding. They argue, nonetheless, that the findings are sufficiently plausible that they must be factored into decisions about whether or not to use DDT for malaria control.

Coincidentally, this article appeared just as an op-ed was published in the New York Times recommending that DDT be used in the US to fight West Nile Virus (WNV). While Chen and Rogan do not address WNV in their paper, the analysis they advance is equally, if not even more relevant here. In this case, the increases in infant mortality due to DDT use would dramatically outweigh any possible benefits that might result from using DDT to fight West Nile Virus, because the number of deaths due to WNV is so low.

What did they do? Chen and Rogan began with data from African studies of DDT/DDE levels in people documenting the increase in DDT/DDE levels that are caused by living in homes treated in malaria control programs. They then calculated, based on the elevation in DDE levels produced by treatment, how much infant mortality would be expected to increase because of maternal DDT exposure based on the following findings from the scientific literature.

	The risk of preterm birth increases with the level of DDE measured in the mother's serum. This finding, reported by a team of scientists from the National Institute of Environmental Health Sciences and they US Centers for Disease Control, found over a 3-fold increase in preterm birth in the most exposed group. Preterm birth, in turn, has a strong link to infant mortality. Babies born before term are significantly more likely to die. [They also are at risk to a range of adverse health conditions through life (a cost not included in Chen and Rogan's calculation.]
	The duration of lactation decreases as serum DDE levels increase. Two studies (data combined in graph at right, from Chen and Rogan) both show that mothers with higher serum DDE levels nurse their babies 40%-50% less than mothers with little or no DDE.

As shown in the graph to the left (from Chen and Rogan; data from World Health Organization) An infant that is not breast feeding at under two months of age is 5.8 times more likely to die than an infant that is breast feeding.

Several factors contribute to this pattern. For example, children still breast feeding are less likely to contract diarrheal diseases through exposure to polluted water.

What did they find?

Chen and Rogan calculated that infant mortality would increase by 9% because of preterm births, and by 20% because of shortened lactation.

Combining the effect of preterm birth and shortened lactation, Chen and Rogan calculated an overall increase of 20.5 deaths per 1000 infants. By comparison, data from Africa indicate that malaria itself causes 20% of infant deaths in Africa (175 deaths per 1000 infants), with additional deaths caused by maternal transmission of malaria (3-8%).

What does it mean? The increase in infant deaths estimated by Chen and Rogan to result from maternal DDE exposure is smaller but comparable in magnitude to the number of children that die from malaria. Therefore, according to Chen and Rogers, "the side effects of DDT spraying might reduce or abolish its benefits from the control of malaria in infants."

In evaluating this comparison, several factors must be considered.

• Chen and Rogan used conservative assumptions in their calculations. Hence the side effects of DDT spraying could be substantially more adverse, even through just these two mechanims (shortened lactation, increased preterm birth).
• Use of DDT does not eliminate all childhood malaria. Hence the comparison to total infant deaths caused by malaria is somewhat misleading; the actual number of malarial deaths prevented by DDT would be lower.
• Other health effects of DDT/DDE (for example, the recent indication that DDT exposure in the womb reduces fertility in women 30 years later) would decrease DDT's net benefit further.
• Just as they don't include adverse impacts of DDT use other than increased infant mortality, they don't include beneficial impacts of DDT use that result from decreased cases of childhood and adult malaria.
• The studies on which Chen and Rogan base their research were carried out in the US and Mexico. They may not provide an accurate basis for calculations extrapolated to Africa. The extrapolation could be biased in either direction, high or low. As noted above, Chen and Rogan used conservative procedures in the calculations to avoid inflating the possible adverse effects of DDT.
• These studies have not established a causal relationship between DDE serum levels and either preterm birth or shortened lactation. The effects may be causal, and Chen and Rogan's calculations make sense only if they are causal, but the nature of the epidemiological methods used do not allow causal conclusions. This is a common situation in epidemiology.

Chen and Rogan's calculations challenge a prevailing assumption in considerations about using DDT to control malaria, that the adverse effects of DDT use are small compared to the benefits in avoided deaths. Instead, Chen and Rogan show that the adverse effects are plausibly the same order of magnitude as the benefits. Their work, while far from definitive, raises a series of new questions about the wisdom of building malaria control programs on DDT use, when not only are there likely to be substantial adverse effects, but there are also affordable alternatives.

Chen and Rogan's work is also of immediate relevance to a proposal that DDT be used to fight West Nile Virus in the United States. In this case, the possible benefits of DDT use are numerically tiny compared to the likely adverse effects, even if the only adverse effect of DDT is increased infant mortality.

By one calculation, from the lead scientist at the National Institute of Health who carried out the study linking DDE to increased preterm birth, up to 15% of infant mortality in the US during the 1950's and 1960's, may have been caused by DDT. If that estimate is correct, elimination of DDT use in the US is today avoiding approximately 66,000 infant deaths each year. Even if that estimate is high by a factor of 10, that would be 6,600 deaths, and this calculation does not include any contribution of decreased lactation period. By comparison, West Nile Virus last year caused approximately 300 deaths in the US.

Gladen, BC and WJ Rogan. 1995. DDE and shortened duration of lactation in a northern Mexican town. American Journal of Public Health 85:504-508.

Rogan, WJ, BC Gladen, JD McKinney, N Carreras, P Hardy, and J Thullen, J Tingelstad and M Tully. 1987. Polychlorinated biphenyls (PCBs) and dichlorodiphenyl dichloroethene (DDE) in human milk: effects on growth, morbidity, and duration of lactation. American Journal of Public Health 77:1294-1297.