Opler, MGA, AS Brown, J Graziano, M Desai, W Zheng, C Schaefer, P Factor-Litvak and ES Susser. 2004. Prenatal lead exposure, δ-Aminolevulinic acid and schizophrenia. Environmental Health Perspectives doi:10.1289/ehp.6777 (available at http://dx.doi.org/).

Opler et al. open a new chapter in the study of lead's impact on brain development by exploring the association between lead exposure in the womb and adult schizophrenia. Their results suggest a link, but fall just short of statistical significance.

What did they do? Opler et al. analyze maternal serum lead levels in a cohort of newborns born between 1959 and 1976 and enlisted to a study of the prenatal determinants of schizophrenia. Maternal blood samples were taken several times during pregnancy and then had stored frozen at NIH ever since, as the babies developed into adulthood.

The researchers searched through computerized health records of the participants to identify potential cases of schizophrenia. Some 12,000 of the original participants remained in the health plan beyond 1981, when computerization of records began to allow the necessary searches. Possible cases were then contacted and assessed by experienced clinicians for schizophrenia and related conditions.

Forty-three subjects were found with a diagnosis of schizophrenia, plus another 28 with related psychological disorders.

For technical reasons, Opler et al. used an indirect measurement of lead exposure, determining the level of δ-aminolevulinic acid (δ-ALA) in the stored maternal serum. Lead inhibits the metabolic conversion of δ-ALA to another compound. Hence higher δ-ALA indicates higher lead exposure. They also decided to limit the study to individuals for whom maternal serum had been taken during the second trimester of pregnancy. This was done both because of indications from other studies that second trimester events might be linked to later schizophrenia, and because δ-ALA levels fluctuate immediately prior to birth in ways that might obscure an association with lead exposure.

δ-ALA levels from second trimester maternal serum for people diagnosed with schizophrenia (44 cases) and related conditions were then compared with δ-ALA in control patients (75 controls) selected from the remaining patients in the study cohort to match the cases.

What did they find? In their first analysis, not correcting for confounding variables, Opler et al. found the odds ratios comparing individuals whose mothers had elevated δ-ALA levels to those with lower levels was 1.83, with a 95% confidence interval running from 0.85 to 3.95.

They then refined their analysis taking into account a series of confounding variables, including maternal age. The odds ratio rose to 2.4, with 95% confidence intervals running from 0.99 to 5.96.

What does it mean? This is the first prospective study of a link between prenatal chemical exposure and an adult psychiatric disorder. Hovering at the edge of statistical significance, the results are tantalizingly suggestive but far from conclusive.

At the very least, they legitimize, indeed compel, further reseach into the possible contribution of lead exposure in the womb to adult psychiatric disorders. Given the substantial literature on lead impacts on children's neurological development [for example, recent research on delinquent behavior] such contributions should not be unexpected.

The study is limited by several factors: The sample size is small, despite the remarkable investment made in establishing the birth cohort during the 1960s and maintaining health records through to adulthood. The measurement used to assess lead exposure, δ-ALA, is not ideal. And none of the confounding variables involved post-natal factors that might have modified the effects of prenatal lead exposure.