Neuropsychological correlates of manganese exposure: a meta-analysis.

The hypothesized effect of recurrent low-dose manganese (Mn) exposure on neuropsychological function is controversial because of inconsistent findings across three decades of research. We conducted a meta-analysis on 41 variables from nineteen neuropsychological studies of Mn-exposed workers. The results showed: Large effect size (ES) for biological markers of Mn and lead levels; thirteen of 26 neurocognitive measures showing a small average ES; only one of 26 tasks showed a moderate ES; and small to medium ES for confounding/competing variables such as education and aptitude. Tasks with the highest ES included clerical substitution tasks, digit span, tapping endurance, and Swedish Performance Evaluation System "Additions" reaction time, but none exceeded the ES for education or aptitude. The mean ES of dose-response relationships was zero. The data did not support a theory of preclinical ("early") neuromotor or cognitive dysfunction. Overall, the pooled data are more consistent with covariate effect than toxic effect, insofar as the pooled exposure group showed demographics less favorable to neuropsychological performance than the pooled referent groups. Future consideration of demographic and biological covariates is necessary before inferring subtle toxin-induced brain damage because neuropsychological tests are nonspecific.