The impact of solvent mixtures on neurobehavioral performance: conclusions from epidemiological data

The review of epidemiological studies investigating the neurobehavioral effects of occupational exposure to solvent mixtures sought to contribute to the following issues: (1) Identification of affected cognitive and motor functions. (2) Identification of sensitive neuropsychological tests. (3) Analysis of exposure-effect relationships. The approach was based on the meta-analytical method of effect size estimates. Fifty-three groups from occupational studies were included in the meta-analysis. Forty-eight neuropsychological performance variables could be analyzed as they were included in at least three studies. Seventeen articles provided detailed information on the constituents of mixtures, thereby enabling the computation of an exposure index that allowed the comparison of different mixtures. Significant negative effect sizes were obtained for 12 test variables measuring attention, memory, motor performance and constructional abilities. The greatest proportion of lower performance scores in the exposed groups was shown by different tests of attention: significant effect sizes between d=-0.16 and -0.46 were calculated. Tests of cognitive processing speed, response alternation and inhibition seemed to be sensitive tools for the detection of poorer performance. Exposure-effect relationships were mainly characterized by inconsistent patterns. Crude and inappropriately calculated exposure measures were blamed for this outcome. A healthy worker effect was suggested more consistently: studies examining groups with longer exposure duration obtained smaller effect sizes. Indications of confounding were observed; however, they did not seem sufficient to question consistent effect size patterns. Paying greater attention to the measurement of exposure and including measures of confounding is advisable for future studies and would enhance the explanatory power of cross-sectional studies and meta-analyses.