The influence of genetic polymorphisms in GSTM1 and GSTT1 genes on micronucleus frequencies in human peripheral blood lymphocytes was assessed through a pooled analysis of data from seven laboratories that did biomonitoring studies using the in vivo cytokinesis-block micronucleus assay. A total of 301 nonoccupationally exposed individuals (207 males and 94 females) and 343 workers (237 males and 106 females) occupationally exposed to known or suspected genotoxic substances were analyzed by Poisson regression. The results of the pooled analysis indicate that the GSTT1 null subjects had lower micronucleus frequencies than their positive counterparts in the total population (frequency ratio, 0.55; 95% confidence interval, 0.33-0.89). The protective effect of this genotype is reversed with increasing age, with a frequency ratio of 1.33 (95% confidence interval, 1.06-1.68) in subjects aged 60 years. A significant overall increase in micronucleus frequency with age and gender (P < 0.001 and P = 0.024, respectively) was observed, females having higher micronucleus frequencies than males, when occupationally exposed (P = 0.002). Nonoccupationally exposed smokers had lower micronucleus frequencies than nonsmokers (P = 0.001), whereas no significant difference in micronucleus level was observed between smokers and nonsmokers in the occupationally exposed group (P = 0.79). This study confirms that pooled analyses, by increasing the statistical power, are adequate for assessing the involvement of genetic variants on genome stability and for resolving discrepancies among individual studies. (Cancer Epidemiol Biomarkers Prev 2006;15(5):1038 -42)