Objective: We conducted a meta-analysis of epidemiological studies investigating the association between maternal age and autism. Method: Using recommended guidelines for performing meta-analyses, we systematically selected, and extracted results from, epidemiological scientific studies reported before January 2012. We calculated pooled risk estimates comparing categories of advancing maternal age with and without adjusting for possible confounding factors. We investigated the influence of gender ratio among cases, ratio of infantile autism to autism spectrum disorder (ASD), and median year of diagnosis as effect moderators in mixed-effect meta-regression. Results: We found 16 epidemiological papers fulfilling the a priori search criteria. The meta-analysis included 25,687 ASD cases and 8,655,576 control subjects. Comparing mothers greater than or equal to 35 years with mothers 25 to 29 years old, the crude relative risk (RR) for autism in the offspring was 1.52 (95% confidence interval [CI] = 1.12-1.92). Comparing mothers greater than or equal to 35 with mothers 25 to 29, the adjusted relative risk (RR) for autism in the offspring was 1.52 (95% CI = 1.12-1.92). For mothers less than 20 compared with mothers 25 to 29 years old, there was a statistically significant decrease in risk (RR = 0.76; 95% confidence interval = 0.60-0.97). Almost all studies showed a dose-response effect of maternal age on risk of autism. The meta-regression suggested a stronger maternal age effect in the studies with more male offspring and for children diagnosed in later years. Conclusions: The results of this meta-analysis support an association between advancing maternal age and risk of autism. The RR increased monotonically with increasing maternal age. The association persisted after the effects of paternal age and other potential confounders had been considered, supporting an independent relation between higher maternal age and autism. (Contains 4 figures and 3 tables.)