Summary Aims The main objective of this study was to see whether diabetes is associated with an increased collision risk and to test the effect of age and gender on the overall collision risk for diabetes drivers. Materials and methods Twenty‐eight studies were included in meta‐analysis, using mean age, gender, continent and the prevalence of fatal road incidents as covariates. Results The collision risk for diabetes drivers was small and not statistically significant – RR = 1.11 (1.01–1.23) with a prediction interval (PI) or 0.77‐1.65. Age and gender were not associated with an increased overall risk. Insulin‐dependent diabetes patients had a slightly increased effect size compared with the overall diabetes population, but the effect was not statistically significant. European diabetes drivers had a lower collision risk compared with their North American counterparts, the main cause being the difference of collision risk in the countries in which the studies were performed. Conclusions Overall, diabetes patients do not have a statistically significant increased risk for unfavourable traffic events. Old age and insulin‐dependent patients tend to have a higher risk. Advances in diabetes care, associated with advances in road safety regulations, and automotive industry have not decreased significantly the collision risk in the last 50 years for drivers with diabetes.