OBJECTIVE: We assessed the incidence of diabetes in an ambulatory population treated for the first time with either olanzapine or risperidone. METHODS: We conducted a population-based cohort study using data from the Quebec drug benefit plan. Included in the cohort, all people who received a first prescription of olanzapine or risperidone between 1/1/1997 and 8/31/1999, who were eligible for the drug plan, had not been prescribed an antidiabetic drug or any atypical antipsychotic for the six-month period preceding the first olanzapine or risperidone prescription. Person-months of follow-up were calculated as the amount of time from the date of the first olanzapine or risperidone prescription to the date of the first antidiabetic drug prescription. Those who had a prescription for an antidiabetic drug were considered as having diabetes. Subjects who discontinued olanzapine or risperidone, who became non-eligible for the drug plan and those who reached the end of the follow-up period (8/31/2000) were censored at the event date. We used a proportional hazard model to compute the age- and sex-adjusted incidence-rate ratio (IRR) of having diabetes among olanzapine users compared to risperidone users. RESULTS: In all, 28,142 people were included in the cohort (12,945 olanzapine users and 15,197 risperidone users). The risk of developing diabetes was higher for those exposed to olanzapine than for those exposed to risperidone, but this increase in risk was marginally statistically significant (IRR: 1.209, 95 % CI: 1.001-1.460). CONCLUSION: Compared to risperidone users, olanzapine users have a slightly higher risk of developing diabetes. There is a need to extend this study over a longer observation period, as the risk of developing diabetes after exposure to antipsychotic drugs is likely to increase with time.