Purpose Patients with epilepsy or psychiatric diseases have increased risk of suicide, but whether the risk is influenced by antiepileptic drug (AED) treatment is unclear. Studies have suggested that AEDs in general increase the risk of suicidal behaviour shortly after initiation. This study investigated possible differences in suicide risk associated with different AEDs. Methods The use of AEDs in the Danish population from 1997 to 2006 was determined by prescription claims. The risk of suicide associated with use of AEDs was estimated by case-crossover analyses, where each case serves at its own control during different periods. For sensitivity, the risk of suicide was estimated by a time-dependent Cox proportional-hazard analysis in AED treatment-nave patients. Results There were 6780 cases committing suicide in the 10-year study period, of which 422 received AED treatment at the time of suicide. The case-crossover analysis estimated AED treatment initiation to increase the risk of suicide (odds ratio (OR): 1.84, 95% confidence interval (CI): 1.36-2.49). Clonazepam (OR: 2.01, CI: 1.25-3.25), valproate (OR: 2.08, CI: 1.01-4.16), lamotrigine (OR: 3.15, CI: 1.35-7.34) and phenobarbital (OR: 1.96, CI: 1.02-3.75) were associated with a significant increased risk, while the remaining examined AEDs did not significantly influence the risk. In the cohort comprising of 169 725 AED treatment-nave patients, the Cox proportional-hazard analysis yielded similar results. Conclusions This study suggests that clonazepam, valproate, lamotrigine and phenobarbital relatively shortly after treatment initiation may increase the risk of suicide. The increased risk of suicide associated with these AEDs appears to be a consistent finding. Copyright (C) 2010 John Wiley & Sons, Ltd