Record-linkage studies of the coexistence of epilepsy and bipolar disorder.

PURPOSE: Interest is growing in a possible link between epilepsy and bipolar disorder (BPD). We used two large datasets of hospital admission data to determine whether epilepsy and BPD occur together in the same individuals more commonly than expected. METHODS: We undertook retrospective cohort studies using the Oxford Record Linkage Study (ORLS) and English national linked Hospital Episode Statistics. We constructed a cohort of people in each dataset admitted with epilepsy (without prior admission for BPD), and a control cohort (without prior admission for BPD), and compared their subsequent admission rates for BPD. Conversely, we constructed a cohort of people in each dataset admitted with BPD and a control cohort (both without prior admission for epilepsy), and compared their subsequent admission rates for epilepsy. RESULTS: In the epilepsy cohort, compared with the control cohort, the rate ratio (RR) for BPD was significantly high at 3.0 (95 % confidence interval 1.7-5.1) in the ORLS and 3.6 (3.3-3.9) in the all-England dataset. In the BPD cohort, the RR for epilepsy was 2.2 (1.2-3.7) in the ORLS and 4.2 (4.0-4.4) in the all-England cohort. We found no significant differences between RRs for males and females. CONCLUSIONS: Epilepsy and BPD occur together in individuals more frequently than expected by chance.