Because randomized clinical trials in bipolar disorder include restricted study populations, the possibilities for generalizing to real-world bipolar patients are limited. Naturalistic long-term data can add valuable information about the diversity of treatment, outcome and risk factors in bipolar disorder.

After discharge from a psychiatric community hospital, 300 consecutively admitted ICD-10 bipolar I (n=158) and II (n=142) patients were followed-up naturalistically during for 4 years. Patients were assessed as to time to relapse, relapse in relation to index episode, prophylactic effects of prescribed medication, and risk factors for relapses such as prescribing attitudes, medication adherence, life events and alcohol use disorders.

204 (68%) of 300 patients relapsed within 4 years, with a mean of 208 days (SD=356.2) until the next affective episode. Relapses correlated in a statistically significant manner with the index episode. Using a Kaplan survival analysis, only lithium delayed time to the next affective relapse in a statistically significant way. Survival was reduced in a statistically significant manner when prophylactic medication was replaced by the psychiatrist or stopped by the patient. In a sub-analysis of this cohort life events (n=222) and alcohol use disorders (n=284) were associated with more depressive episodes in bipolar I patients.

Even though lithium seems more protective than other commonly used drugs, bipolar patients still suffer from a high relapse rate. Bad adherence, life events and alcohol use disorders are hereby major risk factors. The results urge for further and more holistic treatment approaches in bipolar disorder.