Medical comorbidity, acute medical care use in late-life bipolar disorder: a comparison of lithium, valproate, and other pharmacotherapies.

Objective: Bipolar disorder is associated with high rates of medical comorbidity, particularly in late life. Little is known about medical health service utilization and potential effects of bipolar pharmacotherapy. We hypothesized that lithium use would not be associated with higher rates of medical hospitalization.

Methods: Population-based retrospective cohort study of 1388 bipolar disorder patients aged ≥66years discharged from a psychiatric hospitalization in Ontario, Canada, between 2006 and 2012. Patients were divided into lithium users, valproate users, and non-lithium/non-valproate users. The main outcome was acute non-psychiatric, medical/surgical hospitalization during 1-year follow-up.

Results: The rate of medical hospitalizations was 0.22 per patient-year. Time-to-medical hospitalization did not differ among lithium, valproate, and non-lithium/non-valproate users after adjusting for age, sex, past medical hospitalization, and antipsychotic use. Lithium, valproate, and non-lithium/non-valproate users did not differ markedly in terms of reason for medical hospitalization, 1-year acute medical health utilization outcomes, and medical comorbidity rates.

Conclusion: There were high rates of health service use for medical conditions among older adults with bipolar disorder, but this did not appear to be associated with lithium use, compared to valproate and other medication use (e.g., antipsychotics). A proactive collaborative care approach may prevent medical service utilization in severe late-life bipolar disorder.