Incidence of and risk factors for alcohol dependence in bipolar disorder: A population-based cohort and nested case–control study

Abstract

Objectives: Although alcohol dependence is highly prevalent in patients with bipolar disorder, the causal relationship is not yet well-established. This study estimated the incidence of alcohol dependence in a nationwide bipolar disorder cohort and examined risk factors for alcohol dependence. Methods: Patients aged 15–65 years with consistent bipolar disorder who had their first psychiatric admission between 1999 and 2012 ( n = 21,791) were enrolled from the National Health Insurance Research Database in Taiwan. We calculated the adjusted incidence rate ratio of alcohol dependence in the bipolar cohort relative to the general population after stratification by age and sex. In the nested case–control study, we included patients with incident alcohol dependence as cases and four age- and sex-matched controls for each case to analyze health care utilization, comorbidities and concomitant medications between them. Results: We identified 1261 patients with bipolar disorder with incident alcohol dependence. Relative to the general population, the adjusted incidence rate ratio of alcohol dependence was 9.20 in the bipolar cohort. All adjusted incidence rate ratios were high across all age subgroups. Cases had higher psychiatric and nonpsychiatric health care utilization than did controls. Multivariate analysis revealed that cases tended to have cardiovascular disease, diabetes mellitus, chronic hepatic disease, pneumonia and delirium before alcohol dependence diagnosis. Cases had higher psychiatric comorbidities, namely drug-induced mental disorders, anxiety disorder, personality disorder, adjustment disorder and sleep disorder. Conclusion: The bipolar cohort had a higher incidence of alcohol dependence. We identified specific groups with a high risk of alcohol dependence. Additional strategies for early detection, treatment and intervention for alcohol dependence should be developed.