Comparing Medicine Combinations Used to Treat Schizophrenia [Internet]

Excerpt

Background: Schizophrenia is a serious mental health disorder that has a lifetime population risk of nearly 1%. It typically begins in early adulthood and is a major cause of disability. Schizophrenia outcomes vary, ranging from slight to severe disability. For affected individuals and their families, schizophrenia can be devastating. Repeated hospitalizations, crises, and serious medical illnesses commonly disrupt the lives of people with schizophrenia. Age-specific all-cause mortality rates in schizophrenia are more than 3 times higher than in the general population. Many people with schizophrenia have substantial social and functional impairments that contribute to high rates of homelessness, unemployment, disability, and incarceration. Globally, schizophrenia is among the top 10 causes of years lost due to disability.

Antipsychotic medications are a cornerstone of schizophrenia treatment. However, because most people with schizophrenia respond only partially or not at all to an antipsychotic medication, these medications are changed frequently and combined, or different classes of psychotropic medications are added. Although the US Food & Drug Administration has approved only antipsychotics for the treatment of schizophrenia, the vast majority of patients with schizophrenia take multiple classes of psychotropic medications. Evidence is scarce for effectiveness of psychotropic medications other than antipsychotics in people with schizophrenia.

Methods: For this retrospective cohort study, we used 10 years of US Medicaid data from 44 states to examine treatment outcomes among 81 921 patients with schizophrenia who were stably treated with a single (baseline) antipsychotic. We then initiated a second drug: an antidepressant (N = 31 117), a benzodiazepine (N = 11 941), a mood stabilizer (N = 12 849), or another antipsychotic medication (N = 26 014). The reference group was the one that started another antipsychotic medication. Patients were required to have an active prescription for the baseline antipsychotic at the time of initiation, but no refills were required.

We used multinomial logistic regression models to estimate propensity scores to balance covariates across the 4 medication groups. We used weighted Cox proportional hazards models to compare treatment outcomes over 365 days on an intent-to-treat basis for each patient. Outcomes included time to hospitalization for a mental health disorder (primary), emergency department (ED) visits for a mental health disorder, and all-cause mortality. We conducted multiple sensitivity analyses to examine the robustness of the primary results. In addition, secondary analyses examined possible heterogeneity of treatment effects according to age or presence of a substance use disorder and compared the effectiveness of the treatments within clinical subgroups.

Results: Compared with initiating another antipsychotic, initiating an antidepressant was associated with a lower risk (0.84; 95% CI, 0.80-0.88) of psychiatric hospitalization, initiating a benzodiazepine was associated with a higher risk (1.08; 95% CI, 1.02-1.15), and initiating a mood stabilizer was not significantly different (1.01; 95% CI, 0.97-1.06). For psychiatric ED visits, we observed a similar pattern of associations for initiating an antidepressant (0.92; 95% CI, 0.88-0.96), a benzodiazepine (1.12; 95% CI, 1.07-1.19), and a mood stabilizer (1.00; 95% CI, 0.96-1.05). Initiating a mood stabilizer was associated with an increased risk of mortality (1.31; 95% CI, 1.04-1.55). Sensitivity analyses supported the main findings.

Conclusion: In the treatment of schizophrenia, initiating adjunctive treatment with an antidepressant was associated with reduced risk of psychiatric hospitalization and ED visits compared with initiating other psychotropic medications. Associations of benzodiazepines and mood stabilizers with poorer outcomes warrant clinical caution and further investigation.

Limitations: The data on medication use are based on prescriptions filled, but no information was available on actual medication ingestion. The data do not include information on what prompted the changes in medications or which symptoms were targeted. Because this is an observational study, the results must be interpreted with caution. Although many features of the study design, including propensity score adjustment and weighting, enhanced our ability to make causal inferences, residual confounding cannot be completely ruled out as an explanation for the associations.