RESUMO
BACKGROUND: Increased long-term prescription of opioids and/or benzodiazepines necessitates evaluating risks associated with their receipt. We sought to evaluate the association between long-term opioids and/or benzodiazepines and mortality in HIV-infected patients receiving antiretroviral therapy and uninfected patients. METHODS: Prospective analysis of all-cause mortality using multivariable methods and propensity score matching among HIV-infected patients receiving antiretroviral therapy and uninfected patients. RESULTS: Of 64,602 available patients (16,989 HIV-infected and 47,613 uninfected), 27,128 (exposed and unexposed to long-term opioids and/or benzodiazepines) were 1:1 matched by propensity score. The hazard ratio for death was 1.40 [95% confidence interval (CI): 1.22 to 1.61] for long-term opioid receipt, 1.26 (95% CI: 1.08 to 1.48) for long-term benzodiazepine receipt, and 1.56 (95% CI: 1.26 to 1.92) for long-term opioid and benzodiazepine receipt. There was an interaction (P = 0.01) between long-term opioid receipt and HIV status with mortality. For long-term opioid receipt, the hazard ratio was 1.46 (95% CI: 1.15 to 1.87) among HIV-infected patients, and 1.25 (95% CI: 1.05 to 1.49) among uninfected patients. Mortality risk was increased for patients receiving both long-term opioids and benzodiazepines when opioid doses were ≥ 20 mg morphine-equivalent daily dose and for patients receiving long-term opioids alone when doses were ≥ 50 mg morphine-equivalent daily dose. CONCLUSIONS: Long-term opioid receipt was associated with an increased risk of death; especially with long-term benzodiazepine receipt, higher opioid doses, and among HIV-infected patients. Long-term benzodiazepine receipt was associated with an increased risk of death regardless of opioid receipt. Strategies to mitigate risks associated with these medications, and caution when they are coprescribed, are needed particularly in HIV-infected populations.
