Association between Preoperative Opioid and Benzodiazepine Prescription Patterns and Mortality after Noncardiac Surgery

Abstract

Importance: The number of patients prescribed long-term opioids and benzodiazepines and complications from their long-term use have increased. Information regarding the perioperative outcomes of patients prescribed these medications before surgery is limited. Objective: To determine whether patients prescribed opioids and/or benzodiazepines within 6 months preoperatively would have greater short- and long-term mortality and increased opioid consumption postoperatively. Design, Setting, and Participants: This retrospective, single-center, population-based cohort study included all patients 18 years or older, undergoing noncardiac surgical procedures at a national hospital in Iceland from December 12, 2005, to December 31, 2015, with follow-up through May 20, 2016. A propensity score-matched control cohort was generated using individuals from the group that received prescriptions for neither medication class within 6 months preoperatively. Data analysis was performed from April 10, 2018, to March 9, 2019. Exposures: Patients who filled prescriptions for opioids only, benzodiazepines only, both opioids and benzodiazepines, or neither medication within 6 months preoperatively. Main Outcomes and Measures: Long-term survival compared with propensity score-matched controls. Secondary outcomes were 30-day survival and persistent postoperative opioid consumption, defined as a prescription filled more than 3 months postoperatively. Results: Among 41170 noncardiac surgical cases in 27787 individuals (16004 women [57.6%]; mean [SD] age, 56.3 [18.8] years), a preoperative prescription for opioids only was filled for 7460 cases (17.7%), benzodiazepines only for 3121 (7.4%), and both for 2633 (6.2%). Patients who filled preoperative prescriptions for either medication class had a greater comorbidity burden compared with patients receiving neither medication class (Elixhauser comorbidity index >0 for 16% of patients filling prescriptions for opioids only, 22% for benzodiazepines only, and 21% for both medications compared with 14% for patients filling neither). There was no difference in 30-day (opioids only: 1.3% vs 1.0%; P =.23; benzodiazepines only: 1.9% vs 1.5%; P =.32) or long-term (opioids only: hazard ratio [HR], 1.12 [95% CI, 1.01-1.24]; P =.03; benzodiazepines only: HR, 1.11 [95% CI, 0.98-1.26]; P =.11) survival among the patients receiving opioids or benzodiazepines only compared with controls. However, patients prescribed both opioids and benzodiazepines had greater 30-day mortality (3.2% vs 1.8%; P =.004) and a greater hazard of long-term mortality (HR, 1.41; 95% CI, 1.22-1.64; P