Abstract
Background
Cohort studies of surgery compared with medical treatment (MT) on endocarditis mortality are conflicting. We conducted a population-based study to estimate associations between treatment and mortality.
Methods
1,381 patients with substance use disorder (SUD) and 5,053 without (NSUD) hospitalized for endocarditis were included. Treatment was modeled as a time-dependent variable: patients who underwent surgery after admission were classified as MT until surgery occurred and surgically treated thereafter. Patients without surgery were classified as MT. Adjusted hazard ratios (aHR) between treatment and death (in-hospital, 30 days, one, two, five years) by SUD status were estimated.
Results
Among SUD patients, there was a trend towards reduction in in-hospital death with surgery vs. MT (aHR 0.61 [95% CI: 0.35–1.04]), but no difference at 30 days (aHR 0.79 [95% CI: 0.42–1.48]). Mortality was higher in SUD patients who underwent surgery compared with MT at one (aHR 1.30 [95% CI: 0.95–1.76]), two (aHR 1.27 [95% CI: 0.97–1.65]), and five years (aHR 1.37 [95% CI: 1.09–1.72]). In NSUD patients, in-hospital mortality (aHR 0.93 [95% CI 0.76–1.16]) did not differ, but 30 day mortality (aHR 1.36 [95% CI 1.04–1.77]) was higher with surgery versus MT, and lower at one (aHR 0.87 [95% CI: 0.73–1.03]), two (aHR 0.75 [95% CI: 0.64–0.88]), and five years (aHR 0.70 [95% CI: 0.61–0.81]).
Kaplan-Meier Survival Curves of Patients
Interpretation
Surgery compared with MT conferred no long-term survival benefit in SUD patients. In NSUD patients, surgery was associated with an initial increased risk of early death followed by a lower risk after one year.
Acknowledgement/Funding
Grant from Department of Surgery, Queen's University