Objective While opioid use disorder (OUD) has been previously associated with increased morbidity and resource use in cardiac operations, its impact on readmissions is understudied.

Methods Patients undergoing coronary artery bypass grafting and valve repair or replacement, excluding infective endocarditis, were identified in the 2010–16 Nationwide Readmissions Database. Using International Classification of Diseases 9/10, we tabulated OUD and other characteristics. Multivariable regression was used to adjust for differences.

Results Of an estimated 1 978 276 patients who had cardiac surgery, 5707 (0.3%) had OUD. During the study period, the prevalence of OUD increased threefold (0.15% in 2010 vs 0.53% in 2016, parametric trend<0.001). Patients with OUD were more likely to be younger (54.0 vs 66.0 years, p<0.001), insured by Medicaid (28.2 vs 6.2%, p<0.001) and of the lowest income quartile (33.6 vs 27.1%, p<0.001). After multivariable adjustment, OUD was associated with decreased mortality (1.5 vs 2.7%, p=0.001). Although these patients had similar rates of overall complications (36.1 vs 35.1%, p=0.363), they had increased thromboembolic (1.3 vs 0.8%, p<0.001) and infectious (4.1 vs 2.8%, p<0.001) events, as well as readmission at 30 days (19.0 vs 13.2%, p<0.001). While patients with OUD had similar hospitalisation costs ($50 766 vs $50 759, p=0.994), they did have longer hospitalisations (11.4 vs 10.3 days, p<0.001).

Conclusion The prevalence of OUD among cardiac surgical patients has steeply increased over the past decade. Although the presence of OUD was not associated with excess mortality at index hospitalisation, it was predictive of 30-day readmission. Increased attention to predischarge interventions and care coordination may improve outcomes in this population.