Cannabis use disorder and perioperative outcomes in vascular surgery

Heavy cannabis use is known to have an adverse impact on cardiovascular and cerebrovascular outcomes in the general population and in patients presenting for surgery. However, there have been no studies that have focused on patients undergoing vascular surgical procedures. The objective of this study was to determine the perioperative risk of cannabis use disorder (CUD), primarily cardiovascular risk, in perioperative vascular surgery patients.

Using the National Inpatient Sample from 2006 to 2015, we conducted a retrospective cohort study involving those undergoing one of six elective and emergent vascular surgical procedures (carotid endarterectomy [CEA], infrainguinal bypasses, open abdominal aortic aneurysm repair, aortobifemoral bypass, endovascular aortic aneurysm repair, or peripheral arterial endovascular procedures). Patients with CUD identified by the International Classification of Diseases, 9th edition, were matched with patients without CUD in a 1:1 ratio using propensity scores. The primary outcome was perioperative myocardial infarction (MI). Secondary outcomes include stroke, sepsis, deep vein thrombosis, pulmonary embolus, acute kidney injury requiring dialysis, respiratory failure, in-hospital mortality, total cost, and length of stay.

We identified a total cohort of 510,007 patients. Over the study period, the recorded prevalence of CUD increased from 1.3/1000 to 10.3/1000 admissions (P < .001). After propensity score matching the cohort consisted of 4684 patients. Those with CUD had a higher incidence of perioperative MI (3.3% vs 2.1%; odds ratio [OR], 1.56; 95% confidence interval [CI], 1.09-2.24; P = .016) and perioperative stroke (5.5% vs 3.5%; OR, 1.59; 95% CI, 1.20-2.12; P = .0013) than patients without CUD. In a sensitivity analysis, where the risk was evaluated separately by type of procedure, the higher incidence of perioperative stroke was primarily seen among those undergoing CEA. Patients with CUD had a lower incidence of sepsis (3.3% vs 5.1%; OR, 0.64; 95% CI, 0.47-0.85; P = .0024). We obtained similar results in a sensitivity analysis that included all patients in the complete unmatched cohort and adjusted for confounding using logistic regression models accounting for the survey design, although the findings of sepsis and stroke failed to reach statistical significance after correcting for multiple testing (MI P = .001; stroke P = .031; sepsis P = .009).

CUD was associated with a significantly higher incidence of perioperative MI in vascular surgery patients. Those with CUD had a greater incidence of diagnosis of acute perioperative stroke when undergoing CEA. Owing to limitations in administrative data, it is unclear if this represents a true effect or selection bias. These findings warrant further investigation in a prospective cohort.