LAUSR

OBJECTIVE Whereas bleeding complications requiring a return to the operating room (OR) after carotid endarterectomy (CEA) are infrequent (1%), they are associated with an increased 30-day combined postoperative stroke or death rate. Drain placement after CEA varies among vascular surgeons, and there are limited data to support the practice. The goal of this study was to evaluate factors leading to drain placement and the effect of drains on postoperative outcomes including return to OR for bleeding, stroke, and death. METHODS There were 47,752 patients identified undergoing CEA using the Vascular Quality Initiative registry from 2011 to 2015. Demographic, preoperative, and intraoperative factors between patients who underwent CEA with (n = 19,425) and without (n = 28,327) drain placement were compared. End points included return to the OR for bleeding, stroke, death, postoperative wound infection, and hospital length of stay. We then compared postoperative outcomes between groups using mixed effect logistic regression models to control for correlation within center. Similar methods were used to show relationship between return to OR for bleeding and other variables. Subgroup analysis of patients with drain placement was compared among centers with high (>66.7% of cases), medium (33.3%-66.7%), and low (<33.3%) use. RESULTS Patients with drain placement were more likely to be taking a preoperative P2Y12 antagonist (P < .001), to have prior CEA or carotid artery stenting (P < .001), to use dextran (P < .001), and to have a concomitant procedure or coronary artery bypass graft (P < .001) and less likely to use protamine (P < .001) compared with those without drain placement. Drain placement did not prevent return to the OR for bleeding (P < .22). Re-exploration of the carotid artery after closure in the OR (P < .001), preoperative P2Y12 antagonist use (P < .001), and no protamine use (P < .001) were predictors for return to the OR for bleeding among those with drain placement. Of patients requiring return to the OR for bleeding, drain placement did not influence 30-day stroke (P = .82), 30-day mortality (P = .43), or 30-day combined stroke/mortality (P = .42) compared with those without drain placement. Drain placement did not influence postoperative wound infection (P < .3). Hospital length of stay was increased in patients with drain placement (P < .001). Return to the OR for bleeding (P = .24), wound infection (P = .16), and length of stay (P = .94) did not differ between the groups of high, medium, and low drain use. CONCLUSIONS Drain placement after CEA does not reduce return to the OR for bleeding, nor does it reduce perioperative stroke or death. Drain placement is associated with increased length of stay.