LAUSR

comparison to CEA. Patients enrolled in this project were compared with those undergoing CEA during the same period (2016-2017). The primary outcome was a composite of in-hospital stroke and death. Average treatment effects were estimated by augmented inverse-probability weighting. Additional analysis was performed using multivariable logistic regression as well as various matching techniques, such as propensity score matching and coarsened exact matching. Adjusted analysis accounted for age, sex, race, insurance status, coronary artery disease, congestive heart failure, chronic kidney disease, chronic obstructive pulmonary disease, symptomatic status, restenosis, prior vascular procedures, degree of ipsilateral stenosis, and preoperative medication use. Results: A total of 637 patients underwent TCAR compared with 12,049 patients who underwent CEA. Patients undergoing TCAR were older, more likely to be symptomatic, and had more medical comorbidities, such as coronary artery disease, congestive heart failure, chronic obstructive pulmonary disease, chronic kidney disease, and prior vascular procedures, compared with CEA patients (Table). The majority of TCAR procedures were done under local/regional anesthesia (95.3% vs 9.7% in CEA; P < .001). On average, TCAR was 36.7 minutes shorter than CEA (78.0 6 33.9 vs 114.7 6 42.5 minutes; P < .001). On univariate analysis, there were no differences in the rates of in-hospital stroke/death (1.3% vs 1.7%; P 1⁄4 .42), overall neurologic events (2.0% vs 1.9%; P 1⁄4 .83), in-hospital myocardial infarction (0.7% vs 1.1%; P 1⁄4 .31), and 30-day mortality (0.5% vs 0.9%; P 1⁄4 .08) between CEA and TCAR, respectively. Patients undergoing CEA had higher rates of cranial nerve injury (2.8% vs 0.8%; P < .01) and postoperative hypertension (18.3% vs 11.6%; P < .001) compared with TCAR patients. On multivariable analysis and using different matching methods, there were no differences in overall stroke, stroke/death, or overall neurologic events (Fig). The absolute difference in adjusted stroke/death rates between the two groups was 0.3% (95% confidence interval, 1.7% to 1.0%; P 1⁄4 .64). Conclusions: Despite a substantially higher medical risk in patients undergoing TCAR, analysis of the preliminary results from the SVS Vascular Quality Initiative TCAR Surveillance Project showed similar in-hospital stroke/death rates between TCAR and CEA after multivariable adjustment and rigorous matching. Further studies with larger sample sizes and longer follow-up will be needed to establish the equivalence of TCAR compared with CEA.