The effect of surgeon's specialty and volume on the perioperative outcome of carotid endarterectomy.

BACKGROUND Several studies have demonstrated better outcomes for carotid endarterectomy (CEA) at high-volume hospitals and providers. However, only a few studies have reported on the impact of surgeons' specialty and volume on the perioperative outcome of CEA. METHODS This is a retrospective analysis of CEA during a recent 2-year period. Surgeons' specialties were classified according to their Board specialties into general surgeons (GS), cardiothoracic surgeons (CT), and vascular surgeons (VS). Surgeons' annual volume was categorized into low volume (<10 CEAs), medium volume (10 to <30 CEAs), and high volume (≥30 CEAs). The primary outcome was 30-day perioperative stroke and/or death; however, other perioperative complications were analyzed. Both univariate and multivariate analyses were done to predict the effect of specialty/volume and any other patient risk factors on stroke outcome. RESULTS Nine hundred and fifty-three CEAs were performed by 24 surgeons: 122 by seven GS, 383 by 13 CT, and 448 by 4 VS. Patients' demographics/clinical characteristics were similar between specialties, except the incidence of coronary artery disease, which was higher for CT (P < .0001). The indications for CEA were symptomatic disease in 38% for VS, 31% for GS, and 23% for CT (P < .0001). The perioperative stroke and death rates were 4.1%, 2.9%, and 1.3% for GS, CT, and VS, respectively (P = .126). A subgroup analysis showed that the perioperative stroke rates for symptomatic patients were 5.3%, 2.3%, and 2.3% (P = .511) and for asymptomatic patients were 3.6%, 3%, and 0.72% (P = .099) for GS, CT, and VS, respectively. Perioperative stroke rates were significantly higher for nonvascular surgeons (GS and CT combined) vs VS in asymptomatic patients (3.2% vs 0.72%; P = .033). Perioperative stroke/death was also significantly lower for high-volume surgeons: 1.3% vs 4.1% and 4.3% for medium- and low-volume surgeons (P = .019) (1.3% vs 4.15% for high vs low/medium combined; P = .005). More CEAs were done for asymptomatic patients in the low/medium-volume surgeons (78%) vs high-volume surgeons (64%; P < .0001) with a stroke rate of 4.6% for low/medium-volume surgeons vs 0.51% for high-volume surgeons (P = .0005). A univariate logistic analysis showed that the odds ratio of having a perioperative stroke was 0.3 (95% confidence interval [CI], 0.13-0.73; P =.008) for high-volume surgeons vs low/medium-volume surgeons, 0.4 (95% CI, 0.16-1.07; P = .069) for VS vs CT/GS and 0.2 (95% CI, 0.06-0.45; P = .0004) when patching was used. A multivariate analysis showed that the odds ratio of having a perioperative stroke for CT VS was 2.1 (95% CI, 0.71-5.92; P = .183); for GS vs VS, 1.8 (95% CI, 0.49-6.90; P = .3709); for low-volume surgeons (vs high-volume) 3.4 (95% CI, 0.96-11.77; P = .0581); medium- vs high-volume surgeons 2.2 (95% CI, 0.75-6.42; P = .1509). CONCLUSIONS High-volume surgeons had significantly better perioperative stroke/death rates for CEA than low/medium-volume surgeons. Perioperative stroke/death rates were also higher for nonvascular surgeons in asymptomatic patients.