Esophagectomies for malignancy among general and thoracic surgeons: A propensity score-matched NSQIP analysis stratified by surgical approach.

Abstract

166 Background: Training of general and thoracic surgeons continues to diverge, especially with the increasing role for minimally invasive surgical (MIS) approaches. Previous studies of esophagectomy outcomes by specialty do not adequately address malignancy or surgical approach. We sought to evaluate perioperative outcomes of esophagectomy for malignancy stratified by surgical specialty and approach using a national database. Methods: The National Surgical Quality Improvement Program (NSQIP) Targeted Esophagectomy Dataset was queried for esophagectomies for malignancy and grouped by surgeon specialty: thoracic surgery (TS) or general surgery (GS). Those with missing data were excluded (n = 6). To account for confounding due to specialty selection bias, we performed propensity score matching (PSM) by age, body mass index, ethnicity, American Society of Anesthesiologists class ³ 3, and surgical approach in a 1:1 ratio. An absolute standardized difference of ≤ 0.1 was considered an appropriate balance. The primary outcome was mortality and secondary outcomes were anastomotic leak, Clavien-Dindo grade ≥ 3 and positive margin rate. Univariate logistic regression analysis was performed for these outcomes on the matched cohort, with stratification by surgical approach (open vs. MIS). Results: A total of 1463 patients met inclusion criteria (512 GS, 951 TS). After PSM each group was comprised of 512 patients with similar demographics, neoadjuvant chemotherapy and radiation rates, and preoperative stage. The TS group consisted of 169 (33.0%) open and 343 (67.0%) MIS cases, while the GS group consisted of 177 (34.6%) open and 335 (65.4%) MIS cases. Postoperative complications, including surgical site infection, pneumonia, pulmonary embolism, stroke, and myocardial infarction were similar between matched groups, and remained similar when stratified by surgical approach. Mortality rates were similar between the TS and GS groups, both overall (14 (2.7%) vs. 10 (2.0%)) and when stratified by surgical approach (MIS: 11 (3.2%) vs. 10 (3.0%), open: 3 (1.8%) vs. 0 (0%)). By univariate analysis of the matched cohort stratified by surgical approach, TS patients had similar odds as GS patients of anastomotic leak (open: adjusted odds ratio (AOR) = 1.11, 95% confidence interval (95%CI) = 0.58 – 2.15, p = 0.75; MIS: AOR = 0.70, 95%CI = 0.47 – 1.04, p = 0.08), Clavien-Dindo grade ≥ 3 (open: AOR = 1.27, 95%CI = 0.79 – 2.06, p = 0.32; MIS: AOR = 1.01, 95%CI = 0.73 – 1.39, p = 0.97), positive surgical margins (open: AOR = 0.75, 95%CI = 0.33 – 1.68, p = 0.49; MIS: AOR = 0.62, 95%CI = 0.35 – 1.07, p = 0.09), and mortality (open: unable to be calculated due to 0 deaths in the GS group; MIS: AOR = 1.08, 95%CI = 0.45 – 2.62, p = 0.87). Conclusions: Esophagectomy for malignancy had a similar perioperative safety profile and positive margin rate among general and thoracic surgeons, regardless of surgical approach.