C. F. Janowak2, L. Taylor2, J. Blasberg1, J. Maloney1, R. Macke1 1University Of Wisconsin,Division Of Cardiothoracic Surgery,Madison, WI, USA 2University Of Wisconsin,Department Of Surgery,Madison, WI, USA
Introduction: Most postoperative assessments and triage decisions are based on subjective evaluation of a patient’s risk factors and overall condition. The Surgical Apgar Score (SAS) is a validated prognostic tool used to predict postoperative morbidity and mortality in a wide variety of surgical patients. The esophagectomy population is a unique subset of surgical patients who are high risk for post-operative complication and disposition resources. An objective prognostic metric is an appealing and efficient way to allocate limited care resources to the sickest of postoperative patients. Although other more complex risk calculators have been developed, the SAS is a simple, bedside usable, model that has been validated in a variety of surgical populations. We evaluated the reliability of the SAS in a major esophageal surgery population.
Methods: A retrospective review of a prospectively collected and internally validated database of cardiothoracic operations was performed for consecutive esophagectomies from 2009 to 2013. Basic demographics, comorbidities, post-operative complications, and intraoperative variables were collected for all patients. The primary outcomes studied were mortality and NSQIP-defined in-hospital major complication; secondary outcomes were prolonged length of hospital stay (LOS) greater than 10 days and post-operative disposition. We used descriptive statistics, receiver operating characteristics (ROC) and Pearson Chi-Square analysis to analyze primary and secondary outcome prediction efficacy of SAS. Preoperative comorbid conditions were also analyzed for association with post-operative outcomes prognostication using odds ratio (OR) analysis.
Results: A total of 172 consecutive esophageal resections over four years were reviewed. Overall mortality was 5 deaths (2.9%) with 4 occurring within 30 days of surgery, 1 after discharge within 30 days, and 1 after 90 days of hospitalization. Overall SAS 9-10, n=16; SAS 7-8, n=113; SAS 5-6, n= 42; and SAS ≤ 4, n=1. Of these, 34.3% had a major complication, 27.3% had a prolonged LOS, and 12.2% were discharged to a care facility other than home. No significant correlation was demonstrated between complication, LOS, or discharge disposition and the SAS with respective ROC of 0.44, 0.43, and 0.44. Of the preoperative comorbid conditions analyzed, only neoadjuvant chemoradiation significantly increased the risk of any outcome, with an OR of 3.59 (95% CI 1.38-9.37, p < 0.01) risk of discharge to care other than home.
Conclusion: The perioperative performance measure of the SAS does not appear to have a good ability to predict major post-operative adverse outcomes in a major esophageal surgery population.
