A NSQIP Analysis of MELD and Perioperative Outcomes in General Surgery

Abstract

It is well known that liver disease has an adverse effect on postoperative outcomes. However, what is still unknown is how to appropriately risk stratify this patient population based on the degree of liver failure. Because data are limited, specifically in general surgery practice, we analyzed the model of end-stage liver disease (MELD) in terms of predicting postoperative complications after one of three general surgery operations: inguinal hernia repair (IHR), umbilical hernia repair (UHR), and colon resection (CRXN). National Surgical Quality Improvement Program data on 17,812 total patients undergoing one of three general surgery operations from 2008 to 2012 were analyzed retrospectively. There were 7402 patients undergoing IHR; 5014 patients undergoing UHR; 5396 patients undergoing CRXN. MELD score was calculated using international normalized ratio, total bilirubin, and creatinine. The primary end point was any postoperative complication. The statistical method used was logistic regression. For IHR, UHR, and CRXN, the overall complication rates were 3.4, 6.4, and 45.9 per cent, respectively. The mean MELD scores were 8.6, 8.5, and 8.5, respectively. For every 1-point increase greater than the mean MELD score, there was a 7.8, 13.8, and 11.6 per cent increase in any postoperative complication. The overall 30-day mortality rate was 0.9 per cent. In conclusion, the MELD score continuum adequately predicts patients’ increased risk of postoperative complications after IHR, UHR, and CRXN. Therefore, MELD could be used for preoperative risk stratification and guide clinical decision making for general surgery in the cirrhotic patient.