Perioperative mortality after non-hepatic general surgery in patients with liver cirrhosis: an analysis of 138 operations in the 2000s using Child and MELD scores.

INTRODUCTION: Despite of advances in modern surgical and intensive care treatment, perioperative mortality remains high in patients with liver cirrhosis undergoing nonhepatic general surgery. In the few existing articles, mortality was reported to be as high as 70% in patients with poor liver function (high Child or model for end-stage liver disease (MELD) score). Since data are limited, we analyzed our recent experience with cirrhotic patients undergoing emergent or elective nonhepatic general surgery at a German university hospital.
METHODS: Since 2000, 138 nonhepatic general surgical procedures (99 intra-abdominal, 39 abdominal wall) were performed in patients with liver cirrhosis. Liver cirrhosis was preoperatively classified according to the Child (41 Child A, 59 B, 38 C) and the MELD score (MELD median 13). Sixty-eight (49%) of the patients underwent emergent operations. Most abdominal wall operations were for hernias. Intra-abdominal operations consisted of GI tract procedures (n=53), cholecystectomies (n=15), and various others (n=31). Perioperative data were gained by retrospective analysis.
RESULTS: Overall perioperative mortality in all 138 cases was 28% (9% in elective surgery, 47% in emergent surgery; p<0.001). Perioperative mortality was higher after intra-abdominal than after abdominal wall operations (35% vs. 8%; p=0.001) or in patients requiring transfusions (43% vs. 5% without transfusions; p<0.001). Perioperative mortality increased with the Child score (10% Child A, 17% Child B, 63% Child C; p<0.01) and the MELD score (9% MELD <10, 19% MELD 10–15, 54% MELD >15; p<0.001). Univariately, further factors like American Society of Anesthesiologists (ASA) score and various preoperative laboratory values were also associated with perioperative mortality. By multivariate analysis of all 138 operations, the Child and ASA classifications, intraoperative transfusions, and a preoperative sodium <130 mmol/l, but not the MELD score, were independent prognostic factors. Analysis of elective operations revealed only a preoperatively increased creatinine as risk factor for perioperative mortality. In emergent operations again, Child class, blood transfusions, and low sodium level, but not the MELD score, predicted postoperative mortality.
CONCLUSIONS: Our results demonstrate that perioperative mortality remains high in patients with liver cirrhosisundergoing general surgery, especially in emergent situations. Patients with poor liver function and/or need for blood transfusions even had a very high mortality. In our experience, the Child score (together with other variables) independently correlates with perioperative mortality in emergent operations whereas the MELD score was inferior in predicting the outcome.