Pancreatoduodenectomy in the cirrhotic patient: predictors of outcome

Abstract

Listen

Translate article

Background: Hepatic cirrhosis is a formidable risk for mortality in gastrointestinal surgery. Though cirrhosis has been considered to be a contraindication to pancreatoduodenectomy (PD), few data are available in this patient population. We analyzed a large population of cirrhotic patients undergoing PD, seeking to identify preoperative predictors of morbidity and mortality. Methods: Patients undergoing PD with biopsy-proved cirrhosis were evaluated. Primary endpoints were mortality and morbidity. Child score, MELD score, and radiographic evidence of portal hypertension were assessed for accuracy in preoperative risk stratification. Results: Between 2005 and 2013, 36 cirrhotic patients underwent PD; follow-up was complete. Median survival was 37.3 months (range 0.2–116). Overall perioperative (90-day) mortality was 13.9%. Median Child score was 6 (range 5–10); median MELD score was 9 (range 7–18). Age at time of surgery, smoking, operative duration, perioperative transfusion requirement did not predict mortality. MELD ≥ 10 was associated with increased postoperative mortality (23% vs 8.7%, p = 0.004). Radiographic evidence of portal hypertension (pHTN) was observed in 16/29 (55%). Irrespective of Child or MELD score, those with pHTN had poor outcomes including significantly greater intraoperative blood loss, increased transfusion requirement, increased incidence of major complication, and length of stay. Postoperative mortality was significantly higher with pHTN (19% vs 7.7%, p = 0.01). Conclusion: Pancreatoduodenectomy may be considered in carefully selected cirrhotic patients. MELD ≥ 10 predicts postoperative mortality. Specific attention should be afforded to patients with preoperative radiographic evidence of portal hypertension as this group experiences poor outcomes irrespective of MELD or Child score.