RESUMO
PURPOSE: We aimed to analyze the predictive value of hyperamylasemia after pancreatectomy for morbidity and for the decision to perform rescue completion pancreatectomy (CP) in a retrospective cohort study. METHODS: Data were extracted from a retrospective clinical database. Postoperative hyperamylasemia (POH) and postoperative hyperlipasemia (POHL) were defined by values greater than those accepted as the upper limit at our institution on postoperative day 1 (POD1). The endpoints of the study were the association of POH with postoperative morbidity and the possible predictors for postpancreatectomy acute pancreatitis (PPAP) and severe complications such as the necessity for rescue CP. RESULTS: We analyzed 437 patients who underwent pancreaticoduodenectomy over a period of 7 years. Among them, 219 (52.3%) patients had POH and 200 (47.7%) had normal postoperative amylase (non-POH) levels. A soft pancreatic texture (odds ratio [OR] 3.86) and POH on POD1 (OR 8.2) were independent predictors of postoperative pancreatic fistula (POPF), and POH on POD1 (OR 6.38) was an independent predictor of rescue CP. The clinically relevant POPF (49.5% vs. 11.4%, p < 0.001), intraabdominal abscess (38.3% vs. 15.3%, p < 0.001), postoperative hemorrhage (22.8% vs. 5.1%, p < 0.001), major complications (Clavien-Dindo classification > 2) (52.5% vs. 25.6%, p < 0.001), and CP (13% vs. 1.8%, p < 0.001) occurred significantly more often in the POH group than in the non-POH group. CONCLUSION: Although POH on POD1 occurs frequently, in addition to other risk factors, it has a predictive value for the development of postoperative morbidity associated with PPAP and CP.