Postoperative liver dysfunction and future remnant liver: where is the limit? Results of a prospective study.

The future remnant liver (FRL) limit for safe major hepatectomy with low risk of postoperative liver failure has not yet been well defined. Between April 2000 and September 2004, every patient scheduled for major hepatectomy in our institution underwent CT-volumetry of FRL. Patients with FRL <25% underwent portal vein embolization (PVE). Exclusion criteria were PVE, associated vascular resection and liver cirrhosis. The FRL was correlated with short-term results in patients with normal liver (group A) and those with impaired liver function secondary to neoadjuvant chemotherapy or cholestasis (bilirubin >2 mg/100 ml) (group B). Liver dysfunction was defined as both PT <50% and serum bilirubin level >5 mg/100 ml for three or more consecutive days. A total of 119 patients were analyzed, 72 in group A and 47 in group B. The FRL value was the only significant risk factor for postoperative liver dysfunction in the univariate and multivariate analysis (p = 0.009). The FRL did not correlate with postoperative mortality and morbidity. Bilirubin and prothrombin time (PT) on days 3 and 7 were significantly correlated to FRL in both groups. In group A, patients with postoperative liver dysfunction had a FRL<30% (3 versus 0; p = 0.005). According to receiving operator characteristic (ROC) curve analysis, a FRL value of 26.5% predicted postoperative liver dysfunction with 66.7% sensitivity, 97.1% specificity, 50% positive predictive value (PPV), and 98.5% negative predictive value (NPV). In group B, patients with postoperative liver dysfunction had a FRL <35% (4 versus 0; p = 0.027). According to ROC curve analysis, a FRL value of 31.05% predicted postoperative liver dysfunction with 75% sensitivity, 79.1% specificity, 25% PPV, and 97.1% NPV. Hepatectomy can be considered safe when FRL is >26.5% in patients with healthy liver and >31% in patients with impaired liver function.