Background: To establish a reliable definition of postoperative hepatic insufficiency (PHI) in noncirrhotic patients undergoing major hepatectomy. No standard definition of PHI has been established, but one is essential for meaningful comparison of outcomes data across studies.

Methods: Data from 1,059 noncirrhotic patients who underwent major hepatectomy (3 or more liver segments) at 3 centers from 1995 to 2005 were analyzed. Receiver operating characteristics (ROC) analysis of peak postoperative bilirubin ((Peak)Bil) and international normalized ratio ((Peak)INR) were used to define PHI.

Results: A total of 669 patients (63%) underwent resection of 3 to 4 liver segments; 390 (37%) underwent resection of 5 or more segments. Complications occurred in 453 (43%). The 90-day all-cause mortality rate was 4.7%, which is 47% higher than the 30-day rate (3.2%). Twenty (1.9%) patients died of causes unrelated to the liver. Of the remaining 1,039 patients, 30 (2.8%) died a median 36 days from liver-related causes (liver failure with or without multiorgan failure). ROC analysis revealed cut-offs that predict liver-related death are (Peak)Bil 7.0 mg/dL (area under the curve 0.982; sensitivity 93.3%; specificity 94.3%) and (Peak)INR 2.0 (area under the curve 0.846; sensitivity 76.7%; specificity 82.0%). (Peak)Bil > 7.0 mg/dL was the most powerful predictor of any (odds ratio [OR] = 83.3) or major complication (OR = 10.0), 90-day mortality (OR = 10.8), and 90-day liver-related mortality (OR = 250) (all p < 0.0001).

Conclusions: PHI defined as (Peak)Bil > 7.0 mg/dL accurately predicts liver-related death and worse outcomes after major hepatectomy. Standardized reporting of complications, PHI, and 90-day mortality is essential to accurately determine the risk of major hepatectomy and to compare outcomes data.