Thirty-day mortality leads to underestimation of postoperative death after liver resection: A novel method to define the acute postoperative period

BackgroundPostoperative mortality commonly is defined as death occurring within 30 days of surgery or during hospitalization. After resection for liver malignancies, this definition may result in underreporting, because mortality caused by postoperative complications can be delayed as the result of improved critical care. The aim of this study was to estimate statistically the acute postoperative period (APP) after partial hepatectomy and to compare mortality within this phase to standard timestamps.MethodsFrom a prospective database, 784 patients undergoing resection for primary and secondary hepatic malignancies between 2003 and 2013 were reviewed. For estimation of APP, a novel statistical method applying tests for a constant postoperative hazard was implemented. Multivariable mortality analysis was performed.ResultsThe APP was determined to last for 80 postoperative days (95% confidence interval 40–100 days). Within this period, 55 patients died (7.0%; 80-day mortality). In comparison, 30-day mortality (N = 32, 4.0%) and in-hospital death (N = 39, 5.0%) were relevantly less. No patient died between postoperative days 80 and 90. The causes of mortality within 30 days and from days 30–80 did not greatly differ, especially regarding posthepatectomy liver failure (44% vs 39%, P = .787). Septic complications, however, tended to cause late deaths more frequently (43% vs 25%, P = .255). Comorbidities (Charlson comorbidity index ≥3; P = .046), increased preoperative alanine aminotransferase activity (P = .030), and major liver resection (P = .035) were independent risk factors of 80-day mortality.ConclusionAfter liver resection for primary and secondary malignancies, 90-day rather than 30-day or in-hospital mortality should be used to avoid underreporting of deaths.