Development and validation of a difficulty score to predict intraoperative complications during laparoscopic liver resection

Abstract Background Previous studies have demonstrated that patient, surgical, tumour and operative variables affect the complexity of laparoscopic liver resections. However, current difficulty scoring systems address only tumour factors. The aim of this study was to develop and validate a predictive model for the risk of intraoperative complications during laparoscopic liver resections. Methods The prospectively maintained databases of seven European tertiary referral liver centres were compiled. Data from two-thirds of the patients were used for development and one-third for validation of the model. Intraoperative complications were based on a modified Satava classification. Using the methodology of the Framingham Heart Study, developed to identify risk factors that contribute to the development of cardiovascular disease, factors found to predict intraoperative complications independently were assigned points, and grouped into low-, moderate-, high- and extremely high-risk groups based on the likelihood of intraoperative complications. Results A total of 2856 patients were included. Neoadjuvant chemotherapy, lesion type and size, classification of resection and previous open liver resection were found to be independent predictors of intraoperative complications. Patients with intraoperative complications had a longer duration of hospital stay (5 versus 4 days; P < 0·001), higher complication rates (32·5 versus 15·5 per cent; P < 0·001), and higher 30-day (3·0 versus 0·3 per cent; P < 0·001) and 90-day (3·8 versus 0·8 per cent; P < 0·001) mortality rates than those who did not. The model was able to predict intraoperative complications (area under the receiver operating characteristic (ROC) curve (AUC) 0·677, 95 per cent c.i. 0·647 to 0·706) as well as postoperative 90-day mortality (AUC 0·769, 0·681 to 0·858). Conclusion This comprehensive scoring system, based on patient, surgical and tumour factors, and developed and validated using a large multicentre European database, helped estimate the risk of intraoperative complications.