Predicting the difficult laparoscopic cholecystectomy: development and validation of a pre-operative risk score using an objective operative difficulty grading system

Abstract

Background: The prediction of a difficult cholecystectomy has traditionally been based on certain pre-operative clinical and imaging factors. Most of the previous literature reported small patient cohorts and have not used an objective measure of operative difficulty. The aim of this study was to develop a pre-operative score to predict difficult cholecystectomy, as defined by a validated intra-operative difficulty grading scale. Method: Two cohorts from prospectively maintained databases of patients who underwent laparoscopic cholecystectomy were analysed: the CholeS Study (8755 patients) and a single surgeon series (4089 patients). Factors potentially predictive of difficulty were correlated to the Nassar intra-operative difficulty scale. A multivariable binary logistic regression analysis was then used to identify factors that were independently associated with difficult laparoscopic cholecystectomy, defined as operative difficulty grades 3 to 5. The resulting model was then converted to a risk score, and validated on both internal and external datasets. Result: Increasing age and ASA classification, male gender, diagnosis of CBD stone or cholecystitis, thick-walled gallbladders, CBD dilation, use of pre-operative ERCP and non-elective operations were found to be significant independent predictors of difficult cases. A risk score based on these factors returned an area under the ROC curve of 0.789 (95% CI 0.773–0.806, p ' 0.001) on external validation, with 11.0% versus 80.0% of patients classified as low versus high risk having difficult surgeries. Conclusion: We have developed and validated a pre-operative scoring system that uses easily available pre-operative variables to predict difficult laparoscopic cholecystectomies. This scoring system should assist in patient selection for day case surgery, optimising pre-operative surgical planning (e.g. allocation of the procedure to a suitably trained surgeon) and counselling patients during the consent process. The score could also be used to risk adjust outcomes in future research.