Risk factors of surgical mortality in patients with Clostridium difficile colitis. A novel scoring system

Background The purpose of the study is to identify the risk factors of mortality and develop a risk scoring system in patients who underwent colectomy due to Clostridium difficile colitis (CD-C). Methods Patient information was extracted using the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) data from 2012 to 2016. All adult patients who underwent colectomy for CD-C were included in the study. The data were split into training and testing data sets. A multiple logistic regression model was developed by backward deletion methods for risk assessment. To test the performance of the prediction model for 30-day mortality, a receiver operating characteristic (ROC) curve was generated and an area under the curve (AUC) was created. Results The training data set consisted of 434 (80%) patients, and the testing data set consisted of 91 (20%) patients. The overall mortality was 35%. No significant differences were found between the training and testing data sets for patient characteristics, comorbidities and mortality. The final model of the logistic regression model revealed a highly significant 30-day mortality for an age of >= 75 years old, ventilator dependency, Septic shock prior to surgery and a history of steroid use. The AUC value was 0.745 (95% CI 0.660-0.826). The risk of mortality scores range from 0 to 37. The highest score of 37 was related to an 83.9% predicted mortality. Conclusion Older age, septic shock, ventilator dependency requiring supportive care and a history of chronic steroid use were highly associated with mortality. A nomogram showing the scores and their relationship to mortality may provide guidance to point of care physicians for deciding the goal of care.

Automated summary

The study aimed to identify risk factors and develop a risk scoring system for mortality in patients undergoing colectomy for Clostridium difficile colitis. A logistic regression model was developed, revealing older age, ventilator dependency, septic shock, and history of steroid use as significant factors for 30-day mortality. The risk scoring system showed a strong correlation between scores and predicted mortality rates.