A novel nephrectomy-specific respiratory failure index using the ACS-NSQIP dataset

Abstract

Purpose: Post-operative pulmonary failure is a major complication of nephrectomy that may lead to severe morbidity and mortality. Hence, we aimed to derive a nephrectomy-specific post-operative respiratory failure index. Methods: Our cohort was derived from The American College of Surgeons—National Surgical Quality Improvement Program database between 2005 and 2019. The outcome of interest was post-operative respiratory failure (PRF) defined as any incidence of unplanned intubation post-operatively or requiring mechanical ventilation post-operatively for a period > 48 h. A multivariable logistic regression model was constructed, and model calibration and performance were assessed using a ROC analysis and the Hosmer–Lemeshow test. Finally, we derived the nephrectomy-specific respiratory failure (NSRF) index and compared it to Gupta’s index. Results: Seventy-nine thousand five hundred and twenty-three patients underwent nephrectomy between the years 2005 and 2019 of which nine hundred and sixty-two patients developed PRF. The final NSRF model encompassed ten variables: age, smoking status, American society of anesthesiology class, abnormal creatinine (≥ 1.5 mg/dL), anemia (< 36%), functional health status, chronic obstructive pulmonary disease, surgical approach, emergency case, and obesity (≥ 40 kg/m2). The NSRF ROC analysis provided C-statistic = 0.78, calibration R2 = 0.99, and proper goodness of fit. In comparison, the C-statistics of Gupta’s index was found to be 0.71 (p value < 0.001). Conclusion: The NSRF is a procedure tailored index for predicting post-operative respiratory failure. It is a valuable tool in the pre-operative evaluation setting that can help identify high-risk patients who will require additional respiratory evaluation and preparation for their surgery.