Abstract
Introduction: Partial nephrectomy (PN) is associated with a non-negligible risk of postoperative cardiovascular morbidity and mortality. Identification of high-risk patients may enable optimization of perioperative management and consideration of alternative approaches. The authors aim to develop a procedure-specific cardiovascular risk index for PN patients and compare its performance to the widely used revised cardiac risk index (RCRI) and AUB-HAS2 cardiovascular risk index.
Methods: The cohort was derived from the American College of Surgeons - National Surgical Quality Improvement Program (ACS-NSQIP) database. The primary outcome was the incidence of major adverse cardiovascular events (MACE), defined as 30-day postoperative incidence of myocardial infarction, stroke, or mortality. A multivariate logistic regression model was constructed; performance and calibration were evaluated using an ROC analysis and the Hosmer-Lemeshow test and compared to the RCRI and the AUB-HAS2 index.
Results: In a cohort of 4795 patients, MACE occurred in 52 (1.1%) patients. A univariate analysis yielded 13 eligible variables for entry into the multivariate model. The final PN-A 4 CH model utilized six variables: Age >= 75years, ASA class >2, Anemia, surgical Approach, Creatinine >1.5, and history of Heart disease. Index ROC analysis provided a C-statistic of 0.81, calibration R-2 was 0.99, and sensitivity was 85%. In comparison, the RCRI and AUB-HAS2 C-statistics were 0.59 and 0.68, respectively.
Conclusion: This study proposes a novel procedure-specific cardiovascular risk index. The PN-A(4)CH index demonstrated good predictive ability and excellent calibration using a large national database and may enable further individualization of patient care and optimization of patient selection.