Revised cardiac risk index in predicting cardiovascular complications in patients receiving chronic kidney replacement therapy undergoing elective general surgery

Abstract

Abstract Introduction The Revised Cardiac Risk Index (RCRI) is a six-parameter model that is commonly used in assessing individual 30-day perioperative cardiovascular risk before general surgery, but its use in patients on chronic kidney replacement therapy (KRT) is unvalidated. This study aimed to externally validate RCRI in this patient group over a 15-year period. Methods Data linkage was used between the the Australia and New Zealand Dialysis and Transplant (ANZDATA) Registry and jurisdictional hospital admisisons data across Australia and New Zealand to identify all incident and prevalent patients on chronic KRT between 2000 and 2015 who underwent elective abdominal surgery. Chronic KRT was categorised as haemodialysis (HD), peritoneal dialysis (PD), home haemodialysis (HHD) and kidney transplant. The outcome of interest was major adverse cardiovascular event (MACE) which was defined as nonfatal myocardial infarction, nonfatal stroke, non-fatal cardiac arrest and cardiovascular mortality at 30 days. Logistic regression was used with the RCRI score included as a continuous variable to estimate discrimination by area under the receiver operating curve (AUROC). Calibration was evaluated using a calibration plot. Clinical utility was assessed using a decision curve analysis to determine the net benefit. Results A total of 5094 elective surgeries were undertaken, and MACE occurred in 153 individuals (3.0%). Overall, RCRI had poor discrimination in patients on chronic KRT undergoing elective surgery (AUROC 0.67), particularly in patients aged greater than 65 years (AUROC 0.591). A calibration plot showed that RCRI overestimated risk of MACE. The expected-to-observed outcome ratio was 6.0, 5.1 and 2.5 for those with RCRI scores of 1, 2 and ≥ 3, respectively. Discrimination was moderate in patients under 65 years and in kidney transplant recipients, with AUROC values of 0.740 and 0.718, respectively. Overestimation was common but less so for kidney transplant recipients. Decision curve analysis showed that there was no net benefit of using the tool in neither the overall cohort nor patients under 65 years, but a slight benefit associated with threshold probability > 5.5% in kidney transplant recipients. Conclusions The RCRI tool performed poorly and overestimated risk in patients on chronic dialysis, potentially misinforming patients and clinicians about the risk of elective surgery. Further research is needed to define a more comprehensive means of estimating risk in this unique population.