Can Anticoagulation be Reduced After Mechanical Valve Replacement? Interim Results From the Proact Study

Abstract

cardiac surgery. Methods: The study included 28,422 cardiac surgery patients who had no pre-operative renal dialysis between June 2001 and June 2009 in 18 hospitals, from the ANZSCTS database. Logistic regression analyses were undertaken to identify the best combination of risk factors for predicting AKI. Two models were developed, one including pre-operative risk factors and another including pre-, periand early post-operative risk factors. The area under the receiver operating characteristic curve, AUC, was calculated, using splitsample internal validation, to assess model discrimination. Results: The incidence of acute kidney injury, AKI, was 5.8% (1642 patients). Mortality for patients who experienced AKI was 17.4% vs. 1.6% for patients who did not have AKI. Upon validation, the AUC for the pre-operative model was 0.77, and the Hosmer–Lemeshow, H–L, goodness-of-fit pvalue was 0.06. For the preand post-operative model the AUC = 0.81 and H–L p-value = 0.6. Both models had good discrimination and satisfactory calibration. Conclusion: AKI following cardiac surgery can be predicted using only pre-operative risk factors or, more accurately, using a combination of pre-, peri-, and early postoperative risk factors. The ability to identify high-risk individuals can be useful in pre-operative patient management/ selection for surgery, and for recruitment of appropriate patients to clinical trials. Prediction in the early stages of post-operative care can be used to guide intensive care and also as a retrospective performance audit tool.