Predictivity of acute kidney injury risk scores for late kidney injury in patients with chronic coronary syndrome

Late rather than acute kidney injury after percutaneous coronary intervention (PCI) has been recently recognized as a predictor of future adverse events in patient with coronary artery disease. The risk-predicting models for acute kidney injury reported by Mehran et al., Bartholomew et al., and Tsai et al. were derived from a large cohort and externally validated, although the applicability of these models for predicting late kidney injury is unknown. A total of 327 patients undergoing elective PCI procedures were included. We calculated the three scores and tested their diagnostic ability for predicting late kidney injury (> 6 months after PCI), defined as an increase in creatinine levels >= 0.3 mg/dl or >= 50% from baseline. During the median follow-up period of 28 months, 27 (8.3%) patients had late kidney injury. All three scores significantly predicted late kidney injury, among which the score by Tsai et al. had a better diagnostic ability (area under the curve 0.83, best cut-off value 14, p < 0.001). With the best cut-off value, patients with Tsai score >= 14 had a significantly higher rate of late kidney injury than their counterpart (27.4% vs. 2.8%, p < 0.001). In conclusion, established risk scores for acute kidney injury may be useful for predicting late kidney injury after PCI in patients with chronic coronary syndrome.