Development and Validation of a Risk Prediction Model for Acute Kidney Injury After the First Course of Cisplatin

Abstract

Purpose Cisplatin-associated acute kidney injury (C-AKI) is common. We sought to develop and validate a predictive model for C-AKI after the first course of cisplatin. Methods Clinical and demographic data were collected on patients who received cisplatin between 2000 and 2016 at two cancer centers. C-AKI was defined as a 0.3 mg/dL rise in serum creatinine within 14 days of receiving cisplatin. Using multivariable logistic regression models with C-AKI as the primary outcome, we created a scoring model from the development cohort (DC) and tested it in the validation cohort (VC). Results C-AKI occurred in 13.6% of 2,118 patients in the DC and in 11.6% of 2,363 patients in the VC. Factors significantly associated with C-AKI included age 61 to 70 years (odds ratio [OR], 1.64 [95% CI, 1.21 to 2.23]; P = .001) and 71 to 90 years (OR, 2.97 [95% CI, 2.06 to 4.28]; P < .001) compared with ≤ 60 years; cisplatin dose 101 to 150 mg (OR, 1.58 [95% CI, 1.14 to 2.19]; P = .007) and > 150 mg (OR, 3.73 [95% CI, 2.68 to 5.20]; P < .001) compared with ≤ 100 mg; a history of hypertension (OR, 2.10 [95% CI, 1.54 to 2.72]; P < .001) compared with no hypertension; and serum albumin 2.0 to 3.5 g/dL (OR, 2.21 [95% CI, 1.62 to 3.03]; P < .001) compared with > 3.5 g/dL. The baseline estimated glomerular filtration rate was not significantly associated with the risk of C-AKI. The c-statistics of the score-based model in the DC and the VC were 0.72 (95% CI, 0.69 to 0.75) and 0.70 (95% CI, 0.67 to 0.73), respectively. Scores of 0, 3.5, and 8.5 were associated with a probability of C-AKI of 0.03 (95% CI, 0.03 to 0.05), 0.12 (95% CI, 0.11 to 0.14), and 0.51 (95% CI, 0.43 to 0.60), respectively. Conclusion A score-based model created by using the patient’s age, cisplatin dose, hypertension, and serum albumin is predictive of C-AKI.