Prediction of acute kidney injury after cardiac surgery: model development using a Chinese electronic health record dataset

Table 1 Multivariate logistic regression model for predicting acute kidney injury after cardiac surgery based on discovery set

Risk factor	β	OR (95% CI)	P value
Age, years	0.0170	1.017 (1.001–1.035)	0.048
Male	0.4076	1.503 (1.027–2.209)	0.036
Preoperative Scr, mg/dL	1.0381	2.824 (1.427–5.859)	0.004
Preoperative NLR	0.1333	1.143 (1.065–1.233)	< 0.001
Preoperative blood glucose, mmol/L	0.1441	1.155 (1.048–1.273)	0.003
Preoperative HDL, mmol/L	-0.6207	0.537 (0.299–0.947)	0.034
Intraoperative urine output, mL/kg/hr	-0.0602	0.942 (0.889–0.990)	0.027
CUF, mL/kg	0.0205	1.021 (1.009–1.033)	< 0.001
CVP, cmH₂O (trajectory group)
2 vs. 1	0.5632	1.756 (0.778–4.427)	0.199
3 vs. 1	1.4666	4.334 (1.838–11.314)	0.001
Perfusion flow, L/min/m² (trajectory group)
2 vs. 1	0.3693	1.447 (0.937–2.227)	0.094
3 vs. 1	0.4780	1.613 (1.091–2.389)	0.016
Intubated PaO₂/FiO₂ ratio	− 0.0035	0.996 (0.995–0.998)	< 0.001
Postoperative hemoglobin, g/L	− 0.0505	0.951 (0.938–0.963)	< 0.001
Postoperative serum potassium, mmol/L	0.6649	1.944 (1.448–2.623)	< 0.001
Postoperative LDH, U/L	0.0037	1.005 (1.003–1.006)	< 0.001
Intercept	− 2.8021

OR, odds ratio; CI, confidence interval; Scr, serum creatinine; NLR, neutrophil to lymphocyte ratio; HDL, high density lipoprotein; CUF, conventional ultrafiltration; CVP, central venous pressure; LDH, lactic dehydrogenase

ISSN: 1479-5876