Development of a prognostic prediction model for patients with severe acute kidney injury based on quantitative urinary protein， kidney injury molecule-1， serum creatinine， and albumin

ObjectiveTo construct a Nomogram prediction model for the prognosis of patients with severe acute kidney injury （AKI） based on quantitative urinary protein， kidney injury molecule-1 （KIM-1）， serum creatinine （Scr）， and albumin expression， so as to provide evidence for improving the prognosis of patients with severe AKI.MethodsThis retrospective study included 230 patients with severe AKI admitted to the Department of Critical Care Medicine， the No.4 People's Hospital of Hengshui from May 10， 2021 to December 10， 2023 as the model development cohort. Prognostic outcomes were recorded， and the patients were divided into a good prognosis group （143 cases） and a poor prognosis group （87 cases）. Clinical data were compared between the two groups. Based on quantitative urinary protein， KIM-1， Scr， and albumin expression， factors influencing mortality in patients with severe AKI were analyzed， and a Nomogram prediction model for prognosis was constructed. In addition， 98 patients with severe AKI during the same period were selected as an external validation cohort to evaluate the performance of the model.ResultsIn the poor prognosis group， age， Acute Physiology and Chronic Health Evaluation Ⅱ （APACHE Ⅱ） score， comorbid cardiovascular disease， comorbid multiple organ dysfunction syndrome （MODS）， mechanical ventilation， quantitative urinary protein， KIM-1， and Scr were （66.53±5.84） years， （18.73±5.88） points， 41.38% （36/87）， 33.33% （29/87）， 49.43% （43/87）， （2.16±0.68） g/24 h， （3 610.37±741.18） ng/L， and （242.67±32.98） μmol/L， respectively， all of which were higher than those in the good prognosis group： （64.79±6.27） years， （13.65±3.90） points， 28.67% （41/143）， 14.69% （21/143）， 26.57% （38/143）， （1.96±0.61） g/24 h， （1 120.15±289.52） ng/L， and （198.23±26.25） μmol/L. Albumin was （32.21±6.08） g/L， which was lower than that in the good prognosis group （［34.74±7.39］ g/L）. The differences were statistically significant （P<0.05）. The area under the curve for predicting prognosis in patients with severe AKI was 0.745 for APACHE Ⅱ score， 0.760 for comorbid MODS， 0.620 for mechanical ventilation， 0.769 for quantitative urinary protein， 0.788 for KIM-1， 0.765 for Scr， and 0.718 for albumin. Multivariate Logistic regression analysis showed that an APACHE Ⅱ score ≥16.70 points， comorbid MODS， mechanical ventilation， quantitative urinary protein ≥1.76 g/24 h， KIM-1 ≥3152.26 ng/L， and Scr≥223.64 μmol/L were independent risk factors for poor prognosis in patients with severe AKI， whereas albumin≥33.48 g/L was an independent protective factor （P<0.05）. A Nomogram model was constructed using APACHE Ⅱ score， comorbid MODS， mechanical ventilation， quantitative urinary protein， KIM-1， Scr， and albumin as predictive factors. The calibration curve of this Nomogram model for predicting poor prognosis in patients with severe AKI was close to the ideal curve， and decision curve analysis showed significant clinical net benefit when the risk threshold was 0.2-0.8. In external validation， the calibration curve and decision curve indicated that the model had high predictive performance for poor prognosis in patients with severe AKI and could provide significant clinical net benefit.ConclusionThe Nomogram prediction model for the prognosis of patients with severe AKI， constructed based on APACHE Ⅱscore， comorbid MODS， mechanical ventilation， quantitative urinary protein， KIM-1， Scr， and albumin， is of important value for prognostic prediction in patients with severe AKI and can effectively predict poor outcomes.