Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 Jan 2;16(1):68.
doi: 10.1038/s41467-024-55629-5.

Development and validation of a real-time prediction model for acute kidney injury in hospitalized patients

Yuhui Zhang #  1   2   3 , Damin Xu #  1   2   3 , Jianwei Gao  4 , Ruiguo Wang  4 , Kun Yan  5 , Hong Liang  6 , Juan Xu  4 , Youlu Zhao  1   2   3 , Xizi Zheng  1   2   3 , Lingyi Xu  1   2   3 , Jinwei Wang  1   2   3 , Fude Zhou  1   2   3 , Guopeng Zhou  7 , Qingqing Zhou  1   2   3 , Zhao Yang  8 , Xiaoli Chen  9 , Yulan Shen  10 , Tianrong Ji  11   12 , Yunlin Feng  13   14 , Ping Wang  15   16 , Jundong Jiao  17   18 , Li Wang  19   20 , Jicheng Lv  21   22   23 , Li Yang  24   25   26
Affiliations

Development and validation of a real-time prediction model for acute kidney injury in hospitalized patients

Yuhui Zhang et al. Nat Commun. .

Abstract

Early prediction of acute kidney injury (AKI) may provide a crucial opportunity for AKI prevention. To date, no prediction model targeting AKI among general hospitalized patients in developing countries has been published. Here we show a simple, real-time, interpretable AKI prediction model for general hospitalized patients developed from a large tertiary hospital in China, which has been validated across five independent, geographically distinct, different tiered hospitals. The model containing 20 readily available variables demonstrates consistent, high levels of predictive discrimination in validation cohort, with AUCs for serum creatinine-based AKI and severe AKI within 48 h ranging from 0.74-0.85 and 0.83-0.90 for transported models and from 0.81-0.90 and 0.88-0.95 for refitted models, respectively. With optimal probability cutoffs, the refitted model could predict AKI at a median of 72 (24-198) hours in advance in internal validation, and 54-90 h in advance in external validation. Broad application of the model in the future may provide an effective, convenient and cost-effective approach for AKI prevention.

PubMed Disclaimer

Conflict of interest statement

Competing interests: The authors declare no competing interests. Inclusion & ethics statement: The research has included local researchers throughout the research process; The research is locally relevant and has been determined in collaboration with local partners; Roles and responsibilities were agreed amongst collaborators ahead of research; This research has not been restricted or prohibited in the setting of the researchers; The study has been approved by all local ethics review committees; The research was undertaken to higher standards of biorisk-related regulations in the local research setting; The research does not result in any risk to participants; We have taken local research relevant to our study into account in citations.

Figures

Fig. 1
Fig. 1. Flow diagram for the Derivation Cohort and Validation Cohort.
A Flow diagram for the derivation cohort; B Flow diagram for the validation cohort; C Construction of derivation, internal validation, and external validation cohorts.
Fig. 2
Fig. 2. Receiver-operating characteristic curves of refitted models for AKI and severe AKI in validation cohorts.
A AUC curves for prediction AKI within 24, 48, and 72 h; B AUC curves for prediction severe AKI within 24, 48, and 72 h; Site 1, Peking University First Hospital; Site 2, Sichuan Provincial People’s Hospital; Site 3, the Second Affiliated Hospital of Harbin Medical University; Site 4, Beijing Miyun District Hospital; Site 5, Taiyuan Central Hospital. Source data are provided as a Source Data file.
Fig. 3
Fig. 3. Model interpretability analysis of features included in the prediction model for AKI within 48 h.
A SHAP summary plot of 20 features included in the prediction model for AKI. SHAP value represents the impact of each feature on model output, with SHAP values over zero representing an increased risk of AKI and higher SHAP value representing higher risk of AKI. Each sample in the internal validation dataset is represented as a dot per feature. Dots are colored according to feature values for each sample and accumulated vertically to depict density. B The mean absolute SHAP value summary plot, representing the average impact of each feature on the prediction of AKI. Scr serum creatinine, ICU intensive care unit; NE, neutrophilic granulocyte; BUN, blood urea nitrogen; SHAP, Shapley Additive exPlanations. Source data are provided as a Source Data file.
Fig. 4
Fig. 4. The SHAP dependence plot for features included in the prediction model for AKI within 48 h.
The plot shows the relationship between feature values and SHAP values. The y-axis indicates SHAP values of features, corresponding to feature values on the x-axis. The dataset was from timepoints in the predictive data timeframe, which was before the AKI identification for patients with AKI, and before hospital discharge for patients without AKI. Scr serum creatinine, ICU intensive care unit, BUN blood urea nitrogen, NE neutrophilic granulocyte, ALB albumin, RBC red blood cell, PLT platelet, SHAP Shapley Additive exPlanations. Source data are provided as a Source Data file.
Fig. 5
Fig. 5. Force Plots showing feature contributions to model output in the prediction model for AKI within 48 h.
A Force plot of a patient with AKI, the model generated a predicted AKI probability of 67%. Admission to the ICU, a lower percentage of lymphocytes, elevated neutrophilic granulocyte-to-albumin ratio and heart rate are major features contributing to a higher predicted risk of AKI. B Force plot of a patient without AKI, the model generated a predicted AKI probability of 9%. The absence of diuretics, negative latest changing rate of Scr, a normal lowest Scr, and no hospitalization in the ICU are major features that drive the patient toward a lower predicted risk of AKI. NE neutrophilic granulocyte, ALB albumin, C intensive care unit, Scr serum creatinine. Source data are provided as a Source Data file.
See this image and copyright information in PMC

References

    1. Lameire, N. H. et al. Acute kidney injury: an increasing global concern. Lancet (Lond., Engl.)382, 170–179 (2013). - PubMed
    1. Griffin, B. R., Liu, K. D. & Teixeira, J. P. Critical care nephrology: core curriculum 2020. Am. J. Kidney Dis.: Off. J. Natl Kidney Found.75, 435–452 (2020). - PMC - PubMed
    1. Pickkers, P. et al. Acute kidney injury in the critically ill: an updated review on pathophysiology and management. Intensive care Med.47, 835–850 (2021). - PMC - PubMed
    1. Mehta, R. L. et al. International Society of Nephrology’s 0by25 initiative for acute kidney injury (zero preventable deaths by 2025): a human rights case for nephrology. Lancet (Lond., Engl.)385, 2616–2643 (2015). - PubMed
    1. Lewington, A. J., Cerdá, J. & Mehta, R. L. Raising awareness of acute kidney injury: a global perspective of a silent killer. Kidney Int.84, 457–467 (2013). - PMC - PubMed

Publication types

LinkOut - more resources