Development and validation of a nomogram for predicting hyperphosphatemia in non-dialysis patients with chronic kidney disease.

Abstract

Background: Elevated serum phosphate levels are strongly associated with an increased risk of all-cause mortality in patients with chronic kidney disease (CKD). The aim of this study was to identify independent risk factors for hyperphosphatemia in patients with non-dialysis CKD and use the findings to develop and validate a predictive model for assessing hyperphosphatemia risk.

Methods: Data of patients with CKD discharged from the Department of Nephrology between January 2021 and December 2023 were retrospectively analyzed. Potential predictors were screened from an array of clinical variables using least absolute shrinkage and selection operator regression in conjunction with 10-fold cross-validation. A multivariate logistic regression model was constructed to identify independent risk factors for predicting hyperphosphatemia. The C-index, receiver operating characteristic curve, calibration curve, and decision curve analysis were used to evaluate model predictive power, discriminability, accuracy, and clinical utility. Internal validation was implemented through a comparison of results from a validation set and the entire dataset.

Results: This study included 216 patients, with 134 (62.04%) individuals who developed hyperphosphatemia. Logistic regression revealed that hemoglobin, blood urea nitrogen, serum creatinine, and parathyroid hormone were independently correlated with hyperphosphatemia. The nomogram C-index was 0.916 (95% confidence interval [CI]: 0.872-0.961). The model demonstrated excellent discriminative ability in the independent validation set (area under the curve [AUC] = 0.953, 95% CI: 0.909-0.998), with the full dataset analysis showing concordant results (AUC = 0.923, 95% CI: 0.889-0.958). The decision and clinical impact curves showed the clinical value of our nomogram for patients with CKD and hyperphosphatemia.

Conclusions: The nomogram model was highly accurate in identifying CKD subpopulations at an elevated risk of serum phosphorus metabolic disorders. Our model can be utilized for prospective monitoring and preventive intervention. Furthermore, through individualized risk assessments, the model can contribute to the development of customized treatment strategies that have the potential to markedly improve long-term prognosis.