Predicting Nocturnal Hypertension in CKD Through Table-Value Diffusion Model.

Area of Science:

• Nephrology
• Cardiology
• Medical Informatics

Background:

• Ambulatory blood pressure monitoring (ABPM) is crucial for diagnosing nocturnal hypertension (NH) in chronic kidney disease (CKD) patients.
• Current ABPM methods are costly and time-consuming, necessitating efficient screening tools for high-risk NH.
• Predictive models for NH are urgently needed to optimize patient management in CKD.

Purpose of the Study:

• To develop and validate a predictive model for identifying patients with nocturnal hypertension (NH) among those with nondialysis chronic kidney disease (CKD).
• To assess the efficacy of a novel table-value diffusion model in predicting NH risk.
• To establish a risk stratification system for targeted ABPM screening in CKD.

Main Methods:

• A large cohort of 5769 nondialysis CKD patients was analyzed, with data split into training, internal, and external test sets.
• A table-value diffusion model was developed to predict NH probability using 7 core clinical variables.
• Model performance was evaluated using area under the curve (AUC) and Cohen kappa index, with secondary validation on 1006 patients.

Main Results:

• The 7-variable diffusion model demonstrated superior predictive performance compared to a logistic model in both internal (AUC 0.870 vs. 0.807) and external (AUC 0.854 vs. 0.792) test sets.
• The model showed stability in secondary validation (AUC=0.869, Cohen kappa=0.560).
• Risk stratification identified high-risk patients (>0.692) with a significantly higher incidence of adverse renal/cardiovascular events.

Conclusions:

• A concise 7-variable diffusion model was successfully developed for predicting NH in CKD patients.
• This diffusion model supports targeted ABPM screening, enhancing diagnostic efficiency.
• The model facilitates risk stratification, enabling focused monitoring for adverse outcomes in high-risk CKD individuals.