AIM To investigate the risk factors associated with blood pressure variability (BPV) in patients with sleep disorders and essential hypertension and to develop a prediction model for patients with increased BPV.

METHODS Patients with sleep disorders and essential hypertension admitted to the Chinese PLA General Hospital and its collaborating hospitals from 2023 to 2024 were included in the study. Data collected included 24-hour ambulatory blood pressure monitoring, 24-hour ambulatory electrocardiogram monitoring, laboratory tests, PSQI scores, GAD-7 scores, PHQ-9 scores, PHQ-15 scores, and clinical information related to hospitalization and disease management. BPV was quantified using mean true variability (ARV). Patients were divided into an ARV increased group (n=228) and a non-ARV increased group (n=228) based on their ambulatory blood pressure results. 456 patients were randomly divided into a validation group (n=182) and a training group (n=274). Univariate and multivariate logistic regression analyses were performed to identify the influencing factors of BPV. In multivariate analysis, a bidirectional stepwise regression method was used to establish the prediction model with independent variables having a p-value less than 0.05. After the regression equation was established, the prediction model was visualized using a nomogram and evaluated.

RESULTS Compared with the non ARV elevation group, the ARV elevation group had higher age (P<0.05), higher BMI (P<0.05), higher proportion of hypertension history (P<0.05), higher PSQI (P<0.01), higher 24-hour mean systolic blood pressure (P<0.05), higher mean arterial pressure (P<0.05), higher dynamic arteriosclerosis index (P<0.01), lower proportion of spoon shaped and inverted spoon shaped curves, and higher proportion of non spoon shaped and super spoon shaped curves, all P<0.01. SDNNindx was low (P<0.01), SDANN and RMSSD were high (both P<0.05), pNN50 was low (P<0.01), and ALT was high (P<0.05); There was no significant difference between the validation group and the training group in terms of each item; BMI (OR=1.14, 95% CI: 1.05～1.24, P<0.01), PSQI (OR=1.14, 95% CI: 1.05～1.22, P<0.01), ALT (OR=1.05, 95% CI: 1.01～1.08, P<0.05), SDNNindx (OR=0.98, 95% CI: 0.97～0.99, P<0.01), and pNN50 (OR=0.95, 95% CI: 0.93～0.98, P<0.01) are closely related to the increase of ARV in various projects. Incorporating the above factors into the prediction model, ROC curve analysis evaluated the performance of the model, showing an AUC value of 0.822 for the training set and 0.745 for the validation set. The calibration curve indicates that the model has good consistency in predicting probabilities and actual observations on the training set, demonstrating good calibration of the model. Decision Curve Analysis (DCA) indicates that the model performs well on both the training and validation sets, and can benefit clinically.

CONCLUSION PSQI is closely positively correlated with systolic blood pressure ARV in patients with sleep disorders and primary hypertension. BMI and ALT are risk factors for increased systolic blood pressure ARV, while SDNNindx and PNN50 are protective factors. Based on the above factors, a predictive model is established, which has good predictive performance.