Interaction between sleep disorder and depression on early onset hypertension: a machine learning modelling approach

睡眠障碍与抑郁症对早发性高血压的交互作用:一种机器学习建模方法

阅读:4

Abstract

OBJECTIVE: The interrelationship between sleep disorders, depression, and hypertension remains unclear. This study aimed to investigate the associations between these conditions. METHODS: The National Health and Nutrition Examination Survey data categorized patients into hypertensive and non-hypertensive groups, with propensity score matching (PSM) applied for demographic adjustment. Weighted multivariate adjusted logistic regression identified independent risk factors for hypertension, specifically focusing on early-onset hypertension in the subgroup analysis. Restricted cubic spline analysis assessed the relationship between sleep duration, PHQ-9 score, and hypertension. An additive interaction model evaluated the combined effects of sleep disorders and depression on hypertension risk. Additionally, seven machine learning models were established to evaluate their predictive capacity for hypertension incidence. Finally, two-sample Mendelian randomization (MR) analysis determined the causal relationships between sleep disorders, depression, and hypertension. RESULTS: Of the 21,453 participants, 9,055 from each group remained after PSM. Weighted multivariate logistic regression showed that short and long sleep durations and sleep disorders independently increased hypertension risk by 17.2%, 35.8%, and 48.6%, respectively (p < 0.001). The association between sleep duration and hypertension was nonlinear, with an optimal duration of 6.92 h (p < 0.001). Depression was associated with a 44.1% increased hypertension risk (p < 0.001), showing no nonlinear relationship with the PHQ-9 score (p = 0.132). Compared with the overall hypertensive population, early-onset hypertension demonstrated stronger associations with short sleep duration, sleep disorders, and depression. The additive interaction model indicated that sleep disorders and depression combinations accounted for 21.4% of hypertension cases. The XGBoost, DT, LR, and RF models exhibited strong predictive capabilities for hypertension incidence. Two-sample MR analysis supported causal relationships between sleep disorders (OR = 1.497, 95% CI: 1.258–1.782), depression (OR = 1.266, 95% CI: 1.122–1.428), and hypertension. CONCLUSIONS: This study demonstrates the complex associations between sleep disorders, depression, and hypertension, particularly in early-onset cases. The findings indicate that sleep disturbance and depression independently cause hypertension, and their interaction contributes significantly to the overall hypertension burden. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s13690-025-01778-y.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。