Renal Failure · Nov 17, 2026

Association of questionnaire-defined probable obstructive sleep apnea with advanced cardiovascular-kidney-metabolic syndrome in US adults: a NHANES study with machine learning analysis

By Li Y, Ge Y, Jing S et al.

Cardiovascular-kidney-metabolic (CKM) syndrome integrates cardiovascular disease, kidney disease, and metabolic dysfunction within a prevention-oriented framework, but the population-level association between obstructive sleep apnea and advanced CKM remains unclear. This cross-sectional study included 8,523 adults from the National Health and Nutrition Examination Survey. Survey-weighted logistic regression models were used to evaluate the association between questionnaire-defined probable OSA (pOSA) and advanced CKM, with subgroup analyses and interaction tests performed across predefined population groups. Six machine learning models were also developed for exploratory classification of advanced CKM, and Shapley Additive Explanations were used to interpret the selected model. pOSA was more common among participants with advanced CKM than among those with non-advanced stages. In the core multivariable-adjusted survey-weighted model, pOSA was associated with higher odds of advanced CKM; however, after additional adjustment for body mass index categories, the association was attenuated and was no longer statistically significant (OR = 1.17, 95% CI: 0.92–1.49, p = 0.197). In the machine learning analysis, XGBoost had the highest area under the receiver operating characteristic curve (0.876), whereas the random forest model showed favorable calibration with an area under the curve of 0.868 and was selected for interpretation. Age contributed most to the model output, followed by estimated glomerular filtration rate, smoking, urinary albumin-to-creatinine ratio, glycated hemoglobin, and mean systolic blood pressure. These findings suggest a potential association between pOSA and advanced CKM syndrome among United States adults, while highlighting the importance of considering adiposity when interpreting this association.

doi.org/10.1080/0886022x.2026.2757912Open accessPubMed 43046633Populations

AgingResearch.News