Micronutrients and ICU-Acquired Weakness: A Narrative Review.

Cardiometabolic multimorbidity (CMM) refers to the co-occurrence of two or more of the following conditions: hypertension, type 2 diabetes mellitus, coronary heart disease, and stroke (). This complex chronic condition is a major burden on adults worldwide. It leads to higher mortality, disability, and healthcare use (–). This is especially true in large, diverse countries such as China, the United Kingdom, and the United States (, ). Identifying modifiable early risk factors and developing scalable prevention strategies are important for delaying the onset and progression of CMM. Sleep disturbances are getting more attention in cardiometabolic research. They are common and can be changed. In addition to known links with hypertension and diabetes, disrupted sleep is also linked to adiposity, insulin resistance, and dyslipidemia (, ). Each adds to the cumulative cardiometabolic burden. At the mechanistic level, insufficient or excessive nighttime sleep, poor sleep quality, insomnia symptoms, and circadian misalignment may contribute to metabolic and vascular problems (–). This occurs through several pathways: neuroendocrine activation, sympathetic overactivity, chronic low-grade inflammation, oxidative stress, and changes in appetite- and energy-related hormones such as leptin and ghrelin (–). Despite these mechanisms, it remains unclear whether nighttime sleep duration, sleep quality, and daytime napping each independently and consistently predict CMM, especially across different sociocultural settings. A main limitation of existing research is its focus on single sleep measures, specific outcomes, or composite indicators. This makes it hard to study the range of sleep patterns and their possible links with CMM. At the methodological level, machine learning provides a distinct advantage for disentangling the complex associations between sleep and CMM. Unlike conventional approaches, machine learning can capture nonlinear relationships and higher-order interactions across high-dimensional feature spaces (, ), making it particularly well suited to modeling risk structures arising from the interplay among multiple sleep dimensions, lifestyle factors, and sociodemographic variables. Adding interpretability methods may help explain the role of sleep-related variables in the model and improve the practical value of the findings. These methods may also help identify sleep patterns and participant groups that could benefit most from targeted prevention.