Association Between Metabolic Score for Insulin Resistance (METS-IR) and Risk of Obstructive Sleep Apnea: Analysis of NHANES Database and a Chinese Cohort

Introduction

Obstructive sleep apnea (OSA) is a prevalent but often overlooked chronic condition. Its pathogenesis involves recurrent upper airway collapse during sleep, leading to apnea and intermittent hypoxia.¹ This cyclical process of hypoxia and reoxygenation activates various cellular and molecular pathways, including oxidative stress, inflammation, and sympathetic nervous system activation. These mechanisms create a vicious cycle that elevates the risk of cardiovascular and metabolic diseases associated with OSA.^2,3

Metabolic syndrome is defined as a convergent state of a group of metabolic abnormalities, including central obesity, insulin resistance (IR), hypertension, and dyslipidemia.⁴ Obesity is one of the most important risk factors for OSA and can directly promote the occurrence of OSA by increasing pharyngeal fat deposition and airway resistance.⁵ Due to abnormal insulin and leptin levels caused by sleep fragmentation and circadian rhythm disturbance, patients with OSA may develop more appetite and further aggravate obesity.⁶ Body mass index (BMI) is commonly used as a measure of obesity in an individual, and studies have shown that the risk of severe OSA increases two to four times for each unit increase in BMI.⁷ Abnormal lipid and glucose metabolism are the two main components of metabolic disorders, and they are often closely intertwined with the clinical outcomes associated with OSA. Studies have shown that serum high-density lipoprotein cholesterol (HDL-C) is negatively correlated with the risk of OSA, while elevated serum triglyceride (TG) and fasting blood glucose (FBG) may increase the risk of OSA.⁸ Intermittent hypoxia may also induce dyslipidemia through mechanisms such as increased lipolysis, up-regulation of hepatic triglyceride lipid biosynthesis, and inhibition of lipoprotein clearance.⁹ It has been reported that up to 64.3% of OSA patients have insulin resistance.¹⁰ On one hand, OSA induces oxidative stress and inflammation through intermittent hypoxia, which damages the insulin signaling pathway and induces IR.¹¹ Conversely, IR may aggravate the pathological process of OSA by promoting obesity, systemic inflammation and metabolic disorders.^6,12

In recent years, triglyceride glucose (TyG) index, as a surrogate index for evaluating IR, has been shown to have a strong correlation with OSA.¹³ However, TyG mainly depends on fasting plasma glucose and triglyceride levels, which may not fully reflect obesity and the multidimensional characteristics of insulin resistance. Obesity, the most common manifestation of metabolic syndrome, holds significant importance in clinical diagnosis.¹⁴ It serves as a crucial link between OSA and metabolic syndrome.¹⁵ Both BMI and TyG serve as potential predictors of OSA, yet neither is without limitations.¹⁶ Therefore, it is of great clinical significance to explore a more comprehensive and sensitive surrogate indicator of insulin resistance for the risk assessment and early screening of OSA. Metabolic score for insulin resistance (METS-IR), a novel assessment tool for insulin resistance that combines FBG, TG, BMI, and HDL-C,¹⁷ has been shown to have good application value in predicting cardiovascular disease, hypertension, and other metabolic disorders.^18,19 This study aimed to evaluate the association between METS-IR and OSA using data from both American and Chinese populations. It further compared METS-IR with traditional predictors, such as BMI and TyG, to explore the potential value of METS-IR in OSA risk prediction.

Materials and Methods

Study Population

This study first identified the relationship between METS-IR and high risk of OSA using data from the NHANES American population. To further validate this association, we examined it in the Chinese cohort, where OSA was confirmed through sleep monitoring. This objective assessment provided a more reliable evaluation of the association between METS-IR and OSA. We utilized cross-sectional data from the NHANES (National Health and Nutrition Examination Survey) cycles 2005–2006, 2007–2008, 2015–2016, and 2017–2018, encompassing a total of 39722 American population. Participants were excluded if they met any of the following criteria: (1) missing sleep-related questionnaire data; (2) lack of information on METS-IR index factors (including fasting blood glucose, high-density lipoprotein, triglycerides, and body mass index); (3) age <20 years; and (4) missing data on other covariates such as alcohol consumption (Figure 1). Ultimately, the study included 8297 participants. Additionally, In the model using METS-IR to predict OSA, the expected model performance is 0.7. The prevalence of OSA among hospitalized patients is 80%, with a shrinkage coefficient of 0.9. Based on these parameters, the estimated minimum sample size is 246, including 197 OSA events. Finally, the study incorporated data from 581 consecutively hospitalized patients who underwent sleep monitoring at the Department of Respiratory Medicine, Renmin Hospital of Wuhan University, China, from January to June 2024. These patients are hospitalized in respiratory, cardiology and endocrinology departments for conditions such as obstructive pulmonary disease (COPD), asthma, pneumonia, hypertension, coronary heart disease (CHD) and diabetes mellitus. During the stable phase of their condition, patients underwent sleep monitoring due to complaints of snoring, daytime sleepiness, or unexplained hypertension. Inclusion criteria were as follows: availability of complete sleep monitoring data, comprehensive clinical information, and biochemical marker data. Exclusion criteria included: age <18 years, receiving treatment for OSA (continuous positive airway pressure, surgery), or using anxiolytics, antidepressants, hypnotics, or antipsychotics. The results of sleep monitoring and important clinical information were collected from the electronic medical records, including details of the patient’s basic information (such as sex, age, height, weight, smoking, and alcohol consumption), medical history such as hypertension, diabetes mellitus, COPD, CHD, chronic kidney disease (CKD), and medication use (such as antihypertensive and lipid-lowering medications). The study was approved by the Ethics Committee of Renmin Hospital of Wuhan University, and informed consent was waived due to its retrospective nature.

Figure 1 Flowchart of the screening process for the American population.

Assessment of OSA in the NHANES

In this study, high risk of OSA in the American population was identified based on self-reported responses to specific questions related to sleep disorders in the NHANES survey.²⁰ Participants were classified as having a high risk of OSA if they answered “yes” to any of the following three binary questions: (1) Despite obtaining at least 7 hours of sleep on workdays, do you still experience excessive daytime sleepiness 16 to 30 times per month? (2) Do you experience breathing difficulties, snoring, or gasping for air at least three nights per week? (3) Do you snore at least three nights per week? This definition was first introduced in the NHANES 2005–2008 dataset and has since been validated and widely applied in epidemiological studies to explore the association between OSA and health outcomes.²¹

Assessment of METS-IR, TyG, TyG-BMI and Hyperlipidemia

METS-IR was calculated as ln[(2 × FBG (mg/dL) + TG (mg/dL)] × BMI (kg/m²))/(ln[HDL-C (mg/dL)]). Additionally, the TyG index was calculated as ln [TG (mg/dL) × FBG (mg/dL) /2]. BMI was calculated as weight (kg)/height (m²). TyG-BMI (triglyceride glucose-body mass index) was calculated as the TyG index × BMI. Hyperlipidemia was defined according to any of the following criteria: (1) TG ≥ 150 mg/dL, (2) total cholesterol (TC) ≥ 200 mg/dL, (3) low-density lipoprotein cholesterol (LDL-C) ≥ 130 mg/dL, or (4) HDL < 40 mg/dL for men or < 50 mg/dL for women.²²

Assessment of Sleep Parameters and OSA in Chinese Population

Each patient underwent sleep assessment with a Zhaoguan ultra-wideband biological-radar sleep-screening (UWB, Hangzhou Megasens Technology Co., LTD., China). The noncontact monitor consisted of a radar transmitter (ZG-S01D) for tracking breathing and body movement and a finger plethysmograph (ZG-P11F) for continuous oxygen saturation recording. Our team has conducted multiple studies using this device.^23–25 To minimize interference with sleep quality, patients were instructed to avoid alcohol, caffeine, and sedatives for at least 24 hours before monitoring. On the night of monitoring, the operator positioned the radar transmitter approximately 1 meter away from the bedside and 0.5 meters above the mattress, ensuring it was parallel to the patient’s chest and directed toward the torso. The monitoring device operated continuously throughout the overnight sleep period, with each patient was monitored for at least 7 hours. The next day, after monitoring ended, the operator downloaded the recorded data to a computer for processing, graphical visualization, and generation of a final report.^26,27 The report contains detailed sleep indicators, including total sleep time, sleep efficiency, sleep stages, lowest pulse oximetry (min SpO2), mean SpO2, percentage of sleep time with SpO2 <90% (T90), oxygen desaturation index (ODI; number of times per hour of sleep that the blood oxygen level drops by ≥4% from baseline), and apnea-hypopnea index (AHI). The severity of sleep apnea was measured using the AHI and was categorized as mild (AHI 5–14.9 events/h), moderate (AHI 15–29.9 events/h), or severe (AHI≥30 events/h) according to conventional cutoffs.

NHANES Covariates

To understand the effect of confounding factors in the multivariable adjusted model on the relationship between the METS-IR and OSA, we considered the following covariates: (1) Demographic characteristics: age, gender, race; (2) Physical measurement indicators: BMI; (3) Lifestyle factors: smoking status, drinking status; (4) Disease status: hypertension, diabetes mellitus, CHD, stroke, COPD, CKD, hyperlipidemia. All (except hyperlipidemia) were defined based on self-reported prior clinical diagnoses.

Statistical Analysis

Fasting subsample weights were used according to NCHS guidance during the analysis of the American population in this study. Continuous variables are described as medians [interquartile range (IQR)], and categorical variables as counts and percentages. The American population was divided into four groups according to METS-IR quartiles. Chinese patients were divided into four groups based on the severity of OSA or quartiles of METS-IR. Comparisons were made using the Kruskal–Wallis test or Pearson’s chi-square test. A multivariate logistic regression model was constructed to explore the association between METS-IR and the risk of OSA in the American population. And multivariate logistic regression and multiple linear regression model were constructed in Chinese patients to test the correlation between METS-IR and OSA, AHI, ODI, and T90. Subgroup analyses were performed to examine variations in the association between METS-IR and the prevalence of OSA in the Chinese population. P-values for interaction were calculated to assess the interaction effect of subgroups in the association between METS-IR and OSA. Receiver operating characteristic (ROC) curve analysis was used to evaluate diagnostic value and cutoff values along with sensitivity and specificity were evaluated in the Chinese population. In addition to METS-IR, BMI, TyG, and TyG-BMI were also included for comparison with METS-IR. Prognostic power was determined by calculating the area under the curve (AUC). The DeLong test was employed to compare the AUCs of the different models to assess their relative performance. Furthermore, the cost-effectiveness of METS-IR was also calculated. All statistical analyses were performed using R software version 4.3.0 (http://www.r-project.org), and double-sided P < 0.05 was defined as statistical significance.

Results

Baseline Characteristics of the American Population

Table 1 presents the weighted demographic and clinical characteristics of the subjects according to the quartiles of the METS-IR. The study included a total of 8,297 participants, with 51.6% men and 48.4% women. The median age was 47 years, and the prevalence of OSA was 49.8%. Higher METS-IR quartiles were associated with a greater proportion of individuals who smoked, had diabetes mellitus, hypertension, CHD, COPD, hyperlipidemia, or OSA (P < 0.001).

Table 1 Characteristics of the American Population Across METS-IR Quartiles

Relationship Between METS-IR and Higher Risk of OSA in the American Population

To better explore the relationship between the METS-IR and the higher risk of OSA, we analyzed three logistic regression models. In the unadjusted model, no variables were adjusted, and the results showed that METS-IR was positively correlated with OSA risk (OR= 1.046; 95% CI: 1.039–1.053; P < 0.001). Similarly, in Model I, the association between METS-IR and OSA remained significant after accounting for age, sex, and race (OR= 1.044; 95% CI: 1.037–1.050; P < 0.001). In the fully adjusted Model II, the risk of OSA increased significantly by 4.4% for each unit increase in the METS-IR. Compared with the lowest quartile group, higher risk of OSA in the fourth quartile group increased by 302.8% (OR= 4.028; 95% CI: 3.261–4.975; P < 0.001) (Table 2).

Table 2 Multivariate Regression Analysis of METS-IR and OSA in the American Population

Baseline Characteristics of the Chinese Population

A total of 581 patients who underwent sleep monitoring in Renmin Hospital of Wuhan University were included in this study. The median age of the patients was 60 years, with 71.4% being male and 28.6% female. The prevalence of OSA among these patients was 87.3%. Compared to non-OSA patients, those with OSA exhibited significantly higher TG and BMI, significantly HDL, and a significantly higher prevalence of hypertension and CHD. Notably, METS-IR was significantly elevated in OSA patients and increased with the severity of OSA (P<0.001) (Table 3).

Table 3 Characteristics of the Chinese Population by Different OSA Severity Levels

Characteristics of the Chinese Population Stratified by METS-IR Quartiles

Patients were divided into four groups according to METS-IR quartile. Significant differences were observed in age, gender, AHI, ODI, T90, FBG, TG, HDL, smoking, drinking, OSA, hypertension, diabetes mellitus, COPD, CKD, hyperlipidemia, and the use of antihypertensive, lipid-lowering, and antidiabetic medications (P<0.05). As METS-IR increased, the age of patients decreased, and the proportions of men and alcohol drinkers rose. The prevalence of OSA, hypertension, diabetes mellitus, and hyperlipidemia also increased. Additionally, patients with higher METS-IR exhibited worse sleep parameters and more common use of antihypertensive, lipid-lowering, and antidiabetic drugs (Table 4).

Table 4 Characteristics of the Chinese Population Across METS-IR Quartiles

Association Between METS-IR and OSA Prevalence, AHI, ODI, and T90 in the Chinese Population

After adjusting for all confounding factors, models revealed a significant positive correlation between METS-IR and OSA, AHI, ODI, and T90. For each 1-unit increase in METS-IR, the prevalence of OSA increased by 6.7% (OR= 1.067; 95% CI: 1.035–1.103), AHI increased by 0.732 (β= 0.732; 95% CI: 0.573–0.732), ODI increased by 0.777 (β= 0.777; 95% CI: 0.614–0.940), and T90 increased by 0.389 (β= 0.389; 95% CI: 0.285–0.492). Compared with the Q1 group, AHI, ODI, and T90 were higher in the Q3 and Q4 groups, with the prevalence of OSA increasing by 109.5% and 877.4%, respectively. As the METS-IR quartile increased, the probability of OSA, as well as AHI, ODI, and T90, rose significantly (P<0.001) (Table 5).

Table 5 Association of METS-IR with OSA, AHI, ODI and T90 in the Chinese Population

Subgroup Analysis of the Chinese Population

Subgroup analyses were conducted to determine if the relationship between METS-IR and OSA prevalence remained consistent across various demographic groups. The positive correlation between METS-IR and OSA prevalence was observed in all groups except for women, individuals with a BMI ≥ 28 kg/m², and those without hyperlipidemia, hypertension, coronary heart disease, COPD, or those not taking antiplatelet drugs. Additionally, gender significantly influenced the relationship between METS-IR and OSA (P for interaction = 0.047) (Figure 2).

Figure 2 Subgroup analysis of the Chinese population. Abbreviations: OSA, obstructive sleep apnea; AHI, apnea hypopnea index; ODI, oxygen desaturation index; T90, the percentage of time with oxygen saturation less than 90%; FBG, fasting blood glucose; TC, total cholesterol; TG, triglyceride; HDL, high density lipoprotein-cholesterol; LDL, low density lipoprotein-cholesterol; EGFR, estimated glomerular filtration rate; TyG, triglyceride glucose index; BMI, body mass index; TyG-BMI, triglyceride glucose-body mass index; METS-IR, metabolic score for insulin resistance; CHD, coronary heart disease; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease.

ROC Curve Analysis of the Chinese Population

The ROC curve analysis demonstrated the diagnostic effectiveness of BMI, TyG, TyG-BMI, and METS-IR for OSA. In the unadjusted model, the AUC for METS-IR was 0.703, which was higher than that for TyG-BMI (AUC: 0.697), BMI (AUC: 0.668) and TyG (AUC: 0.649) (Figure 3A). After adjusting for gender, age, smoking and drinking status, COPD, hypertension, diabetes mellitus, CKD, CHD, cerebrovascular disease, and the use of antihypertensive, antidiabetic, lipid-lowering, and antiplatelet drugs, TyG-BMI and METS-IR showed superior diagnostic performance, each with an AUC of 0.777, compared to BMI and TyG alone. TyG had the lowest diagnostic performance (AUC: 0.749), followed by BMI (AUC: 0.769). Furthermore, TyG-BMI was significantly more effective in diagnosing OSA than TyG (P = 0.044) (Figure 3B). The optimal cutoff points for BMI, TyG, TyG-BMI, and METS-IR were 25.34, 8.19, 225.20, and 41.97 respectively. Using these cutoff points, the obtained sensitivity values were 0.548, 0.771, 0.517, and 0.503, while the specificity values were 0.743, 0.486, 0.838, and 0.838.

Figure 3 Receiver-operating characteristic curve of BMI, TyG, TyG-BMI and METS-IR for predicting OSA in the Chinese population. (A) Unadjusted model. (B) The model was adjusted for gender, age, smoking status, drinking status, COPD, hypertension, diabetes mellitus, CKD, CHD, cerebrovascular disease, antihypertensive agents, antidiabetic drug, lipid-lowering drug, antiplatelet drug. Abbreviations: ROC, receiver operating characteristic; AUC, the area under the curve; TyG, triglyceride glucose index; BMI, body mass index; TyG-BMI, triglyceride.

Cost-Effectiveness Analysis of METS-IR in Evaluating OSA in the Chinese Population

Next, the cost-effectiveness of METS-IR and UWB in the diagnosis of OSA was evaluated. According to prior research, UWB demonstrated a sensitivity of 100% and a specificity of 70% for OSA diagnosis.²⁶ The respective costs per test for UWB, METS-IR, TyG, and TyG-BMI were 150, 20, 11, and 11 yuan. The average cost-effectiveness ratios (ACERs) for these tests were 171.8, 45.5, 16.3, and 24.4 yuan. When comparing METS-IR to UWB, TyG, and TyG-BMI, the incremental cost-effectiveness ratios (ICERs) were 299.6 yuan, −38.5 yuan, and −736.4 yuan, respectively.

Discussion

This study represents a preliminary investigation into the association between METS-IR and OSA risk in Chinese and American populations. The results indicate that METS-IR is associated with an increased risk of OSA after adjusting for confounding factors. Specifically, the risk of OSA significantly increases with higher METS-IR levels. Therefore, METS-IR is likely to be an important predictor of OSA.

Obesity, particularly visceral obesity, is widely recognized as a core driver of metabolic abnormalities.²⁸ Excess adipose tissue disrupts metabolic homeostasis by altering adipokine secretion and activating pro-inflammatory pathways.²⁹ Studies have shown that visceral obesity and neck circumference are associated with the occurrence of OSA, and that neck circumference has an independent correlation with both metabolic syndrome and OSA.³⁰ When obesity and OSA coexist, they have potential synergistic effects. Obesity increases the effort required to maintain airway patency, especially during REM sleep, leading to OSA. Furthermore, long-term obesity can cause related conditions such as hypertension, insulin resistance, and dyslipidemia, which can exacerbate OSA. Conversely, OSA impairs insulin sensitivity and increases inflammation, leading to obesity and poor glycemic control.³¹ Therefore, given obesity’s critical role in linking metabolic disorders and OSA, weight management is of paramount importance.

IR is physiologically described as a diminished sensitivity and reaction of target tissues to insulin. This condition significantly impacts various adverse clinical outcomes, including metabolic syndrome, non-alcoholic fatty liver disease, atherosclerosis, and type 2 diabetes mellitus (T2DM). The underlying mechanisms of IR are generally believed to include ectopic lipid accumulation, endoplasmic reticulum (ER) stress, and inflammation in the liver and skeletal muscle.³² A randomized crossover trial found that chronic intermittent hypoxia (CIH) induces high sympathetic tone, lipolysis, and reduced sensitivity of free fatty acids to insulin in healthy individuals. This may contribute to the development of systemic insulin resistance and diabetes in OSA patients. Moreover, IR can regulate CIH-induced lipolysis.³³ Some scholars have proposed that systemic “lipotoxicity” caused by excessive lipolysis during sleep in OSA is the root cause of metabolic dysfunction.³⁴ Recent studies have shown increased intramyocellular and extramyocellular lipids in OSA patients, particularly in non-obese individuals. This is related to inappropriate lipolysis triggered by OSA-related intermittent hypoxia, which in turn triggers a free fatty acid storm. Treatment with continuous positive airway pressure significantly improves these characteristics.³⁵

Insulin-based measures of IR are often limited by the need for invasive procedures. In contrast, non-insulin-based measures, including the TyG index, TyG-BMI index, and METS-IR, offer advantages such as cost-effectiveness and simplified acquisition techniques. Additionally, these measures have a stronger association with the gold standard protocol for measuring insulin resistance.³⁶ Recent studies have demonstrated that METS-IR is positively associated with the risk of various diseases, including lung cancer, heart failure, T2DM, and neurological impairment.^37–40 METS-IR is emerging as a potential tool for assessing disease risk. Similarly, TyG related indices have shown remarkable predictive power. With each 1-unit rise in TyG-BMI, the likelihood of OSA rose by 54%.⁴¹ Additionally, an increase in the TyG-waist circumference (TyG-WC) is linked to a 4.29-fold increase in the risk of the first myocardial infarction in OSA patients within the highest quartile of TyG-WC.⁴² Furthermore, some studies have employed machine learning methods to differentiate between OSA and non-OSA individuals, with results indicating that the TyG index is the most reliable predictor compared to BMI.¹⁶

This study evaluated the predictive value of METS-IR against BMI and TyG-related measures for assessing OSA risk. The AUC in ROC analysis was as follows: METS-IR = TyG-BMI > BMI > TyG. Although the improvement in AUC for METS-IR relative to BMI was modest, the added value of METS-IR may be particularly beneficial in clinical settings where a more nuanced assessment of metabolic dysfunction is required to identify patients at risk for OSA, particularly those who have not yet shown overt symptoms but have underlying metabolic abnormalities. Additionally, Gender was identified as the only variable with an interaction effect between METS-IR and OSA. This may be related to sex differences in fat distribution and propensity for lipid metabolism. Studies have shown that men generally accumulate more visceral fat in the abdomen, while women accumulate more subcutaneous fat in the hips and thighs.⁴³ Compared to subcutaneous fat, visceral fat is easier to mobilize and dissolve, and is more sensitive to the anti-lipolytic effect of insulin.⁴⁴ Moreover, men are more likely to have high levels of very-low-density lipoprotein lipids, particularly triglycerides, whereas women typically have higher HDL levels.⁴⁵ As women enter menopause, their estrogen levels decrease, increasing the risk of lipid metabolic disorders, which is manifested by the accumulation of visceral fat and aggravation of insulin resistance. The fact that women have less abdominal fat accumulation may contribute to their lower risks for cardiovascular disease, metabolic syndrome, and diabetes.^43,46 More caution is needed when applying this indicator to predict the risk of OSA in females. In the cost-effectiveness analysis, METS-IR proved to be more cost-effective compared to UWB, with a lower cost per identified true positive case. TyG-BMI demonstrated the highest cost-effectiveness, offering greater benefits at a lower cost. Moreover, TyG-BMI showed a clear advantage over both METS-IR and TyG in terms of cost-effectiveness.

This study has several strengths. Firstly, it is the first to assess the link between METS-IR and OSA risk, highlighting METS-IR as a potential predictive marker for OSA. Secondly, OSA diagnosis in the Chinese population was based on actual sleep monitoring using the AHI, rather than questionnaires or subjective reports. Subgroup analysis was used to enhance the validity and reliability of our findings. Nevertheless, this study has several limitations that warrant consideration. The cross-sectional design precludes the establishment of causal relationships between METS-IR and OSA. While a significant association was observed, it remains unclear whether elevated METS-IR contributes to the development of OSA or whether OSA may influence METS-IR levels. Large-scale epidemiological studies like NHANES rely on questionnaires, as sleep monitoring is impractical for thousands of participants. This approach may introduce misclassification bias, potentially failing to identify mild or asymptomatic OSA cases. Another important limitation lies in the hospital-based recruitment and single-center nature of this study. Our Chinese cohort comprised hospitalized patients with suspected OSA, which may limit the generalizability of the findings to asymptomatic individuals or those with mild OSA typically managed in primary care. Additionally, as the study was conducted at a single institution, the results may not apply to populations with different demographic, clinical, or environmental characteristics. Our study did not systematically analyze night/day symptom patterns. Future research should combine METS-IR with symptom scales to explore how metabolic features interact with specific OSA phenotypes. Incorporating METS-IR as a routine screening tool for patients at risk for OSA, especially those with metabolic disorders, is essential. Early identification of high-risk individuals allows for targeted interventions, such as lifestyle modifications and continuous positive airway pressure therapy, to reduce OSA incidence and severity. For instance, non-obese patients with metabolic dysregulation (elevated TG, low HDL-C) could be flagged for OSA screening even if their BMI is normal, addressing the limitations of BMI-centric approaches. As a simple, non-invasive, and readily calculable index (requiring routine laboratory tests), METS-IR could be integrated into primary care settings to prioritize high-risk patients for costly and time-consuming sleep studies like polysomnography. This is particularly valuable in resource-limited regions where OSA remains underdiagnosed. METS-IR’s ability to quantify metabolic dysfunction in OSA patients may help subclassify OSA phenotypes, such as “metabolically unhealthy OSA” vs “metabolically healthy OSA”, enabling precision medicine strategies. Additionally, given the association between METS-IR and OSA, clinicians should consider incorporating metabolic health assessments into the evaluation and management of OSA patients. By combining METS-IR with traditional OSA screening tools (such as the STOP-Bang Questionnaire and NoSAS Score), the sensitivity and specificity of screenings can be enhanced. This approach can aid in the early identification of individuals with metabolic abnormalities who may not yet exhibit obvious OSA symptoms, potentially improving health outcomes and reducing the risk of long-term complications.

Conclusions

The results suggest that an elevated METS-IR correlates with a heightened risk of OSA. Early identification and intervention for individuals with higher METS-IR may help prevent the onset of OSA and mitigate its negative impact on health.