Correlation between METS-IR index and obstructive sleep apnea in non- diabetic adults: evidence from NHANES 2001–2018

doi:10.21203/rs.3.rs-4913542/v1

Background: Insulin resistance (IR) is strongly associated with obstructive sleep apnea (OSA). Whereas, few studies have focused on the potential correlation between the Metabolic Score for Insulin Resistance (METS-IR), a novel non-insulin-dependent IR index, and OSA.

Methods: Subjects from the National Health and Nutrition Examination Survey (NHANES) spanning 2005-2008 and 2015-2018 were recruited. The potential relationship between METS-IR and other IR indices with OSA was explored through three logistic regression analysis models and restricted cubic spline (RCS) curves. Receiver operating characteristic (ROC) curves were used to assess the diagnostic value of these indicators for OSA. On the basis of age, sex, race, body mass index (BMI), hypertension, and cardiovascular disease (CVD), subgroup analyses were conducted to test the robustness of the METS-IR and OSA relationship.

Results: In all of 6,633 non-diabetic participants were enrolled, with an OSA prevalence of 28.40%. After adjusting for potential confounders, METS-IR, triglyceride to high-density lipoprotein cholesterol (TG/HDL-C) ratio, triglyceride glucose Index (TyG), and homeostatic model assessment of insulin resistance (HOMA-IR) indices showed positive associations with OSA prevalence. In the highest tertile of METS-IR, TG/HDL-C, TyG index, and HOMA-IR indices, OSA prevalence was 3.22-fold, 1.42-fold, 1.45-fold, and 1.51-fold higher, respectively, compared to the lowest tertile (METS-IR: OR = 3.22, 95% CI: 2.67, 3.89, P < 0.0001; TG/HDL-C: OR = 1.42, 95% CI: 1.15, 1.76, P =0.002; TyG index: OR = 1.45, 95% CI: 1.18, 1.78, P<0.001; HOMA-IR: OR = 1.51, 95% CI: 1.24, 1.85, P <0.001). ROC analysis revealed that METS-IR had the highest diagnostic accuracy for OSA (AUC = 0.65). The relationship between METS-IR and OSA did not show significant interaction across all subgroups (P for interaction > 0.05).

Conclusion: Compared with HOMA-IR, TG/HDL-C and TyG, METS-IR index was positively correlated with OSA prevalence and had superior diagnostic accuracy.

insulin resistance

METS-IR

OSA

NHANES

cross-sectional study

The recent inclusion of sleep as one of the criteria for assessing cardiovascular health by the American Heart Association has drawn increased attention to sleep-related issues among researchers(1). Obstructive sleep apnea (OSA) ranks among the most prevalent sleep disorders. The global prevalence of OSA has been reported to affect up to 936 million individuals aged 30–69 years, with nearly half experiencing moderate to severe forms(2), particularly among male and obese populations(3). During sleep, individuals with OSA suffer from recurrent upper airway collapses that obstruct airflow, leading to hypoventilation or apnea, characterized by intermittent hypoxemia (IH) and disrupted sleep patterns(4). OSA has been identified as an independent risk factor for several non-communicable diseases, such as cardiovascular disease (CVD) and metabolic disorders(5–7), and is related to higher all-cause mortality risk(8, 9). Moreover, daytime sleepiness significantly impairs work productivity, contributing to a higher incidence of workplace and traffic accidents(10). Thus, early identification of high-risk OSA groups is crucial for implementing appropriate interventions.

Insulin resistance (IR) refers to a pathophysiological condition where peripheral tissues exhibit reduced responsiveness to insulin due to a decrease in the number or binding capacity of insulin receptors, or diminished insulin sensitivity(11). The high insulin normoglycemic clamp technique (HEC) is widely regarded as the most accurate method for assessing insulin sensitivity in peripheral tissues(12). However, the high price and complexity of the procedure hinder its widespread clinical application. The classical homeostasis model assessment of insulin resistance (HOMA-IR) appears to have relative advantages, but its calculation depends on fasting insulin (FINS) levels, limiting its utility in cases of β-cell insufficiency or exogenous insulin use(13). Other readily available insulin-independent non-invasive indices, such as the triglyceride to high-density lipoprotein cholesterol (TG/HDL-C) ratio and the triglyceride glucose (TyG) index, have demonstrated superiority over traditional indices (such as HOMA-IR) in predicting CVD and its mortality(14, 15). However, these indices fail to take into account the significant role of body mass index (BMI) in IR. Therefore, Bello-Chavolla et al. developed a novel indicator of IR, METS-IR, based on fasting blood glucose (FBG), high-density lipoprotein cholesterol (HDL-C), triglycerides (TG), and BMI(16–18). Based on its simplicity and validity, researchers have observed correlations between METS-IR and conditions such as CVD, hypertension, and depression(19). Notably, METS-IR has demonstrated greater accuracy than TyG, TG/HDL-C and even HOMA-IR in diagnosing testosterone deficiency, hyperuricemia, and predicting all-cause and cardiovascular mortality (20–22).

OSA appears to exhibit a bidirectional association with diabetes and metabolic syndrome (MetS)(23–25). IR is recognized as a key contributor to both MetS and diabetes, and is implicated in the development of CVD(26–28). Therefore, the relationship between OSA and IR may also be bidirectional. Mechanistically, IR could potentially increase the risk of OSA by reducing carotid body (CB) sensitivity and impairing autonomic function, thereby disrupting ventilation(24, 25). Conversely, intermittent hypoxia (IH) and fragmented sleep patterns characteristic of OSA may exacerbate IR(29–32). Clinically, several studies have reported correlations between IR indices such as the TyG index, Visceral Adiposity Index (VAI), and HOMA-IR, and the risk of OSA(33–36). Conversely, OSA has been suggested to be associated with increased IR(30, 37). Whereas, there do not exist studies focusing on the relationship between METS-IR and OSA, leaving uncertain whether a relationship exists between them.

Given the significant role of IR in OSA and the diagnostic advantages of METS-IR, this study will utilize the National Health and Nutrition Examination Survey (NHANES) to explore further the correlation between METS-IR and OSA, comparing it with other IR indicators. The aim is to provide a simpler and more effective method to identify individuals at high risk of OSA early.

Data sources

NHANES conducted a comprehensive assessment of nutrition and health status across the entire U.S. population through physical examinations and interviews. This survey involved demographic information, dietary assessments, medical history interviews, as well as laboratory and physical examinations. NHANES was approved by the Ethics Review Board of the National Center for Health Statistics (NCHS) and adhered to the ethical standards outlined in the Declaration of Helsinki and its subsequent amendments. In addition, each participant provided written informed consent voluntarily.

Study participants

Due to limitations in the sleep questionnaire, participants were recruited from the NHANES database for four cycles: 2005–2008 and 2015–2018. A total of 39,722 individuals were included, from which 22,202 participants older than 20 years were selected. Diabetes was diagnosed if participants met any of the following criteria: 1) self-reported history of diabetes; 2) use of hypoglycemic medication to maintain normal blood glucose levels; 3) FBG ≥ 7.0 mmol/l; or 4) glycosylated hemoglobin A1c (HbA1c) ≥ 6.5%. What we can see in Fig. 1, a final cohort of 6,633 participants was enrolled after exclusions based on the following criteria: 1) pregnant women (N = 521) or participants with abnormal energy intake (total energy intake < 500 kcal or ≥ 5000 kcal per day, N = 248); 2) those lacking a diagnosis of OSA (N = 10) or with missing data on HDL-C, TG, BMI, and FINS levels (N = 12,577); 3) participants with a weight equal to zero (N = 524) or diagnosed with diabetes (N = 1,676); and 4) individuals missing data on physical activity (PA), alcohol consumption and family income to poverty ratio (PIR) were included and categorized as the "unknown" group due to their higher frequency. Participants missing other covariates were excluded (N = 13).

Evaluation of METS-IR and other insulin resistance surrogate indices

In this study, the exposure variables included four alternative indicators of IR: HOMA-IR, TG/HDL-C ratio, TyG index, and METS-IR. METS-IR, TyG index, and TG/HDL-C ratio were calculated on account of FBG, TG, HDL-C levels, and BMI, independent of FINS levels. The calculation of HOMA-IR, however, required both FINS and FBG levels. The specific calculation formulas were as follows:

METS-IR = (Ln[(2*FBG) + TG]*BMI)/[Ln(HDL-C)](16)

TyG index = Ln[TG*FBG/2](38)

TG/HDL-C ratio = TG/HDL-C(39)

(The units for the above formulas of FBG, TG, and HDL-C are mg/dL, while BMI is measured in kg/m²)

HOMA-IR = FPG (mmol/L)*FINS (mIU/L) /22.5(40)

Diagnosis of obstructive sleep apnea

According to previous studies, participants were diagnosed with OSA if they met more than one of the following criteria: 1) self-reported snoring more than 3 nights per week; 2) feeling excessively sleepy 16–30 times per month despite having ≥ 7 hours of nighttime sleep, whether on weekdays or weekends; and 3) self-reported episodes of apnea, wheezing, or snoring more than 3 nights per week(35, 41).

Covariates

We refer to the covariates of previous similar studies. Sociodemographic variables included age (continuous), sex (men and women), race (non-Hispanic White, non-Hispanic Black, Mexican American, other), marital status (married/living with partner, divorced/separated/widowed, never married), PIR (< 1.3, 1.3–3.5, ≥ 3.5, unknown), and educational attainment (less than high school, high school graduate, higher than high school). BMI was treated as a continuous variable. Lifestyle habits included smoking status (never smoking, previous smoking, current smoking) and drinking status categorized as never drinker, past drinker, current drinker, and unknown. Current drinkers were further categorized by alcohol intake: mild (men: ≤ 2 cups/day, women: ≤ 1 cup/day or ≤ 1 binge per month), moderate (men: ≤ 3 cups/day, women: ≤ 2 cups/day or 2–5 binge per month), and heavy (men: ≥ 4 cups/day, women: ≥ 3 cups/day or ≥ 5 binge per month). Metabolic equivalents (MET) were used to quantify PA, calculated as recommended MET*total weekly activity time. Participants were classified into three groups using a cutoff of 600 MET-min/week: low activity, high activity, and unknown.

A diagnosis of hypertension was considered if the average blood pressure, measured at the Mobile Examination Center, met the criteria for hypertension (average systolic blood pressure ≥ 140 mmHg or diastolic blood pressure ≥ 90 mmHg), or if they reported being informed by a professional of hypertension or were taking antihypertensive medication. The diagnosis of CVD depended on self-reported history including stroke, coronary artery disease, angina pectoris, myocardial infarction, heart failure and so on.

.

Statistical analysis

Following NHANES practice guidelines, appropriate weights were applied to account for the complex stratified sampling approach, aiming to generalize statistical results to the entire U.S. noninstitutionalized population. The fasting subsample sampling weight (wtsaf2 year) was adjusted as wtsaf2 year/4 to adhere to the principle of least-subsample weighting. Independent sample t-tests (for continuous variables) and chi-square tests (for categorical variables) were used to show the baseline characteristics of participants with and without OSA. Continuous variables were reported as means with standard errors, while categorical variables were presented as frequencies and percentages. Multifactorial logistic regression analysis was employed to investigate the relationship between METS-IR and other indicators of IR with the prevalence of OSA. Three models were constructed: Model 1 was a crude adjusted model; Model 2 was partially adjusted for age, sex and race,; and Model 3 was fully adjusted, incorporating covariates such as race, gender, age, PIR, marital status, smoking and drinking status, education level, PA, BMI (excluding METS-IR), CVD, and hypertension. Nonlinear relationships were assessed using restricted cubic spline (RCS) curves based on covariates adjusted in Model 3. The area under the curve (AUC) represented the diagnostic value of these IR indicators for OSA, in the receiver operating characteristic (ROC) curves. On the basis of age, sex, race, body mass index (BMI), hypertension, and cardiovascular disease (CVD), subgroup analyses were performed to test the robustness of the relationship between METS-IR and OSA.

Comparison of Baseline Characteristics

A total of 6,633 non-diabetic participants (3,234 men and 3,399 women) were included. The mean METS-IR was 41.48 (0.25), and the prevalence of OSA amounted to 28.40%. As shown in Table 1, participants were stratified into OSA and non-OSA groups to compare their baseline characteristics. Individuals in the OSA group exhibited higher levels of METS-IR, TG/HDL-C, TyG index and HOMA-IR compared to the non-OSA group. Additionally, significant differences were observed in variables such as age, BMI, gender, marital status, education level, smoking status, alcohol consumption, and hypertension between the two groups.

Table 1

Weighted comparison of baseline characteristics
Variables	Total (N = 6,633)	Non-OSA (N = 4,749)	OSA (N = 1,884)	P-value
Age (years)	45.60(0.34)	45.16(0.40)	46.67(0.44)	0.004
Gender (%)				< 0.0001
Female	3399(51.68)	2530(54.53)	869(44.78)
Male	3234(48.32)	2219(45.48)	1015(55.22)
Race (%)				0.551
Non-Hispanic Black	1310(10.39)	935(10.31)	375(10.57)
Non-Hispanic White	2864(67.91)	2009(67.52)	855(68.84)
Mexican American	1076(8.23)	794(8.45)	282(7.71)
Other	1383(13.47)	1011(13.71)	372(12.88)
Marital status (%)				< 0.0001
Never married	1237(18.59)	942(20.23)	295(14.63)
Separated	1346(17.18)	997(17.81)	349(15.64)
Married	4050(64.23)	2810(61.96)	1240(69.73)
Education (%)				0.019
Below high school	635( 5.05)	478(5.28)	157(4.48)
High school graduate	2416(33.88)	1712(32.43)	704(37.40)
Above high school	3582(61.07)	2559(62.29)	1023(58.12)
PIR (%)				0.051
<1.3	1660(17.39)	1187(17.23)	473(17.79)
1.3–3.5	2417(34.05)	1692(32.92)	725(36.78)
≥3.5	1992(42.12)	1443(43.07)	549(39.80)
Unknown	564( 6.45)	427(6.78)	137(5.64)
PA (%)				0.099
<600	1343(21.21)	947(20.44)	396(23.08)
≥600	3770(60.69)	2711(61.62)	1059(58.45)
Unknown	1520(18.10)	1091(17.95)	429(18.47)
Drinking status (%)				0.031
Never	786(9.17)	612(10.10)	174(6.91)
Former	803(9.92)	577(10.02)	226(9.69)
Mild	2140(35.45)	1501(34.95)	639(36.65)
Moderate	981(16.36)	701(16.63)	280(15.69)
Heavy	1264(20.73)	884(19.98)	380(22.55)
Unknown	659(8.37)	474(8.31)	185(8.52)
Smoking status (%)				< 0.0001
Never	3698(54.47)	2765(57.11)	933(48.08)
Former	1574(24.66)	1106(24.21)	468(25.78)
Now	1361(20.86)	878(18.69)	483(26.15)
BMI (kg/m²)	28.41(0.14)	27.54(0.16)	30.52(0.19)	< 0.0001
Hypertension (%)				< 0.0001
No	4221(67.89)	3132(70.63)	1089(61.24)
Yes	2412(32.11)	1617(29.37)	795(38.76)
CVD (%)				0.054
No	6071(93.39)	4389(93.87)	1682(92.22)
Yes	562(6.61)	360(6.13)	202(7.78)
HDL-C (mg/dL)	55.79(0.34)	57.27(0.40)	52.21(0.50)	< 0.0001
TG (mg/dL)	117.82(1.32)	111.23(1.50)	133.79(2.39)	< 0.0001
METS-IR	41.48(0.25)	39.76(0.28)	45.65(0.36)	< 0.0001
TG/HDL-C	2.53(0.04)	2.31(0.04)	3.05(0.08)	< 0.0001
TyG index	8.49(0.01)	8.43(0.01)	8.63(0.02)	< 0.0001
HOMA-IR	2.74(0.05)	2.49(0.05)	3.36(0.08)	< 0.0001
Comparisons of categorical variables using the chi-square test. Continuous variables were compared using t-test.
BMI: body mass index; PIR: poverty income ratio; PA: physical activity; CVD: cardiovascular disease; OSA: obstructive sleep apnea; HDL-C: high-density lipoprotein cholesterol; TG: triglyceride; METS-IR: metabolic score for insulin resistance; TyG index: triglyceride glucose index; HOMA-IR: homeostatic model assessment of insulin resistance.
.

Association of METS-IR and other surrogate indices of insulin resistance with OSA

As we can see in Table 2, all four IR indices were positively related to the prevalence of OSA across three multifactorial logistic regression models. In model 3, for each unit increase in METS-IR, TyG index, TG/HDL-C, and HOMA-IR, the prevalence of OSA increased by 4%, 31%, 4%, and 5%, respectively (METS-IR: OR = 1.04, 95% CI: 1.04, 1.05, P-value < 0.0001; TyG index: OR = 1.31, 95% CI: 1.14, 1.51, P-value < 0.001; TG/HDL-C: OR = 1.04, 95% CI: 1.01, 1.07, P-value = 0.005; HOMA-IR: OR = 1.05, 95% CI: 1.02, 1.09, P-value = 0.002). After adjusting for all potential confounders, participants in the highest tertile of METS-IR had a prevalence of OSA 3.22 times higher than those in the lowest tertile (METS-IR: OR = 3.22, 95% CI: 2.67, 3.89, P-value < 0.0001). Similarly, participants in the highest tertiles of TyG index, TG/HDL-C, and HOMA-IR had ORs for OSA of 1.45, 1.42, and 1.51, respectively, compared to those in the lowest tertiles (TyG index: OR = 1.45, 95% CI: 1.18, 1.78, P-value < 0.001; TG/HDL-C: OR = 1.42, 95% CI: 1.15, 1.76, P-value = 0.002; HOMA-IR: OR = 1.51, 95% CI: 1.24, 1.85, P-value < 0.001). These findings indicate that METS-IR exhibits the strongest correlation with OSA prevalence.

Table 2

Weighted logistic regression for association between METS-IR, TG/HDL-C, TyG index, HOMA-IR and the prevalence of OSA
Exposures	Model1 [OR (95% CI) P-value]	Model2 [OR (95% CI) P-value]	Model3 [OR (95% CI) P-value]
METS-IR (Continuous)	1.04(1.04,1.05) < 0.0001	1.04(1.04,1.05) < 0.0001	1.04(1.04,1.05) < 0.0001
METS-IR (Tertiles)
T1 (≤ 35.73)	ref	ref	ref
T2 (35.73–45.04)	1.92(1.57,2.35) < 0.0001	1.81(1.48,2.23) < 0.0001	1.82(1.47,2.24) < 0.0001
T3 (> 45.04)	3.40(2.80,4.11) < 0.0001	3.28(2.70,3.97) < 0.0001	3.22(2.67,3.89) < 0.0001
P for trend	< 0.0001	< 0.0001	< 0.0001
TG/HDL-C (Continuous)	1.10(1.07,1.13) < 0.0001	1.09(1.05,1.12) < 0.0001	1.04(1.01,1.07) 0.005
TG/HDL-C (Tertiles)
T1 (≤ 1.29)	ref	ref	ref
T2 (1.29–2.50)	1.63(1.35,1.98) < 0.0001	1.59(1.30,1.94) < 0.0001	1.28(1.05,1.56) 0.015
T3 (> 2.50)	2.24(1.88,2.67) < 0.0001	2.13(1.75,2.58) < 0.0001	1.42(1.15,1.76) 0.002
P for trend	< 0.0001	< 0.0001	0.002
TyG index (Continuous)	1.74(1.56,1.94) < 0.0001	1.68(1.49,1.90) < 0.0001	1.31(1.14,1.51) < 0.001
TyG index (Tertiles)
T1 (≤ 8.22)	ref	ref	ref
T2 (8.22–8.73)	1.63(1.35,1.96) < 0.0001	1.58(1.31,1.91) < 0.0001	1.26(1.03,1.55) 0.026
T3 (> 8.73)	2.21(1.88,2.60) < 0.0001	2.11(1.77,2.52) < 0.0001	1.45(1.18,1.78) < 0.001
P for trend	< 0.0001	< 0.0001	< 0.001
HOMA-IR (Continuous)	1.15(1.12,1.19) < 0.0001	1.15(1.12,1.18) < 0.0001	1.05(1.02,1.09) 0.002
HOMA-IR (Tertiles)
T1 (≤ 1.61)	ref	ref	ref
T2 (1.61–2.98)	1.51(1.24,1.84) < 0.0001	1.50(1.24,1.83) < 0.001	1.25(1.00,1.55) 0.048
T3 (> 2.98)	2.48(2.11,2.92) < 0.0001	2.43(2.06,2.87) < 0.0001	1.51(1.24,1.85) < 0.001
P for trend	< 0.0001	< 0.0001	< 0.001
Model 1: Adjusted for no variables.
Model 2: Adjusted for race, gender, and age.
Model 3: Adjusted for gender, age, race, PIR, alcohol consumption, smoking status, marital status, PA, BMI (exception of METS-IR), education level, hypertension, and CVD.

In the RCS curves adjusting for all potential confounders, the associations of METS-IR, HOMA-IR, and TG/HDL-C with OSA were nonlinear (P for non-linearity < 0.05) (Fig. 2A, B, D), while the association of TyG index with OSA was linear (P for non-linearity > 0.05) (Fig. 2C).

Comparing the diagnostic value of different surrogate indices of IR for OSA

Further weighted ROC curves were plotted to assess the diagnostic value of various surrogate indicators of IR for OSA. In diagnosing OSA, METS-IR exhibited the highest accuracy with an AUC of 0.65. The diagnostic values of the other indices in descending order were HOMA-IR (AUC = 0.61), TG/HDL-C (AUC = 0.60), and TyG index (AUC = 0.59) (Fig. 3).

Subgroup analysis

Subgroup analyses and interaction tests examinate the robustness of the correlation between METS-IR and OSA. As depicted in Fig. 4, the correlation between METS-IR and OSA was not influenced by sex, age, race, BMI, hypertension, or CVD (P for interaction > 0.05).

This groundbreaking study investigates, for the first time, the correlation between METS-IR and OSA. We conducted large cross-sectional study of U.S. adults and found that increased levels of METS-IR were related to a higher prevalence of OSA, surpassing the correlations observed with the TyG index, HOMA-IR and TG/HDL-C. The RCS curves demonstrated a nonlinear positive correlation between METS-IR levels and OSA. METS-IR also exhibited superior diagnostic accuracy for OSA compared to the other three surrogate indicators of IR. Furthermore, there was no significant interaction of METS-IR with gender, age, race, BMI, hypertension, and CVD.

OSA is a prevalent worldwide sleep disorder that poses a significant public health threat. There are evidence suggesting a bidirectional association between OSA and MetS as well as diabetes(23–25). Large retrospective cohort studies have indicated that diabetic patients face a 48–53% increased risk of OSA compared to non-diabetic individuals(42, 43). Conversely, the risk of diabetes among those with OSA is 2.06 times higher than in those without OSA(43). IR has an important effect on the pathogenesis of MetS and diabetes, sparking interest among researchers in exploring its potential relationship with OSA. In prospective studies involving diabetic patients, those requiring insulin to maintain normoglycemia demonstrated a higher risk of OSA compared to those not needing insulin therapy(42, 43). Moreover, intensive glucose-lowering therapy has shown benefit in reducing OSA severity, as evidenced by decreased apnea-hypopnea index (AHI) and reduced sleep time with oxygen saturation below 90%(44). These findings suggest that IR may contribute to the risk of OSA. A prospective cohort study highlighted that increased FINS levels, but not HbA1c, correlated with an increased risk of OSA(45). A study conducted by Balkau et al. found that each unit increase in FINS or HOMA-IR was connected with a 31% increase in the risk of incident OSA after logarithmic transformation(36). These studies demonstrated that IR may have a significant influence on the pathogenesis of OSA. Minimum oxyhemoglobin saturation (SO2) and the AHI are commonly used to assess OSA severity. A study involving Japanese subjects identified a positive correlation between HOMA-IR and AHI and SO2(37). In addition, fragmented sleep, a hallmark of OSA, has been linked to decreased IR(30). Moderate to severe OSA patients receiving continuous positive airway pressure (CPAP) treatment can reduce the risk of developing type 2 diabetes from 2.51 times to 1.35 times.(46). Taken together, these results reveal a potential bidirectional relationship between OSA and IR.

FINS presents challenges for generalization in developing and remote areas, and is susceptible to exogenous insulin and islet insufficiency, leading to the limitations of HOMA-IR(13). Consequently, researchers have turned their attention to IR assessment metrics independent of insulin levels. The TyG index, a novel non-insulin-dependent metric, has gained popularity for predicting CVD and its prognosis. Wang et al. suggest a significant positive association between the TyG index and OSA(33), a finding consistent across Korean and non-obese, non-diabetic populations(34, 47). Another non-insulin IR metric, VAI, has emerged as a potential predictor of OSA, particularly among individuals under 60 years old(35). In 2018, Bello-Chavolla et al. introduced the METS-IR concept, integrating TG, HDL-C, FBG, and BMI, demonstrating its superiority in predicting diabetes risk through prospective cohort studies(16). METS-IR is increasingly recognized for its efficiency, simplicity, and affordability in IR assessment. Zeng et al. observed a 2.89-fold increase in hypertension prevalence among participants in the highest quartile of METS-IR compared to the lowest quartile, suggesting its potential utility in assessing hypertension risk(17). Additionally, several studies have independently linked higher METS-IR levels with cancer, depression, and CVD(18, 48, 49). METS-IR demonstrates superior diagnostic efficacy over TyG and TG/HDL-C index by effectively accounting for the impact of overnutrition on IR(16). A recent prospective cohort study conducted in the U.S. revealed stronger associations of METS-IR with cardiovascular and all-cause mortality in general population compared to the TyG index, HOMA-IR, and TG/HDL-C (21). METS-IR illustrated superior predictive accuracy for major adverse cardiovascular events relative to the TyG index and its variants (TyG-BMI, TyG-WC) as well as TG/HDL-C, especially in patients diagnosed with both MetS and heart failure(50). In non-diabetic populations, METS-IR shows the strongest correlation with hyperuricemia and performs with optimal diagnostic accuracy(22). Therefore, the present study focused on METS-IR. Consistent with findings from previous studies, our research confirms a potential positive association between METS-IR and OSA. Moreover, METS-IR exhibited superior diagnostic utility compared to the TyG index, TG/HDL-C, and HOMA-IR. Additionally, variables such as sex, age, race, BMI, hypertension, and CVD did not show significant interactions, highlighting the broad applicability of METS-IR in non-diabetic populations. These findings indicate that improving IR may contribute to reducing the risk of OSA to some extent. Furthermore, METS-IR provides a simpler, inexpensive and more accurate method to early identify individuals at high risk of OSA in clinical practice.

However, the specific mechanisms underlying the increased prevalence of IR and OSA remain unclear and may involve several aspects. On the one side, IR potentially increases the risk of OSA. At the animal level, mice with diabetes or concurrent IR exhibit long-term chronic hyperglycemia, which desensitizes the CB to hypoxic and hypercapnic stimuli, thereby reducing respiration and impairing ventilation(51, 52). This ventilation abnormality can be effectively reversed with medications that counteract IR, such as metformin(53). Clinically, in non-diabetic obese women, the severity of IR relates with the degree of upper airway collapse during sleep(54). In addition, the prevalence of OSA in diabetic patients with autonomic neuropathy (AN) was higher than that in those without AN(55). The severity of IH in obese diabetic patients also associates with AN(56), suggesting that chronic hyperglycemia and AN in the context of IR may contribute to an increased risk of OSA. On the other side, OSA can improve IR and abnormalities in glucose-lipid metabolism. IH and fragmented sleep patterns are hallmark features of OSA and are considered primary contributors to IR in these patients(57–59). Specifically, nocturnal IH inhibits insulin receptor expression in peripheral tissues such as adipose and skeletal muscle cells(60), triggering pathological processes like oxidative stress and inflammation through repeated cycles of hypoxia and reoxygenation. These processes can lead to apoptosis of pancreatic islet β-cells, ultimately resulting in decreased insulin sensitivity and elevated glucose levels(61–63). Furthermore, fragmented sleep exacerbates IR by stimulating sympathetic nerves and activating the adrenocortical axis(30). Hypoxia-inducible factor 1 (HIF-1) serves as a critical regulator in the body's adaptive response to low oxygen levels(64). Hyperglycemia or hyperlipidemia create a hypoxic environment within cells(65, 66), inhibiting the stability and transcriptional activity of HIF-1α, leading to impaired adaptation to hypoxic conditions and compromised ventilation(67–69). Conversely, knocking down HIF-1α in adipocytes significantly improves IR in mice(66).

Limitations and strengths

The participants in this study were exclusively from NHANES, ensuring a large sample size and reliable sample quality. Additionally, we weighted all samples to ensure the findings are applicable to a broad spectrum of U.S. noninstitutionalized residents. Further subgroup analyses could help assess the generalizability of this association across different populations. However, there existed serval limitations in our study. First of all, this was cross-sectional study and it is difficult to identify the causal relationship between METS-IR and OSA. Secondly, the diagnosis of OSA relied on a sleep questionnaire, which may introduce bias; however, previous research on OSA has shown the reliability of this diagnostic approach. Lastly, the study was confined to adult Americans, necessitating validation in ethnically diverse populations for broader generalization of the findings.

In conclusion, the METS-IR index identified a positive association with the prevalence of OSA and demonstrated superior diagnostic value for OSA compared to the HOMA-IR, TyG index, and TG/HDL-C. Therefore, active management of IR may deserve to be recommended for individuals at high risk of OSA. In comparison with other IR indices, the METS-IR index appears to be more suitable for early identifying individuals at high risk of OSA in clinical practice.

OSA: obstructive sleep apnea; IH: intermittent hypoxemia; CVD: cardiovascular disease; IR: insulin resistance; HEC: high insulin normoglycemic clamp; HOMA-IR: classical homeostasis model assessment of insulin resistance; FINS: fasting insulin; TG/HDL-C: triglyceride to high-density lipoprotein cholesterol; TyG: triglyceride glucose; BMI: body mass index; FBG: fasting blood glucose; HDL-C: high-density lipoprotein cholesterol; TG: triglycerides; MetS: metabolic syndrome; CB: carotid body; IH: intermittent hypoxia; VAI: Visceral Adiposity Index; NHANES: National Health and Nutrition Examination Survey; NCHS: National Center for Health Statistics; HbA1c: glycosylated hemoglobin A1c; PIR: family income to poverty ratio; PA: physical activity; MET: metabolic equivalents; wtsaf2yr: fasting subsample sampling weight; RCS: restricted cubic spline; ROC: receiver operating characteristic; AUC: area under the curve; AHI: apnea-hypopnea index; SO2: oxyhemoglobin saturation; CPAP: continuous positive airway pressure; AN: autonomic neuropathy; HIF-1: Hypoxia-inducible factor.

Data Availability

The raw data for this article are publicly available in the NHANES repository (http://www.cdc.gov/nchs/nhanes/).

Author Contributions

H.Y.Y and W.H were responsible for processing and visualizing the data, and drafting the original manuscript. B.J.Y contributed to the design of the study and reviewed the manuscript. All authors read and approved the final manuscript.

Funding

None.

Acknowledgments

We sincerely appreciate the staff and participants involved in NHANES.

Conflicts of Interest

The authors declare no conflict of interests.

Ethical Approval

NHANES is a publicly available database, approved by the Institutional Review Board of the National Center for Health Statistics. All participants provided written informed consent when conducting a national survey in the United States. Since this study was a secondary analysis, ethical review and approval were exempted.

Lloyd-Jones DM, Allen NB, Anderson CAM, Black T, Brewer LC, Foraker RE, et al. Life's Essential 8: Updating and Enhancing the American Heart Association's Construct of Cardiovascular Health: A Presidential Advisory From the American Heart Association. Circulation. 2022;146(5):e18-e43.
Benjafield AV, Ayas NT, Eastwood PR, Heinzer R, Ip MSM, Morrell MJ, et al. Estimation of the global prevalence and burden of obstructive sleep apnoea: a literature-based analysis. Lancet Respir Med. 2019;7(8):687-98.
Senaratna CV, Perret JL, Lodge CJ, Lowe AJ, Campbell BE, Matheson MC, et al. Prevalence of obstructive sleep apnea in the general population: A systematic review. Sleep Med Rev. 2017;34:70-81.
Osman AM, Carter SG, Carberry JC, Eckert DJ. Obstructive sleep apnea: current perspectives. Nat Sci Sleep. 2018;10:21-34.
Redline S, Azarbarzin A, Peker Y. Obstructive sleep apnoea heterogeneity and cardiovascular disease. Nat Rev Cardiol. 2023;20(8):560-73.
Hirotsu C, Haba-Rubio J, Togeiro SM, Marques-Vidal P, Drager LF, Vollenweider P, et al. Obstructive sleep apnoea as a risk factor for incident metabolic syndrome: a joined Episono and HypnoLaus prospective cohorts study. Eur Respir J. 2018;52(5).
Kendzerska T, Gershon AS, Hawker G, Tomlinson G, Leung RS. Obstructive sleep apnea and incident diabetes. A historical cohort study. Am J Respir Crit Care Med. 2014;190(2):218-25.
Lechat B, Appleton S, Melaku YA, Hansen K, McEvoy RD, Adams R, et al. Comorbid insomnia and sleep apnoea is associated with all-cause mortality. Eur Respir J. 2022;60(1).
Trzepizur W, Blanchard M, Ganem T, Balusson F, Feuilloy M, Girault JM, et al. Sleep Apnea-Specific Hypoxic Burden, Symptom Subtypes, and Risk of Cardiovascular Events and All-Cause Mortality. Am J Resp Crit Care. 2022;205(1):108-17.
Teran-Santos J, Jimenez-Gomez A, Cordero-Guevara J. The association between sleep apnea and the risk of traffic accidents. Cooperative Group Burgos-Santander. N Engl J Med. 1999;340(11):847-51.
Brown AE, Walker M. Genetics of Insulin Resistance and the Metabolic Syndrome. Curr Cardiol Rep. 2016;18(8):75.
DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol. 1979;237(3):E214-23.
Minh HV, Tien HA, Sinh CT, Thang DC, Chen CH, Tay JC, et al. Assessment of preferred methods to measure insulin resistance in Asian patients with hypertension. J Clin Hypertens. 2021;23(3):529-37.
Wang SJ, Shi J, Peng Y, Fang QH, Mu Q, Gu WQ, et al. Stronger association of triglyceride glucose index than the HOMA-IR with arterial stiffness in patients with type 2 diabetes: a real-world single-centre study. Cardiovasc Diabetol. 2021;20(1).
Chen JQ, Wu KX, Lin YY, Huang MY, Xie SH. Association of triglyceride glucose index with all-cause and cardiovascular mortality in the general population. Cardiovasc Diabetol. 2023;22(1).
Bello-Chavolla OY, Almeda-Valdes P, Gomez-Velasco D, Viveros-Ruiz T, Cruz-Bautista I, Romo-Romo A, et al. METS-IR, a novel score to evaluate insulin sensitivity, is predictive of visceral adiposity and incident type 2 diabetes. Eur J Endocrinol. 2018;178(5):533-44.
Zeng J, Zhang TT, Yang Y, Wang JJ, Zheng D, Hou YW, et al. Association between a metabolic score for insulin resistance and hypertension: results from National Health and Nutrition Examination Survey 2007-2016 analyses. Front Endocrinol. 2024;15.
Huang Q, Wang DH, Chen SS, Tang L, Ma CY. Association of METS-IR index with depressive symptoms in US adults: A cross-sectional study. J Affect Disorders. 2024;355:355-62.
Qian TM, Sheng XM, Shen PY, Fang YT, Deng Y, Zou GL. Mets-IR as a predictor of cardiovascular events in the middle-aged and elderly population and mediator role of blood lipids. Front Endocrinol. 2023;14.
Li CM, Xu J. Negative correlation between metabolic score for insulin resistance index and testosterone in male adults. Diabetol Metab Syndr. 2024;16(1).
Duan M, Zhao X, Li S, Miao G, Bai L, Zhang Q, et al. Metabolic score for insulin resistance (METS-IR) predicts all-cause and cardiovascular mortality in the general population: evidence from NHANES 2001-2018. Cardiovasc Diabetol. 2024;23(1):243.
Wang H, Zhang J, Pu YZ, Qin SM, Liu H, Tian YM, et al. Comparison of different insulin resistance surrogates to predict hyperuricemia among US non-diabetic adults. Front Endocrinol. 2022;13.
Reutrakul S, Mokhlesi B. Obstructive Sleep Apnea and Diabetes A State of the Art Review. Chest. 2017;152(5):1070-86.
Song SO, He K, Narla RR, Kang HG, Ryu HU, Boyko EJ. Metabolic Consequences of Obstructive Sleep Apnea Especially Pertaining to Diabetes Mellitus and Insulin Sensitivity. Diabetes Metab J. 2019;43(2):144-55.
Heffernan A, Duplancic D, Kumric M, Kurir TT, Bozic J. Metabolic Crossroads: Unveiling the Complex Interactions between Obstructive Sleep Apnoea and Metabolic Syndrome. International Journal of Molecular Sciences. 2024;25(6).
Lee SH, Park SY, Choi CS. Insulin Resistance: From Mechanisms to Therapeutic Strategies. Diabetes Metab J. 2022;46(1):15-37.
Lebovitz HE. Insulin resistance: definition and consequences. Exp Clin Endocr Diab. 2001;109:S135-S48.
Hill MA, Yang Y, Zhang LP, Sun Z, Jia GH, Parrish AR, et al. Insulin resistance, cardiovascular stiffening and cardiovascular disease. Metabolism-Clinical and Experimental. 2021;119.
Prabhakar NR, Peng YJ, Nanduri J. Hypoxia-inducible factors and obstructive sleep apnea. J Clin Invest. 2020;130(10):5042-51.
Stamatakis KA, Punjabi NM. Effects of sleep fragmentation on glucose metabolism in normal subjects. Chest. 2010;137(1):95-101.
Herzog N, Jauch-Chara K, Hyzy F, Richter A, Friedrich A, Benedict C, et al. Selective slow wave sleep but not rapid eye movement sleep suppression impairs morning glucose tolerance in healthy men. Psychoneuroendocrinology. 2013;38(10):2075-82.
Mangas-Moro A, Casitas R, Sanchez-Sanchez B, Fernandez-Navarro I, Fernandez-Lahera J, Galera R, et al. Characteristics of obstructive sleep apnea related to insulin resistance. Sleep Breath. 2024.
Wang C, Shi MD, Lin CS, Wang JY, Xie LZ, Li Y. Association between the triglyceride glucose index and obstructive sleep apnea and its symptoms: results from the NHANES. Lipids in Health and Disease. 2024;23(1).
Bikov A, Frent S, Meszaros M, Kunos L, Mathioudakis A, Negru AG, et al. Triglyceride-glucose index in non-diabetic, non-obese patients with obstructive sleep apnoea. Eur Respir J. 2021;58.
Zhou TF, Chen SH, Mao JS, Zhu P, Yu XR, Lin RY. Association between obstructive sleep apnea and visceral adiposity index and lipid accumulation product: NHANES 2015-2018. Lipids in Health and Disease. 2024;23(1).
Balkau B, Vol S, Loko S, Andriamboavonjy T, Lantieri O, Gusto G, et al. High baseline insulin levels associated with 6-year incident observed sleep apnea. Diabetes Care. 2010;33(5):1044-9.
Tomo Y, Naito R, Tomita Y, Kasagi S, Sato T, Kasai T. The Correlation between the Severity of Obstructive Sleep Apnea and Insulin Resistance in a Japanese Population. J Clin Med. 2024;13(11).
Ramdas Nayak VK, Satheesh P, Shenoy MT, Kalra S. Triglyceride Glucose (TyG) Index: A surrogate biomarker of insulin resistance. J Pak Med Assoc. 2022;72(5):986-8.
Abbasi F, Reaven GM. Comparison of two methods using plasma triglyceride concentration as a surrogate estimate of insulin action in nondiabetic subjects: triglycerides x glucose versus triglyceride/high-density lipoprotein cholesterol. Metabolism-Clinical and Experimental. 2011;60(12):1673-6.
Meyer C, Szoke E, Pimenta W, Mitrakou A, Woerle HJ, Gerich J, et al. Different mechanisms for impaired fasting glucose and impaired postprandial glucose tolerance in humans. Diabetes Care. 2006;29(8):1909-14.
Cavallino V, Rankin E, Popescu A, Gopang M, Hale L, Meliker JR. Antimony and sleep health outcomes: NHANES 2009-2016. Sleep Health. 2022;8(4):373-9.
Subramanian A, Adderley NJ, Tracy A, Taverner T, Hanif W, Toulis KA, et al. Risk of Incident Obstructive Sleep Apnea Among Patients With Type 2 Diabetes. Diabetes Care. 2019;42(5):954-63.
Huang TY, Lin BM, Stampfer MJ, Tworoger SS, Hu FB, Redline S. A Population-Based Study of the Bidirectional Association Between Obstructive Sleep Apnea and Type 2 Diabetes in Three Prospective US Cohorts. Diabetes Care. 2018;41(10):2111-9.
Gutiérrez-Carrasquilla L, López-Cano C, Sánchez E, Barbé F, Dalmases M, Hernández M, et al. Effect of Glucose Improvement on Nocturnal Sleep Breathing Parameters in Patients with Type 2 Diabetes: The Candy Dreams Study. J Clin Med. 2020;9(4).
Huang TY, Sands SA, Stampfer MJ, Tworoger SS, Hu FB, Redline S. Insulin Resistance, Hyperglycemia, and Risk of Developing Obstructive Sleep Apnea in Men and Women in the United States. Ann Am Thorac Soc. 2022;19(10):1740-9.
Xu PH, Hui CKM, Lui MMS, Lam DCL, Fong DYT, Ip MSM. Incident Type 2 Diabetes in OSA and Effect of CPAP Treatment: A Retrospective Clinic Cohort Study. Chest. 2019;156(4):743-53.
Kang HH, Kim SW, Lee SH. Association between triglyceride glucose index and obstructive sleep apnea risk in Korean adults: a cross-sectional cohort study. Lipids in Health and Disease. 2020;19(1).
Liu GL. Association between the metabolic score for insulin resistance (METS-IR) and arterial stiffness among health check-up population in Japan: a retrospective cross-sectional study. Front Endocrinol. 2024;14.
Son M, Moon SY, Koh M, Kang YW, Lee JY. Association between Surrogate Markers of Insulin Resistance and the Incidence of Colorectal Cancer in Korea: A Nationwide Population-Based Study. J Clin Med. 2024;13(6).
Xie Q, Liu C, Liu F, Zhang XH, Zhang ZY, An X, et al. Predictive Effect of Alternative Insulin Resistance Indexes on Adverse Cardiovascular Events in Patients with Metabolic Syndrome with Heart Failure. Diabet Metab Synd Ob. 2024;17:2347-56.
Polotsky VY, Wilson JA, Haines AS, Scharf MT, Soutiere SE, Tankersley CG, et al. The impact of insulin-dependent diabetes on ventilatory control in the mouse. Am J Resp Crit Care. 2001;163(3):624-32.
Weisbrod CJ, Eastwood PR, O'Driscoll G, Green DJ. Abnormal ventilatory responses to hypoxia in Type 2 diabetes. Diabetic Med. 2005;22(5):563-8.
Ramadan W, Dewasmes G, Petitjean M, Wiernsperger N, Delanaud S, Geloen A, et al. Sleep apnea is induced by a high-fat diet and reversed and prevented by metformin in non obese rats. Obesity. 2007;15(6):1409-18.
Llanos OL, Galiatsatos P, Guzmán-Vélez E, Patil SP, Smith PL, Magnuson T, et al. Pharyngeal collapsibility during sleep is elevated in insulin-resistant females with morbid obesity. Eur Respir J. 2016;47(6):1718-26.
Keller T, Hader C, De Zeeuw J, Rasche K. Obstructive sleep apnea syndrome: The effect of diabetes and autonomic neuropathy. J Physiol Pharmacol. 2007;58:313-8.
Bottini P, Redolfi S, Dottorini ML, Tantucci C. Autonomic neuropathy increases the risk of obstructive sleep apnea in obese diabetics. Respiration. 2008;75(3):265-71.
Lindberg E, Theorell-Haglöw J, Svensson M, Gislason T, Berne C, Janson C. Sleep Apnea and Glucose Metabolism A Long-term Follow-up in a Community-Based Sample. Chest. 2012;142(4):935-42.
Gabryelska A, Karuga FF, Szmyd B, Bialasiewicz P. HIF-1α as a Mediator of Insulin Resistance, T2DM, and Its Complications: Potential Links With Obstructive Sleep Apnea. Front Physiol. 2020;11.
Meszaros M, Bikov A. Obstructive Sleep Apnoea and Lipid Metabolism: The Summary of Evidence and Future Perspectives in the Pathophysiology of OSA-Associated Dyslipidaemia. Biomedicines. 2022;10(11).
Sacramento JF, Ribeiro MJ, Rodrigues T, Guarino MP, Diogo LN, Seiça R, et al. Insulin resistance is associated with tissue-specific regulation of HIF-1α and HIF-2α during mild chronic intermittent hypoxia. Resp Physiol Neurobi. 2016;228:30-8.
Polotsky VY, Li JG, Punjabi NM, Rubin AE, Smith PL, Schwartz AR, et al. Intermittent hypoxia increases insulin resistance in genetically obese mice. J Physiol-London. 2003;552(1):253-64.
Xu W, Chi L, Row BW, Xu R, Ke Y, Xu B, et al. Increased oxidative stress is associated with chronic intermittent hypoxia-mediated brain cortical neuronal cell apoptosis in a mouse model of sleep APNEA. Neuroscience. 2004;126(2):313-23.
Zeng S, Wang YY, Ai L, Huang LW, Liu ZJ, He CX, et al. Chronic intermittent hypoxia-induced oxidative stress activates TRB3 and phosphorylated JNK to mediate insulin resistance and cell apoptosis in the pancreas. Clin Exp Pharmacol P. 2024;51(3).
Semenza GL. Pharmacologic Targeting of Hypoxia-Inducible Factors. Annu Rev Pharmacol. 2019;59:379-403.
Sato Y, Endo H, Okuyama H, Takeda T, Iwahashi H, Imagawa A, et al. Cellular Hypoxia of Pancreatic β-Cells Due to High Levels of Oxygen Consumption for Insulin Secretion. Journal of Biological Chemistry. 2011;286(14).
Lee YS, Kim JW, Osborne O, Oh DY, Sasik R, Schenk S, et al. Increased Adipocyte O Consumption Triggers HIF-1α, Causing Inflammation and Insulin Resistance in Obesity. Cell. 2014;157(6):1339-52.
Dodd MS, Fialho MDS, Aparicio CNM, Kerr M, Timm KN, Griffin JL, et al. Fatty Acids Prevent Hypoxia-Inducible Factor-1α Signaling Through Decreased Succinate in Diabetes. Jacc-Basic Transl Sc. 2018;3(4):485-98.
Gu HF, Zheng XW, Abu Seman N, Gu TW, Botusan IR, Sunkari VG, et al. Impact of the Hypoxia-Inducible Factor-1 α (HIF1A) Pro582Ser Polymorphism on Diabetes Nephropathy. Diabetes Care. 2013;36(2):415-21.
Catrina SB, Okamoto K, Pereira T, Brismar K, Poellinger L. Hyperglycemia regulates hypoxia-inducible factor-1alpha protein stability and function. Diabetes. 2004;53(12):3226-32.

No competing interests reported.

Correlation between METS-IR index and obstructive sleep apnea in non- diabetic adults: evidence from NHANES 2001–2018

Status:

Version 1

Abstract

Figures

Introduction

Methods

Data sources

Study participants

Evaluation of METS-IR and other insulin resistance surrogate indices

Diagnosis of obstructive sleep apnea

Covariates

Statistical analysis

Results

Comparison of Baseline Characteristics

Association of METS-IR and other surrogate indices of insulin resistance with OSA

Comparing the diagnostic value of different surrogate indices of IR for OSA

Subgroup analysis

Discussion

Limitations and strengths

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Status:

Version 1