Association between weight-adjusted waist-index and symptoms of sleep apnea in US adults: results from 2015–2018 national health and nutrition examination survey

Abstract

Background

Literature shows that traditional measures of obesity such as body mass index (BMI), waist-to-height ratio (WtHR), waist-to-hip ratio (WHR), and waist circumference (WC) may not be precise measures of obesity compared to weight-adjusted waist-index (WWI), a relatively new obesity index. BMI, WC, WHR, and WtHR were found to be associated with sleep apnea. However, the association between WWI and sleep apnea symptoms is not known. This study investigated the association between WWI and symptoms of sleep apnea among United States adults and compared WWI’s discriminative power with BMI, WtHR, WHR, and WC in the evaluation of symptoms of sleep apnea.

Methods

Data from the 2015–2018 National Health and Nutrition Examination Surveys were analyzed. WWI was calculated as WC (cm)/√weight (kg). A weighted multivariable logistic regression model and smooth curve fitting were used to investigate the association between WWI and sleep apnea symptoms. We generated the area under the receiver operating characteristics curve (ROC-AUC) to compare different obesity indices’ ability to distinguish those with and without sleep apnea symptoms.

Results

We analyzed data from 9,566 participants whose mean ± SD WWI was 11.03 ± 2.37 cm/√kg. The prevalence of sleep apnea symptoms was 13.10%. In the fully adjusted model, WWI was positively associated with symptoms of sleep apnea [Adjusted Odds Ratio = 1.71 95% Confidence Interval (CI): 1.32—2.21, p < 0.01]. Smoothing curve fitting showed a non-linear positive association between WWI and symptoms of sleep apnea. The ROC-AUCs were 0.63, 0.64, 0.63, 0.66, and 67 for WWI, BMI, WtHR, WC, and WHR, respectively. The association between WWI and symptoms of sleep apnea was consistent across different age groups, gender, asthma, and physical activity groups.

Conclusion

WWI was positively associated with increased odds of sleep apnea symptoms and the association showed a dose–response relationship. Our findings suggest that WWI can be used in the evaluation of symptoms of sleep apnea; however, prospective studies may be needed to confirm our findings.

Background

Sleep disorders such as sleep apnea are health conditions affecting sleep quality, duration, and timing. These conditions impact an individual’s ability to function daily while influencing health and longevity (Colten et al. 2006). Despite the evidence showing how sleep conditions contribute to chronic medical conditions (Ramos et al. 2023), these conditions have received less attention. This is concerning given that they have a major impact on overall health and quality of life (Ramar et al. 2021; Zhao et al. 2020). Sleep apnea is the second leading type of sleep disorder, with about 936 million people between the ages of 30- and 69 years having sleep apnea worldwide (Benjafield et al. 2019). The disease is common in developed nations, with prevalence estimated to be between 9 to 38% in most European and North American adults (Senaratna et al. 2017). Sleep apnea is a sleep-disordered breathing (SDB) condition in which an individual’s breathing is interrupted during sleep due to airway blockage, resulting in reduced tissue perfusion (National Library of Medicine 2021). Clinically, individuals with apnea–hypopnea-index (AHI) greater than or equal to 5 and who experience excessive daytime sleepiness, and/or breathing-related symptoms during the night are considered to have sleep apnea (American Academy of Sleep Medicine 2014).

Obesity is one of the major modifiable risk factors for sleep apnea (Peppard et al. 2000; Addo et al. 2024). Although many health conditions are associated with an increased likelihood of developing sleep apnea, the condition is most prevalent in people who are overweight or obese (Young et al. 2004). A 10% increase in weight increases the risk of sleep apnea by six-fold (Peppard et al. 2000). The prevalence of sleep apnea is estimated to be between 50 and 98% in severely obese individuals (Pillar and Shehadeh 2008), and it is likely to increase due to the obesity epidemic. Fat deposition in the upper airway reduces muscle activity, narrowing the airway, which can cause airway collapse and apnea (Pillar and Shehadeh 2008). According to the World Health Organization, 2.5 billion of the adult global population are overweight or obese (World Health Organization 2024).

Adiposity is commonly determined by body mass index [BMI (kg/m²)], with individuals with BMI ≥ 30 being considered obese (Weir and Jan 2023). However, BMI may not be an accurate anthropometric measure of adiposity because of its inability to accurately distinguish between muscle and lean mass (Kim et al. 2021; Ding et al. 2023). Moreover, its association with some chronic health conditions is not stable across different age groups (Yu et al. 2018), sex (Wang et al. 2020), and racial/ethnic groups (Stommel and Schoenborn 2010). Park et al. suggested a new anthropometric measure of obesity called weight-adjusted waist index (WWI), that reflects central/abdominal obesity as it incorporates waist circumference (WC) in centimeters (cm) and body weight in kilograms (kg) (Park et al. 2018). Studies have shown WWI to be an important predictor of several health outcomes, including hyperuricemia (Ding et al. 2023), female infertility (Wen and Li 2023), kidney stones (Lin et al. 2023), hypertension (Wang et al. 2023), non-alcoholic fatty liver disease (Yu et al. 2024), and stress urinary incontinence (Sun et al. 2024). However, to our knowledge, the association of WWI with sleep apnea has not been investigated. The aim of the present study is to 1) investigate the association of WWI with symptoms of sleep apnea using a representative sample of United States (US) adults, and 2) compare the ability of WWI to distinguish those with and without sleep apnea symptoms compared with other measures of adiposity (BMI, waist-to-height ratio (WtHR), waist-to-hip ratio (WHR), and waist circumference (WC)).

Methods

Data source

Data from the National Health and Nutrition Examination Survey (NHANES), a population-based series of cross-sectional studies of a representative sample of the US population, were used to investigate the hypothesized association between WWI and symptoms of sleep apnea. NHANES survey provides health and nutritional-related data collected through interviews and physical and laboratory examinations of US non-institutionalized civilians.

Study population

The study population consists of adults 20 years or older who participated in the survey between 2015 to 2018. The initial sample size comprised 19,225 participants. After excluding participants under 20 years (n = 7,937), pregnant participants (n = 94), participants with missing data for WC (n = 1,225), weight (n = 11), and sleep apnea symptoms (n = 392), the final sample size was 9,566 participants (Fig. 1). Pregnant women were excluded from the study considering the association of sleep disorders with pregnancy (Silvestri and Aricò 2019).

Fig. 1

Flow chart of NHANES participants selection criteria

Outcome variable

The outcome variable of interest was symptoms of sleep apnea, defined according to the Healthy People 2030 guideline and prior studies which used four NHANES sleep health questions. Participants were asked: (1) “Number of hours usually sleep on weekdays or workdays.”; (2) “In the past 12 months, how often did {you/SP} snore while {you were/s/he was} sleeping?”; (3) “In the past 12 months, how often did {you/SP} snort, gasp, or stop breathing while {you were/s/he was} asleep?”, and (4) “In the past month, how often did you feel excessively or overly sleepy during the day?”. Participants who reported “snoring 3 or more nights per week in the past 12 months” OR “snort, gasp or stop breathing 3 or more nights per week in the past 12 months” OR “felt excessively sleepy during the day almost always 16–30 times per month AND usually sleep 7 or more hours per night on weekdays or worknights” were considered to have sleep apnea symptoms (U.S. Department of Health and Human Services 2024; Cai et al. 2024).

Exposure variable

The main exposure variable of interest was WWI, calculated by dividing WC (in cm) by the square root of weight (in kg) (Park et al. 2018). Participants’ WC, height, hip circumference (HC), and body weight measurements were taken at the mobile exam center (MEC) by trained health staff using a steel measuring tape and a digital weight scale (Centers for Disease Control and Prevention 2017). We also examined traditional indicators of obesity (WtHR, WHR, BMI, and WC) to test and compare their ability to distinguish individuals with and without symptoms of sleep apnea with that of WWI. WtHR was calculated by dividing WC by height (cm/cm). WHR was calculated by dividing WC by HC (cm/cm).

Covariates

We selected covariates based on a previous study that investigated obesity and sleep problems (Palm et al. 2015). We considered age (years); gender (males/females); education (Less than High School, High School, Some college/AA degree, College graduate or above); race/ethnicity (Hispanic, Non-Hispanic White, Non-Hispanic Black, Other); asthma (Yes/No); triglycerides (mg/dL)(Tang et al. 2022); smoking status (Current, Former, Never) and the average number of alcoholic drinks/day consumed in the past 12 months. Multiracial and other races were grouped into one category called Other. Smoking status was defined as “current smoker” (smoked at least 100 cigarettes in a lifetime and currently smoke every day or some days), “former smoker” (smoked at least 100 cigarettes in a lifetime but are currently not smoking), and “never smoker” (never smoked 100 cigarettes in a lifetime). Drug abuse (cocaine/heroin/methamphetamine) was defined as a binary (yes/no) variable. We also considered physical activity as a potential confounding variable, with those spending at least 150 min of moderate aerobic activity a week considered physically active (U.S. Department of Health and Human Services 2008).

Statistical analysis

For analysis, we accounted for the NHANES complex survey design. Means and standard deviations (SD), or median and interquartile (IQR) range, were reported for continuous variables as appropriate. Frequencies (or percentages) were reported for categorical variables. We calculated WWI and categorized it by tertiles for sensitivity analysis as follows: Tertile 1 (< 10.77 cm/√kg), Tertile 2 (10.77—11.51 cm/√kg), and Tertile 3 (> 11.51 cm/√kg). WWI was analyzed both as a continuous and categorical variable. The crude association of categorical variables with WWI tertiles was tested using a weighted chi-square test. We used a weighted linear regression to test for the mean difference of continuous variables across tertiles of WWI. A weighted multivariable logistic regression was used to examine the association of WWI with sleep apnea symptoms in three separate models: Model I, only WWI was included; Model II was adjusted for age, gender, education, and race, while Model III further adjusted for asthma, drug abuse, smoking, physical activity, triglycerides, and alcohol consumption after testing for interaction and confounding. Confounding was assessed using the 10% rule, that is, if a variable changes the magnitude of the association between WWI and sleep apnea symptoms by at least 10% it was considered a confounder and therefore included in the multivariable logistic regression model. The non-linear association between WWI and sleep apnea symptoms was also investigated by fitting a smoothing curve. We performed subgroup analysis to identify the consistency of the association of WWI with sleep apnea symptoms across different age groups, gender, education, race, asthma, smoking, drug abuse, and physical activity. For subgroup analysis, we adjusted for all covariates except for other measures of obesity (weight, HC, WC, BMI, WHR, and WtHR). We used the areas under the receiver operating characteristics curve (ROC-AUR) to compare the different measures of obesity in terms of their ability to distinguish between individuals with and without sleep apnea symptoms. We performed data analysis using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) and R studio app (version 2024.04.1 + 748). A p < 0.05 was considered statistically significant.

Results

Baseline characteristics of participants by WWI

Table 1 shows various baseline characteristics of study participants by WWI. The total sample size was 9,566, with a mean WWI of 11.03 ± 2.37 cm/√kg. The prevalence of sleep apnea symptoms was 13.10%, and it increased as WWI increased: [Tertile 1: 8.66%], [Tertile 2: 13.33%] and [Tertile 3: 17.31%]. The crude associations of demographic, behavioral, clinical, and laboratory variables with WWI tertiles are presented. A statistically significant association was observed for all variables with WWI tertiles (p < 0.05 for all) except for drug abuse and asthma (p > 0.05). Individuals in the highest WWI tertile group (Tertile 3) were more likely to be 60 years or older, females, less educated, non-Hispanic white, never cigarette smoker, and less physically active. Participants in the highest WWI tertile group (Tertile 3) also had higher levels of triglycerides, weight, HC, WC, BMI, WtHR, and WHR.

Table 1 Baseline characteristics of participants by tertiles of WWI

Association between WWI and the odds of having sleep apnea symptoms

Table 2 shows that WWI is positively associated with increased odds of having sleep apnea symptoms; Model I, OR = 1.45 95% CI: 1.35, 1.56; Model II, AOR = 1.50 95% CI:1.40, 1.61, and Model III, AOR = 1.71; 95% CI: 1.32, 2.21 (p < 0.05 for all). After categorizing WWI into tertiles, the odds ratio of having sleep apnea symptoms in Tertile 3 vs. Tertile 1 (the reference) was 2.81 [95% CI: 1.71, 4.60, p = 0.01] in the fully adjusted model (Table 2). Results from the smooth curve fitting in Fig. 2 confirm the non-linear positive association between WWI and sleep apnea symptoms.

Table 2 Association of WWI with sleep apnea symptoms

Fig. 2

Illustrates a non-linear association between WWI and sleep apnea symptoms. The solid red line shows the predicted probability of sleep apnea symptoms. The blue dashed lines represent the 95% confidence interval of the probability of sleep apnea symptoms

Subgroup analysis and interaction test

Table 3 shows that the association of WWI with sleep apnea symptoms was not modified by age, gender, education, race, asthma, smoking, drug abuse, or physical activity (p > 0.05 for all). Subgroup analysis revealed a consistent association between WWI and sleep apnea symptoms across different age groups, gender, asthma, and physical activity groups.

Table 3 Subgroup analysis and interaction test for the association of WWI with sleep apnea symptoms

ROC curve analysis for each obesity index in predicting sleep apnea symptoms

Although WWI is significantly associated with an increase in the odds of sleep apnea symptoms, it did not provide better ability to distinguish individuals with and without sleep apnea symptoms than other obesity indices (BMI, WC, WtHR, WHR) as shown in Fig. 3.

Fig. 3

The receiver operating characteristics analysis for predicting sleep apnea symptoms

ROC statistics- unadjusted			ROC statistics-adjusteda
Predictor	AUC	95% CI	AUC	95% CI
WWI	0.59	0.57—0.62	0.63	0.58—0.67
BMI	0.60	0.57—0.62	0.64	0.60—0.69
WtHR	0.61	0.59—0.63	0.63	0.58—0.67
WC	0.61	0.659—0.64	0.66	0.61—0.70
WHR	0.59	0.57—0.62	0.67	0.52—0.71

1. ^aAdjusted for age, gender, education, race, asthma, triglycerides, drug abuse, smoking, physical activity and alcohol consumption. AUC, area under the curve. ROC, receiver operating characteristics

Discussion

To our knowledge, no study has explored the association between WWI and sleep apnea symptoms, and this is the first study to demonstrate the existence of this association. Our study revealed that WWI is significantly associated with an increased odds of sleep apnea symptoms. This association remained the same even after adjusting for age, gender, race, education, triglycerides, asthma, drug abuse, smoking, physical activity, and alcohol consumption. Our findings were also confirmed by a non-linear smooth curve, which showed a dose response relationship between WWI and sleep apnea symptoms. Moreover, the association between WWI and sleep apnea symptoms was consistent in different age groups, gender, asthma, and physical activity groups. Although WWI’s ability to distinguish those with and without symptoms of sleep apnea was slightly lower than that of traditional measures of obesity (BMI, WtHR, WHR, and WC), it is a reliable marker because of the consistency of its association with sleep apnea symptoms across heterogenous populations.

Although studies have reported a positive association between BMI and sleep apnea (Addo et al. 2024; Wiginder et al. 2022; Kishan et al. 2021), the main limitation of BMI, which is its failure to accurately reflect central obesity, should not be ignored. WWI, a new marker of obesity that incorporates WC and body weight, reflects central/abdominal obesity (Kim et al. 2021, 2022). Higher levels of WWI reflect a higher degree of obesity, and results from our study revealed higher odds of sleep apnea symptoms among individuals with higher levels of WWI. Our results concur with findings from prior studies that investigated abdominal obesity and sleep apnea. For instance, a study conducted in Japan reported a positive correlation of visceral fat (r = 0.50; p < 0.01) with the apnea–hypopnea index (Sekizuka et al. 2021). Likewise, a case–control study conducted in India (Patial et al. 2023), reported a larger mean percentage (%) of visceral fat in individuals with sleep apnea compared to their controls (21.18% vs. 10.57%; p < 0.01). Vgontzas et al. also found visceral fat to be correlated with sleep apnea (r = 0.70, p < 0.01) (Vgontzas et al. 2000).

The mechanism by which abdominal obesity, as measured by WWI, could lead to sleep apnea is complex. During sleep, obese individuals may have trouble breathing because of the pressure exerted on the lungs, chest, and diaphragm by abdominal visceral fat (Mafort et al. 2016). The pressure exerted reduces functional residual capacity (Salome et al. 2010), leading to reduced airflow and difficulty breathing, both of which are symptoms common in individuals with obstructive sleep apnea. Visceral fat around the abdomen secretes hormones such as adipokines and cytokines, both of which affect sleep (Wei et al. 2022; Marin et al. 2005). A study that investigated plasma cytokine levels and disorders of excessive daytime sleepiness, one of the symptoms of sleep apnea, found significantly increased levels of plasma cytokines (TNF-α and IL-6) among individuals with sleep apnea (Vgontzas et al. 2000). Sleep apnea is an important public health problem, as it is a modifiable risk factor for several serious conditions, including cardiovascular diseases (Marin et al. 2005), stroke (Valham et al. 2008), and diabetes (Reichmuth et al. 2005).

Strengths and limitations

Our study boasts several strengths that bolster the reliability of our findings. First, our results can be extrapolated to the wider US population, as our sample was drawn from NHANES, a national survey known for its rigorous data collection methods. Second, NHANES employs standardized procedures and strict quality control measures, ensuring the high quality of the collected data. Third, we meticulously adjusted for numerous potential confounders and utilized advanced statistical techniques to analyze the data. However, our study also has some limitations, including the inability to adjust for neck circumference, an important risk factor for sleep apnea in obese individuals, as it was not measured. Also, establishing a temporal relationship between WWI and sleep apnea symptoms is not possible because of the cross-sectional nature of the NHANES survey. Therefore, prospective studies are needed to help confirm our findings. Thirdly, some variables, including symptoms of sleep apnea, were measured by self-report which may have reduced the reliability of the measurements.

Conclusion

Our results support the hypothesis of a positive association of WWI with the odds of sleep apnea symptoms in a representative sample of the US adult population. Although WWI’s ability to distinguish individuals with and without sleep apnea symptoms was slightly lower than that of traditional measures of obesity (BMI, WtHR, WHR, and WC), its association with sleep apnea symptoms was consistent across different subgroup of populations indicating its reliability in the evaluation of sleep apnea symptoms across heterogenous populations. However, prospective studies may be needed to confirm our findings.