Higher waist-to-height ratio was associated with greater odds of obstructive sleep apnea in 2 data sets, but its screening performance differed by population, according to findings published in Frontiers in Medicine.
Researchers analyzed 3,715 adults from the 2015 to 2018 National Health and Nutrition Examination Survey and 200 patients who underwent overnight polysomnography at the Sleep Center of Tianjin Chest Hospital in China. The study compared waist-to-height ratio (WHtR), body mass index (BMI), and waist circumference as anthropometric markers of obstructive sleep apnea (OSA) risk.
The 2 cohorts differed substantially. The NHANES analysis was cross-sectional and identified OSA using self-reported symptoms, including frequent snoring, gasping or breathing cessation during sleep, or excessive daytime sleepiness despite adequate sleep duration. The Tianjin cohort was retrospective, single-center, and included patients referred for suspected OSA who underwent polysomnography.
After adjustment for age and sex or gender, each 0.1-unit increase in WHtR was associated with about 2.8 times the odds of OSA in NHANES and about 12 times the odds of OSA in the clinical cohort.
Restricted cubic spline analyses suggested nonlinear associations in both cohorts, with OSA odds increasing more rapidly around WHtR values of 0.589 in NHANES and 0.523 in the clinical cohort. These values were analytic inflection points rather than validated diagnostic thresholds.
In NHANES, WHtR, BMI, and waist circumference showed broadly similar discrimination for OSA. Waist circumference had a slightly higher area under the curve than WHtR, although the researchers described the absolute difference as small and unlikely to be clinically meaningful. WHtR did not differ significantly from BMI in the population-based analysis.
In the clinical cohort, WHtR showed stronger discrimination than BMI, with an area under the curve of 0.88 compared with 0.59 for BMI and 0.86 for waist circumference. At a Youden-index-derived WHtR cutoff of 0.501 in that cohort, WHtR had 93% sensitivity and 81% specificity for identifying patients with OSA.
The cutoff should be interpreted cautiously. The clinical cohort was composed of patients referred for suspected OSA and included equal numbers of patients with and without OSA, which may affect measures such as positive and negative predictive value. The findings may not apply to primary care or community screening populations with lower disease prevalence.
Subgroup analyses in the clinical cohort showed that WHtR had stronger predictive performance than BMI among male and female patients and among patients younger than 60 years and those aged 60 years or older. However, the researchers noted that the clinical sample size limited the precision of subgroup analyses.
Several limitations warrant consideration. The NHANES analysis could not establish causality and relied on symptom-based OSA classification rather than polysomnography, raising the possibility of disease misclassification. The analysis also excluded many participants because of missing data and did not include apnea-hypopnea index data, preventing assessment of OSA severity in the population sample.
In addition, adjustment was limited to age and sex or gender, leaving the possibility of residual confounding from comorbidities, smoking, alcohol use, metabolic factors, and other variables. The clinical cohort was small, retrospective, and drawn from a single center in China, and the WHtR cutoff requires external validation before broader clinical use.
Taken together, the findings support WHtR as a promising marker of OSA risk in sleep-center referral populations, but not yet as a replacement for BMI in existing screening questionnaires or for broader population-based screening. Whether WHtR should replace or supplement BMI in tools such as STOP-Bang, Berlin, or NoSAS remains uncertain and requires prospective head-to-head studies evaluating screening accuracy, referral decisions, and patient outcomes.
Disclosures: The researchers reported no commercial or financial conflicts of interest.
Source: Frontiers in Medicine
