Development of a Body Mass Index-Independent (SWET) Score to Predict Moderate or Severe Obstructive Sleep Apnea in Indian Population

Although closely associated with obesity, non-obese patients are also at significant risk of obstructive sleep apnea (OSA). Burden of OSA is grossly underestimated in non-obese Asian population. Most risk scores incorporate body mass index (BMI). This study tried to develop BMI-independent score to predict high-risk OSA in Indian population.

This retrospective observational study analyzed anthropo-demographic, clinical and polysomnography data of 1898 patients. Coefficients of variables from multivariate regression analysis were used to create a BMI-independent score. Receiver operator characteristic curve analysis used to determine optimal cut-off; performance reported using sensitivity, specificity, predictive values and likelihood ratios.

Mean age was 45.9 years; 73.0% were males. Most patients had severe OSA (54.7%) and high-risk Epworth Sleepiness Scale (ESS) score (62.0%). Mean ESS value was higher for high-risk compared to mild OSA group (11.8 ± 5.6 vs. 9.2 ± 5.1, p < 0.001). Area under receiver operator curve (AUROC) to predict high-risk OSA by ESS and MBQ was 0.59 and 0.57 respectively. Sub scapular skin fold thickness (SSFT), waist-hip ratio (WHR), Epworth sleepiness score (ESS) and triceps skin fold thickness (TSFT) were significantly associated with high risk OSA and included in SWET score. SWET score ≥ 37.2 had 68.2% sensitivity and 81.8% positive predictive value to predict moderate or severe OSA with AUROC of 0.62.

The novel SWET score shows significant promise to predict risk of OSA, independent of BMI. The performance exceeded that of traditional tools like ESS and MBQ in our study population. However, it requires further validation in large prospective cohorts.