Increasing OSA severity was associated with increasing markers of obesity.
Visceral adiposity (VA) is strongly associated with obstructive sleep apnea (OSA) in elderly patients, while general adiposity is strongly associated with OSA in non-elderly patients, according to new research.
Researchers found that VA over 128 cm2 resulted in an area under the curve (AUC) of 0.580 with a sensitivity of 0.382 and specificity of 0.790 in predicting moderate-to-severe-OSA in elderly patients. In non-elderly patients, body mass index (BMI) over 24.7 kg/m2 resulted in an AUC of 0.550 with a sensitivity of 0.883 and specificity of 0.484 in predictive power.
"Although some subjects may have never been clinically categorized as overweight or obese, if they have increased VA, they may be at increased risk of OSA and its adverse health outcomes. However, this hypothesis needs yet to be tested in the elderly,” wrote senior author Xinmin Liu, MD, president and respiratory specialist, Peking University First Hospital, and colleagues.
Liu and colleagues analyzed data from 169 patients aged at least 65 years (elderly) and 142 patients aged less than 65 years (non-elderly) referred for overnight polysomnography (PSG). More than 80% of participants were men (82.3% in elderly and 84.5% in non-elderly). Mild or no OSA was present in 69 participants (41% of total; 38 in elderly and 31 in non-elderly), moderate OSA was present in 104 (62% of total; 47 in elderly and 57 in non-elderly) and severe OSA was present in 138 (82% of total; 84 in elderly and 534 in non-elderly).
Patients reported smoking (25.1%) and drinking (23.2%), as well as comorbidities such as dyslipidemia (n = 263), hypertension (n = 229), thyroid diseases (n = 171), fatty liver (n =139), coronary heart disease (n = 122), diabetes (n = 105), gastroesophageal reflux disease (n = 69), hyperuricemia (n = 78), stroke (n = 66), chronic obstructive pulmonary disease (n = 25), cancer (n = 24), and chronic kidney disease (n = 20).
Increasing OSA severity was associated with increasing markers of obesity such as BMI, waist circumference (WC), hip circumference (HC), WC/HC ratio, NC/height (H) ratio, WC/H ratio, conicity index, SA, and VA (all P <.02) but was not associated with neck circumference (NC) and VA/subcutaneous adiposity (SA) ratio.
The researchers found a significant interaction between OSA severity and age only for VA/SA (P <.005), which increased with OSA severity in the non-elderly and decreased in the elderly, although both VA and SA increased with OSA severity. Compared with the non-elderly, the elderly showed higher conicity index. VA/SA, lower BMI, NC, WC, HC and SA. WC/HC, NC/H, WC/H and VA increased with OSA severity across both groups (all P <.05).
Univariate analyses showed that BMI, WC, HC, WC/HC, WC/H, conicity index, VA and SA were significantly associated with apnea-hypopnea index (AHI) in the elderly, while BMI, NC, WC, HC, WC/HC, NC/H, WC/H, VA, SA and VA/SA were significantly associated with AHI in the non-elderly.
The strongest correlations were WC/H with AHI in the elderly (β = 0.296; P <.05) 0.5 and VA (β = 0.422; P <.01), BMI (β = 0.395; P <.01), and WC/H (β = 0.376; P <.01) in the non-elderly. After adjusting for age, sex, cigarette smoking, alcohol drinking and main comorbidities, BMI, VA and VA/SA, explained 25.9% of AHI variability in the non-elderly and 17.2% variability in the elderly.
All told, Liu and colleagues found that VA over 128 cm2 resulted in an AUC of 0.580 with a sensitivity of 0.382 and specificity of 0.790 in predicting moderate-to-severe-OSA in elderly patients. In non-elderly patients, BMI over 24.7 kg/m2 resulted in an AUC of 0.550 with a sensitivity of 0.883 and specificity of 0.484 in predictive power.
VA/SA ratio less than 0.41 resulted in an AUC of 0.553 with a sensitivity of 0.176 and specificity of 0.947 in predictive power in elderly patients. In non-elderly patients, WA/SA predicted moderate-to-severe OSA with an AUC of 0.667, sensitivity of 0.550, and specificity of 0.710.
The differences in OSA associations “may suggest the need for age-specific screening and therapeutic strategies. However, our results should be considered preliminary and point to the need for future prospective longitudinal studies in a large cohort to elucidate this issue and to explore possible etiological differences between the elderly and the non-elderly,” Liu and colleagues concluded.