Craniofacial Load Modifies the Interaction of Obesity and Apnea Hypopnea Index Severity in Adult Males

Objectives: Multiple physiological, skeletal structural, and BMI variables determine obstructive sleep apnea (OSA) severity. To better understand how craniofacial, hyoid, and obesity measures interact, data from a population-based study was re-evaluated. Methods: Lateral facial cephalometric and polysomnographic data (2% desaturation criteria for hypopnea) were reanalyzed from a population-based study and enhanced clinic population (Wisconsin Cohort Study). Cephalometric x-rays were performed awake, upright, at end expiration. Data on craniofacial structure (45 craniofacial, 10 hyoid related), age, and obesity were re-evaluated using linear and non-linear logarithmic methods. Residual variance was assessed with hierarchical cluster analysis and classified with linear discriminant analysis. Results: 203 male patients (141 population based and 62 clinic) were included (female numbers were inadequate and were excluded from analysis). Linear analysis of all variables explained 41.3% of AHI variance with two primary contributors of BMI (28.5%) and lower facial width (12.8%). AHI at higher levels remained unexplained. Subsequent non-linear analysis identified 3 statistical groups differing in cephalometric, age, and BMI variables contributing to AHI. Clinically, the 3 groups differed in craniofacial and hyoid (1.4%, 42%, and 67%) and BMI contributions (87.6%, 40.4%, and 19.3%) and described low, intermediate, and high levels of craniofacial load. BMI effects differed by craniofacial load. For moderate or severe levels of AHI, subjects having low craniofacial load required morbid obesity, subjects with intermediate load required a combined contribution of skeletal and BMI variables, while subjects with high load did not require obesity. Conclusions: In adult males, multiple craniofacial cephalometric, hyoid, age, and BMI measures interact in a non-linear manner to create AHI severity. Craniofacial and hyoid variables can be described as craniofacial load. Groups with differing craniofacial load interact differently with increasing BMI. This study supports a model using craniofacial load to better understand AHI severity rather than individual traits. Level of Evidence: 4.