Obstructive Sleep Apnea Is a Distinct Physiological Endotype in Individuals with Comorbid Insomnia and Sleep Apnea

Rationale: With up to 40% of individuals with either insomnia or obstructive sleep apnea (OSA) demonstrating clinically significant symptoms of the other disorder, the high degree of comorbidity among the two most common sleep disorders suggests a bidirectional relationship and/or shared underpinnings. Although the presence of insomnia disorder is believed to influence the underlying pathophysiology of OSA, this influence is yet to be examined directly. Objectives: To investigate whether the four OSA endotypes (upper airway collapsibility, muscle compensation, loop gain, and the arousal threshold) are different in patients with OSA with and without comorbid insomnia disorder. Methods: Using the ventilatory flow pattern captured from routine polysomnography, the four OSA endotypes were measured in 34 patients with OSA who met the diagnostic criteria for insomnia disorder (COMISA) and 34 patients with OSA without insomnia (OSA only). Patients demonstrated mild-to-severe OSA (apnea-hypopnea index, 25.8 +/- 2.0 events/h) and were individually matched according to age (50.2 +/- 1.5 yr), sex (42 male: 26 female), and body mass index (29.3 +/- 0.6 kg/m(2)). Results: Compared with patients with OSA without comorbid insomnia, patients with COMISA demonstrated significantly lower respiratory arousal thresholds (128.9 [118.1 to 137.1] vs. 147.7 [132.3 to 165.0] % eupneic ventilation ((V) overbar(eupnea)); U= 261; 95% confidence interval [CI], 238.3 to 213.9; d= 1.1; P < 0.001), less collapsible upper airways (88.2 [85.5 to 94.6] vs. 72.9 [64.7 to 79.2] %(V) overbar(eupnea); U= 1081; 95% CI, 14.0 to 26.7; d= 2.3; P < 0.001), and more stable ventilatory control (i.e., lower loop gain: 0.51 [0.44 to 0.56] vs. 0.58 [0.49 to 0.70]; U= 402; 95% CI, 20.2 to 20.01; d= 0.05; P= 0.03). Muscle compensation was similar between groups. Moderated linear regression revealed that the arousal threshold moderated the relationship between collapsibility and OSA severity in patients with COMISA but not in patients with OSA only. Conclusions: A low arousal threshold is an overrepresented endotypic trait in individuals with COMISA and may exhibit a greater relative contribution to OSA pathogenesis in these patients. Contrastingly, the prevalence of a highly collapsible upper airway in COMISA was low, suggesting that anatomical predisposition may contribute less to OSA development in COMISA. Based on our findings, we theorize that conditioned hyperarousal perpetuating insomnia may translate to a reduced arousal threshold to respiratory events, thereby increasing the risk or severity of OSA. Therapies that target increased nocturnal hyperarousal (e.g., through cognitive behavior therapy for insomnia) may be effective in individuals with COMISA.