To the Editor,

Obstructive sleep apnea (OSA) is a common co-morbidity in surgical patients, with a prevalence of 7-10%.1,2 Patients with OSA − especially those with undiagnosed and untreated OSA − are at increased risk of postoperative complications.1-3 The risk of cardiovascular complications, primarily cardiac arrest and shock, was significantly higher in patients with undiagnosed OSA than in those with diagnosed OSA who had been given a prescription for continuous positive airway therapy.4

The exact mechanism for the increased risk of postoperative complications associated with OSA is unknown. In a recent letter to the Journal, Deflandre et al. showed that surgical patients with OSA of similar severity may have different degrees of hypoxia.5 Among their patients with severe OSA (i.e., apnea hypopnea index [AHI] > 30 events per hour), 33% had no or only mild nocturnal hypoxia [defined as an oxygen desaturation index (ODI) of < 5 or 5-15 events per hour], and 29% and 38% had moderate or severe nocturnal hypoxia (ODI 15-30 or > 30 events per hour).5 This variation in the severity of hypoxia may be related to different OSA phenotypes and variations in the underlying OSA pathophysiology.6

Deflandre et al. proposed the hypothesis that hypoxia-mediated inflammatory modifications could increase the risk of postoperative complications.5 At present, however, there are no direct data indicating that hypoxia-mediated inflammation in a patient with OSA increases the risk of postoperative complications. Indirect data, however, show that preoperative nocturnal hypoxia is associated with increased postoperative complications.

We have shown that ODI is a sensitive, specific predictor of OSA. In 573 patients (37.1% with moderate/severe OSA and an AHI of >15 events per hour), preoperative nocturnal hypoxia was associated with the incidence of postoperative complications.7 In the same study, among patients with a mean oxygen saturation (SpO2) of < 92.7%, an ODI of > 28.5 events per hour, and/or an accumulated overnight duration of oxygen saturation of < 90%, > 7.2% were at higher risk for postoperative adverse events.7 In addition, in a longitudinal study (5.3 years) of 10,701 patients with OSA, sudden cardiac death was associated with an AHI of > 20 events per hour [hazard ratio (HR), 1.60], a mean nocturnal SpO2 of < 93% (HR, 2.93), and a lowest nocturnal SpO2 of < 78% (HR, 2.60) (all P < 0.001).8 The lowest nocturnal SpO2 predicted an 81% increase in sudden cardiac death.8

Various potential mechanisms for increased cardiovascular complications have been proposed, including increased vascular sympathetic activity and serum catecholamines, cardiac autonomic dysfunction, and an ineffective arousal mechanism related to impaired chemosensitivity.8 If the hypothesis that nocturnal hypoxia plays a key role in mediating complications is confirmed, supplemental oxygen and/or continuous positive airway therapy during the postoperative period may prevent hypoxic episodes and thus reduce sudden cardiac death.9 At present, limited evidence from our previous study7 supports the idea that nocturnal hypoxia may play a key role in mediating complications. Further trials are needed to investigate this hypothesis.