Sleep-disordered breathing-related neurocognitive impairment, time to think beyond hypoxia and sleep fragmentation?

Abstract

It has been well accepted that sleep-disordered breathing (SDB) is associated with impaired neurocognitive and behavioral performance and may consequently cause motor vehicle accidents and occupational injuries [1–3]. However, the relevant pathophysiological mechanism causing this is unclear. In particular, the key determinants of the neurocognitive/ behavioral impairment in SDB are not known. Many candidate factors such as hypoxia, sleep fragmentation, apneahypopnea index (AHI), BMI, sleep time, and metabolic factors have been studied, but none show a strong correlation with neurocognitive/behavioral impairment [4, 5]. Of these factors, intermittent hypoxia and sleep fragmentation has the strongest evidence and may play a more important role than other factors [4–6]. However, lack of robust findings means that it is difficult to claim these two parameters as “key determinants”. In a landmark US study in this field (APPLES study), Quan et al. tested long-term neurocognitive outcomes of continuous positive airway pressure (CPAP) on 1,204 obstructive sleep apnea (OSA) patients, using a robust randomized, double-blind, and sham-controlled design [7]. They found that severity of O2 desaturation was the strongest predictor, but explaining only <2 % of the variance of neurocognitive performance. Sleep fragmentation was not a significant predictor [7]. In a multicenter European study, 1,649 OSA patients with excessive daytime sleepiness (EDS) were compared with 1,233 OSA patients without EDS [8]. The two groups were found to have only 1 % difference in SpO2 nadir, no difference in mean SpO2, and only four events per hour difference in Arousal index (37/h in EDS and 33/h in non-EDS group) [8]. Given the above context, it is not surprising that in the current issue of the journal, Olaithe et al. found that neither hypoxia nor sleep fragmentation significantly predict cognitive performances in any domain [9]. The authors thoroughly assessed different domains of cognitive function and overnight polysomnography (PSG) in 150 Australians ranging from no-OSA to severe OSA. They examined the potential relationships between hypoxia/sleep fragmentation and cognitive function using structured equation modeling and carefully controlled for the confounding factors of IQ and age. They found that hypoxia was unrelated to any cognitive measures. Although modest relationships between sleep fragmentation and certain domains of cognition were found, the significance was lost after controlling for age [9]. Indeed, given that daytime cognition is affected by so many coexisting factors, any claim of relationship without a rigorous control of confounding factors would be misleading. From this aspect, the well-controlled study from Olaithe et al. can certainly add weight in the literature [9]. Importantly, data from Olaithe et al. and other studies may imply that we should not limit our thoughts solely to hypoxia and sleep fragmentation when considering the mechanisms of SDB-related cognitive impairment. Recent data have suggested other lines of investigation. For example, sleep fragmentation is usually quantified by arousal index from overnight EEG/PSG. Recent data suggest that certain parameters from EEG spectral analyses are related to daytime neurocognitive/behavioral function [10, 11]. Quantitative individual EEG band power/percentage or certain ratios showing overall slowing of EEG (such as Delta/Alpha ratio, Delta+ D. Wang (*) :B. J. Yee Department of Respiratory and Sleep Medicine, Royal Prince Alfred Hospital, Missenden Rd, Camperdown, NSW 2050, Australia e-mail: david.wang@sydney.edu.au