Contrarily, positional OSA, which is a form of OSA noted predominantly during supine sleep, was reported with higher prevalence among Asians. The first data on positional OSA by Cartwright et al. using criteria for positional OSA as 50 % or more in apneic index between back and side sleep positions reported a prevalence of 58.3 % (14 in 24 subjects) [15]. Pevernagie et al. demonstrated overall positional OSA prevalence of 58.2 % using positional OSA criteria of apnea-hypopnea index (AHI)-O (off back)/AHI-B (back) ≤0.5 (similar to Cartwright criteria) [16]. Oksenberg et al. subsequently reported a prevalence of 55.9 % (321 in 574 subjects) for positional OSA using criteria for positional OSA as a ratio of supine RDI to the lateral RDI ≥2 and required sleep time more than 30 min in either supine or lateral position [17]. Mador et al. published a paper using a more rigid positional OSA criteria which was >50 % reduction in the AHI between the supine and non-supine position and the AHI in the non-supine position <5 with data of at least 15 minutes of both the supine and non-supine position [18]. They reported an overall prevalence of 27.4 % (68 in 248 subjects). For each severity of OSA, positional OSA prevalence was reported as 49.5 %, 19.4 %, and 6.5 % in mild, moderate, and severe OSA, respectively. Richard et al. reported overall positional OSA prevalence of 55.8 % using positional OSA criteria as at least two times higher AHI in supine position than the average AHI in the other positions (similar to Cartwright criteria) [19]. All these combined reports represent an overall prevalence of positional OSA in the Western population of approximately 27.4–58.3 %.

However, positional OSA among Asians was reported with higher prevalence ranging between 49 and 74.5 % [20–27] (Table 1). The prevalence of positional OSA was observed to be less as the severity of OSA increases [26, 27]. For clinical characteristics, most of the studies did not show differences in terms of age; however, two studies reported that the positional OSA group was older [23, 26]. Half of the studies with available information on BMI reported that positional OSA group was found to have lower BMI [23, 26, 27]. Two out of three studies with available information on neck circumference reported positional OSA to have smaller neck size [25, 27]. Excessive daytime sleepiness determined by Epworth Sleepiness Scale (ESS) was observed to be no different between the two groups. However, one study reported the positional OSA group to be less sleepy compared to the non-positional OSA group [23] (Table 2). For polysomnographic characteristics, all studies reported the AHI or RDI to be lower in the positional OSA group. Mean and nadir oxygen saturations were observed to be higher in positional OSA group [25, 26]. Only two papers on positional OSA in Asians described the details on sleep architecture. Both papers did not find a difference in sleep efficiency between the two groups. One study found less N1 and increase in REM sleep with no difference in NREM3 in the positional OSA group, compared to the non-positional OSA group [23]. Another study found the positional OSA group to have an increase in N3 with no difference in N1 or REM sleep [25]. Two studies reported information on sleep position duration [23, 26]. One out of these two papers reported that the patients with positional OSA spent less time in supine position when compared to the non-positional OSA patients [23] (Table 3). Tanaka et al. [23] reported that the lateral AHI but not the supine AHI significantly correlated with excessive daytime sleepiness using ESS (r = 0.102; p < 0.05). For cephalometric measure, Chang et al. found the MP-H to be correlated with ESS [22]. Predictors for positional OSA were studied by Teerapraipruk et al. [25]. Low snoring frequency (less than 20 % of total sleep time) was a significant predictor for positional OSA (odd ratio of 3.27; p = 0.011). Low mean oxygen saturation <95 % was found to be a negative predictor (odd ratio of 0.31; p = 0.009). Low RDI (<15) was a significant predictor for normalization of RDI to less than 5 in non-supine position (odd ratio of 8.77; p < 0.001). Furthermore, Nakano et al. also reported an interesting finding that positional dependency of snoring was observed in terms of snoring intensity but not in terms of % of snoring time [21]. Position change in snoring was significantly correlated with AHI, especially in supine position.