Obesity hypoventilation syndrome in J apan and independent determinants of arterial carbon dioxide levels

Background and objective Obesity hypoventilation syndrome ( OHS ) prevalence was previously estimated at 9% in patients with obstructive sleep apnoea ( OSA ) in J apan. However, the definition of OSA in that study was based on an apnoea‐hypopnoea index ( AHI ) of ≥ 20/h rather than ≥ 5/h. Therefore, the prevalence of OHS in OSA was not measured in the same way as for Western countries. Our study objectives were to investigate the characteristics of Japanese patients with OHS . Methods Nine hundred eighty‐one consecutive patients investigated for suspected OSA were enrolled. At least 90% of them were from urban areas, including 162 with obese OSA (body mass index ( BMI ) ≥ 30 kg/m 2 and AHI  ≥ 5/h). Results The prevalence of OHS ( BMI 36.7 ± 4.9 kg/m 2 ) in OSA and that in obese OSA were 2.3% and 12.3%, respectively. Multiple regression analysis revealed that independent of age and BMI, arterial oxygen pressure (contribution rate (R 2 ) = 7.7%), 4% oxygen desaturation index (R 2  = 8.9%), carbon monoxide diffusing capacity/alveolar volume (R 2  = 8.3%), haemoglobin concentration (R 2  = 4.9%) and waist circumference (R 2  = 4.9%) were independently associated with arterial carbon dioxide pressure. After 12.3 ± 4.6 months of CPAP treatment, more than 60% of OHS patients no longer had hypercapnia. Conclusions The prevalence of OHS in OSA in J apan was 2.3%. The mean BMI of patients with OHS in J apan was lower than that in Western countries (36.7 kg/m 2 vs 44.0 kg/m 2 ).