Predictors of sleep hypoventilation in severe COPD. preliminary results from the australian trial of noninvasive ventilation in chronic airflow limitation (AVCAL)

Introduction:Slecp hypoventilation is proposed to contribute to daytime hypercapnia in severe COPD. We sought to assess factors predictive of severity of sleep hypoventilation, and whether severity was predictive of night-to-moming change in pCC 2. Methods: Full polysomnography including transcutaneous C02 (TcC02) was performed in 78 subjects with stable severe COPD (mean (± SD) FEV1 24±10 % predicted, Pa02 on air 51.8±9.2 mmHg, PaC02 55.4i7.4 mmHg ) without sleep apnea (AHI 20/h). All were on long term 02 therapy, and were studied on their usual 02 flow rate. Evening and morning ABG were measured and TcC02 sensor drift corrected by correlation with simultaneous PaCU2. Detailed lung function assessment was performed and BMI calculated. Severity of sleep hypoventilation was assessed by i) % of total sleep time spent at TcCO₂ >10mmHg above baseline (%TST>10) ii) maximal increment in TcCO₂ above baseline (maxincr), iii) slope of rise in TcCO₂ on entry into REM (REMslo). Results: BMI (but not lung function or baseline ABG) was correlated with Maxincr (r=0.3) and remained in the model on multivariate analysis. REMslo correlated weakly with pO₂ on air, with pCO₂ and with O₂ flow rate. Only O₂ flow remained in the model on multivariate analysis. None of the variables was predictive of TST%>10. PaCO₂ rose significantly between evening and morning (59.8±IO,3 to 65.0 ±12.2 p<0.001). The evening to morning increase in PaCO₂ correlated with baseline PaC02 on air (r=0.38), with %TST>10 (r=0.56), and with maxincr (r=0.56) (both p0.0001), but not with REMslo. On multivariate analysis only pCO₂ on air and TST%>10 remained in the model. Conclusion: These results suggest sleep hypoventilation in hypercapnic COPD increases with increasing BMI and, in REM sleep, with increasing FIO₂. Sleep hypoventilation is associated with a rise in night-to-morning PaCO₂ Analysis is proceeding to determine the influence of inspiratory flow limitation and alcohol consumption on sleep hypoventilation.