Upper airway obstruction induced by non-invasive ventilation using an oronasal interface.

BACKGROUND: On initiation of long-term non-invasive ventilation (NIV), intermittent upper airway obstruction has rarely been described as possibly treatment-induced. Inspiratory pressure effects and the use of an oronasal interface may promote obstructive events in some patients with neuromuscular disease (NMD) and amyotrophic lateral sclerosis (ALS) in particular.

METHODS: We evaluated clinical data from 212 patients in whom NIV was initiated using an oronasal mask. Treatment-induced upper airway obstruction (TAO) was defined as an AHI > 5/h along with a relative increase of the AHI in the first treatment night compared to diagnostic sleep studies.

RESULTS: Prevalence of TAO was 14.2% in the entire cohort, 17.0% in patients with NMD (n = 165), 20.4% in the ALS subgroup (n = 93), and 4.3% in non-NMD patients (n = 47). Fixed expiratory positive airway pressure (EPAP, n = 192) was significantly correlated with AHI reduction (r = 0.50; p < 0.001). The inspiratory-expiratory pressure interval (∆PAP, n = 191) showed inverse correlation with the AHI change achieved in the first treatment night (r = - 0.28; p < 0.001). However, ∆PAP and the effective pressure range between EPAP and the highest inspiratory PAP achieved were not predictive of TAO. In patients with ALS, TAO was associated with better bulbar function. Study results were limited by initial EPAP being significantly lower in NMD patients reflecting that sleep apnea was less frequent and severe in this subgroup.

CONCLUSIONS: Initiation of NIV using an oronasal interface may be associated with TAO in a subset of patients. Since both EPAP and ∆PAP appear to play a causative role, careful titration of ventilator settings is recommended.