Residual Sleep-Related Respiratory Events During Early Versus Long-Term NIV Follow-Up: Burden and Relationship with Gas Exchange

Background: A recent study published by our group demonstrated that patient–ventilator asynchrony (PVA) is independently associated with impaired gas exchange during long-term non-invasive ventilation (NIV). We aimed to determine whether the burden and phenotype of residual PVA vary according to disease category, NIV initiation context and follow-up timing, and whether the relationship between PVA and gas exchange differs according to follow-up timing. Methods: We performed a secondary analysis of 123 NIV-treated patients included in the primary study. PVA burden, upper-airway obstructive events and leaks were compared across disease categories and follow-up settings. Multivariable models for gas exchange outcomes were extended by introducing interaction terms between PVA burden and follow-up category. Patients underwent either long-term follow-up performed >3 months after NIV initiation (n = 74), early acute follow-up, i.e., after stabilization following NIV initiation in acute respiratory failure (n = 30), or early planned follow-up, i.e., within 3 months after NIV initiation in stable chronic respiratory failure (CRF) (n = 19). Results: Overall, leaks, upper airway obstructions (UAOs) and PVA burden did not significantly differ across disease categories, and neither leaks nor UAO differed across follow-up categories. In contrast, patients undergoing early follow-up after acute NIV initiation exhibited a significantly greater burden of residual PVA than those initiated in stable CRF, predominantly because of ineffective efforts. Within the subgroup of patients undergoing early follow-up, total PVA burden was more than doubled after acute NIV initiation (4.5% versus 1.5% of recording time; 6.2 versus 2.5 events·h−1). The association between PVA burden and post-NIV PaCO2 did not differ according to follow-up timing. Conversely, in an exploratory analysis limited to 40 patients, the relationship between PVA burden and nocturnal TcPCO2 was significantly attenuated during early compared with long-term follow-up (interaction β = −1.04, 95% CI −1.68 to −0.40; p = 0.003). However, this interaction was attenuated after adjustment for daytime PaCO2 (β = −0.55, 95% CI −1.12 to 0.03; p = 0.076). Stratified analyses demonstrated significant associations between PVA burden and both post-NIV PaCO2 (β = 0.38, p = 0.034) and nocturnal TcPCO2 (β = 0.80, p < 0.001) during routine long-term follow-up. Conclusions: Patients evaluated after acute NIV initiation exhibited the highest burden of residual PVA, predominantly ineffective efforts. However, PVA was independently associated with nocturnal hypercapnia primarily during established long-term NIV. These findings should be considered hypothesis-generating and require confirmation in larger prospectively monitored cohorts.