Pendelluft in hypoxemic patients resuming spontaneous breathing: proportional modes versus pressure support ventilation

Abstract

Background: Internal redistribution of gas, referred to as pendelluft, is a new potential mechanism of effort-dependent lung injury. Neurally-adjusted ventilatory assist (NAVA) and proportional assist ventilation (PAV +) follow the patient's respiratory effort and improve synchrony compared with pressure support ventilation (PSV). Whether these modes could prevent the development of pendelluft compared with PSV is unknown. We aimed to compare pendelluft magnitude during PAV + and NAVA versus PSV in patients with resolving acute respiratory distress syndrome (ARDS).

Methods: Patients received either NAVA, PAV + , or PSV in a crossover trial for 20-min using comparable assistance levels after controlled ventilation (> 72 h). We assessed pendelluft (the percentage of lost volume from the non-dependent lung region displaced to the dependent region during inspiration), drive (as the delta esophageal swing of the first 100 ms [ΔP_{es 100 ms}]) and inspiratory effort (as the esophageal pressure-time product per minute [PTP_min]). We performed repeated measures analysis with post-hoc tests and mixed-effects models.

Results: Twenty patients mechanically ventilated for 9 [5-14] days were monitored. Despite matching for a similar tidal volume, respiratory drive and inspiratory effort were slightly higher with NAVA and PAV + compared with PSV (ΔP_{es 100 ms} of -2.8 [-3.8--1.9] cm H₂O, -3.6 [-3.9--2.4] cm H₂O and -2.1 [-2.5--1.1] cm H₂O, respectively, p < 0.001 for both comparisons; PTP_min of 155 [118-209] cm H₂O s/min, 197 [145-269] cm H₂O s/min, and 134 [93-169] cm H₂O s/min, respectively, p < 0.001 for both comparisons). Pendelluft magnitude was higher in NAVA (12 ± 7%) and PAV + (13 ± 7%) compared with PSV (8 ± 6%), p < 0.001. Pendelluft magnitude was strongly associated with respiratory drive (β = -2.771, p-value < 0.001) and inspiratory effort (β = 0.026, p < 0.001), independent of the ventilatory mode. A higher magnitude of pendelluft in proportional modes compared with PSV existed after adjusting for PTP_min (β = 2.606, p = 0.010 for NAVA, and β = 3.360, p = 0.004 for PAV +), and only for PAV + when adjusted for respiratory drive (β = 2.643, p = 0.009 for PAV +).

Conclusions: Pendelluft magnitude is associated with respiratory drive and inspiratory effort. Proportional modes do not prevent its occurrence in resolving ARDS compared with PSV.

Keywords: Acute respiratory distress syndrome; Neurally-adjusted ventilatory assist; Pendelluft; Pressure support ventilation; Proportional assist ventilation.

Fig. 1

Illustrations of pendelluft magnitude assessment in NAVA, PAV + and PSV modes from a representative patient (#8). Panel A represents the change of lung volume (0–100%) during inspiration and expiration for non-dependent (blue line) and dependent lung regions (black line). The dashed vertical line delimits the inspiration time. In these representative ventilatory cycles with pendelluft, non-dependent region loses volume at the early stage on inspiration while dependent region starts inflation, producing that the wave of the non-dependent region would lag behind the dependent region. The phase angle visualization (panel B) allows to evidence the lost volume from non-dependent region with concomitant gain volume in dependent region during inspiration and not just inflation delay. Panel C illustrates the volume displacement between ventral and dorsal regions, which is calculated as the average difference between non-dependent and dependent volume (%) throughout the inspiration (solid blue line in panel C). Note that the greater the negative swing the greater the magnitude of pendelluft, as indicated by the red arrows

Fig. 2

Respiratory drive and inspiratory effort variables during partial support ventilation modes. Compared with PSV, patients during proportional modes (NAVA and PAV +) presented a higher pressure–time product [PTP] per minute (panel A), a higher delta P_es at the first 100 ms from the onset of inspiration [ΔP_{es 100 ms}] (panel B), and a higher PTP during the first 300 ms [PTP_300ms] (panel C) (*p < 0,05; ^†p < 0,001; and, ^‡p < 0,0001 all compared with PSV)

Fig. 3

Comparison of pendelluft magnitude and the frequency of pendelluft with magnitudes above specific thresholds between proportional modes and pressure support ventilation. Compared with PSV, patients during NAVA and PAV + exhibited a higher mean pendelluft magnitude (Panel A). When analyzing pendelluft at different magnitude cutoffs, both NAVA and PSV + had a higher frequency at 15% magnitude (Panel B), but only PAV + was higher at 20% magnitude and NAVA at 25% magnitude compared with PSV (Panel C and D, respectively) (*p < 0,05; and, ^†p < 0,001 all compared with PSV)

Fig. 4

Correlations between pendelluft magnitude and ΔP_{es 100 ms} or PTP_min. Each color represents a subject. Solid circles represent the average of ΔP_{es 100 ms} or PTP_min, and pendelluft magnitude from each patient in each mode. In panel A, solid lines represent the slope of the simple regressions of pendelluft with ΔP_{es 100 ms} (left) or PTP_min, (right) by patient. In panel B, solid lines represent the slope of the unadjusted regressions from repeated measures analysis with linear mixed-effects models for each patient. Black solid line corresponds to the regression model representative of all patients

Fig. 5

Drive and effort at the beginning of inspiration in representative ventilatory cycles with similar esophageal swing from selected patients in NAVA, PAV + and PSV. Ten representative ventilatory cycles in each ventilatory mode (NAVA, PAV + and PSV) with similar magnitude of esophageal swing from 5 selected patients were analyzed. No differences were observed in esophageal swings and pressure–time product [PTP] per cycle between proportional modes and PSV (A and C, respectively). By contrast, delta P_es during the first 100 ms [ΔP_{es 100 ms}] (B) and PTP during the first 300 ms [PTP_300ms] (D) were significantly higher in proportional modes than in PSV. Representative ventilatory tracings from patient #8 are shown in letter E. The values of V_T, ΔP_es per cycle, ΔP_{es 100 ms} and PTP_300ms during NAVA, PAV + and PSV are 347, 339 and 348 ml; −8.8, −9,2 and −8.8 cm H₂O; −4.3, −5.3 and −1.7 cm H₂O, and 1.26, 1.38 and 0.85 cm H₂O*s, respectively (*p < 0,05; and, ^†p < 0,001 all compared with PSV)