Introduction

Reduction of tidal volume (V_T) to 3–4 mL/kg of predicted body weight (PBW) and of end-inspiratory plateau pressure (P_PLAT) to ≤25 cmH₂O integrated by extracorporeal carbon dioxide removal (ECCO₂R) has been proposed in patients with acute respiratory distress syndrome (ARDS).1

We recently reported results of the SUPERNOVA trial, a phase II study that assessed feasibility and safety of ECCO₂R in patients with moderate ARDS.2 The study was conducted using lower (cross-sectional area of membrane lung 0.59 m², blood flow 300–500 mL/min; Hemolung, ALung Technologies, Pittsburgh, Pennsylvania, USA) and higher (cross-sectional area of membrane lung 1.30 m², blood flow between 800 and 1000 mL/min; iLA, Novalung, Heilbronn, Germany and Cardiohelp, Getinge, Rastatt, Germany) CO₂ extraction devices.2 We perform a retrospective analysis of the SUPERNOVA trial to explore the hypothesis that efficacy and safety of ECCO₂R to allow V_T of 4 mL/kg and P_PLAT ≤25 of cmH₂O may vary between lower vs higher CO₂ extraction devices.

Methods

Patients older than 18 years with moderate ARDS3 were included. Inclusion and exclusion criteria have been previously reported.2

V_T and positive end-expiratory pressure (PEEP) were in September as previously described (baseline).4 V_T was reduced to 4 mL/kg PBW, titrating PEEP to target P_PLAT of 23–25 cmH₂O and ECCO₂R hence was initiated.

Effectiveness was assessed as the proportion of patients who achieved a V_T of 4 mL/kg with PaCO₂ not increasing more than 20% from baseline (with arterial pH >7.30). Safety was assessed as the number of patients experiencing severe and ECCO₂R-related adverse events. Physiological variables were recorded at baseline, 8 and 24 hours.

Proportions and continuous variables between lower and higher extracting devices were compared using χ², Fisher exact tests, Student t-test or Wilcoxon rank-sum test. Continuous variables assessed at 8 and 24 hours were compared with previously described statistical analysis plans.2

Results

Thirty-three patients (35%) were treated with the lower and 62 patients (65%) with the higher CO₂ extraction devices. At baseline, values ​​of PaO₂/FiO₂ were higher and use of recruiting manoeuvres was less frequent in patients on lower vs higher CO₂ extraction devices (table 1) (p<0.05).

A 15.5 Fr catheter was used with the lower extraction device; an 18 (18–20) Fr catheter was used in patients on higher extraction devices (p=0.0001). At 12 hours after inclusion, doses of heparin and activated partial thromboplastin time ratio were similar between lower and higher extraction devices (21 000 (18 000; 27 950) vs 20 000 (14 000; 26 400), p=0.425 and 49.1±14.9 vs 57.6±21.6, p=0.074, respectively).

Respiratory rate and minute ventilation were higher in patients on the lower versus higher extraction devices (p<0.001) (figure 1). Time course of ventilator and blood gas parameters are shown in table 1. Compared with baseline, P_PLAT decreased by 10%–20% in both lower and higher extraction groups (p<0.001), and at 24 hours was significantly lower in patients treated with lower than in patients with higher extraction devices (p=0.003). Compared with baseline, PaCO₂ decreased at 8 and 24 (p<0.01) hours in patients treated with the higher extraction devices while it increased at 8 hours in patients treated with the lower extraction system (p=0.01). PaCO₂ at 8 hours was lower with the higher than with the lower extraction system (p=0.008).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Time course of respiratory variables. *p<0.001 lower vs higher CO₂ extraction; #p<0.01 vs baseline.

Percentage of patients reaching a V_T of 4 mL/kg PBW was 55% and 64% with the lower versus 90% and 92% with higher extraction devices at 8 and 24 hours, respectively (p<0.001).

Two severe adverse events were attributed to ECCO₂R (right frontal massive parenchymal haematoma with the lower and pneumothorax at insertion cannula in the internal jugular vein with the higher extraction devices). Percentage of patients experiencing ECCO₂R-related adverse events was 48% with the lower and 34% with the higher extraction devices (p=0.242). Percentage of patients experiencing episodes of haemolysis and bleeding was higher with lower than with higher extraction devices (p<0.05) (table 2).

Discussion

Analysis of the SUPERNOVA trial suggests that systems using higher flow (800–1000 mL/min) and a larger membrane (1.30 m²) are more effective in facilitating ultraprotective ventilator settings than using lower flow (<500 mL/min) and a smaller membrane (0.59 m²). Since there is no evidence of a safe upper limit for protective ventilator settings,5 outcomes may be improved by aggressively lowering V_T using devices that can remove more CO₂, allowing lower respiratory rates, which were shown to be lung protective6 by decreasing mechanical power.7 It should be acknowledged that clearance and total amount of CO₂ removed by ECCO₂R with lower and higher extraction devices were not quantified.

Fitzgerald and coworkers recently performed a comprehensive systematic review that included 14 studies with 495 patients treated with ECCO₂R and found that complication rates ranged from 0% to 25%.8 Clotting/membrane malfunction, bleeding and increasing the requirement for blood transfusion were the most common complications.8 We have found that, although the lower extraction device used smaller cannulas and lower blood flow than the ones used with higher extraction devices, the percentage of patients with haemolysis and bleeding episodes was significantly higher in the former (table 2). Moreover, it should be noted that heparose and activated partial thromboplastin time ratio were similar between lower and higher extraction devices. This observation may be particularly relevant since (1) the balance between excessive anticoagulation (bleeding, haemolysis) and insufficient anticoagulation (clotting) is crucial for the safety and feasibility of ECCO₂R, and (2) the importance of monitoring coagulation during ECCO₂R is confirmed. Moreover, since in the present study contraindications for systemic anticoagulation and bleeding disorders were observed in 30% of patients,2 feasibility and safety of randomised clinical trials using lower CO₂ extraction devices may be problematic.

In conclusion, these data suggest that reductions of V_T to 4 mL/kg and of P_PLAT to ≤25 of cmH₂O are consistently achievable only with the higher extraction devices. Future randomised clinical trials assessing the overall benefit and harm of ultraprotective ventilation should be carried out with ECCO₂R devices equipped with a larger membrane lung and blood flows between 800 and 1000 mL/min.