Supplementary Material for: Prediction of long-term mortality in acute hypercapnic respiratory failure with use of low-flow veno-venous extracorporeal CO2 removal (ECCO2R): A retrospective single-center study

Introduction Hypercapnic respiratory failure is associated with high morbidity and mortality. Low-flow extracorporeal CO2 removal (ECCO2R) has been shown to facilitate lung protective ventilation or spontaneous breathing. However, three multicenter randomized trials have failed to show benefit which could potentially be a result of patient selection. In this study, we aimed to characterize prognostic scores developed for extracorporeal membrane oxygenation therapy which could potentially assist with the selection of patients for ECCO2R. Methods 70 patients admitted to the ICU at the University Hospital of Zurich between 10/2009 and 02/2017 with hypercapnic respiratory failure were treated with ECCO2R if pH ≤ 7.25 and/or PaCO2 ≥ 9kPa experiencing respiratory exhaustion during spontaneous breathing in obstructive lung disease or reaching the limits of lung protective ventilation (n=22 and n=48) in patients with restrictive lung pathologies. Data including baseline characteristics and respiratory parameters were collected prospectively. Scores were calculated retrospectively. Results The underlying diseases were ARDS (n=27), COPD (n=12), bronchiolitis obliterans syndrome (n=9), cystic fibrosis (n=10), pulmonary fibrosis (n=8) and other causes (n=4). 180-day mortality was 45.7% with the highest rate observed in PF and BOS patients as well as in patients who had been mechanically ventilated > 6 days before initiation of ECCO2R. The modified PRedicting dEath for SEvere hypercapnic Respiratory failure on vv-ECCO2R (PRESERVE-CO2) score differentiated well between survivors and non-survivors (4.3 ± 2.2 vs 6.9 ± 2.6, p < 0.01), whereas the modified Respiratory ECMO Survival Prediction (RESP-CO2) score showed no significant distinction. Receiver operating characteristics analysis of the PRESERVE-CO2 score revealed an area under the curve of 0.78, suggesting a cut-off of 7 points. Conclusion Careful selection of patients for ECCO2R therapy may help to improve outcomes. The proposed PRESERVE-CO2 score may serve as a guide. A score of 7 points or higher is associated with an unfavorable outcome regarding the 180-day mortality in the specific patient cohort of this study, but future studies to externally validate this score are required.