Background There is an increasing frequency of oncology and hematopoietic stem cell transplant (HSCT) patients seen in the intensive care unit and requiring extracorporeal membrane oxygenation (ECMO), however, prognosis of this population over time is unclear. This study aims to determine the mortality trends and complication rate in oncology and HSCT patients receiving ECMO by conducting a systematic review and meta-analysis.

Objectives The main outcome was all-cause mortality and studies reporting mortality within any timeframe (e.g. survival to decannulation, ICU, hospital, 28-, 60- and 90-day mortality, etc.) were included. However, hospital mortality was the most clinically relevant and was used as the primary outcome in the analyses. Secondary outcomes included bleeding, mechanical, cardiovascular, pulmonary, neurological and renal complications on ECMO, defined by the ELSO registry database definitions, and duration of ECMO, mechanical ventilation (MV) and ICU stay.

Methods MEDLINE, EMBASE, Cochrane and Web of Science were searched from earliest publication until April 10, 2020 for studies to determine the mortality trend over time in oncology and HSCT patients requiring ECMO. Primary outcome was hospital mortality. Random-effects meta-analysis model was used to obtain pooled estimates of mortality and 95% confidence intervals. A priori subgroup metanalysis compared adult versus pediatric, oncology versus HSCT, hematological malignancy versus solid tumor, allogeneic versus autologous HSCT, and veno-arterial versus veno-venous ECMO populations. Multivariable meta-regression was also performed for hospital mortality to account for year of study and HSCT population.

Results 17 eligible observational studies (n=1109 patients) were included. Overall pooled hospital mortality was 72% (95% CI: 65, 78). In the subgroup analysis, only HSCT was associated with a higher hospital mortality compared to oncology subgroup [84% (95% CI: 70, 93) vs. 66% (95% CI: 56, 74); p=0·021]. Meta-regression showed that HSCT was associated with increased mortality [adjusted odds ratio (aOR) 3.84 (95% CI 1.77, 8.31)] but a later year of study was associated with decreased mortality [adjusted odds ratio (aOR) 0.92 (95% CI: 0.85, 0.99)].

Conclusions This study reports a high overall hospital mortality in oncology and HSCT patients on ECMO which improved over time. The presence of HSCT portends almost a fourfold increased risk of mortality and this finding may need to be taken into consideration during patient selection for ECMO.