Dear Editor:

Survival for childhood cancer has dramatically improved, particularly for acute lymphoblastic leukemia, reaching over 90% overall survival in industrialized countries [1]. However, some patients may encounter severe adverse events, limiting this high success rate. ARF is one of the most serious complications and is associated with high mortality if conventional therapy fails [2]. Escalation to ECMO has rarely been used in patients with malignancy due to its limited success rates and higher risk for infectious and bleeding complications [3,4,5].

We report on a single centre experience of ECMO on patients with childhood leukemia and ARF. This retrospective study was approved by the local research ethics committee. Nine patients with childhood leukemia received ECMO in induction treatment (8/9 at first remission, 1/9 at second remission) between January 2004 and June 2017. Details on these patients are provided in Table 1. ARF resulted from pulmonary infections (two patients with Candida albicans, one patient with Aspergillus terreus, four patients with no organisms identified) and pulmonary non-infectious complications (one patient with transfusion-related acute lung injury and one patient with leukemic infiltration). Median duration of mechanical ventilation before ECMO was 3 days (range 0.4–14). The median duration of ECMO support was 14 days (range 2–24). Five (56%) patients survived ECMO und four (44%) survived to hospital discharge. When compared to survivors, non-survivors had a significantly higher vasoactive inotrope score (VIS) at ECMO initiation (85 vs. 11; p = 0.032), including two patients requiring veno-arterial cannulation. Time on ECMO support was shorter (5 vs. 15 days; p = 0.032) in non-survivors and was stopped because of multiorgan failure (22%), intracranial bleeding (11%) and progressive leukemia (11%). One patient (11%) recovered from hematopoietic stem cell transplantation performed on ECMO, but died two months later of septic shock. Moreover, non-survivors had significantly lower platelet count on ECMO (30 ×  103/μL vs 98 × 103/μL; p = 0.041). Eight (89%) patients received chemotherapy in the four weeks prior to and five (56%) were neutropenic at ECMO cannulation. Neutropenic patients did not have higher mortality compared to those without neutropenia (3/5 vs 2/4).

Table 1 Clinical characteristics and demographics of patients on ECMO

All four survivors are in complete oncologic remission at a median follow-up of 8.4 years (range 1.8–13.1), are restored to full health, and are all engaged to full-time study or work. Our data is limited by a small sample size and by its retrospective analysis. Nevertheless, it indicates that ECMO provides an effective rescue therapy in childhood leukemic patients with ARF.