A comparison of the mortality rates in patients receiving ECMO therapy for ARDS from the ELSO registry with the data in this study revealed a large gap. In 2016, the mortality rate among all ECLS patients with ARDS was 55.6% in Germany, while it was only 45% in adult patients in the ELSO registry [
3]. According to the latest data from the ELSO registry, mortality in adult ARDS patients receiving ECLS was only 39% over the period from 2015 to 25.1.2020 [
9]. One possible explanation is bias because very few hospitals worldwide submit their data to ELSO, and most of the submitting hospitals are experienced, high-volume centres that actively contribute their data. The median for the annual centre volume from the European ELSO registry in 2019 was 18 cases per centre [
9]. In contrast, Germany had a distinctly lower median of 4 cases per centre in 2018. In Germany, only 21 out of 1,160 hospitals with intensive care units are registered with ELSO [
10,
11]. In 2018, in the European ELSO Registry, only 81 centres with 2,233 cases were reported. In contrast, in Germany, there were 231 centres with 7,317 cases reported, which shows a high number of cases unreported to the registry. There is no minimum standard for ECLS therapy in Germany by law, and registration in a programme, such as ELSO or the German ARDS network (79 registered hospitals), does not obligate hospitals to implement or adhere to specific standards. This finding highlights the importance of treating patients in specialized centres to provide the survival benefit seen in ELSO centres to all patients. International and national recommendations for the minimum standards for ECLS therapy have existed for a long time, but their implementation has been lacking [
12,
13]. The shuttle-spoke model, which was already recommended in 2014 by the International ECMO Network (ECMONet), should also be reconsidered as a possible solution for Germany, in which each federal state establishes a central ECLS centre (hub) and all regional hospitals transfer their patients to the hub [
14]. Comparing the hospital mortality rates in 2011 (63%) with data from the US health care system, Sauer et al. showed a similarly high mortality rate of 61% among all ECLS patients [
15]. In contrast, the mortality rates in the two largest randomized studies, the CESAR trial and the EOLIA trial, were relatively low for ARDS patients treated with ECLS (37% and 38%) [
16,
17]. Again, treatment was mainly administered at specialized centres, and there were strict inclusion criteria, which may have contributed to these mortality rates. Treatment in specialized centres clearly results in reduced mortality. Similarly, Muguruma et al. showed a significantly lower mortality rate in patients treated in high-volume ECMO centres (50.4%) than in those treated in low-volume centres (62.5%) [
18]. Therapy with ECMO is a multidisciplinary process and requires a high level of experience for all persons involved. Therefore, locally adapted standard operating procedures with regard to the indication for ECLS, selection of the mode of ECLS, catheterization strategy, therapy parameters and weaning protocols are recommended, and the implementation of these standards, including the introduction of a minimum number of ECLS therapies through state regulations, should be reconsidered given the high mortality rates.
The mortality rates varied substantially with treatment duration. The highest hospital mortality rate (69.7%) was observed in the group treated for less than 2 days; conversely, the lowest rate (43%) was observed in the group treated between 6 and 8 days. This shows that patients who benefit from VV-ECMO therapy need approximately one week before they can be successfully weaned from ECLS with favourable outcomes. In the case of the patients who died after less than 2 days of therapy, it should be considered whether the indication for ECMO therapy was identified too late in these patients, becoming an exclusion criterion for ECLS therapy, or whether the indication for the initiation of ECLS was questionable. A similar pattern was observed in patients receiving VA-ECMO. Again, the highest mortality was in the group that received treatment for less than 2 days, and the lowest mortality rate was in the group treated for 4–6 days (73.9% vs 53.4%). Compared to VV-ECMO, however, patients receiving VA-ECMO had a significantly higher mortality over the treatment period, possibly due to the main indication for VA-ECMO, cardiogenic shock, which is associated with increased mortality.