Trial: ECMO Doesn’t Save Lives in Cardiogenic Shock

AMSTERDAM — Venoarterial extracorporeal membrane oxygenation (VA ECMO) didn’t improve survival with cardiogenic shock, the ECLS-SHOCK trial showed.

This form of extracorporeal life support (ECLS) yielded similar 30-day all-cause mortality to that with usual medical therapy alone (47.8% vs 49.0%, RR 0.98, 95% CI 0.80-1.19, P=0.81), reported Holger Thiele, MD, of the Heart Center Leipzig at the University of Leipzig in Germany, at the European Society of Cardiology (ESC) conference.

Nor were there any advantages in the key secondary endpoint of arterial lactate at any time point or in renal function or severity of disease as measured by SAPS-II in the findings simultaneously published in the New England Journal of Medicine (NEJM).

However, there were harms, with over two-fold more moderate or severe bleeding (23.4% vs 9.6% in the control group) and peripheral vascular complications warranting intervention (11.0% vs 3.8%), both statistically significant.

“These findings challenge current guideline recommendations regarding active mechanical circulatory support and probably also clinical practice with more than tenfold increase of active mechanical circulatory support in patients with cardiogenic shock [over the last 10 years],” Thiele said at the ESC hotline session.

“This is a disappointing result,” he told MedPage Today at an ESC press conference. “We believe that putting in a machine is saving lives, but we see we put in large-scale devices, which induce bleeding, which induce ischemia, this may also be harmful for the patient. We have to rethink.”

But there’s still a chance with another type of mechanical circulatory support (MCS) — the Impella device — for cardiogenic shock in acute MI, he noted.

“VA ECMO may be more invasive in comparison to Impella. On the other hand, ECMO also has many advantages: It’s providing more support; it’s also providing full support to the lung,” Thiele noted at the press conference. Pointing to data expected next year from the DanGer trial with that left ventricular assist device, he said, “We will have to wait until we have additional data.”

Meanwhile, “the best course may be to reserve the early initiation of ECLS for those patients with infarct-related cardiogenic shock in whom the likely benefits more clearly outweigh the potential harms,” wrote Jane A. Leopold, MD, of Brigham and Women’s Hospital in Boston, and Darren B. Taichman, MD, PhD, of the University of Pennsylvania in Philadelphia, in an accompanying NEJM editorial.

However, ESC session study discussant Alaide Chieffo, MD, of Vita-Salute San Raffaele University and IRCCS San Raffaele Scientific Institute in Milan, ripped the trial for a variety of enrollment and protocol choices.

The trial included a largely all-comer population of 420 patients, ages 18 to 80, with acute MI complicated by cardiogenic shock for whom early revascularization was planned at 44 centers in Germany and Slovenia — an average of just three patients per year per center, which Chieffo questioned on volume, while noting that half the patients assessed for enrollment were excluded.

Enrollment criteria were agnostic to ST-segment elevation for the MI but encompassed SCAI shock categories C, D, and E, with some 30% in extremis. Arterial lactate had to be over 3 mmol/L (median 6.8). Patients could have been resuscitated, which 77% were, for up to 45 minutes, with a median 20 minutes of resuscitation — both of which Chieffo called quite long.

Active left ventricular unloading was “very low” in the ECMO group, at 5.8%, compared with 31.6% in the control group, she added. “There are publication on that showing that this is impacting our mortality … If you unload the ventricle within 2 hours you can have a 36% lower risk of 30-day mortality.”

Chieffo also noted that almost 9% of the patients randomly assigned to ECMO didn’t receive it (largely due to death before initiation), while 12.5% of those randomized to medical therapy alone crossed over to ECMO and another 15.4% got other mechanical circulatory support, mainly Impella devices.

“You end up with 25.9% of the patient in the control group that received at least one MCS,” Chieffo said. “And clearly I think that this can impact on the outcome of this study.”

Leopold and Taichman acknowledged but downplayed this limitation, writing: “Although substantial crossover might have biased the results toward the null, the findings of a prespecified as-treated analysis were consistent with the primary results.”

Thiele also pointed to the consistent lack of apparent mortality benefit across multiple subgroup analyses, including by sex, age, the presence or absence of diabetes, STEMI or non-STEMI, anterior MI, a lactate level over 6 mmol/L, or receipt of cardiopulmonary resuscitation.

“There will be some patients in this population for whom ECLS is necessary and lifesaving, but the results of the ECLS-SHOCK trial do not tell us which ones,” the editorialists concluded. “We need further studies to tell us who they are.”

Individual-patient data meta-analysis with a total of nearly 600 patients that Thiele’s group reported in The Lancet during the meeting, combining the ECLS-SHOCK data with prior trials, provided what Thiele called “robust evidence” of no advantage to VA ECMO and reaffirmed that across all subgroups.

“Maybe in the end only very few patients may benefit,” he said at the press conference. “But honestly, based on all the subgroups we have been looking for, I currently cannot identify really which patients this will be.”

“And I’m not 100% sure if the way of VA ECMO plus the combination of an additional mechanical circulatory support device is the way to go,” he stated. “We only have observational data on this…I personally believe we have to do less. Less is probably more.”

Please enable JavaScript to view the

comments powered by Disqus.