Temporary mechanical circulatory support in infarct-related cardiogenic shock: an individual patient data meta-analysis of randomised trials with 6-month follow-up

Background: Percutaneous active mechanical circulatory support (MCS) devices are being increasingly used in the treatment of acute myocardial infarction-related cardiogenic shock (AMICS) despite conflicting evidence regarding their effect on mortality. We aimed to ascertain the effect of early routine active percutaneous MCS versus control treatment on 6-month all-cause mortality in patients with AMICS.

Methods: In this individual patient data meta-analysis, randomised controlled trials of potential interest were identified, without language restriction, by querying the electronic databases MEDLINE via PubMed, Cochrane Central Register of Controlled Trials, and Embase, as well as ClinicalTrials.gov, up to Jan 26, 2024. All randomised trials with 6-month mortality data comparing early routine active MCS (directly in the catheterisation laboratory after randomisation) versus control in patients with AMICS were included. The primary outcome was 6-month all-cause mortality in patients with AMICS treated with early routine active percutaneous MCS versus control, with a focus on device type (loading, such as venoarterial extracorporeal membrane oxygenation [VA-ECMO] vs unloading) and patient selection. Hazard ratios (HRs) of the primary outcome measure were calculated using Cox regression models. This study is registered with PROSPERO, CRD42024504295.

Findings: Nine reports of randomised controlled trials (n=1114 patients) were evaluated in detail. Overall, four randomised controlled trials (n=611 patients) compared VA-ECMO with a control treatment and five randomised controlled trials (n=503 patients) compared left ventricular unloading devices with a control treatment. Two randomised controlled trials also included patients who did not have AMICS, who were excluded (55 patients [44 who were treated with VA-ECMO and 11 who were treated with a left ventricular unloading device]). The median patient age was 65 years (IQR 57-73); 845 (79·9%) of 1058 patients with data were male and 213 (20·1%) were female. No significant benefit of early unselected MCS use on 6-month mortality was noted (HR 0·87 [95% CI 0·74-1·03]; p=0·10). No significant differences were observed for left ventricular unloading devices versus control (0·80 [0·62-1·02]; p=0·075), and loading devices also had no effect on mortality (0·93 [0·75-1·17]; p=0·55). Patients with ST-elevation cardiogenic shock without risk of hypoxic brain injury had a reduction in mortality with MCS use (0·77 [0·61-0·97]; p=0·024). Major bleeding (odds ratio 2·64 [95% CI 1·91-3·65]) and vascular complications (4·43 [2·37-8·26]) were more frequent with MCS use than with control.

Interpretation: The use of active MCS devices in patients with AMICS did not reduce 6-month mortality (regardless of the device used) and increased major bleeding and vascular complications. However, patients with ST-elevation cardiogenic shock without risk of hypoxic brain injury had a reduction in mortality after MCS use. Therefore, the use of MCS should be restricted to certain patients only.

Funding: The Heart Center Leipzig at Leipzig University and the Foundation Institut für Herzinfarktforschung.