Cardiogenic shock (CS) remains one of the most devastating complications of acute cardiac failure, with mortality rates approaching 40–50%. As pharmacologic therapy often proves insufficient in severe cases, temporary mechanical circulatory support (MCS) devices have become central to modern shock management. Among these, veno-arterial extracorporeal membrane oxygenation (VA-ECMO) and the Impella percutaneous ventricular assist device are two widely utilized strategies. This systematic review and meta-analysis by Soliman et al. (2026) provides a comprehensive comparison of these devices in adult patients with cardiogenic shock .
The investigators conducted a systematic search of CENTRAL, MEDLINE, and Scopus through October 2025, adhering to PRISMA guidelines and registering the protocol in PROSPERO (CRD420251158340). Ten comparative observational studies published between 2011 and 2023 were included, enrolling 5,364 patients—3,546 treated with Impella and 1,818 treated with VA-ECMO. As outlined in the PRISMA flow diagram (Figure 1, page 4), 827 records were screened, ultimately narrowing to 10 eligible studies.
Mechanistic Differences Between VA-ECMO and Impella
Understanding the physiological distinction between devices is essential. VA-ECMO provides both cardiac and respiratory support by diverting venous blood through an oxygenator and returning it to the arterial circulation. However, this retrograde aortic flow increases left ventricular (LV) afterload and may impair myocardial recovery.
In contrast, the Impella device is a transvalvular microaxial pump that unloads the LV by directly transferring blood into the ascending aorta. This reduces wall stress and myocardial oxygen demand while improving forward flow. These divergent mechanisms create clinical equipoise regarding optimal device selection in cardiogenic shock.
Primary Outcome: Short-Term Mortality
All 10 studies reported short-term mortality (in-hospital or 30-day). The pooled random-effects meta-analysis demonstrated no statistically significant mortality difference between Impella and VA-ECMO (RR 0.92; 95% CI 0.76–1.10; p=0.30). However, substantial heterogeneity was present (I² = 64.5%), as illustrated in the forest plot (Figure 4, page 7).
Subgroup analysis (Figure 10, page 10) compared adjusted versus unadjusted cohorts. Adjusted studies yielded RR 0.92 (95% CI 0.66–1.28), while unadjusted cohorts showed a modest apparent mortality reduction with Impella (RR 0.91; 95% CI 0.86–0.95). Importantly, the subgroup difference was not statistically significant (p=0.92), suggesting that adjustment methodology did not meaningfully alter conclusions.
A leave-one-out sensitivity analysis (Figure 9, page 9) confirmed the stability of the primary finding, with no single study driving results.
Secondary Outcomes: Safety Profile
While survival was neutral, safety outcomes favored Impella.
Stroke: Five studies were included in the stroke analysis. Impella significantly reduced stroke risk (RR 0.52; 95% CI 0.36–0.75), with low heterogeneity (I² = 24.4%) (Figure 5, page 7).
Major Bleeding: Seven studies contributed to this outcome. Impella was associated with a substantially lower risk of major bleeding (RR 0.53; 95% CI 0.49–0.57), with no observed heterogeneity (I² = 0%) (Figure 6, page 8).
Limb Ischemia: Five studies were analyzed, demonstrating significantly reduced limb ischemia with Impella (RR 0.55; 95% CI 0.45–0.68), again with no heterogeneity (I² = 0%) (Figure 7, page 8).
These safety findings likely reflect device-specific characteristics. VA-ECMO requires large-bore arterial cannulation and higher anticoagulation targets, increasing bleeding and vascular complications. The retrograde flow may also promote thrombus formation and embolic stroke risk. Conversely, Impella’s smaller access profile and antegrade physiologic flow may mitigate these risks.
Risk of Bias and Evidence Certainty
Despite compelling safety signals, the certainty of evidence was rated very low. As shown in the ROBINS-I assessment summary (Figures 2 and 3, pages 5–6), all included studies had an overall serious risk of bias. Confounding by indication was universal: patients with more severe shock—such as biventricular failure or post-cardiac arrest states—were preferentially treated with VA-ECMO.
This systematic baseline imbalance likely biased comparisons in favor of Impella. Even propensity matching and inverse probability weighting can only adjust for measured confounders, leaving residual bias from unmeasured clinical severity variables.
The funnel plot (Figure 8, page 9) and Egger’s test (p=0.76) suggested no significant publication bias.
Clinical Implications
The findings support an individualized approach to mechanical circulatory support in cardiogenic shock. For patients with isolated LV failure (SCAI Stage C or early D), Impella may offer adequate hemodynamic support with a more favorable complication profile.
However, in patients with profound circulatory collapse, severe hypoxemia, or biventricular failure (SCAI Stage E), VA-ECMO’s comprehensive cardiopulmonary support remains indispensable despite higher complication risks.
Importantly, the study underscores the urgent need for adequately powered randomized controlled trials directly comparing these devices. Until such trials are completed, high-quality registries and standardized outcome definitions are critical to improving evidence clarity.
Conclusion
This updated meta-analysis demonstrates no mortality difference between Impella and VA-ECMO in cardiogenic shock but identifies a significantly improved safety profile with Impella. Due to pervasive confounding in observational data, these findings should guide—but not dictate—clinical decision-making. Definitive randomized trials remain the gold standard for establishing optimal mechanical circulatory support strategies in this high-risk population .
