Critical review · Mar 29, 2026
CHIP-BCIS3: Impella CP versus no mechanical support in high-risk elective PCI
Perera D, Ryan M, Ezad SM, Khan SQ, Webb I, O'Kane PD, et al. Left Ventricular Unloading in High-Risk Percutaneous Coronary Intervention. N Engl J Med. 2026 (published online Mar 29, 2026).
doi:10.1056/NEJMoa2515704 · Open-label randomized trial with blinded outcome adjudication (United Kingdom; NCT05003817), n = 300
Summary
CHIP-BCIS3 randomized 300 patients with severe left ventricular dysfunction and extensive coronary disease undergoing planned complex PCI. 148 were assigned to elective unloading with a microaxial flow pump (Impella CP) and 152 to standard care. Median LVEF was 27% and median SYNTAX score 38. The trial was funded by the UK NIHR.
The primary endpoint was hierarchical (death, disabling stroke, spontaneous MI, cardiovascular hospitalization and periprocedural myocardial injury), analyzed by win ratio, with a median follow-up of 22 months.
| Result | Value |
|---|---|
| Win ratio | 0.85 (95% CI 0.63 to 1.15) |
| Difference in the death tier | −7.0 percentage points, favoring standard care |
| Difference in the myocardial injury tier | −3.9 percentage points, favoring standard care |
| All-cause death | 47 vs 33 (HR 1.54; 95% CI 0.99 to 2.41) |
| Cardiovascular death | HR 1.91 (95% CI 1.11 to 3.30) |
| Revascularization index | 67% vs 67% |
Strengths
- Investigator-initiated, publicly funded, with no industry sponsorship.
- Blinded outcome adjudication.
- The highest-risk elective PCI population randomized to date.
- The win ratio shows in which tiers each arm won.
Methodological issues
- Cardiovascular death is secondary. The HR of 1.91 is not adjusted for multiplicity. Still, all-cause death, periprocedural injury and the win ratio breakdown point the same way.
Opinion
The two most clinically important tiers, death and myocardial injury, favored standard care.
The proposed mechanism for Impella, enabling more complete revascularization, was not confirmed: the revascularization index was 67% in both arms. Physiological plausibility is a necessary condition, but it does not prove benefit.
DanGer Shock showed benefit in cardiogenic shock. That is a different disease and physiology, and the result was different. Protection in elective PCI and rescue in shock are different scenarios and should not be conflated.
In my view, CHIP-BCIS3 is clean evidence against elective Impella use in high-risk PCI.