BRIGHT-4 subanalysis: bivalirudin's benefit was clearest in low-bleeding-risk STEMI

Design

Prespecified subgroup analysis of a randomised trial (BRIGHT-4)

Population

6,016 patients with STEMI undergoing radial primary PCI; 76.1% low bleeding risk

Primary outcome

30-day all-cause death or BARC 3 to 5 bleeding, by baseline bleeding risk

Effect

Low risk 1.4% vs 2.9%, HR 0.49 (0.32 to 0.75); higher risk 8.1% vs 9.2%, HR 0.88 (0.62 to 1.26)

In the main BRIGHT-4 trial of 6,016 patients with STEMI undergoing primary PCI by the radial route, bivalirudin with a 2 to 4 hour post-PCI high-dose infusion reduced the 30-day composite of all-cause death or major bleeding compared with heparin alone. This prespecified analysis asked whether that holds across bleeding risk, with low bleeding risk defined as a CRUSADE score under 30.

About three quarters (4,581, 76.1%) were low risk. Higher-risk patients had far more events (8.6% vs 2.2%; HR 4.08). In higher-risk patients the primary outcome occurred in 8.1% with bivalirudin and 9.2% with heparin (HR 0.88, 95% CI 0.62 to 1.26). In low-risk patients it occurred in 1.4% vs 2.9% (HR 0.49, 0.32 to 0.75).

The interaction P value for the absolute difference was 0.81, and 0.04 for the relative difference. The authors conclude that the benefit applies in low-risk and higher-risk groups. The data are better read as a clear benefit in low-risk patients and an imprecise result in higher-risk patients.

• In low-bleeding-risk STEMI, bivalirudin with a post-PCI infusion lowered the 30-day composite versus heparin alone.
• In higher-risk patients the confidence interval crossed 1 (HR 0.88, 0.62 to 1.26), so benefit is not shown.
• Estimate bleeding risk before the procedure; a CRUSADE score of 30 or more defined the higher-risk group.
• Regimen details matter: the trial used radial access and a 2 to 4 hour high-dose infusion after PCI.

Why it matters

It questions whether a trial-level benefit applies equally to the patients most likely to bleed.

Don't overread it

This is a subgroup analysis of one trial, run in radial-access PCI, and higher-risk patients were fewer, so the estimate is imprecise.

The statistics, in plain English

The absolute and relative interaction tests disagree (P 0.81 vs 0.04), so whether the effect truly differs by bleeding risk is unsettled. A confidence interval that includes 1, as in the higher-risk group, means the data are compatible with benefit, no effect or modest harm.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.