SELUTION DeNovo HBR: sirolimus-eluting balloon versus DES in high bleeding risk

Summary

SELUTION DeNovo randomized 3323 patients with de novo coronary lesions to a sirolimus-eluting balloon (SEB) strategy with provisional stenting or to systematic DES. Of the 3064 with bleeding risk data, 522 (17.0%) met ARC-HBR criteria: 269 in the SEB arm and 253 in the DES arm. In the SEB arm, 17.1% received a bailout stent. About 68% were discharged on oral anticoagulation.

In the parent trial, TVF was 5.3% (88/1661) with SEB and 4.4% (73/1662) with DES. Difference 0.91 points (95% CI −0.55 to 2.38), noninferior against a 2.44-point margin (one-sided p = 0.02). In the per-protocol analysis the difference was 1.09 (−0.45 to 2.63), and noninferiority was not confirmed. Clinically driven TVR was 3.3% vs 2.1% (difference 1.22; 0.11 to 2.33). In patients without HBR, TVF was 5.7% vs 4.0% (difference 1.79; 0.12 to 3.47).

Strengths

• Prespecified subgroup of a large randomized trial.
• The authors report the interaction test for each outcome and the breakdown of components.
• The study describes the SEB strategy result in HBR with precision: TVF 4.9% (95% CI 2.8 to 8.1%) and lesion thrombosis 0.4%.

Methodological issues

• Too few events to compare. TVF had 33 events. RR 0.61 (0.31 to 1.20): the interval is compatible with a 69% relative reduction and with a 20% increase. With 33 events, the smallest effect detectable at 80% power is an HR of about 0.38. Detecting 4.9% against 7.9% at 80% power would take about 1040 patients per arm, four times this subgroup.
• No bleeding signal. RR 0.71 (0.25 to 2.00): a 75% reduction and a doubling fit 6 versus 8 events equally well. NACE: RR 0.63 (0.35 to 1.11). None of the three intervals excludes harm, and none excludes a large benefit.
• Interaction without a mechanism. The interaction p was 0.04 for TVF and 0.49 for bleeding. Bleeding is the only outcome with a plausible HBR-specific mechanism. Less antithrombotic therapy should raise lesion thrombosis or spontaneous MI, but thrombosis (1 vs 2; definite 1 vs 1) and target vessel MI (8 vs 11) were similar.
• Bleeding in the opposite direction to the expected one. In the first 30 days, under identical regimens, there was 1 BARC 3–5 bleed with SEB and 7 with DES. Afterwards, 5 vs 1.
• Death classification. The protocol classifies unwitnessed and unknown-cause deaths as cardiac. All-cause death was 14 vs 13. Stroke (5 vs 1) was not part of the primary outcomes.
• Selected patients. Operators enrolled lesions with an anticipated bailout stenting probability below 30%.
• Contradictory conclusion. The Methods state that the intervals “should not be used to infer definitive treatment effects.” The Impact box states that the data support “its use as an alternative revascularisation strategy in HBR patients.”