Review of the PREVENT trial

PCI versus medical therapy alone for vulnerable coronary plaques

The Lancet Volume 403, Issue 10438 p1753-1765May 04, 2024

Background Cardiologist have long been taught that acute plaque rupture leading to myocardial infarction is more likely to come from non-flow-limiting lipid-rich atherosclerotic plaques. The concept of the vulnerable plaque is surely one of the reasons that revascularization of high-grade stable coronary artery disease does not reduce myocardial infraction or death over optimal medical therapy.

The search for and treatment of the vulnerable plaque remains one of the important research areas in modern cardiology. South Korean investigators set out to test whether PCI of non-flow-limiting, high-risk vulnerable plaques identified by intracoronary imaging would reduce major adverse cardiac outcomes over medical therapy in the Preventive percutaneous coronary intervention versus optimal medical therapy alone for the treatment of vulnerable atherosclerotic coronary plaques (PREVENT).

Patients Patients were recruited in the cardiac catheterization lab. Both ACS and stable CAD patients were included. FFR was done to exclude significant flow limitations. Clinically relevant lesions with an FFR ≤ 0.80 underwent PCI with a drug-eluting stent before randomization.

All untreated, non-culprit lesions (ie, those that were clearly not responsible for the presenting clinical syndrome) with an angiographic diameter stenosis of 50% or more by site visual estimation were functionally assessed by fractional flow reserve.

Then, intermediate non-flow-limiting lesions (FFR >0.80) were assessed by intra-coronary imaging—with one of four techniques (at the discretion of the operator). These included grey-scale intravascular ultrasonography (IVUS), radiofrequency intravascular ultrasonography, a combination of grey-scale intravascular ultrasonography and near-infrared spectroscopy, or optical coherence tomography (OCT).

Vulnerable plaques were defined as lesions possessing at least two of the following four characteristics: a minimal lumen area of less than 4·0 mm² by IVUS or OCT; a plaque burden of more than 70% by IVUS; a lipid-rich plaque by near-infrared spectroscopy (defined as maximum lipid core burden index within any 4 mm pullback length [maxLCBI_4mm] >315); or a thin-cap fibroatheroma detected by radiofrequency intravascular ultrasonography or optical coherence tomography (defined as a ≥10% confluent necrotic core with >30° abutting the lumen in three consecutive frames on radiofrequency intravascular ultrasonography or as a lipid plaque with arc >90° and fibrous cap thickness <65 μm on optical coherence tomography).

Major exclusion criteria included previous coronary-artery bypass grafting, target-lesions previously stented, patients with three and more target lesions or two target lesions in the same coronary artery, heavily calcified or angulated lesions, or bifurcation lesions requiring two-stent techniques.

Baseline Characteristics Trialists screened 5627 patients and randomized 1606 in 1:1 fashion. Approximately 2000 patients were excluded because all lesions had FFR ≤ 0.80 and were referred for PCI. Another 2000 were excluded for not meeting the imaging criteria of a vulnerable plaque. The median age was 65 years, 73% were men and about a third had diabetes. More patients had stable CAD (84%) then unstable syndromes. PCI of non-target lesions was performed in 36% of patients while 64% had PCI of only target lesions. The mean left ventricular ejection fraction was 63% in both groups.

The median fractional flow reserve of the 1672 target lesions was 0·86 (IQR 0·83–0·90) and their mean diameter stenosis was 54·5%. For the predefined criteria for plaque vulnerability per patient, 1562 (97%) of 1606 patients qualified with minimal luminal area less than 4·0 mm², 1533 (96%) qualified with plaque burden greater than 70%, 186 (11%) qualified with maxLCBI_4mm greater than 315, and 77 (5%) qualified as thin-cap fibroatheromas.

PCI of non-flow-limiting lesions was performed in 729 (91%) of the 803 patients assigned to preventive PCI, with bioresorbable vascular scaffolds (in 237 [33%] of 729 patients) or cobalt–chromium everolimus-eluting metallic stents (in 491 [67%]. In the preventive PCI group, 74 (9%) patients crossed over to medical therapy alone. In the medical therapy group, 791 (99%) received medical therapy alone and 12 (1%) patients crossed over to percutaneous coronary intervention. The most common reason for cross-over was patient or physician preference.

Use of dual antiplatelet therapy was greater in the percutaneous coronary intervention group than the optimal medical therapy alone group. More than half of patients in both groups were taking high-intensity statins or moderate-intensity statins plus ezetimibe during the entire follow-up period

Procedures Initially, patients in the PCI arm had bioresorbable vascular scaffolds, but when these were withdrawn from the market, PCI was done with cobalt–chromium everolimus-eluting metallic stents (Xience; Abbott, Santa Clara, CA, USA) were used. Intravascular imaging of all target lesions in enrolled patients was performed to guide percutaneous coronary intervention. After percutaneous coronary intervention, patients received dual antiplatelet therapy for at least 6 or 12 months according to clinical presentation and anatomical complexity.

Optimal medical therapy in both groups included lifestyle modification, and intensive pharmacological therapy per guidelines. Clinical follow-up was done at 1, 6, 12, and 24 months after randomization and every year thereafter. Follow-up continued annually in all enrolled patients until the last enrolled patient reached 2 years after randomization.

Trialists assumed an incidence of the primary outcome at 2 years of 8·5% for the preventive percutaneous coronary intervention group and 12·0% for the medical therapy alone group, which corresponds to a 30% relative risk reduction. A sample size of 1600 patients provided 80% power at a two-sided significance level of 5%, assuming a 7% loss to follow-up and crossover rate.

Endpoints The primary outcome was a composite of cardiovascular (CV) death, target-vessel myocardial infarction (MI), ischemia-driven revascularization or hospitalization for unstable or progressive angina. Secondary outcomes included all the original components of the primary composite, as well as, death from any cause, any MI, any revascularization, definite stent thrombosis, stroke, bleeding events, angina status, procedural complications, and the composite all-cause death, all MI, or any repeat revascularizations.

Results The trial median follow-up was 4.3 years, but the primary outcome was measured at 2 years. The primary outcome occurred in three (0·4%) patients in the preventive percutaneous coronary intervention group and in 27 (3·4%) patients in the optimal medical therapy group (absolute difference –3·0 percentage points [95% CI –4·4 to –1·8] p=0·0003.

The primary endpoint hazard curves favoring percutaneous coronary intervention diverged through 2 years of follow-up and were thereafter parallel.

All components of the primary outcome favored PCI as did the secondary composite endpoint of death, MI, or any revascularization (3 vs 5.2% at 2 years; HR 0·69 (0·50 to 0·95).

Procedural safety included the observation that 4 of 741 patients had a total of 5 acute adverse events.

The benefit of PCI were similar in most subgroups though there was a stronger effect with cobalt stents over bioresorbable scaffold as well as when the median diameter stenosis was more than 55% vs less than 55%.

Conclusion The trialists concluded that

In patients with non-flow-limiting vulnerable coronary plaques, preventive percutaneous coronary intervention reduced major adverse cardiac events arising from high-risk vulnerable plaques, compared with optimal medical therapy alone.

They then added this sentence:

Given that PREVENT is the first large trial to show the potential effect of the focal treatment for vulnerable plaques, these findings support consideration to expand indications for percutaneous coronary intervention to include non-flow-limiting, high-risk vulnerable.

Our conclusions are different. We laud the authors for testing a unique method of identifying high-risk plaques.

But offer this list of our concerns:

This was a highly select group of patients. More than 5600 patients were screened and only 1600 patients were randomized.

There were small numbers of events 3 vs 27, mostly non-fatal events. In a trial of 1600. Note that all-cause MI rates were only 3.5% in the control arm at 7 years. This does not suggest that vulnerable plaques were that vulnerable (or, alternatively, medical therapy was quite good.)

The significant results at 2 years did not remain significant at 4 and 7 years. You’d think that if this intervention was effective, the accrual of more events would strengthen not weaken the effect size.

The other issue is the open-label design. Two of the endpoints are susceptible to human decision making. Namely, subtraction anxiety. Those patients who did not receive a stent (or get fixed) may have been more apt to be admitted for UA, or have a ‘target vessel revascularization.’

There were also more patients taking P2Y12 inhibitors in the PCI arm. Another reason they may have had fewer events.

There was missing data in 50 patients; that’s not a small number given that the delta was only 24 events.

Finally, and perhaps most importantly, there is the matter of translating this data to real life. Imaging of coronary lesions is not as common in some geographies. What’s more, 2 of the 4 imaging techniques that they used involve complicated devices, that are not universally available to many IC.

In sum, PREVENT is a start. We would want to see imaging-guided identification and treatment of vulnerable plaques replicated in other studies.