Multivessel PCI Helps Most in STEMI Patients With Severe Nonculprit Lesions
This analysis of the COMPLETE trial sets the stage for future investigations into the best ways to guide nonculprit PCI.
Complete revascularization has greater benefits over culprit-only PCI in patients with STEMI and multivessel disease whose nonculprit lesions have a higher degree of stenosis as determined by quantitative coronary angiography (QCA), an analysis from the COMPLETE trial shows.
Intervening on nonculprit lesions was associated with a reduction in the coprimary endpoint of CV death or new MI in patients with QCA stenosis of at least 60% (severe lesions), but not in those with QCA stenosis of less than 60% (moderate lesions).
For the other coprimary endpoint of CV death, new MI, or ischemia-driven revascularization, complete revascularization was associated with a benefit regardless of the degree of QCA stenosis, although the risk reduction was greater among patients with QCA stenosis of at least 60%.
Speaking with TCTMD, lead author Tej Sheth, MD, and senior author Shamir Mehta, MD (both of the Population Health Research Institute and McMaster University, Hamilton, Canada), said the results lay the groundwork for future research into how best to select nonculprit lesions for multivessel PCI in STEMI, a practice that has now been shown to be safe and effective in the COMPLETE trial.
“The results of the COMPLETE trial speak for themselves. They clearly have established the role of nonculprit PCI in the setting of STEMI, and they’ve established that based on an angiographic threshold which is based on visual estimation by the operator,” Sheth said.
“The reason for doing this subgroup analysis,” he explained, “was to see if we could understand a little bit more about the treatment effect by angiographic severity and really open the door to the question: can this approach be refined in any way? Do we need to be treating all of the lesions that were treated in COMPLETE or is it possible that this could be reduced and be a more-targeted approach?”
In the trial, nonculprit lesions were selected for treatment if the stenosis was at least 70% based on visual estimation of the angiogram. However, there is interest in exploring whether other modalities—such as fractional flow reserve (FFR), which was used infrequently in COMPLETE—can improve that selection process.
“The real value of the analysis is that it demonstrates some of the weaknesses with an angiography-guided approach and areas where we can improve,” Mehta said. “Further improvement may include a physiology-based strategy, whether that be an invasive or noninvasive physiology-based strategy, and intracardiac imaging with optical coherence tomography [OCT], or perhaps a combination of these modalities. And that was the reason we published this analysis.”
The real value of the analysis is that it demonstrates some of the weaknesses with an angiography-guided approach and areas where we can improve. Shamir Mehta
Published in the September 15, 2020, issue of the Journal of the American College of Cardiology, the analysis included 3,851 patients (about 95% of the COMPLETE population) who had nonculprit lesion stenosis severity measured using QCA in the angiographic core lab at McMaster University and Hamilton Health Sciences; 5,355 lesions were evaluated. In most patients (64%), QCA stenosis was 60% or greater.
QCA is the “gold standard for assessment of diameter stenosis severity,” the authors note in the paper. “Visual evaluation tends to overestimate diameter stenosis compared with QCA for moderate-to-severe lesions; QCA values typically range from 50% to 70% for visually estimated stenoses of ≥ 70%.”
For the first coprimary endpoint of CV death or new MI, complete revascularization had a significant benefit in patients with QCA stenosis of 60% or greater (2.5% vs 4.2% per year; HR 0.61; 95% CI 0.47-0.79) but not in those with QCA stenosis of less than 60% (3.0% vs 2.9% per year; HR 1.04; 95% CI 0.72-1.50). The interaction was significant (P = 0.02).
For the second coprimary endpoint of CV death, new MI, or ischemia-driven revascularization, complete revascularization had a greater benefit in those with more-severe nonculprit lesions (P = 0.04 for interaction)—events were reduced both in patients with QCA stenosis of 60% or higher (2.9% vs 6.9% per year; HR 0.43; 95% CI 0.34-0.54) and in those with QCA stenosis of less than 60% (3.3% vs 5.2% per year; HR 0.65; 95% CI 0.47-0.89).
In an accompanying editorial, Sanjay Kaul, MD (Cedars-Sinai Medical Center, Los Angeles, CA), notes that the main treatment effect in COMPLETE “is large, statistically persuasive, and clinically important,” albeit with the benefits driven mostly by reductions in MI and repeat revascularization rather than CV death. “No significant increase in risk of major bleeding or contrast-associated acute kidney injury was observed, thereby conferring a desirable benefit-risk ratio in favor of complete revascularization.”
He then delves into whether subgroup analyses from COMPLETE—including the current one and a prior analysis looking at the timing of the nonculprit interventions—are credible, raising questions about statistical reliability, support from prior research, biological plausibility, and relevance to clinical practice.
“Like most subgroup analyses, this one fails to meet best practices for subgroup tests, including prespecification, stratified randomization, adjustment for multiple testing, and careful consideration of statistical power and prior probabilities,” Kaul concludes. “Collectively, these limitations adversely affect the credibility of the subgroup finding in COMPLETE.”
Mehta strongly disputed the assertion that the current analysis was not prespecified, stating that it was included in the initial design paper. Moreover, Kaul’s in-depth discussion about the limitations of subgroup analyses was misplaced because this study was intended to provide information to help highlight opportunities for improvement rather than to guide clinical practice, he said, noting that QCA cannot be used by operators in real time.
I think it does some service to the field for us to publish this . . . to put some boundaries around where the benefit is likely to be. Tej Sheth
The findings, Sheth said, provide insights into which patients are most likely to benefit from complete revascularization. “Now that nonculprit PCI has been established, there is the potential that enthusiastic operators are going to be treating all kinds of lesions in nonculprit PCI,” he said. “So I think it does some service to the field for us to publish this . . . to put some boundaries around where the benefit is likely to be. I think that’s informative for practitioners out there and sets the stage for how we can refine our approach.”
Mehta added, “In other words, it puts the brakes a little bit on nonculprit-lesion PCI.”
Future studies will explore whether other approaches to identifying nonculprit lesions suitable for intervention—such as FFR, OCT, or a combination of the two—are superior to an angiography-guided approach, Mehta said. He noted that a COMPLETE 2 trial comparing a physiology-guided approach with an angiography-based strategy has received funding and is in the planning stages.
For David Cohen, MD (Kansas City, MO), this new COMPLETE analysis “certainly points in the direction that the more severe the lesion is in the nonculprit vessel, the greater the benefit is. . . . This can help us to decide who are the patients we really need to do multivessel PCI on either at the same time or in a staged setting.”
Sheth T, Pinilla-Echeverri N, Moreno R, et al. Nonculprit lesion severity and outcome of revascularization in patients with STEMI and multivessel coronary disease. J Am Coll Cardiol. 2020;76:1277-1286.
Kaul S. On the credibility of subgroup analyses in the COMPLETE trial. J Am Coll Cardiol. 2020;76:1287-1290.
- COMPLETE was funded by the Canadian Institutes of Health Research with additional support from AstraZeneca, Boston Scientific, and the Population Health Research Institute.
- Sheth and Kaul report no relevant conflicts of interest.
- Mehta reports having received grants and other support from AstraZeneca and grants from Boston Scientific during the conduct of the study, as well as grants from AstraZeneca outside of the submitted work.