New Analysis Bolsters Evidence Backing the Safety of Multivessel PCI in STEMI
After accounting for potential biases, accumulated observational evidence supports the safety—at least in the short term—of multivessel PCI in hemodynamically stable patients with STEMI, a new analysis shows. That contradicts earlier recommendations against intervening on nonculprit lesions because of possible harm and is consistent with more recent guidelines that have opened the door to the practice.
In a meta-analysis of studies published between 2002 and 2014, the rate of in-hospital mortality was 5.1% in patients undergoing multivessel PCI and 5.3% in those undergoing revascularization of the infarct-related artery only (OR 0.87; 95% CI 0.65-1.17), Saurav Chatterjee, MD (Mount Sinai St. Luke’s-Roosevelt Hospitals, New York, NY), and colleagues report in a study published online June 24, 2016, ahead of print in Catheterization and Cardiovascular Interventions.
Moreover, in a sequential analysis looking at how the totality of the evidence evolved over that stretch, there was no point in time at which the threshold to indicate harm from performing multivessel PCI was crossed, thus calling into question the class III recommendation (indicating harm) against the practice that was removed in a focused update in October 2015 after several smaller trials provided favorable results.
“We decided to play Monday morning quarterback a little bit,” senior author Jay Giri, MD, MPH (Hospital of the University of Pennsylvania, Philadelphia), told TCTMD. But, he added, “it brings up the question [of] how are we evaluating bias and how are we doing our own statistical analyses as guideline writers to verify that our guidelines are based on the data that’s most rigorously analyzed.”
The difference between the current analysis indicating the early safety of nonculprit PCI and prior observational studies suggesting harm is in the handling of various biases, particularly treatment selection bias and immortal time bias (in which patients who survive to the second part of a staged procedure have automatic survival benefits), Giri said. He pointed out that his team made efforts to mitigate the effects of those biases by excluding patients with cardiac arrest and cardiogenic shock, by using propensity matching, and by looking only at short-term outcomes.
Commenting on the study for TCTMD, Stephen Ellis, MD (Cleveland Clinic, Cleveland, OH), agreed that biases contributed to the signal of harm seen in earlier observational analyses, adding that it is challenging in a nonrandomized study to account for all of the factors—such as the severity of the disease in nonculprit arteries and the presence of major comorbidities—that physicians consider when deciding between multivessel and culprit-artery-only PCI.
“To be able to correct for that in a nonrandomized fashion, particularly when you don’t have a whole lot of granular detail . . . is I think fraught with problems,” Ellis said.
That said, the fact that the results of the study by Giri et al are consistent with the recent randomized trials is reassuring. “It’s further information [pointing] to the fact that at least strong consideration ought to be given to more complete revascularization in patients presenting with a STEMI,” he said.
More definitive evidence, however, will come with the conclusion of the COMPLETE trial, which will randomize an estimated 3,900 patients with STEMI to complete or culprit-only revascularization and is scheduled to wrap up in December 2018, Ellis said.
Safety in the Short Term
Giri and colleagues set out to assess the impact of the previous class III recommendation on use of multivessel PCI and to take a comprehensive look at the existing literature.
For the first objective, they looked at data from the Nationwide Inpatient Sample on 168,465 hemodynamically stable patients with STEMI who underwent primary PCI between 2009 and 2012. Of those, only 7.3% received a multivessel intervention during the index hospitalization, reflecting adherence to the earlier guidelines.
After propensity matching, multivessel PCI was associated with a lower rate of in-hospital mortality compared with culprit-only PCI (1.91% vs 5.32%; OR 0.28; 95% CI 0.24-0.32). The authors acknowledge, however, that “statistical adjustment was likely inadequate to balance confounders among such fundamentally different populations.”
They then performed the meta-analysis, which included 19 studies encompassing a total of 76,399 patients. The lack of difference in in-hospital mortality based on the completeness of revascularization seen in the main analysis remained in a sensitivity analysis accounting for potential publication bias.
Based on more recent studies, Ellis said, “I think it’s fair to say that [multivessel PCI] doesn’t seem to be contraindicated like we used to think, but which patients actually benefit and when it should be done—should it be done actually at the same setting you open up the culprit artery, same hospitalization, within a month—we don’t know.”
More trials are needed to work out the specifics of treating nonculprit arteries, so the new class IIb recommendation stating that multivessel PCI “may be considered” in hemodynamically stable patients with STEMI is appropriate, Ellis said.
Giri agreed that the strength of the recommendation is fair considering the lack of definitive evidence from randomized trials. He and his colleagues also note in their paper that further research should be conducted to determine the “usefulness of fractional flow reserve, intravascular ultrasound, optical coherence tomography, and near-infrared spectroscopy as advanced physiologic and anatomic assessments of coronary lesions during multivessel intervention in STEMI patients.”
Chatterjee S, Yeh RW, Sardar P, et al. Is multivessel intervention in ST-elevation myocardial infarction associated with early harm? Catheter Cardiovasc Interv. 2016;Epub ahead of print.
- Chatterjee and Giri report no relevant conflicts of interest.
- Ellis reports consulting for all major US stent manufacturers.