Complete Versus Lesion-Only Primary PCI

Background Complete revascularization may improve outcomes compared with an infarct-related artery (IRA)-only strategy in patients being treated with primary percutaneous coronary intervention (PPCI) who have multivessel disease presenting with ST-segment elevation myocardial infarction (STEMI). However, there is concern that non-IRA PCI may cause additional non-IRA myocardial infarction (MI). Objectives This study sought to determine whether in-hospital complete revascularization was associated with increased total infarct size compared with an IRA-only strategy. Methods This multicenter prospective, randomized, open-label, blinded endpoint clinical trial evaluated STEMI patients with multivessel disease having PPCI within 12 h of symptom onset. Patients were randomized to either IRA-only PCI or complete in-hospital revascularization. Contrast-enhanced cardiovascular magnetic resonance (CMR) was performed following PPCI (median day 3) and stress CMR at 9 months. The pre-specified primary endpoint was infarct size on pre-discharge CMR. The study had 80% power to detect a 4% difference in infarct size with 100 patients per group. Results Of the 296 patients in the main trial, 205 participated in the CMR substudy, and 203 patients (98 complete revascularization and 105 IRA-only) completed the pre-discharge CMR. The groups were well-matched. Total infarct size (median, interquartile range) was similar to IRA-only revascularization: 13.5% (6.2% to 21.9%) versus complete revascularization, 12.6% (7.2% to 22.6%) of left ventricular mass, p = 0.57 (95% confidence interval for difference in geometric means 0.82 to 1.41). The complete revascularization group had an increase in non-IRA MI on the pre-discharge CMR (22 of 98 vs. 11 of 105, p = 0.02). There was no difference in total infarct size or ischemic burden between treatment groups at follow-up CMR. Conclusions Multivessel PCI in the setting of STEMI leads to a small increase in CMR-detected non-IRA MI, but total infarct size was not significantly different from an IRA-only revascularization strategy. (Complete Versus Lesion-Only Primary PCI Pilot Study [CvLPRIT]; ISRCTN70913605)

The mechanisms leading to improved clinical outcomes are currently unclear. However, there is concern that PCI to non-IRAs may be associated with additional procedural-related infarction (5). These well-described type 4a myocardial infarctions (MIs) (6) cannot be detected by conventional enzymatic markers at the time of PPCI because the associated increases are relatively small compared with the large rise in enzymes caused by the STEMI itself. Cardiovascular magnetic resonance (CMR) is able to precisely characterize areas of myocardial injury following myocardial ischemia. The myocardium at risk becomes edematous (7), and late gadolinium-enhanced (LGE) imaging allows the accurate detection and quantification of infarct size and microvascular obstruction (MVO) (8).
Infarct size (9) and MVO (10) measured on CMR are both strong medium-term prognostic markers following PPCI. There are no CMR data as yet in the literature on patients undergoing complete revascularization for multivessel disease at the time of PPCI.
The primary aim of the current pre-specified substudy was to assess whether a complete revascularization strategy, due to causing additional infarcts in the non-IRA territories, was associated with greater infarct size than an IRA-only strategy in patients randomized in CvLPRIT. Additionally, we aimed to assess whether myocardial salvage and myocardial ischemia at follow-up CMR were different in the 2 groups.

METHODS
STUDY DESIGN. The design and rationale of the study have been published previously (11). Briefly, CvLPRIT  McCann et al. The CvLPRIT CMR Substudy D E C E M B E R 2 2 , 2 0 1 5 : 2 7 1 3 -2 4 RANDOMIZATION AND TREATMENT. Eligible patients presenting with STEMI within 12 h were randomized, after verbal assent and coronary angiography, but before PCI to the culprit lesion, to either IRA-only or in-hospital complete revascularization.
Randomization was stratified by infarct location (anterior/non-anterior MI) and time to presentation (>3 or #3 h). PCI was performed according to current guidelines. Written informed consent for continued participation in the study was obtained on the day(s) following the PPCI, once the patient was able to understand and retain the information.    quences, if no regional variation in signal intensity within the myocardium was seen, these patients were excluded from analysis of the area at risk (AAR).
LV volumes and mass were calculated from cine images as previously described using QMass v7.1 (Medis) (15). The presence of LGE was assessed by 2 observers (G.P.M., J.N.K.) and was quantitated with cvi42 (Cardiovascular Imaging, Calgary, Alberta, Canada) using the full-width half-maximum technique (16). If infarction was seen in more than 1 coronary territory in the pre-discharge CMR, this was recorded as being in the IRA territory (associated edema and/or MVO) or the non-IRA territory with the   Normally distributed continuous variables were expressed as mean AE SD, and comparison was with Student t tests. Non-normally distributed data were expressed as median (25th to 75th quartiles) and analyzed using independent Student t testing where log transformation normalized data, and using Mann-Whitney testing were the degree of skew rendered data nontransformable. Categorical variables were compared using chi-square testing. Clinical outcomes were assessed using time-to-first-event survival analysis (log-rank test with right censoring), and Cox proportional hazard models were fitted to estimate hazard ratios and 95% confidence intervals for treatment comparisons. One hundred patients in each arm gave 81% power to detect a 4% absolute difference in infarct size, assuming a mean of 20% of LV mass and standard deviation of 10% (19,20), using a 2-tailed test with alpha ¼ 0.05. New infarct comprising 4% of LV mass is associated with adverse prognosis in patients with revascularization-related injury (21).

PATIENTS.
In the CMR substudy, 205 consented to participate. Of these, 2 patients were excluded: 1 patient did not complete the early CMR, and in 1 patient, the LGE images were not analyzable. The The CvLPRIT CMR Substudy complete revascularization and IRA-only groups in the CMR substudy were well-matched for characteristics, with no statistically significant differences between groups, although there was a trend for more women in the IRA-only group ( Table 1).
ANGIOGRAPHIC AND PCI DETAILS. Data are shown in Table 2 Table 3. Pre-discharge CMR was undertaken at a me- The prevalence of multiple territory infarcts in the complete revascularization group was double that of the IRA-only group and the number of acute non-IRA infarcts was increased 3-fold in those undergoing complete revascularization (Table 3). Examples, with corresponding edema images, are shown in Figure 3,  Table 2). Values are n/N (%), median (interquartile range), or mean AE SD. The bold type indicates statistically significant p values. Additional antianginal medication includes calcium-channel blockers, nitrates, or nicorandil. *Non-normally distributed data: analyzed after log transformation with independent Student t testing.
McCann et al. MVO was present in more than one-half of all patients, although quantitatively, the amount was very low (median <0.2% of LV mass). In 52 patients (26%), AAR could not be quantified: no artifact, but no edema discernable (n ¼ 33); not performed due to arrhythmia or suboptimal breath-holding (n ¼ 14); or severe artifact (n ¼ 5). AAR and MSI were lower, but not significantly, in the complete revascularization group. LV volume, mass, and ejection fraction were similar in both groups.  Values are n/N (%), median (interquartile range), mean AE SD, or n (%), unless otherwise noted. The bold type indicates statistically significant p values. *Nonnormally distributed data: analyzed after log transformation with independent Student t testing. †Nonnormally distributed data: analyzed using Mann-Whitney analysis. ‡Because the median and interquartile range was 0 (0 to 0) for both IRA and CR groups, mean AE SD of the results are presented although the data are nonparametrically distributed. §Analyzable edema imaging available in 75 of the complete revascularization group and 77 of the IRA-only group.

DISCUSSION
This is the first detailed study of pre-discharge and follow-up CMR outcomes in a randomized study of IRA-only versus complete revascularization in multivessel coronary disease at PPCI. The data have confirmed that non-IRA PCI is associated with additional infarction. However, these type 4a MIs (6) are relatively infrequent, generally small, and did not result in an increase in total infarct size. There is mounting evidence from randomized trials that treating multivessel disease with complete revascularization (4,22) leads to a reduction in MACE after PPCI compared with an IRA-only strategy.
The patients in the substudy had similar baseline characteristics to those in the main trial. Because time to revascularization (4) and anterior MI (23) are strongly associated with infarct size, randomization was stratified by these variables. There was a similar reduction in the hazard ratio for MACE in the complete revascularization CMR subgroup as that seen in the main study compared with IRA-only revascularization, and we believe that the CMR substudy population is representative of those in the main study.
It is well-recognized that elective PCI can cause a troponin rise in approximately 30% of patients and approximately 50% undergoing PCI for unstable angina (24). Such type 4a MIs (6) can be detected on CMR and have been associated with adverse prognosis (21,25). In this substudy of CvLPRIT, the prevalence of >1 CMR-detected infarct in patients receiving complete revascularization was double that in the IRA-only arm (23.8% vs. 11.2%), and more than 3-fold for the acute non-IRA infarcts (17.1% vs. 4.8%) (Central Illustration). Previous Q-wave MI was an exclusion criterion in this study, but 4% had a history of previous non-STEMI, and a similar number (6% in the IRA-only and 5% in the complete revascularization groups) had chronic non-IRA MI on the predischarge CMR. Excluding these patients did not significantly affect the results. These data suggest that an additional 12% of patients with multivessel disease who receive complete revascularization at the time of PPCI will have evidence of additional CMRdetectable infarction compared with IRA-only revascularization. However, this proportion is less than might have been expected from previous studies in elective PCI (24), where up to 29% of patients have evidence of new infarction on CMR associated with troponin elevation (25). The extent of acute non-IRA infarction was also smaller (median 2.5% of LV mass) than may have been anticipated from elective PCI data given that average infarct size in those with new late enhancement on CMR was 5.0 AE 4.8% of LV mass (25), despite all patients in that study being pre-treated with clopidogrel for >24 h and given a glycoprotein IIb/IIIa inhibitor periprocedurally.
Importantly, in the present study, total infarct size was not increased in the short term or at follow-up, and there were no significant differences in myocardial salvage, LV volumes, or ejection fraction between the treatment groups. Peak creatine kinase levels were also similar in the 2 groups.
These findings provide reassurance that non-IRA intervention at the time of PPCI does not lead to increased total infarct size. In the main CvLPRIT trial, complete revascularization resulted in a significantly reduced hazard ratio for 12-month MACE despite the greater prevalence of CMR-detected type 4a MIs shown here. There are limited data as to whether revascularization-induced myocardial injury detected by CMR is linked to prognosis (21), and none in patients presenting with STEMI. In an observational study of 152 patients undergoing elective revascularization, 32% had evidence of new LGE, which averaged 5 g (4% of LV mass), and one-half of these patients were treated with coronary artery bypass surgery (21). In that study, patients with new infarction following revascularization had reduced ejection fraction, increased LV volume, increased total infarct size, and a 3-fold increase in MACE at a median of 2.9 years follow-up compared with those without new LGE (21). Given that the complete revascularization group in the current study had no increase in total infarct size, LV volume, or reduced ejection fraction, it seems unlikely that the short-to medium-term clinical benefits of complete revascularization (22) will be offset in the long term by increased heart failure or sudden cardiac deaths. However, longerterm follow-up of patients in this study is needed to confirm this.
We did not observe any significant differences in myocardial salvage between the treatment groups in this study. Non-IRA revascularization at the time of PPCI could increase perfusion to watershed areas by relieving flow-limiting stenoses, resulting in increased myocardial salvage (26). Alternatively, resting myocardial perfusion and flow reserve following PCI may actually be reduced, as has been shown in elective patients as a result of distal embolization, particularly when the PCI is associated with new LGE (26,27). It may be that both effects are seen with non-IRA PCI resulting in no net benefit with regard to myocardial salvage in the PPCI setting.
Unexpectedly, we also observed no difference in ischemic burden between the groups undergoing follow-up stress perfusion CMR. There are several potential explanations for this finding. First, it is well recognized that even severe angiographic stenoses may not cause ischemia (28,29). Second, 11 patients in the IRA-only arm had further PCI before Values are n (%), median (interquartile range), or mean AE SD. A dash indicates that no HR was presentable because 1 or both treatment arms had an incidence of 0.
CI ¼ confidence interval; CVA ¼ cerebrovascular accident; HF ¼ heart failure; HR ¼ hazard ratio; MACE ¼ major adverse cardiovascular events; OR ¼ odds ratio; TIA ¼ transient ischemic attack; other abbreviations as in Table 1. The CvLPRIT CMR Substudy the stress CMR that is likely to have reduced ischemic burden in this group. Third, the small number of crossovers from randomization is likely to have diminished the differences in ischemia between the groups. Finally, the stress CMR was undertaken in patients on optimal medical therapy, which may dramatically reduce post-MI ischemia (30) making it more difficult to detect differences between the groups, especially as there was higher use of a second antianginal medication in the IRAonly group. This may also explain why the overall ischemic burden in our study was small (3% to 4%).
It remains to be determined whether ischemia is prognostically important in the PPCI era, especially because medical therapy may result in similar clinical outcomes to a revascularization strategy even in patients treated with thrombolysis (30). Further insight on this subject will be available from the CvLPRIT nuclear substudy.
STUDY LIMITATIONS. The optimal timing to assess infarct size post-STEMI is uncertain (31

CONCLUSIONS
An in-hospital complete revascularization strategy in patients with multivessel disease at the time of PPCI is associated with a small increase in type 4a MIs in non-IRA territories, but total infarct size was not significantly different compared with an IRA-only strategy.