Percutaneous Coronary Intervention vs Medical Therapy for Coronary Lesions With Positive Fractional Flow Reserve (FFR) but Preserved Pressure-Bounded Coronary Flow Reserve (CFR): A Substudy of the Randomized Compare-Acute Trial

Objectives. Performing percutaneous coronary intervention (PCI) for fractional flow reserve (FFR) positive coronary lesions improves clinical outcomes and is recommended by international guidelines. It has been hypothesized that lesions with a positive FFR but a preserved coronary flow reserve (CFR) are less likely to be flow limiting and might best be treated medically. We investigated the association of CFR in FFR-positive lesions with clinical outcomes when treated medically, as well as the treatment effect of PCI vs medical therapy in FFR-positive lesions and a preserved CFR. Methods. We performed a substudy of the randomized, multicenter Compare-Acute trial, in which stabilized ST-segment elevation myocardial infarction (STEMI) patients with non-culprit lesions were randomized to either FFR-guided PCI or medical therapy. Based on baseline and hyperemic pressure gradients, we computed physiologic limits of CFR, the so-called pressure-bounded CFR (pb-CFR), and classified lesions as low (<2) or preserved (≥2). The primary endpoint was 12-month major adverse cardiac and cerebrovascular event (MACCE) rate, defined as a composite of death from any cause, non-fatal myocardial infarction, revascularization, or cerebrovascular events. Results. A total of 980 lesions from 885 patients were included in this substudy. In lesions with FFR ≤0.80, a total of 249 patients had a pb-CFR<2 and 29 patients had a preserved CFR (pb-CFR ≥2). The rate of MACCE at 1 year was not significantly different between patients with FFR ≤0.80 and pb-CFR <2 vs patients with FFR ≤0.80 and pb-CFR ≥2 (25% vs 17%, respectively; P=.39). Because of randomization, baseline characteristics were well balanced between patients with FFR ≤0.80 and pb-CFR ≥2 treated by either by PCI or medical therapy. Importantly, in patients with FFR ≤0.80 and pb-CFR ≥2, MACCE occurred more frequently in patients treated medically vs patients treated by PCI (44% vs 0%, respectively; P=.01). Conclusions. Preserved or low pb-CFR did not alter clinical outcomes in patients with a positive FFR. Patients with FFR-positive coronary lesions but a preserved CFR had more clinical events when treated medically vs those treated with PCI. 

J INVASIVE CARDIOL 2021;33(7):E557-E564. Epub 2021 June 5.

Key words: coronary artery disease, coronary flow reserve, fractional flow reserve

According to current guidelines, fractional flow reserve (FFR) is recommended to assess the hemodynamic significance of stable coronary lesions.^1,2 In addition, randomized controlled trials, including the Compare-Acute study,³ have shown that percutaneous coronary intervention (PCI) of FFR-positive coronary lesions improves clinical outcomes vs medical therapy for non-culprit lesions in ST-segment elevation myocardial infarction (STEMI).^3-6

While FFR only assesses the epicardial compartment of the coronary tree, coronary flow reserve (CFR) provides combined physiologic information on both epicardial and microvascular function. Considering the frequent multilevel involvement of coronary artery disease, CFR could remain additive to FFR.⁷ It has been hypothesized that lesions with a positive (abnormal) FFR but a preserved CFR — termed “non-flow-limiting” by some⁸ — might be best treated medically. However, the current literature lacks randomized controlled data to confirm or refute this hypothesis.

In the Compare-Acute trial, coronary pressure measurements were performed both during rest (baseline) and maximal hyperemia, allowing calculation of both FFR and pressure-bounded CFR (pb-CFR).^7,9 Accordingly, the aim of this posthoc analysis of the large, randomized, controlled Compare-Acute trial was to investigate whether FFR-positive non-culprit lesions with a preserved pb-CFR have a better clinical outcome after treatment with PCI vs medical therapy. 

Overview of the Compare-Acute trial and study population. The study design, methods,¹⁰ and results of the Compare-Acute trial have been reported previously.³ In brief, the Compare-Acute study (NCT01399736) was a multicenter, prospective, investigator-initiated, interventional trial that evaluated FFR-guided treatment of non-infarct related coronary artery lesions in the acute setting compared with a strategy of infarct-related coronary-artery only treatment in patients with STEMI and multivessel disease. Patients between 18-85 years old were eligible if they presented with STEMI within 12 hours after symptom onset, had an indication for primary PCI, and showed non-culprit lesions with ≥50% diameter stenosis according to quantitative coronary angiography or visual assessment, and were felt to be appropriate candidates for PCI by the interventional cardiologist. Non-infarct related coronary artery lesions were those identified as not being responsible for the acute myocardial infarction on the basis of electrocardiography (ECG) and angiography.

Patients were excluded with left main disease, chronic total occlusion, non-culprit vessel stenosis not amenable for PCI treatment (at the operator’s decision), Thrombolysis in Myocardial Infarction (TIMI) flow grade of ≤2 in the non-infarct related coronary artery, suboptimal result or complications after treatment of an infarct-related coronary artery, severe valve dysfunction, and Killip class III or IV. All events, such as recurrent revascularization, were analyzed and adjudicated by an independent clinical evaluation committee. A revascularization was considered to be clinically indicated if quantitative coronary angiography revealed a stenosis ≥50% in diameter and if ≥1 of the following criteria was met: history of recurrent angina pectoris, presumably related to the target vessel; ischemic changes on ECG at rest or during exercise testing (or the equivalent), presumably related to the target vessel; abnormal results on any invasive functional diagnostic test (eg, a Doppler pattern of flow velocity reserve or FFR during follow-up); or stenosis of ≥70% diameter on quantitative coronary angiography in the absence of signs or symptoms of ischemia. 

Pressure-bounded CFR. In the Compare-Acute study, operators were asked to record the ratio of distal coronary pressure (Pd) to aortic pressure (Pa) both during rest (resting Pd/Pa) and during maximal hyperemia (FFR). Pressure-bounded CFR was calculated as previously described.^7,9 Assuming that Pa remained constant, the possible CFR values can be bounded by: 

√((1–FFR)/(1–Pd/Pa)) ≤ CFR ≤ (1–FFR)/(1–Pd/Pa)

In accordance with previous publications,^7,9 we classified lesions into 3 distinct pb-CFR groups: “low” when the upper boundary of CFR was <2; “preserved” when the lower boundary of CFR was ≥2; and “indeterminate” when the range crossed the value of 2. Additionally, we excluded lesions with a resting Pd/Pa value of 0.99 or 1.00 since pb-CFR becomes imprecise in the presence of small baseline gradients.

Endpoints and definitions. The primary endpoint was rate of major adverse cardiac and cerebrovascular event (MACCE), defined as a composite of all-cause mortality, non-fatal myocardial infarction, any revascularization, or cerebrovascular event. Secondary endpoints were all-cause death, cardiac death, non-fatal myocardial infarction, cerebrovascular event, and any revascularization. Elective, clinically indicated revascularizations performed within 45 days after the primary intervention were not counted as events, in accordance with the protocol. Urgent revascularizations performed within 45 days or further revascularizations performed thereafter counted as events. All endpoints were analyzed on a per-lesion and per-patient basis. For per-patient analysis, the lowest value of FFR and pb-CFR was selected as representative for each patient.

The first question we studied was whether the natural history of patients with low FFR (≤0.80) but preserved pb-CFR (≥2) was indeed more favorable regarding clinical outcomes than patients with low FFR (≤0.80) and low pb-CFR (<2). We studied the natural history in the optimal medical therapy or “culprit-only” group. The second question was whether patients with low FFR but preserved pb-CFR are best treated by PCI or medical therapy. 

Statistical analysis. Standard statistical techniques were performed. Continuous baseline variables are expressed as mean ± standard deviation or median (interquartile range) and compared with the unpaired Student’s t-test or Mann-Whitney U-test, when appropriate. Fisher’s Exact test or Chi-square statistic was used as appropriate to compare categorical variables. Statistical analysis was performed using IBM SPSS Statistics for Macintosh, version 25.0 (IBM). A P-value <.05 was considered statistically significant. 

FFR and pb-CFR. As shown in the study flow chart (Figure 1), a total of 1307 lesions in 885 patients were identified. After excluding lesions with missing FFR or Pd/Pa values, or almost-normal Pd/Pa values of 0.99 or 1.00, a total of 980 lesions were included in the current analysis, of which 462 had an FFR ≤0.80. Of these, 249 lesions had pb-CFR <2, 29 lesions had pb-CFR ≥2, and 184 lesions had an indeterminate pb-CFR. Table 1 compares baseline characteristics between low CFR (<2) and intact CFR (≥2) in the group with low FFR (≤0.80). As expected physiologically, patients with low CFR had significantly lower resting Pd/Pa compared to patients with intact CFR. There were no other statistically significant differences in baseline characteristics between the low and intact pb-CFR arms in patients with FFR ≤0.80.

Natural history according to FFR and CFR. As depicted in Figure 2, medically treated patients had more clinical events with a positive FFR compared with a negative FFR. Nevertheless, there was no significant difference in MACCE for the FFR-positive group when divided into intact vs low pb-CFR subgroups (P=.24). The Kaplan-Meier curves in Figure 3 demonstrate no significant differences in clinical events on a per-lesion or per-patient level (P=.24 and P=.92, respectively), although intact pb-CFR lesions experienced a larger number of events compared with low pb-CFR lesions. In a per-patient analysis, the 1-year MACCE rate was 25% in the group with low pb-CFR and 17% in the group with intact pb-CFR (P=.39) in the patients with FFR ≤0.80 (Table 3). A per-lesion analysis also showed no significant difference in MACCE rate in the FFR positive group when divided into intact vs low pb-CFR subgroups (26% vs 28%, respectively; P=.74). 

Positive FFR and preserved pb-CFR in patients treated medically vs complete revascularization. The flow chart (Figure 1) details the 462 positive FFR lesions, of which 29 had pb-CFR >2; of these lesions, 11 were treated with PCI and 18 were treated with optimal medical therapy. Table 2 compares baseline characteristics between medically treated patients and those who underwent PCI with positive FFR (≤0.80) but with preserved pb-CFR (≥2). As expected, due to randomization of treatment assignment, characteristics were similar for both groups. 

Interestingly, as depicted in Table 4, in the per-patient analysis, there is a non-significant but larger number of clinical events in the medically treated group compared with the PCI group (29% vs 0%, respectively; P=.07). Notably, 3 spontaneous myocardial infarctions occurred in lesions with FFR ≤0.80 but pb-CFR ≥2 when treated medically, compared with 0 corresponding events when treated with PCI. In addition, using a per-lesion analysis, MACCE occurred significantly more often in the medically treated group than in the PCI group (44% vs 0%, respectively; P=.01). These findings are displayed in the Kaplan-Meier curves in Figure 4.

The current study provides the first posthoc analysis of randomized data on the value of pb-CFR in FFR-positive lesions. The 2 key findings of this posthoc substudy of the Compare-Acute trial are as follows: (1) preserved pb-CFR was not associated with improved outcomes vs low pb-CFR in the setting of a positive FFR; and (2) patients and lesions with a low FFR (≤0.80) but preserved pb-CFR (≥2) had more clinical events when treated medically compared with PCI.  

It has been hypothesized that a “benign” subgroup with preserved CFR exists within the larger, higher-risk group of FFR-positive lesions. The underlying theory for this hypothesis is that low FFR might be caused by an exuberant vasodilator response of the coronary vasculature resulting in a large, transstenotic, flow-induced pressure gradient but intact perfusion. The hypothesis claims that when FFR is abnormal but CFR remains intact, the stenosis might not limit flow; therefore, despite a substantial pressure drop across the stenosis, its treatment should be medical.⁸

To date, 3 randomized controlled trials assigned coronary lesions with a positive FFR to either FFR-guided PCI or medical therapy: FAME 2,⁵ DANAMI-3-PRIMULTI,⁶ and Compare-Acute.³ When pooled together, these 3 studies demonstrated a significant reduction in spontaneous myocardial infarction by FFR-guided PCI compared with medical therapy. However, only Compare-Acute captured information on the trans-stenotic pressure gradient during resting (baseline) conditions and, for this reason, uniquely allowed for posthoc calculation of pb-CFR. Pb-CFR provides an easy technique for obtaining information about CFR when measuring intracoronary pressure alone, but without the need to perform additional measurements or purchase special equipment.^7,9 Additionally, the pb-CFR was unknown to the Compare-Acute investigators and patients, thereby providing natural “blinding” that decreases bias and probably explains why unblinded Doppler-derived CFR was more related to future events (mostly repeat revascularization) than pb-CFR in 1 study.¹¹

Predictive value of pb-CFR in medically treated FFR-positive lesions. In our analysis, patients with FFR-positive but reduced pb-CFR lesions did not have a significantly different event rate vs patients who had FFR-positive but preserved pb-CFR lesions. While several studies examined the role of CFR in FFR-negative lesions, only a few studies have included FFR-positive lesions since most of these stenoses receive revascularization in clinical practice. 

In a large, Korean, prospective registry of 2088 coronary lesions, including 288 treated medically with simultaneous FFR and pb-CFR measurement, no incremental utility of pb-CFR was found for clinical events when added to FFR.⁷ Although it is unclear why FFR-positive lesions were deferred in this observational registry, the results are comparable to the current analysis. 

In line with both pb-CFR analyses, a non-randomized Korean study using thermodilution-derived CFR found no difference in event rates between 51 lesions that were FFR positive but had preserved CFR and 32 lesions that were FFR positive but had low CFR (hazard ratio, 1.012; P=.99).¹²

Beyond the presence of ischemia, several other mechanisms might explain why a large pressure drop across a coronary lesion (ie, FFR positive) might cause future events, including high mechanical forces on the plaque itself, the association with vulnerable characteristics, and the relationship with high plaque burden.^13,14 Focusing solely on ischemia, as is done in most FFR vs CFR discussions, neglects these important factors. 

PCI vs medical therapy for patients with FFR-positive but preserved pb-CFR lesions. While its modest sample size prevents definitive conclusions, our data suggest that PCI improves clinical outcomes vs medical therapy in FFR-positive lesions regardless of pb-CFR. Our data therefore contradict the hypothesis that FFR-positive but preserved pb-CFR lesions are best treated medically.⁸ 

In contrast, the large, non-randomized Korean registry found equivalent event rates between FFR-positive lesions with preserved pb-CFR when treated medically vs revascularized (Kaplan-Meier estimates 10.1% vs 10.2%, respectively).⁷ Again, these results should be interpreted with caution, as it is unclear why FFR-positive lesions were deferred in this registry. Patients with reduced FFR but preserved CFR are also being studied in the ongoing prospective DEFINE-FLOW (Distal Evaluation of Functional Performance With Intravascular Sensors to Assess the Narrowing Effect — Combined Pressure and Doppler FLOW Velocity Measurements; NCT02328820) study, in which patients with intact CFR despite reduced FFR receive optimal medical therapy alone.

Study limitations. This study has several limitations that should be acknowledged. First, this study was not a prespecified analysis.  Compare-Acute was designed and powered to address a different question than our current posthoc substudy. Second, in a large minority of subjects, pb-CFR produced an interval that overlapped with the CFR=2 threshold, in line with previous studies using pb-CFR.^7,9,11 Although pb-CFR has been validated against existing publications that measured CFR directly, it shows a reasonable, but not perfect, classification agreement. As such, pb-CFR should be considered an approximate tool rather than a definitive assessment.^7,9 Third, the sample size of patients with FFR-positive coronary lesions but preserved CFR was small, which limits definitive conclusions on clinical outcomes such as myocardial infarction and mortality. Fourth, our study population involves only patients with STEMI and not patients with stable coronary artery disease. Fifth, pb-CFR can only be calculated if adenosine was given to measure FFR. Some lesions had a resting full-cycle Pd/Pa <0.80 and adenosine was not administrated at the discretion of the operator. Finally, as FFR-negative lesions were treated medically in Compare-Acute, regardless of randomization, the current analysis cannot study PCI vs medical therapy in this subgroup. 

In this posthoc analysis of a randomized controlled trial of 885 patients with 980 lesions, preserved vs low pb-CFR was not associated with significantly better clinical outcomes for FFR-positive lesions. Lesions with positive FFR ≤0.80 but a preserved pb-CFR had more clinical events when treated medically vs those treated with PCI. These data suggest that a stenosis with an FFR ≤0.80 should be revascularized, even when pb-CFR is preserved.

From the ¹Department of Cardiology, Medical Center Leeuwarden, the Netherlands; ²Department of Cardiology, Catharina Hospital Eindhoven, Eindhoven, the Netherlands; ³Department of Cardiology, University Heart Center Freiburg-Bad Krozingen, Bad Krozingen, Germany; ⁴Department of Cardiology, Asklepeion General Hospital, Athens, Greece; ⁵Department of Cardiology, Heart Center Leipzig at the University of Leipzig, Leipzig, Germany; ⁶Department of Cardiology, Gothenburg University Hospital, Gothenburg, Sweden; ⁷Department of Cardiology, Maasstad Hospital, Rotterdam, the Netherlands; ⁸Department of Cardiology, Heart Center, Segeberger Kliniken, Bad Segeberg, Germany; and ⁹Department of Medicine, McGovern Medical School at UTHealth, Houston, Texas.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Pijls reports receipt of an institutional research grant from Abbott Vascular (St. Jude Medical) and OpSens, and possesses minor stock positions in Philips, ASML, and Heartflow. Dr Smits reports speaking fees and institutional research grants from Abbott Vascular (St. Jude Medical). Dr Johnson reports internal funding from the Weatherhead PET Center for Preventing and Reversing Atherosclerosis; patents pending on diagnostic methods for quantifying aortic stenosis and TAVI physiology and for correcting pressure signals from fluid-filled catheters; institutional licensing and consulting agreement with Boston Scientific for the smart minimum FFR algorithm; significant institutional research support from St. Jude Medical (CONTRAST, NCT02184117) and Philips Volcano Corporation (DEFINE-FLOW, NCT02328820) for studies using intracoronary pressure and flow sensors. The remaining authors report no conflicts of interest regarding the content herein. 

Manuscript accepted October 29, 2020.

Address for correspondence: Joost D.E. Haeck, MD, PhD, Heart and Vessel Center, Medical Center Leeuwarden, Cardiology Department, Henri Dunantweg 2, 8934AD, Leeuwarden, The Netherlands. Email: joost.haeck@mcl.nl