Significance of Microvascular Function in Visual—Functional Mismatch Between Invasive Coronary Angiography and Fractional Flow Reserve

Background Despite a moderate correlation between angiographical stenosis and physiological significance, the mechanism of discordance has not been fully elucidated, particularly regarding the significance of microvascular function. This study sought to clarify whether microvascular function affects visual‐functional mismatch between quantitative coronary angiography (QCA) and fractional flow reserve (FFR). Methods and Results We assessed QCA, FFR, coronary flow reserve, and the index of microcirculatory resistance in 849 non‐left‐main coronary lesions with visually estimated intermediate stenoses from 532 patients. Clinical and lesion‐specific characteristics and physiological parameters associated with mismatch and reverse mismatch were studied. Coronary flow reserve and index of microcirculatory resistance showed a weak, but significant, correlation with FFR (R=0.306, P<0.001 and R=0.158, P<0.001, respectively). Four hundred twenty‐two lesions were visually nonsignificant (diameter stenosis assessed by QCA [QCA‐DS] ≤50%) and 427 lesions were visually significant (QCA‐DS >50%). Among visually nonsignificant lesions, FFR ≤0.80 (reverse mismatch) was observed in 129 lesions (30.6%). Among visually significant lesions, FFR >0.80 (mismatch) were observed in 179 lesions (41.9%). The significant predictors of reverse mismatch were male sex, nonculprit lesions of acute coronary syndrome, left anterior descending artery location, smaller QCA reference diameter, greater QCA‐DS, lower coronary flow reserve, and lower index of microcirculatory resistance. Mismatch was associated with right coronary artery location, greater QCA reference diameter, smaller QCA‐DS, lesion length, higher coronary flow reserve, and higher index of microcirculatory resistance. Conclusions There was a high prevalence of visual‐functional mismatches between QCA and FFR. The discrepancy was related to clinical characteristics, lesion‐specific factors, and microvascular resistance that was undistinguishable by coronary angiography, thus suggesting the importance of physiological lesion assessment.

I n addition to clinical characteristics and lesion location, microvascular function may play important roles for visualfunctional mismatch and reverse mismatch, which corroborates the importance of coronary flow physiological assessment.
Invasive coronary angiography is a well-accepted method for identifying the presence of flow-limiting epicardial coronary artery stenosis and for guiding revascularization. The seminal study by Gould et al reported that hyperemic flow begins to decline in the presence of stenosis with a reduction in diameter larger than 50%. 1 This cut-off value has been used for the threshold of inducible ischemia; therefore, it is accepted as the gold standard for guiding revascularization, validating noninvasive testing, and evaluating outcomes after revascularization strategies. [2][3][4] However, cumulative evidence suggests that angiographically determined anatomical stenosis severity often underestimates or overestimates the functional significance of lesions. [5][6][7] Fractional flow reserve (FFR) is currently the standard for decision-making regarding revascularization in the catheter laboratory and has become part of the clinical guidelines for the assessment of the physiological significance of epicardial coronary stenosis based on sound concepts and randomized clinical trials. [8][9][10][11] However, FFR evaluation is still underutilized; instead, coronary angiography is widely used as a gatekeeper for decisionmaking of revascularization even in large clinical trials. 2,12 Subanalysis of the FAME (Fractional Flow Reserve Versus Angiography for Multivessel Evaluation) trial has shown that angiography is inaccurate for assessing the functional significance of coronary stenosis when compared with FFR guidance, especially for intermediate stenosis. 6 Although such "visual-functional mismatch" or "reverse visual-functional mismatch" are not uncommon, the mechanism of these phenomena has not been fully elucidated. Recent studies further specifically suggest the relevance of microvascular function to FFR. [13][14][15] Visual-functional mismatch has important implications in overcoming limitations related to angiography-guided decision-making to reduce unnecessary revascularization and avoiding missed appropriate revascularization. Therefore, we sought to identify lesion-specific, patient-related factors associated with visual-functional mismatch using a large cohort of the institutional database that includes FFR, coronary flow reserve (CFR), and the index of microcirculatory resistance (IMR) information with a particular focus on microvascular function.

Study Population
A retrospective analysis of pooled data was performed in the institutional FFR registry listing pressure and ECG tracings between December 2010 and May 2016 that included consecutive 1645 FFR measurements of 815 patients referred for diagnostic or therapeutic catheterization. Institutional indication criteria for coronary physiological assessment including FFR, CFR, and IMR were intermediate coronary stenosis showing 30% to 80% diameter stenosis by visual estimation in patients with stable angina pectoris; in those with suspected asymptomatic ischemia and/or microvascular dysfunction; or in nonculprit lesions of acute coronary syndrome (ACS) more than 2 weeks after onset. We selected 1042 eligible lesions from 623 patients who met the following criteria ( Figure 1): age older than 20 years; detection of an identifiable de novo lesion located at the proximal to middle portion of a native coronary artery; and patients with stable coronary disease or those with ACS in whom the culprit lesions were treated >2 weeks before the examination. Patients with a history of coronary artery bypass surgery, culprit lesions of ACS, left main disease, and in-stent restenosis were excluded from the analysis. Although left main disease has been reported to be an important factor in reverse visual-functional mismatch, 16 the present study excluded those lesions because the relationship between left main disease and microvascular function is not able to be assessed by the IMR. In those 1042 eligible lesions, we further identified lesions with sufficient physiological data for the determination of FFR, CFR, and IMR. Insufficient acquisition of physiological data, such as pressure drift (PD) >3 mm Hg, insufficient waveform during the examination, and repeated FFR measurements irrespective of PD within 3 mm Hg, were excluded. The study was approved by the local ethics committee and conformed to the Declaration of Helsinki statement on research involving human subjects. Informed consent was provided by all participants after a complete explanation of the protocol and potential risks.

Clinical Perspective
What is New?
• High coronary flow reserve and high index of microcirculatory resistance were associated with visual-functional mismatch in which angiographic diameter stenosis exceeds 50% and FFR was not indicative of ischemia (FRR >0.80). • Low CFR and low IMR were associated with visual-functional reverse mismatch in which angiographic diameter stenosis was less than 50% and FFR suggested the presence of ischemia (FFR≤0.80).
What are the Clinical Implications?
• Visual-functional mismatch and reverse mismatch are not rare. • In addition to clinical characteristics and lesion location, microvascular function may play important roles for visualfunctional mismatch and reverse mismatch, which corroborates the importance of coronary flow physiological assessment.

Cardiac Catheterization
Each patient initially underwent standard selective coronary angiography through the radial artery using a 6-Fr system. Quantitative analyses of angiograms were performed using offline analysis software (QAngio XA; Medis Medical Imaging Systems, Leiden, The Netherlands) to measure minimum lumen diameter, (QCA) reference lumen diameter (QCA-RD), QCA percent diameter stenosis (QCA-DS), and lesion length of the target lesion. All patients received a bolus injection of heparin (5000 IU) before the procedure and an additional bolus injection of 2000 IU every hour if the procedure required more than 1 hour. An intracoronary bolus injection of nitroglycerin (0.2 mg) was administered at the beginning of the procedure and repeated every 30 minutes. Intravenous infusion of ATP (150 lg/kg/min) was used to induce hyperemia. According to the manufacturer's specifications, which reported an acceptable pressure error range of 3%, we analyzed FFR measurements with PD ≤3 mm Hg in the present study. Our institutional standard protocol recommends repeated FFR measurements when PD ≥4 mm Hg; however, the final decision to repeat the measurements is at the discretion of the operator. Lesions with PD >3 mm Hg were excluded from the analysis, and FFR was calculated as a corrected value of distal coronary pressure (Pd)/proximal coronary pressure (Pa) after correction of Pd by PD.

Intracoronary Physiological Indices
FFR, CFR, and IMR were measured with a RadiAnalyzer Xpress instrument and a PressureWire Certus (pressure-temperature sensor wire; St. Jude Medical, Uppsala, MN). 5,17,18 The pressure wire was introduced, zeroed, and equalized to the catheter tip pressure before crossing the lesion. Afterward, the pressure sensor was positioned 8 to 10 cm distal to the ostium of the studied artery across the lesion. Baseline pressures were recorded for at least 20 seconds after the guiding catheter was flushed with saline. Thereafter, FFR measurements were performed at stable hyperemia. PD was determined when the pressure sensor was pulled back and reached the tip of the guiding catheter. All pressure and ECG tracings in the catheterization laboratory's monitoring system (RMC-4000 Cardio Master with EP amplifier system JB400G; Nihon Koden, Tokyo, Japan) were submitted to the in-hospital core cardiac physiology and morphological analysis laboratory. Insufficient waveform quality, including absence of pressure signal, catheter-damped waveform, and inappropriate Pd waveform, were excluded. Analyzed FFR data were compared with the original readout values determined by the catheterization laboratory; a consensus reading was agreed on by 2 expert physicians if there were discordant values. In the present study, IMR was calculated as the product of the mean distal coronary pressure during stable hyperemia and mean hyperemic transit time (Tmn) and corrected by using the following formula proposed by Yong et al 19 : IMR=Pa9Tmn9([1.359Pd/Pa]-0.32). In the absence of a validated cutoff to identify abnormally increased hyperemic microvascular resistance and the reported variability of IMR in patients with or without coronary heart disease, IMR values ≥75th percentile (28.0) of the present cohort were arbitrarily assumed as high IMR. 20,21 CFR was also measured simultaneously with FFR using the thermodilution method, as described elsewhere. 18 In this study, low CFR was defined as a value <2.0, consistent with previous studies. 22

Definition of Visual-Functional Reverse Mismatch and Mismatch
The clinical cut-off point of functionally significant ischemia was defined as FFR of 0.80. Visually significant stenosis was defined as QCA-DS >50%.  Table 4).

Determinants of Reverse Mismatch and Mismatch
Univariate and multivariable analyses of factors predicting reverse mismatch in visually nonsignificant lesions are shown in Table 5. Multivariable analysis identified male sex, nonculprit lesion of acute coronary syndrome, lower low-density lipoprotein cholesterol level, left anterior descending coronary artery lesions, smaller QCA-RD, larger QCA-DS, lower CFR, and lower IMR as the significant predictors of reverse mismatch in visually nonsignificant lesions. In visually significant lesions, significant predictors of mismatch were right coronary artery, greater QCA-RD, smaller QCA-DS, shorter QCA lesion length, higher CFR, and higher IMR (Table 6). Figure 3 shows the frequency of reverse mismatch in visually nonsignificant lesions and that of mismatch in visually significant lesions stratified by IMR (>28.0) and CFR (≤2.0). In visually nonsignificant lesions, frequency of reverse mismatch was significantly different among the 4 groups categorized by IMR and CFR and highest in lesions with low IMR and low CFR values. In contrast, in visually significant lesions, frequency of mismatch was highest in the group with high IMR and high CFR (Figure 3).

Discussion
The main findings of the present study were as follows: (1) QCA-derived visual and FFR-defined functional discordance were not rare, with frequencies of mismatch and reverse mismatch being 21.1% and 15.2%, respectively; (2) both mismatch and reverse mismatch were related to the clinical characteristics, lesion-specific factors, and microvascular resistance; (3) high hyperemic microvascular resistance and high CFR were significant predictors of mismatch; and (4) low hyperemic microvascular resistance and low CFR were significant predictors of reverse mismatch. To the best of our knowledge, this is the first study showing that microvascular function affects both visual-functional mismatch and reverse mismatch by using IMR as a specific marker of microvascular resistance. Invasive coronary angiography remains a gatekeeper for decision-making in revascularization because the utilization of FFR is still low, despite the recommendations in the guidelines. 23 The relationship between angiographic stenosis severity and functional significance assessed by FFR and the mismatch between these 2 measures have been extensively investigated to optimize percutaneous coronary intervention (PCI). 16,[24][25][26] Clinical factors, including age, sex, angiographical lesion-related factors, and physiological factors, have been reported to contribute to visual-functional mismatch and reverse mismatch. 15,16,[26][27][28] Our results were in line with these previous studies and further suggested the importance of microvascular function in both mismatch and reverse mismatch. These findings might be related to the fact that FFR as an index is affected by coronary flow volume through the stenosis. Decreased hyperemic microvascular resistance implies that coronary flow might increase for a fixed stenosis and a given driving pressure, and it may dictate a low FFR value despite relatively low resistance to coronary flow induced by the epicardial stenosis and microvascular resistance. In contrast, with increased microvascular resistance, coronary flow through the given fixed stenosis will decrease and FFR will increase without alteration in epicardial resistance to flow resulting from the epicardial stenosis.

Microvascular Resistance and Functional Stenosis Significance
Although the unequivocal benefit of FFR-guided revascularization over angiographic guidance has been established, [9][10][11] it should be noted that FFR is determined under the assumption of minimal and constant microvascular resistance. Our results suggest that microvascular resistance  15 The most important factor related to outcome in  patients with coronary heart disease is the presence or absence of inducible myocardial ischemia and its extent. It has been proposed that the presence of microvascular dysfunction is not an obstacle for making decisions based on FFR, provided that minimal and constant microvascular resistance with induced hyperemia is deemed to be nonreversible or consistent before and after revascularization. 28,29 However, several studies reported the evidence of serial changes in adenosine-induced microvascular resistance before, after, and during follow-up post-PCI. 13,30,31 van de Hoef et al also reported that identification of epicardial disease severity by FFR is partly obscured by microvascular resistance and that FFR increased with increasing hyperemic microvascular resistance to epicardial disease of equivalent severity. 15 Our results shed light on the relationship between FFR and microvascular function and indicate that low hyperemic microvascular resistance most likely influences reverse visual-unctional mismatch for intermediate epicardial stenoses, and that increased microvascular resistance affects visual-functional mismatch. These findings suggest that the mismatch group might include, at least in part, lesions with low pressure gradients attributed to the decreased coronary flow as a consequence of impaired microvascular resistance. 32 In other words, the present study suggests that FFR >0.80 and QCA-DS >50% not only identify vessels without the need for revascularization, but also include vessels with impaired and nonimpaired coronary flow. This potential abnormality in coronary physiology beyond the FFR assumption might provide reasons for why patients with revascularization deferral in randomized trials were not free from long-term cardiac events, such as in DEFER and FAME followup data, why 12% of patients with FFR >0.80 required revascularization within 2 years of deferral [9][10][11] and, conversely, why some patients with FFR ≤0.80 but with preserved CFR have a low rate of major adverse cardiac events at follow-up. 33

CFR and Functional Stenosis Significance
In the present study, in addition to IMR, CFR was associated with the FFR value. CFR is the ratio of maximum stress flow to rest flow for the artery of interest; this fundamentally represents coronary flow capacity modified by the integration of epicardial stenosis, diffuse arterial   narrowing, and microvascular dysfunction. 32 Our results demonstrated that CFR positively correlated with the FFR value in the linear regression analysis. Furthermore, a high CFR value was a significant predictor of visual-functional mismatch in visually significant lesions, and low CFR value was a significant predictor of reverse mismatch in visually nonsignificant lesions. Because CFR is likely to be composed of 3 elements (epicardial stenosis, diffuse narrowing, and microvascular function) and FFR represents the functional significance of epicardial stenosis, our results suggest that epicardial stenosis might influence a significant portion of CFR, rather than diffuse narrowing or microvascular function, in the setting of visual-functional mismatch evaluation because CFR and FFR showed an independent, parallel correlation.

Microvascular Resistance, Coronary Flow, and Functional Stenosis Significance
In the present study, a non-negligible proportion for whom the FFR value was >0.80 showed abnormal coronary microvascular resistance (IMR >28.0; 135 of 472; 28.6%) and disturbed coronary hemodynamics (CFR <2.0; 96 of 472; 20.3%). In those lesions, hyperemic coronary flow might increase if the increased resistance is reduced after revascularization, as suggested by our recent studies. 13,30 The exact mechanism and pathophysiology of increased or decreased hyperemic microvascular resistance were not known and the change in microvascular resistance and its effect on absolute coronary flow after revascularization remain elusive. Further studies are needed to elucidate the relationship between microvascular

Clinical Implication of Mismatch and Reverse Mismatch
Although frequency and determinants of visual-functional mismatch have been investigated in the present and previous studies, 15,16 clinical implication of those lesions has not been elucidated. Specifically, clinical outcomes of the lesions with mismatch or reverse mismatch have not been sufficiently investigated, and, moreover, best therapeutic strategies for those lesions have not been well discussed. If we follow the data from the FAME and FAME2 trials, 10,11 the mismatch group should be treated by optimal medical therapy alone and the reverse-mismatch group should be treated by PCI and optimal medical therapy. However, given the substantial population of patients showing FFR within the gray zone ranging from 0.75 to 0.80 in the reverse mismatch group, superiority of PCI for those lesions may deserve further consideration. 34 Previous studies have suggested the impact of plaque characteristics on FFR values for a given anatomical stenosis in an epicardial coronary artery. 35 The present study reports the association of CFR and IMR with FFR. Although these factors might be represented by a low FFR value as a single variable, nevertheless, integrated information on patient characteristics, coronary flow, and plaque morphology may provide additive predictive values on FFR, which might enable tailored strategies for lesions, such as gray zone lesions, in which the prognostic values of FFR might be suboptimal.

Study Limitations
The results of the present study should be interpreted with the consideration of some limitations. First, this was a retrospective study performed at a single center without a dedicated core laboratory for angiographic analysis. Exclusion of patients with significant left main disease, renal impairment, heart failure, or acute coronary syndrome may have resulted in selection bias. This study enrolled patients with stable angina pectoris based on symptoms and noninvasive test results and those who were referred to the catheter laboratory for treatment or diagnosed by diagnostic catheterization at our institution, which might inevitably involve referral bias. There was no clinically validated or normal range of IMR. Therefore, the quartiles of IMR in the study population were used to define high (highest quartile) and low (lowest quartile) IMR. Although this approach provides a reasonable estimation of an abnormal IMR range for patients with coronary heart disease, our results should be tested in an independent study population. Coronary wedge pressure was not measured because this study cohort included lesions that were not treated with PCI, which might have led to overestimation of IMR in tight stenoses with significant collateral flow. 36

Conclusions
The present data, which are in line with those of previous studies, indicate that coronary angiography underestimates or overestimates physiological stenosis severity in comparison with FFR in non-negligible proportion of visually intermediate lesions.
In addition, our results indicate that these visualfunctional mismatch and reverse mismatch are related to coronary flow and microvascular function, which may emphasize the importance of coronary flow assessment and coronary pressure indices.