Prognostic value of myocardial flow reserve vs corrected myocardial flow reserve in patients without obstructive coronary artery disease

Abstract

Background:

Myocardial flow reserve (MFR) by positron emission tomography (PET) is a validated measure of cardiovascular risk. Elevated resting rate pressure product (RPP = heart rate x systolic blood pressure) can cause high resting myocardial blood flow (MBF), resulting in reduced MFR despite normal/near-normal peak stress MBF. When resting MBF is high, it is not known if RPP-corrected MFR (MFR_corrected) helps reclassify CV risk. We aimed to study this question in patients without obstructive coronary artery disease (CAD).

Methods:

We retrospectively studied patients referred for rest/stress cardiac PET at our center from 2006 to 2020. Patients with abnormal perfusion (summed stress score >3) or prior coronary artery bypass grafting (CABG) were excluded. MFR_corrected was defined as stress MBF/corrected rest MBF where corrected rest MBF = rest MBF x 10,000/RPP. The primary outcome was major cardiovascular events (MACE): cardiovascular death or myocardial infarction. Associations of MFR and MFR_corrected with MACE were assessed using unadjusted and adjusted Cox regression.

Results:

3276 patients were followed for a median of 7 (IQR 3–12) years. 1685 patients (51%) had MFR <2.0, and of those 366 (22%) had an MFR ≥2.0 after RPP correction. MFR <2.0 was associated with an increased absolute risk of MACE (HR 2.24 [1.79–2.81], P < 0.0001). Among patients with MFR <2.0, the risk of MACE was not statistically different between patients with an MFR_corrected ≥2.0 compared with those with MFR_corrected <2.0 (1.9% vs 2.3% MACE/year, HR 0.84 [0.63–1.13], P = 0.26) even after adjustment for confounders (P = 0.66).

Conclusions:

In patients without overt obstructive CAD and MFR< 2.0, there was no significant difference in cardiovascular risk between patients with discordant (≥2.0) and concordant (<2) MFR following RPP correction. This suggests that RPP-corrected MFR may not consistently provide accurate risk stratification in patients with normal perfusion and MFR <2.0. Stress MBF and uncorrected MFR should be reported to more reliably convey cardiovascular risk beyond perfusion results.

INTRODUCTION

Myocardial flow reserve (MFR), measured by positron emission tomography/computed tomography (PET/CT) as the ratio of stress over rest myocardial blood flow (MBF), is a well-validated marker of cardiovascular risk [1,2]. In the absence of obstructive epicardial coronary artery disease (CAD), an MFR value of below 2 is indicative of the cumulative hemodynamic burden of diffuse atherosclerosis and coronary microvascular dysfunction (CMD) [3,4].

Although a reduction in stress MBF is commonly observed in patients with reduced MFR, there are instances where MFR is reduced due to an abnormally high rest MBF. Two common causes for elevated rest MBF are resting tachycardia and/or hypertension. Given that rest MBF is associated with myocardial oxygen demand, a rate pressure product (RPP) adjustment can be applied to obtain a normalized MFR (MFR_corrected) [5,6]. While previous research has demonstrated that the association between MFR and outcomes in large patient cohorts persists even after RPP correction [2,7,8], the clinical implications of this correction remain controversial. Our objective was to assess whether RPP correction of abnormal MFR (<2) helps to refine risk stratification, in patients without overt obstructive epicardial CAD.

METHODS

Study population

We retrospectively included patients referred to the nuclear cardiology laboratory at Brigham and Women’s Hospital (Boston, MA) for rest/stress cardiac PET/CT from 2006 to 2020. Patients with abnormal perfusion (summed stress score >3) or prior CABG were excluded. This resulted in a final cohort of 3276 patients (Figure 1). Patient demographics and clinical history were collected prospectively at the time of PET/CT imaging. The Mass General Brigham Institutional Review Board approved this study with a waiver of informed consent, and data are available on reasonable request.

Figure 1.

Study population.

Position emission tomography/computed tomography imaging

Patients were imaged on a hybrid whole-body PET/CT scanner (Discovery RX, STE, or DMI LightSpeed 64, GE Healthcare) with ¹³N-ammonia (43%) or ⁸²Rubidium (57%) as flow tracers at rest and stress as per standard care. Pharmacologic stress agents included the vasodilators regadenoson (64%), dipyridamole (26%), or adenosine (7%) as well as dobutamine (3%). Summed rest, stress, and difference scores were computed using a standard 17-segment model and a 5-point scoring system. Rest and stress left ventricular ejection fraction (LVEF) were calculated from the gated images using commercially available software (Corridor4DM, INVIA Medical Imaging Solutions).

Myocardial blood flow assessments

Regional and global MBF (in mL/min/g of tissue) was quantified at rest and during maximal hyperemia as recommended by the American Society of Nuclear Cardiology and the Society of Nuclear Medicine and Molecular Imaging [6]. MFR was calculated as the ratio of stress to rest absolute MBF for the entire left ventricle. MFR_corrected was defined as stress MBF/corrected rest MBF where corrected rest MBF = rest MBF x 10,000/RPP. Impaired MFR and MFR_corrected was defined as a ratio <2.0 [2].

Outcomes

The primary outcome was the first occurrence of a major cardiovascular event (MACE), a composite of cardiovascular death or nonfatal myocardial infarction. A secondary outcome was revascularization, defined as percutaneous coronary intervention or coronary artery bypass grafting, and divided into early within 90 days and late after 90 days. Ascertainment of clinical endpoints was performed by a clinical endpoint committee using death certificate data and electronic health records. Nonfatal myocardial infarction was defined as a principal hospitalization diagnosis ICD code, or a non-principal hospitalization diagnosis when associated with a troponin elevation above the reference level [9].

Statistical analysis

Categorical variables are reported as frequencies with percentages. Continuous variables are reported as mean ± standard deviation (SD) or median with interquartile range (IQR) where appropriate. Chi-square and one-way analysis with variance were used to evaluate differences in categorical and continuous variables, respectively. Cox proportional hazard models were used to examine the unadjusted and adjusted association between MFR and MFR_corrected and MACE and create survival curves. Cox proportional hazards assumptions tests based on Schoenfeld residuals were used to verify that proportional hazards assumptions were met. Adjustment factors included age, sex, cardiac risk factors (hypertension, tobacco, dyslipidemia, diabetes, obesity, CKD), history of myocardial infarction, history of percutaneous coronary intervention, history of cardiomyopathy, symptoms at the time of MPI (chest pain, dyspnea, or palpitations), and resting LVEF. Direct adjusted survival curves were obtained from the Cox regression model by computing baseline function estimates at each time using stratified levels of each categorical adjustment variable and the mean values of continuous adjustment variables. Subgroup analyses were performed among patients with ¹³N-ammonia or ⁸²Rubidium radiotracer administration. All tests were 2-sided, and a value of P < 0.05 was considered statistically significant. Statistical analysis was performed using SAS 9.4 (SAS Institute Inc, Cary, NC).

RESULTS

Patient characteristics by MFR and MFR_corrected categories

Of 3276 patients referred to PET MPI without a history of CABG or overt flow-limiting CAD, 1685 (51%) had impaired MFR <2.0 (Figure 1). Patients with impaired MFR were older with a higher burden of cardiometabolic disease (obesity, diabetes, CKD) and prior MI and PCI (Table 1). Among patients with impaired MFR, 366 (22%) had an MFR_corrected ≥ 2. Among patients with preserved MFR, 389 (24%) had an MFR_corrected < 2 (Figure 1). Patients with MFR <2 and discordant MFR_corrected ≥ 2 were more often female, non-white Race, with hypertension and diabetes, and with higher rest MBF, rest RPP, and stress MBF (Table 1). Among all patients, rest MBF and rest RPP were weakly correlated (Pearson correlation 0.18, P < 0.001).

Table 1.

Baseline characteristics stratified by myocardial flow reserve results.

Baseline Characteristics	MFR ≥2 (n = 1591)	MFR <2 (n = 1685)	P-value	Discordant MFR <2 and MFRcorrected ≥ 2 (n = 366)	Concordant MFR and MFRcorrected < 2 (n = 1319)	P-value
Demographics
Age	60 ± 13	64 ± 14	<0.001	63 ± 14	65 ± 14	0.14
Female sex	892 (56%)	962 (56%)	0.55	229 (63%)	732 (56%)	0.02
Non-White race	438 (28%)	542 (32%)	<0.001	144 (39%)	398 (30%)	0.002
Cardiovascular Characteristics and Risk Factors
BMI (kg/m2)	30 ± 7	32 ± 9	<0.001	32 ± 9	32 ± 9	0.47
Hypertension	988 (62%)	1287 (76%)	<0.001	289 (79%)	996 (76%)	0.18
Diabetes	346 (22%)	583 (35%)	<0.001	139 (38%)	444 (34%)	0.13
Dyslipidemia	868 (55%)	1018 (60%)	0.0007	209 (57%)	807 (61%)	0.15
CKD Stage ≥3	230 (14%)	553 (33%)	<0.001	120 (33%)	432 (33%)	0.99
Current tobacco	80 (5%)	117 (7%)	0.02	21 (6%)	96 (7%)	0.30
Prior MI or PCI	234 (15%)	341 (20%)	<0.001	57 (16%)	283 (21%)	0.01
Cardiomyopathy	42 (3%)	58 (3%)	0.18	13 (4%)	45 (3%)	0.90
Symptoms	909 (57%)	1020 (61%)	0.05	225 (61%)	795 (60%)	0.70
Preventive Medications
Antihypertensive use	1055 (66%)	1359 (81%)	<0.001	288 (79%)	1069 (81%)	0.29
Beta blocker use	656 (41%)	953 (57%)	<0.001	184 (50%)	768 (58%)	0.006
Aspirin use	752 (47%)	877 (52%)	0.006	172 (47%)	704 (53%)	0.03
Lipid-lowering therapy use	814 (51%)	969 (58%)	0.003	195 (53%)	772(59%)	0.07
MPI and Hemodynamic Parameters
Rest LVEF, %	59 ± 10	59 ± 19	0.69	59 ± 13	58 ± 20	0.34
Stress LVEF	63 ± 11	62 ± 13	0.03	63 ± 14	62 ± 13	0.33
Rest MBF	0.74 ± 0.22	0.81 ± 0.34	<0.001	0.87 ± 0.36	0.79 ± 0.33	<0.001
Rest rate pressure product (RPP)	9560 ± 2362	10,731 ± 2624	<0.001	13,735 ± 2118	9897 ± 2085	<0.001
Heart rate, bpm	68 ± 12	73 ± 13	<0.001	83 ± 13	70 ± 12	<0.001
Systolic BP, mmHg	140 ± 23	148 ± 26	<0.001	167 ± 22	143 ± 24	<0.001
Hemoglobin (g/dL)	13.2 ± 1.5	12.4 ± 1.8	<0.001	12.6 ± 1.8	12.4 ± 1.7	0.08
Corrected Rest MBF	0.80 ± 0.28	0.79 ± 0.36	0.35	0.65 ± 0.29	0.83 ± 0.37	<0.001
Stress MBF	1.93 ± 0.60	1.23 ± 0.57	<0.001	1.52 ± 0.65	1.15 ± 0.52	<0.001
MFR	2.63 ± 0.58	1.52 ± 0.30	<0.001	1.73 ± 0.20	1.46 ± 0.30	<0.001
Corrected MFR	2.49 ± 0.74	1.63 ± 0.51	<0.001	2.36 ± 0.31	1.42 ± 0.34	<0.001

Values are mean ± standard deviation or number (%).

MFR, myocardial flow reserve; MBF, myocardial blood flow; LVEF, left ventricular ejection fraction; BP, blood pressure; CKD, chronic kidney disease; MI, myocardial infarction; PCI, percutaneous coronary intervention.

Unadjusted and adjusted association of MFR and MFR_corrected with MACE

Patients were followed for a median of 7 (IQR 3–12) years. MACE occurred in 286 (16%) with impaired MFR and in 104 (7%) with preserved MFR. Among patients with MFR <2 (n = 1685), MACE occurred in 231 (18%) with concordant MFR_corrected < 2 and in 55 (15%) with discordant MFR_corrected ≥ 2. Among patients with MFR ≥2 (n = 1591), MACE occurred in 81 (7%) with concordant MFR_corrected ≥ 2 and in 23 (6%) with discordant MFR_corrected < 2.

In unadjusted Cox regression, an MFR <2 was associated with an increased hazard of MACE compared with an MFR ≥2 (2.2% vs 1.0% MACE/year, HR 2.24 [95% CI 1.79–2.81], P < 0.0001, Figure 2A, Table 2). After adjustment, an impaired MFR (<2) remained associated with an increased hazard of MACE (HR 1.52 [95% CI 1.20–1.93], P < 0.001, Figure 3A, Table 2). Additionally, MFR_corrected <2 was associated with increased hazard of MACE in unadjusted and adjusted Cox regression (Supplementary Fig 1).

Figure 2.

Association of (A) MFR<2.0 and MFR≥2.0 with MACE (CV death or nonfatal MI) and (B) MFR_corrected with MACE among patients with MFR<2.0 (C) MFR_corrected with MACE among patients with MFR≥2.0. * Low cardiac risk <1%; Moderate cardiac risk 1% to 3%; High cardiac risk >3% based on ACC/AHA guidelines. ** P-values remained non-significant (P = 0.66 and P = 0.49) in adjusted models (Figure 3B and Figure 3C) that included age, sex, cardiac risk factors (hypertension, tobacco, dyslipidemia, diabetes, obesity, CKD), history of MI, history of PCI, history of cardiomyopathy, symptoms (chest pain, dyspnea, or palpitations) and resting LVEF MFR = myocardial flow reserve; MACE = major adverse cardiovascular event (CV death or nonfatal MI); CV = cardiovascular; MI = myocardial infarction; MFR_corrected = MFR corrected for resting RPP; RPP = heart rate blood pressure product; HR = hazard ratio; PET = positron emission tomography; MPI = myocardial perfusion imaging; CKD = chronic kidney disease; PCI = percutaneous coronary intervention; LVEF = left ventricular ejection fraction.

Table 2.

Unadjusted and adjusted hazard of MACE.

Composite MACE: CV Death or MI
	Unadjusted		Adjusted
	HR (95% CI)	P-value	HR (95% CI)	P-value
MFR ≥2	Ref	–	–	–
MFR <2	2.24 (1.79–2.81)	<0.0001	1.52 (1.20–1.93)	<0.001
Concordant: MFR <2 MFRcorrected < 2	2.32 (1.84–2.93)	<0.0001	1.54 (1.21–1.97)	<0.001
Discordant: MFR <2 MFRcorrected ≥ 2	1.96 (1.41–2.72)	<0.0001	1.46 (1.03–2.06)	0.03
Discordant vs Concordant MFRcorrected	0.84 (0.63–1.13)	0.26	0.93 (0.68–1.27)	0.66
Age (per year)			1.04 (1.03–1.05)	<0.0001
Female			0.86 (0.68–1.07)	0.17
Diabetes			1.41 (1.13–1.76)	0.003
CKD stage 3 or greater			2.26 (1.82–2.81)	<0.0001
Hypertension			1.20 (0.88–1.65)	0.26
Dyslipidemia			0.95 (0.76–1.20)	0.26
Active tobacco use			1.50 (1.03–2.19)	0.04
BMI (per unit kg/m2)			1.004 (0.99–1.02)	0.63
History of CAD (MI or PCI)			1.76 (1.39–2.22)	<0.0001
History of cardiomyopathy			1.33 (0.84–2.09)	0.22
Symptoms (chest pain, dyspnea, palpitations)			1.05 (0.84–1.30)	0.69
LV ejection fraction (per unit %)			0.98 (0.97–0.99)	<0.0001

MFR, myocardial flow reserve; MACE, major adverse cardiovascular event (CV death or nonfatal MI); CV, cardiovascular; MI, myocardial infarction; MFR_corrected, MFR, corrected for resting RPP; HR, hazard ratio; CI, confidence interval; CKD, chronic kidney disease; PCI, percutaneous coronary intervention; LVEF, left ventricular ejection fraction.

Figure 3.

Adjusted association* of (A) MFR<2.0 and MFR≥2.0 with MACE (CV death or nonfatal MI) and (B) MFR_corrected with MACE among patients with MFR<2.0 (C) MFR_corrected with MACE among patients with MFR≥2.0. * Adjusted for age, sex, cardiac risk factors (hypertension, tobacco, dyslipidemia, diabetes, obesity, CKD), history of MI, history of PCI, history of cardiomyopathy, symptoms (chest pain, dyspnea, or palpitations), and resting LVEF MFR = myocardial flow reserve; MACE = major adverse cardiovascular event (CV death or nonfatal MI); CV = cardiovascular; MI = myocardial infarction; MFR_corrected = MFR corrected for resting RPP; RPP = heart rate blood pressure product; HR = hazard ratio; PET = positron emission tomography; MPI = myocardial perfusion imaging; CKD = chronic kidney disease; PCI = percutaneous coronary intervention; LVEF = left ventricular ejection fraction.

Among patients with an MFR <2, there was no statistically significant difference in the hazard of MACE between patients with a discordantly preserved MFR_corrected (≥2) compared with those with a concordantly impaired MFR_corrected (<2) (1.9% vs 2.3% MACE/year, HR 0.84 [95% CI 0.63–1.13], P = 0.26, Figure 2B and Table 2). After adjustment for clinical risk factors, a discordantly preserved MFR_corrected (≥2) was not associated with a statistically significant difference in the hazard of MACE compared to a concordantly impaired MFR_corrected (<2) (HR 0.93 [95% CI 0.68–1.27], P = 0.66, Figure 3B and Table 2).

Among patients with an MFR ≥2, there was no statistically significant difference in the hazard of MACE between patients with a discordantly impaired MFR_corrected (<2) compared with those with a concordantly preserved MFR_corrected (≥2) (0.9% vs 1.0% MACE/year, HR 0.96 [95% CI 0.60–1.53], P = 0.86, Figure 2C). After adjustment for clinical risk factors, a discordantly impaired MFR_corrected (<2) was not associated with a statistically significant difference in the hazard of MACE compared to a concordantly preserved MFR_corrected (≥2) (HR 0.84 [95% CI 0.51–1.38], P = 0.49, Figure 3C).

There were few early and late revascularizations and, overall, no statistically significant differences between MFR and MFR_corrected groups (Supplementary Table 1).

Subgroup analysis

To determine if radiotracer type influenced the results, an unadjusted subgroup analysis was performed. Among patients with an MFR <2, there was no statistically significant difference in the hazard of MACE between patients with a discordantly preserved MFR_corrected (≥2) compared with those with a concordantly impaired MFR_corrected (<2) regardless of radiotracer (⁸²Rubidium: HR 0.88 [95% CI 0.64–1.12], P = 0.39; ¹³N-ammonia: HR 0.71 [0.33–1.51], P = 0.32).

DISCUSSION

There is robust evidence from various studies that myocardial flow reserve (MFR) measured by PET is a potent prognostic marker providing incremental risk stratification [1,2,10]. Specifically, patients with an MFR ≥2 exhibit a substantially lower absolute risk of MACE, including cardiac death, compared to those with an MFR <2, where the risk escalates linearly with decreasing MFR [11]. MFR is the ratio of myocardial blood flow (MBF) under stress to that at rest. Thus, patients with normal or nearly normal maximal stress MBF (>2) may have a reduced MFR (<2) due to increased resting MBF, often in the context of elevated myocardial oxygen demand (e.g., tachycardia, hypertension, anemia). To normalize for abnormally high or low resting MBF, correction for the rate-pressure product (RPP), an index of myocardial oxygen demand, has been proposed [5,6]. However, the clinical implications of this correction remain unexplored and controversial.

Our study found that impaired MFR is associated with an increased risk of MACE, regardless of RPP correction. This indicates that a ‘normalized’ MFR post-RPP correction (≥2) may provide false reassurance regarding cardiovascular risk. Indeed, we observed no significant statistical difference in annualized MACE risk between patients with uncorrected MFR <2 who had a discordantly ‘preserved’ corrected MFR (≥2) and those with persistently impaired MFR post-RPP correction (1.9% vs 2.3%/year, HR 0.84 [0.63–1.13], P = 0.26).

Many factors may contribute to increased resting MBF, including tachycardia, high blood pressure, and abnormal hemoglobin levels among others [12–14], all of them representing conditions that are associated with increased clinical risk. There is significant heterogeneity in resting MBF even among healthy adults, which is not solely explained by RPP [12], which may explain the weak correlation between RPP and resting MBF observed in this study. Consequently, resting MBF likely serves as a non-specific indicator of clinical risk and remains valuable for risk stratification, with elevated levels associated with adverse outcomes [14] and anginal symptoms [15].

Clinically, our data suggest that reporting RPP-corrected MFR may undermine risk stratification, especially when uncorrected MFR is less than 2.0. Our findings advocate for using and reporting both stress MBF and uncorrected MFR to accurately convey cardiovascular risk. In patients with normal or near-normal stress MBF but low MFR due to high resting flow, the stress flow helps to accurately define the presence of flow-limiting CAD [16,17]. It is also important to elucidate and convey reasons for elevated resting MBF when MFR is below 2.0.

The phenotype of patients with impaired MFR due to high resting MBF may differ from those with low hyperemic MBF. In the current study, patients with higher resting MBF were more often female, non-white race, and with hypertension and diabetes. These distinctions suggest that they may benefit from different treatment approaches, warranting further prospective investigation.

Study limitations

The findings of the current study must be interpreted in the context of some limitations. It is a retrospective study, and there may have been differences in preventive therapies or revascularization between patients with discordant and concordant impairment in MFR and MFR_corrected that we were unable to account for which could have accentuated or reduced differences in longitudinal cardiovascular risk. However, revascularization rates were overall low in this population of patients with visually normal PET scans, without statistically significant differences between groups. Finally, we used both ¹³N-ammonia and ⁸²Rubidium during the study period, and there are small differences in the physiologic ranges of MBF between tracers [6].

CONCLUSION

Our study suggests that RPP-corrected MFR may not provide accurate cardiovascular risk stratification in patients with normal perfusion and MFR <2.0, and indeed may provide false reassurance. There was no significant difference in cardiovascular risk between patients with discordant (≥2.0) and concordant (<2) MFR following RPP correction. Hence, stress MBF and uncorrected MFR should be reported to more reliably convey cardiovascular risk beyond perfusion results.

NEW KNOWLEDGE GAINED AND CLINICAL PERSPECTIVE.

What is new?

• Elevated resting myocardial blood flow (MBF) can result in reduced myocardial flow reserve (MFR) by positron emission tomography (PET) despite normal/near-normal peak stress MBF.

• To normalize for abnormally high or low resting MBF, correction for the rate-pressure product (RPP), an index of myocardial oxygen demand, has been proposed; however, the clinical and risk stratification implications are unknown and controversial.

• Our study found that among patients without overt obstructive CAD and MFR< 2.0, there was no significant difference in cardiovascular risk between patients with discordant (≥2.0) and concordant (<2) MFR following RPP correction.

WHAT ARE THE CLINICAL IMPLICATIONS?

• Impaired MFR is associated with an increased risk of MACE among patients without obstructive CAD, regardless of RPP correction.

• A ‘normalized’ MFR post-RPP correction may provide false reassurance regarding cardiovascular risk.

• Instead, stress MBF and uncorrected MFR should be reported to more reliably convey cardiovascular risk beyond perfusion results.

ABBREVIATIONS

CAD

coronary artery disease

CI

confidence interval

CMD

coronary microvascular disease

IQR

interquartile range

MFR

myocardial flow reserve

MFR_corrected

corrected myocardial flow reserve

PET/CT

positron emission tomography/computed tomography