Prognostic value of global myocardial flow reserve in patients with history of coronary artery bypass grafting

Abstract

Aims

It is not well understood whether positron emission tomography (PET)-derived myocardial flow reserve (MFR) is prognostic among patients with prior coronary artery bypass grafting (CABG).

Methods and results

Consecutive patients with a clinical indication for PET were enrolled in the Houston Methodist DeBakey Heart and Vascular Center PET registry and followed prospectively for incident outcomes. The primary outcome was a composite of all-cause death, myocardial infarction (MI)/unplanned revascularization, and heart failure admissions. Cox proportional hazards models were used to study the association between MFR (<2 vs. ≥2) and incident events adjusting for clinical and myocardial perfusion imaging variables. The study population consisted of 836 patients with prior CABG; mean (SD) age 68 (10) years, 53% females, 79% Caucasian, 36% non-Hispanic, and 66% with MFR <2. Over a median (interquartile range [IQR]) follow-up time of 12 (4–24) months, there were 122 incident events (46 HF admissions, 28 all-cause deaths, 23 MI, 22 PCI/3 repeat CABG 90 days after imaging). In adjusted analyses, patients with impaired MFR had a higher risk of the primary outcome [hazard ratio (HR) 2.06; 95% CI 1.23–3.44]. Results were significant for admission for heart failure admissions (HR 2.92; 95% CI 1.11–7.67) but not for all-cause death (HR 2.01, 95% CI 0.85–4.79), or MI/UR (HR 1.93, 95% CI 0.92–4.05).

Conclusion

Among patients with a history of CABG, PET-derived global MFR <2 may identify those with a high risk of subsequent cardiovascular events, especially heart failure, independent of cardiovascular risk factors and perfusion data.

Graphical Abstract

Open in new tabDownload slide

See the editorial comment for this article ‘Significance of myocardial flow reserve after revascularization', by A. Saraste and T. Maaniitty, https://doi.org/10.1093/ehjci/jead151.

Introduction

Patients with a history of coronary artery bypass graft surgery (CABG) remain at high risk of cardiovascular events and death partly due to the high burden of atherosclerosis in their native coronary and bypass graft vessels.^1–3 Residual risk of cardiovascular events in this population may be heterogenous and identification of higher-risk patients is crucial so that appropriate risk mitigation strategies can be instituted.

The presence of inducible ischaemia on stress perfusion cardiac magnetic resonance (CMR) is associated with a higher risk of cardiovascular events among patients after CABG.⁴ The ischaemic burden on positron emission tomography (PET) myocardial perfusion imaging (MPI) also provides incremental prognostic information in predicting all-cause mortality and cardiac deaths in patients with prior CABG.⁵

PET MPI provides an accurate assessment of hemodynamically significant stenoses among patients with suspected or known coronary artery disease (CAD).^6–8 Unique to PET MPI is the accurate and reproducible quantification of myocardial blood flow (MBF) and myocardial flow reserve (MFR: ratio of stress to rest MBF).^9,10 Several prior studies have demonstrated that MFR has independent and incremental value in identifying high-risk patients who may benefit from early revascularization.^11–14 However, the utility of MFR in patients with history of CABG remains less well understood. The American Society of Nuclear Cardiology/Society of Nuclear Medicine and Molecular Imaging did not endorse reporting MFR in patients with prior CABG citing a lack of data supporting its diagnostic or prognostic value.¹⁵

We therefore evaluated the association of global MFR with the risk of cardiovascular events in patients with a history of CABG who had a clinical indication for PET MPI. We hypothesize that MFR is associated with a higher risk of cardiovascular outcomes independent of traditional cardiovascular risk factors and ischaemia.

Methods

Study population

The study population consisted of 3858 patients who were enrolled in the Houston Methodist DeBakey Heart and Vascular Center prospective PET registry including consecutive patients who were referred to clinically indicated PET MPI between August 2019 to July 2022. From this cohort, we identified 836 patients with prior CABG. Among patients with multiple PET MPI, the first MPI was selected, and repeat images were excluded (n = 53). A flow diagram is provided (Supplementary data online, Figure S1).

Approval from the Institutional Review Board at the Houston Methodist Academic Institute was obtained before the start of the study.

Clinical data

Information on sociodemographic variables, cardiovascular comorbidities, and medication use was abstracted through a chart review of electronic health records within 30 days of PET imaging as well as patient interviews at the time of the test.

PET image acquisition

Patients were scanned using a Biograph Vision digital PET/CT scanner (Siemens Healthineers, Knoxville, TN, USA). MPI was performed according to guidelines from the American Society of Nuclear Cardiology.¹⁶ Patients were instructed to fast for 6 h and abstain from coffee and caffeinated beverages for 24 h before the test. Images were acquired in 3D list mode with 7 min of acquisition at rest and 7 min after vasodilator stress. Low-dose non-gated CT scans were obtained for attenuation correction during rest and stress imaging. Patients were also required to hold beta blockers and nitrates on the day of the test.

Maximum hyperaemia was achieved with 0.4 mg/5 mL of Regadenoson infused intravenously over 10 s or 140 µg/kg/minute of Adenosine over 4 min. ⁸²Rubidium (⁸²Rb) was exclusively used as the perfusion radiotracer. Rest and stress images were obtained after intravenous injection of ⁸²Rb (10–30mCi based on body mass index; BMI) at rest and peak hyperaemia. Heart rate, blood pressure, and electrocardiograms were recorded at the baseline and every minute during the stress test. All images were processed for static gated and dynamic imaging and interpreted by expert readers using commercially available software (Corridor4DM; INVIA, Ann Arbor, Michigan).

PET image interpretation

Relative perfusion images were interpreted semi-quantitively using the standard five-point system and 17-segment model. Perfusion abnormalities were graded using the summed stress score (SSS) and summed rest score (SRS). Summed difference score (SDS) was calculated as the difference between SSS and SRS. The percentage of the myocardium with scar (fixed perfusion defect), ischaemia (reversible perfusion defect), and total perfusion defect was determined by dividing the SRS, SDS, and SSS by a maximum score of 68, respectively.

Rest and stress left ventricular ejection fraction (LVEF) was calculated from gated images with 16-frame gating. LVEF reserve was calculated as the difference between stress and rest LVEF. Transient ischaemic dilatation (TID) was obtained automatically from non-gated images.

Myocardial blood flow and flow reserve

Myocardial blood flow (MBF) (mL/min/g) was derived from dynamic images at rest and peak hyperaemia using commercially available software (Corridor4DM; Invia, Ann Arbor, Michigan). MBF was obtained using the one-compartment model of ⁸²Rb kinetics as described by Lortie¹⁷ using time-activity curves to estimate the uptake parameter of the one-compartment model. Global MFR was calculated as the ratio of stress to rest MBF of the left ventricle. MFR was categorized as preserved (MFR ≥2) versus impaired (MFR <2) consistent with prior studies that have demonstrated that an MFR cut-off of 2 is useful for risk prognostication and discrimination of incident cardiovascular outcomes.¹⁴

Follow-up

The primary outcome was a composite of all-cause death, myocardial infarction (MI)/unplanned revascularization (UR)—percutaneous coronary intervention (PCI) or redo coronary artery bypass graft (CABG) occurring more than 90 days after imaging) and admissions for heart failure (HF). While patients with early revascularization (PCI or CABG within 90 days after imaging) were not excluded from the primary analysis, early revascularization was adjusted for in the multivariable model. MI was defined based on the fourth universal definition of MI.¹⁸ Components of the primary outcome were also evaluated separately.

Patients were prospectively followed from the time of PET imaging to incident events. All outcomes were obtained from chart review and adjudicated by expert physicians in a blinded manner. Patients were censored at either the occurrence of the earliest outcome or the last known date of contact noted on their medical records.

Statistical analysis

Study characteristics were stratified by MFR (<2 vs. ≥2). Continuous variables were summarized as mean with standard deviation/medians with interquartile range, and categorical variables were summarized as the number with percentages. Comparisons were made using Student t-test, χ²-test or Wilcoxon rank-sum test as appropriate.

Kaplan–Meier survival curves were generated for each MFR group and compared using log-rank testing. After confirming the proportionality using Schoenfeld residuals, nested multivariable Cox regression models were used to assess the incremental prognostic role by sequentially adding MFR (MFR model) to the PET model (inclusive of clinical variables (age, sex, BMI), cardiovascular risk factors (hypertension, diabetes, and dyslipidemia), early revascularization (PCI or repeat CABG within 90-days of imaging), and PET parameters (SSS, TID, LVEF, and LVEF reserve)]. Variables were selected a priori based on previously published literature and clinical judgment.

In sensitivity analysis, impaired MFR was defined using the median value in the dataset (<1.7 vs. ≥1.7). We performed a sensitivity analysis using minimally adjusted models including age, sex, BMI, hypertension, diabetes, dyslipidemia, early revascularization, and SDS. We also performed a sensitivity analysis accounting for competing risk of death with the primary composite outcome and components of the primary outcome as well (MI/UR and admission for HF). Competing-risk regression models were fit via maximum likelihood according to the method of fine and grey using Statas’ stcrreg command.¹⁹ Subgroup analyses were also performed by age (<65 vs. ≥65 years), sex, BMI (<25 vs. ≥25 kg/m²), normal perfusion, scar, ischaemia, and LVEF (<50% vs. ≥50%). Subgroup effects were tested by including an interaction term between MFR and each of the variables defining the subgroups, one at a time, in adjusted multivariable regression models. A significant interaction term indicates that there is a subgroup effect.

All analyses were performed using Stata 17.0 (StataCorp, College Station, Texas). A P < 0.05 was considered statistically significant.

Results

Baseline characteristics

Demographics and cardiovascular comorbidities are summarized in Tables 1–2. The mean (SD) age of the cohort was 68 (10) years, 53% were females, 79% were Caucasian, and 36% were non-Hispanic. The prevalence of obesity was 45%, hypertension 74%, dyslipidemia 61%, and diabetes 60%. The prevalence of medication use was as follows: statin 91%, beta-blockers 82%, angiotensin-converting enzyme inhibitor/angiotensin receptor blocker 65%, and aspirin or clopidogrel 89%. The median (interquartile range [IQR]) global MFR was 1.7 (1.4–2.2). Compared to patients with preserved MFR, those who had impaired MFR were older and less likely to be female. Conversely, they were more likely to have hypertension and diabetes (all P < 0.05).

PET parameters are summarized in Table 3. Compared to patients who had MFR ≥ 2 (n = 282, 34%), those with MFR < 2 were more likely to have a scar, ischaemia, and total perfusion defects. Rest/stress LVEF, and LVEF reserve were lower among patients with impaired MFR. Rest MBF was higher, stress MBF was lower, and global MFR was lower among patients with impaired MFR (all P < 0.05).

Follow-up and outcomes

Study outcomes are presented in Table 4. Over a median (IQR) follow-up time of 12 (4–24) months, there were 122 incident events (46 HF admissions, 28 all-cause deaths, 23 myocardial infarctions, 22 PCI/3 repeat CABG 90 days after imaging). Compared to patients with preserved MFR, those with impaired MFR were more likely to have MI/UR, HF, and mortality (all P < 0.05). The primary outcome (composite of all-cause death, MI/UR, and HF) was more frequent among those with impaired MFR (18.4% vs. 7.1%; P < 0.001) as shown in Kaplan–Meier curves (Figure 1).

Multivariable Cox regression models are presented in Table 5. In adjusted analyses, patients with impaired MFR had a higher risk of the primary outcome [hazard ratio (HR) 2.06; 95% CI 1.23–3.44]. Results were significant for HF admissions (HR 2.92; 95% CI 1.11–7.67) but not all-cause death (HR 2.01, 95% CI 0.85–4.79) or MI/UR (HR 1.93, 95% CI 0.92–4.05). Similar results were observed in minimally adjusted models (see Supplementary data online, Table S1) and in analyses accounting for competing risk of death (see Supplementary data online, Table S2). Using a cut point of 1.7 for impaired MFR yielded significant results for the primary outcome which was driven by MI/UR and HF but not all-cause mortality (see Supplementary data online, Table S3). There was also a significantly higher risk of outcomes with MFR <2 among those who are >65 years, with BMI ≥25 kg/m², those with perfusion defects (SSS >3), without ischaemia (SDS ≤3), those with or without a scar, and individuals with LVEF ≤50%. Interaction analysis was only significant for BMI (P = 0.02) (see Supplementary data online, Table S2).

Discussion

In a clinical cohort of patients with prior CABG who were referred for PET MPI, impaired MFR was associated with a higher risk of adverse events independent of cardiovascular risk factors and inducible ischaemia.

Assessment of cardiac viability using PET fluorine-18 deoxyglucose has been shown to identify patients who would benefit from CABG.^20,21 The presence of ischaemia in patients post-CABG may also identify those at higher risk of subsequent adverse cardiac events.⁵ Kato et al. demonstrated that coronary flow reserve using CMR had similar prognostic value among patients with known or suspected CAD.²¹ However, the prognostic utility of PET-derived MFR in patients with prior CABG has not been well examined, despite being the most commonly used tool to assess MFR in clinical practice.

In the present analysis of patients with prior CABG, hyperaemic MBF was lower among those with impaired MFR while rest MBF was similar in the two groups. The inability to augment coronary flow during stress can be multifactorial and due to epicardial stenosis, diffuse coronary disease, or microvascular dysfunction.^22–24 MFR may therefore be considered a proxy for vascular health in this high-risk population and can be used for further risk stratification. Patients with CABG and impaired MFR may benefit from early repeat revascularization to reduce the risk of subsequent cardiovascular outcomes.²⁵

MFR may be considered a proxy for vascular health in patients with CABG. Prior studies have found that revascularization was associated with improvement in hyperaemic MBF,^26,27 but only in coronaries with reduced coronary flow capacity at baseline.²⁸ Additionally cardiovascular risk factors, particularly lipid-lowering therapy, should be optimized to mitigate the risk of cardiovascular events.

In the context of high-risk subgroup of patients with CABG, preserved or hyperaemic MFR may indicate a healthy vasculature secondary to patent coronary arteries/bypasses and adequate control of cardiovascular risk factors. MFR >2 can therefore be used to de-risk and defer invasive procedures that may be unlikely to reduce the risk of cardiovascular events.

MFR is likely impacted by time since CABG and the completeness of initial revascularization. These are important considerations given that patency of saphenous vein bypass grafts decreases over time. Adequacy of initial revascularization and atherosclerotic disease burden of native coronary vessels may also impact coronary flow reserve. Further studies are required to determine how these factors can affect the utility of MFR for predicting cardiovascular outcomes.

The prevalence of scar was 56% and ischaemia was noted in 63% of the study population consistent with the high-risk status of this cohort of patients with prior CABG. The prognostic utility of MFR was retained among those without ischaemia (SDS ≤3) and those with or without scar (SRS >3 and ≤3). The hazard ratio for incident outcomes was similar in those with and without ischaemia (1.98 vs. 2.23).

While results were only statistically significant in those without ischaemia, the P-value for interaction was not statistically significant. Similarly, there was no significant interaction between scar and MFR in association with incident outcomes. These results suggest that the prognostic potential of MFR in patients with CABG does not differ according to the presence/absence of scar or ischaemia.

Strengths

To our knowledge, this is the first study demonstrating the prognostic utility of PET-derived global MFR in patients with a history of CABG. Patients were prospectively enrolled and systematically followed for incident outcomes. All PET scans were performed using the latest generation of digital PET scanners with improved image processing and photomultipliers to achieve increased sensitivity.

Limitations

The present results should be interpreted in the context of some limitations. Our patients were enrolled in a single tertiary care centre with unique practice patterns. All patients had a clinical indication for PET MPI and results may therefore not be generalizable to patients with a history of CABG who are otherwise asymptomatic or have no indication for ischaemia evaluation. We do not have information on MFR before CABG and therefore change in MFR could not be assessed. We also did not have complete information on the time of CABG and the adequacy of initial revascularization. The duration of follow-up was short which may have underpowered subgroup analyses. The possibility of residual confounding is still possible despite multivariable adjustment.

Conclusion

Among patients with a history of CABG, PET-derived global MFR <2 may identify those with a high risk of subsequent cardiovascular events, particularly HF. Patients with impaired MFR may benefit from early repeat revascularization and stringent control of cardiovascular risk factors.

Supplementary data

Supplementary data are available at European Heart Journal - Cardiovascular Imaging online.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

References

Author notes