Long-term prognostic value of ischaemia and cardiovascular magnetic resonance-related revascularization for stable coronary disease, irrespective of patient’s sex: a large retrospective study

Abstract

Aims

To assess the sex-specific, long-term prognostic value of myocardial ischaemia induced by stress cardiovascular magnetic resonance (CMR) and early CMR-related revascularization in consecutive patients from a large registry.

Methods and results

Between 2008 and 2010, all consecutive patients referred for stress CMR were followed for the occurrence of major adverse cardiovascular events (MACE), defined by cardiovascular mortality or recurrent non-fatal myocardial infarction (MI). Early CMR-related revascularization was defined as any revascularization within 90 days after CMR. Among 3664 patients (56.9% male, mean age 69.9 ± 11.8 years), 472 (12.9%) had MACE (163 women and 309 men) after a median follow-up of 8.8 (IQR 6.9-9.5) years. Inducible ischaemia and late gadolinium enhancement (LGE) by CMR were associated with MACE in women and men (all P < 0.001). In multivariable Cox regression, inducible ischaemia, LGE, and CMR-related revascularization were independent predictors of MACE both in women [heart rate (HR) 4.79, 95% confidence interval (CI) 2.17–9.10; HR 1.82, 95% CI 1.22–2.71; HR 0.71, 95% CI 0.54–0.92, respectively; all P < 0.001] and men (HR 3.88, 95% CI 2.33–5.98; HR 1.48, 95% CI 1.16–1.89; HR 0.78, 95% CI 0.65–0.97, respectively; all P < 0.001). The addition of CMR-parameters led to improved model discrimination for MACE (C-statistic 0.61 vs. 0.71; NRI = 0.212; IDI = 0.032) for both women and men. CMR-related revascularization was associated with a lower incidence of MACE in patients with left ventricular ejection fraction (LVEF)<50%.

Conclusion

Inducible ischaemia and early CMR-related revascularization were good long-term predictors of MACE irrespective of sex. CMR-related revascularization was associated with a lower MACE incidence in the sole sub-set of patients with LVEF < 50%.

Introduction

Cardiovascular (CV) mortality due to coronary artery disease (CAD) has recently increased, and healthcare costs associated with CAD have been projected to double between 2015 and 2030.¹ Although current guidelines recommend non-invasive cardiac imaging in patients with suspected CAD,² the impact of non-invasive imaging strategies to guide initial coronary revascularization is still controversial. Although the ISCHEMIA trial highlighted the pivotal role of optimized medical therapy and the lack of benefit of an initial invasive strategy,³ sub-group analyses excluding procedural infarctions and performed at longer follow-up suggested that initial coronary revascularization could improve outcome.³ Other studies have suggested a favourable prognostic impact of coronary revascularization in patients with inducible ischaemia.⁴

Women have poorer outcomes after an acute coronary event.⁵ Whether non-invasive functional imaging tests have equal prognostic accuracy in women and men is debated.⁶ A sex-specific analysis of risk stratification through non-invasive imaging testing, along with the prognostic impact of early revascularization, would be clinically useful.

Several studies have demonstrated the cost-effectiveness of stress cardiovascular magnetic resonance (CMR) in patients with known or suspected CAD.^7,8 The use of stress CMR as a first-line strategy in patients with stable angina was shown to be non-inferior in terms of outcomes, with a lower incidence of coronary revascularization compared to an invasive approach with fractional flow reserve.⁹

This study aimed to assess the sex-specific long-term prognostic value of myocardial ischaemia by stress CMR and of CMR-related revascularization in consecutive patients with known or suspected CAD.

Methods

Study population

Between December 2008 and June 2010, we conducted a single-centre study with retrospective enrolment of all consecutive patients referred for vasodilator stress CMR in an EACVI-accredited laboratory. The main exclusion criteria were contraindications to CMR or dipyridamole (detailed list in Supplementary data online, S1). Clinical data were collected according to medical history and a clinical examination on the day of stress CMR. All patients gave informed written consent for a CMR examination and enrolment in the research study. The study was approved by the local ethics committee of our institution and conducted in accordance with the Declaration of Helsinki and STROBE reporting guidelines for cohort studies.

Clinical outcomes

The follow-up consisted of a clinical visit as part of the usual care (73%) or by direct contact with the patient or the referring cardiologist (27%). The data collection ended in June 2020. The primary outcome was the combined major adverse clinical events (MACE) including CV mortality or non-fatal myocardial infarction (MI). The secondary outcomes were late coronary revascularization, hospitalization for heart failure (HF), ventricular arrhythmia, and all-cause mortality (definition of clinical events in Supplementary data online, S2). Patients who underwent PCI or CABG <90 days after the index examination and peri-procedural events were excluded.

CMR protocol

The detailed CMR protocol has been previously described¹⁰ (Supplementary data online, S3). Briefly, CMR was performed in a dedicated CMR laboratory using a 1.5 T scanner (MAGNETOM Espree, Siemens Healthcare, Germany). Vasodilation was induced with dipyridamole injected at 0.84 mg/kg over 3 min. Then, a bolus of 0.1 mmol/kg gadolinium-based contrast agent (Dotarem^®, Guerbet, France) was injected. Stress perfusion imaging was performed using an ECG-triggered saturation-prepared balanced steady-state free-precession sequence. At 10 min after the contrast injection, late gadolinium enhancement (LGE) images were acquired using a breath-hold 3D T1-weighted inversion-recovery gradient-echo sequence.

Left ventricular (LV) volumes and function were quantified on a short-axis cine stack (Syngo.via, Siemens Healthcare). Stress perfusion and LGE images were evaluated according to the AHA 17-segment model.¹¹ The analysis of stress perfusion images was done visually by two experienced cardiologists blinded to the follow-up data. Inducible ischaemia was defined as a sub-endocardial or transmural perfusion defect that (i) occurred in at least one myocardial segment; (ii) persisted for at least three phases beyond peak contrast enhancement; (iii) followed a coronary distribution, and (iv) occurred in the absence of co-location with LGE.^7,8 MI was defined as LGE with ischaemic patterns defined by sub-endocardial or transmural LGE. A segment was considered viable if the LGE thickness was <50% of the myocardial wall. Mild, moderate, and severe ischaemia were defined as the involvement of 1–2, 3–5, and ≥ 6 myocardial segments, respectively.^3,7

Coronary revascularization

CMR-related coronary revascularization was defined as either CABG or PCI performed within 90 days after index CMR exhibiting inducible ischaemia. All patients were treated with optimal medical therapy according to current guidelines.² Decision-making regarding initial coronary revascularization was based on the extent of ischaemia (≥2 contiguous segments), target vessel anatomy (chronic occlusion, vessel diameter, extensive calcifications), co-morbidities, LV function, the presence of symptoms, and the risk of complications due to revascularization.

Statistical analysis

Continuous variables were expressed as mean ± standard deviation (SD), categorical variables as the frequency with a percentage, and follow-up as a median and interquartile range (IQR). Patients with and without inducible ischaemia were compared using the Student’s t-test or the Wilcoxon rank-sum test for the continuous variables and the χ² or Fisher’s exact test for the categorical variables. Cumulative incidence rates of the outcomes were estimated using the Kaplan–Meier method and compared with the log-rank test. Cox proportional hazards methods were used to identify the predictors of MACE among patients with and without inducible ischaemia. The proportional hazard assumption was assessed using Schoenfeld residuals. Martingale residuals were used to detect nonlinearity in continuous variables. To assess the incremental prognostic value of both CMR parameters and the revascularization, different multivariable models were used, as follows:

Model 1: used a stepwise Cox regression strategy to select the strongest parsimonious set of covariates, with a p-value ≤0.2 on univariable screening (without inducible ischaemia, LGE, and revascularization).

Model 2: model 1 with inducible ischaemia and LGE.

Model 3: model 1 with LGE and CMR-related revascularization.

The discriminative capacity of each model for predicting MACE was determined according to Harrell’s C-statistic before and after the addition of inducible ischaemia and LGE. The additional predictive value of each parameter was calculated using Harrell’s C-statistic increment, continuous net reclassification improvement (NRI), and the integrative discrimination index (IDI).

In the competitive risk analysis, cumulative incidence functions were used to display the proportion of patients with the event of interest or the competing event as time progressed. A two-tailed P-value <0.05 was considered statistically significant using R software, version 3.3.1 (R Project for Statistical Computing).

Results

Patient Characteristics

During the inclusion period, 4072 patients were referred for dipyridamole stress CMR (Figure 1). As detailed in the flowchart, 3953 (97.1%) completed the stress test protocol. No patient died during CMR, and detailed safety results are presented in Supplementary data online, S4. Finally, 3664 patients completed the clinical follow-up and constituted our study cohort. Baseline subject characteristics and baseline CMR data are shown in Table 1. Among those 3664 patients (56.9% male, mean age 69.9 ± 11.8 years), 60.2% had hypertension, 56.0% dyslipidaemia, 35.3% obesity, 32.9% diabetes mellitus, 28.3% a family history of CAD, and 22.6% were smokers, with a pre-test probability of obstructive CAD of 17.9 ± 4.9%.¹² Men were more frequently smokers and had a higher rate of diabetes mellitus, dyslipidaemia, and known CAD than women. Women had more frequently hypertension (66.4 vs. 55.4%, P < 0.001), a higher BMI (31.8 ± 8.9 vs. 27.9 ± 4.7 kg/m², P < 0.001). Men presented a higher CV risk than women using the 10-year risk for fatal CAD score¹³ and the Framingham Risk Score >20% of risk of CAD at 10 years.¹⁴ The overall study cohort had a mean left ventricular ejection fraction (LVEF) of 49.8 ± 13.0%. The presence of inducible ischaemia was detected in 11.2% of women and in 26.9% of men (P < 0.001), with a similar ischaemia extent of 2.6 ± 1.7 segments for both (Figure 2). LGE was present in 18.4% of women and 53.8% of men. Of the 738 patients with ischaemia, 623 (84.4%) had a coronary angiography. Among these, 560 (75.7% of women and 75.9% of men) underwent coronary revascularization [497 (88.7%) PCI and 63 (11.3%) CABG]. Among the 225 symptomatic patients with ischaemia, 219 (97.3%) underwent coronary revascularization. Of the 244 patients with ischaemia and LVEF < 40%, 219 (89.8%) underwent coronary revascularization. Among the 233 patients with moderate or severe ischaemia, 221 (94.8%) underwent coronary revascularization. Of note, there were only six peri-procedural events in the group of patients with ischaemia.

CV events

During a median follow-up of 8.8 (IQR 6.9–9.5) years, there were 472 (12.9%) MACE [163 (10.3%) in women and 309 (14.8%) in men], including 305 (8.3%) CV mortality and 167 (4.6%) non-fatal MI. Furthermore, 555 all-cause mortality (15.1%), 209 hospitalizations for HF (5.7%), 193 late coronary revascularizations (5.3%), and 70 sustained documented ventricular arrhythmias (1.9%) were recorded.

Regardless of sex, patients without inducible ischaemia or LGE had the lowest annualized rate of MACE (3.3%/year), whereas it was greater for patients with inducible ischaemia without or with LGE (14.2%/year and 14.3%/year, respectively, both P < 0.001) (Supplementary data online, S5). The annualized rate of MACE was lower in patients without inducible ischaemia compared to patients with mild, moderate, or severe ischaemia (3.3%/year vs. 7.2%/year, 24.8%/year, and 40.4%/year, respectively; p_trend < 0.001), with similar findings in women and men (Figure 3). The negative predictive value of the absence of inducible ischaemia was observed in women and men irrespective of age (all P < 0.001; Supplementary data online, S6). Consistently, the occurrence of MACE was less than 5% at 5 year-follow-up in both women and men (4.2% and 4.5%, respectively).

Long-term Prognostic Value of Stress Cmr Parameters

In the overall population, age, male sex, previous PCI or CABG, history of peripheral atheroma, LVEF value, LV end-diastolic, and end-systolic volumes indexed and the presence and extent of both inducible ischaemia and LGE were all significantly associated with MACE (Supplementary data online, S7). Previous PCI or CABG in men (P = 0.32) and peripheral atheroma in women (P = 0.81) were not significantly associated with MACE. Using the Kaplan–Meier analysis, the presence of inducible ischaemia was associated with the occurrence of MACE in women [heart rate (HR) 4.74, 95% confidence interval (CI) 3.44–6.54; P < 0.001] and in men (HR 3.54, 95% CI 2.83–4.43; P < 0.001; Figure 4A), and the same was observed for LGE (HR 1.37, 95% CI 1.27–1.47; HR 1.32, 95% CI: 1.26–1.39, respectively; both P < 0.001).

In addition, inducible ischaemia was associated with CV mortality in women and men (HR 4.80, 95% CI 3.31–6.97; HR 3.08, 95% CI 2.31–4.12, respectively; both P < 0.001; Figure 4B and Supplementary data online, S8). Irrespective of sex, inducible ischaemia was also associated with non-fatal MI (HR 4.68, 95% CI 3.46–6.34; P < 0.001) and all-cause mortality (HR 2.36, 95% CI 1.98–2.82; P < 0.001). However, inducible ischaemia was associated with late revascularization (P = 0.001) and hospitalization for HF (P = 0.029) only in men (Supplementary data online, S9). The prognostic values of known and unrecognized MI were similar in the overall population and regardless of gender (Supplementary data online, S8 and S9).

In a multivariable stepwise Cox regression, the presence of inducible ischaemia, the number of ischaemic segments, and the presence of LGE were independent predictors of a higher incidence of MACE in women (HR 3.95, 95% CI 2.80–5.56; HR 1.52, 95% CI 1.41–1.64; HR 1.96, 95% CI 1.38–2.76, respectively; all P < 0.001) and in men (HR 3.53, 95% CI 2.81–4.44; HR 1.52, 95% CI 1.45–1.59; HR 1.58, 95% CI 1.23–2.03, respectively; all P < 0.001; Table 2). In a competitive risk analysis, the presence of inducible ischaemia was independently associated with nonfatal MI and CV mortality in women and men (all P < 0.001) (Supplementary data online, S10 and S11).

Long-term Prognostic Value of Cmr-related Coronary Revascularization

Using the Kaplan–Meier analysis in patients with inducible ischaemia, CMR-related revascularization was associated with a lower incidence of MACE in women and men (HR 0.46, 95% CI 0.27–0.79; HR 0.50, 95% CI 0.36–0.69, respectively; P < 0.05) and CV mortality (HR 0.41, 95% CI 0.22–0.75; HR 0.44, 95% CI 0.29–0.67, respectively; P < 0.05). There was a significant reduction of MACE in patients with LVEF <50% (HR 0.28, 95% CI 0.19–0.41; P < 0.0001), whereas revascularization did not reduce MACE in patients with LVEF ≥50% (HR 0.77, 95% CI 0.52–1.14; P = 0.20; Figure 5). In the multivariable stepwise Cox regression (model 3), CMR-related revascularization was an independent predictor of a lower incidence of MACE in women and men (HR 0.71, 95% CI 0.54–0.92; HR 0.78, 95% CI 0.65–0.97, respectively; P < 0.01; Table 2).

Incremental Prognostic Value of Stress Cmr And Cmr-related Coronary Revascularization

For the prediction of MACE, the baseline C-statistic value was 0.61 (95% CI 0.57–0.66) for model 1, with stepwise variable selection. The addition of inducible ischaemia and LGE significantly improved the C-statistic to 0.71 (95% CI 0.66–0.74; C statistic improvement for model 1: 0.10; NRI = 0.212; IDI = 0.032) for both women and men. Furthermore, the addition of CMR-related revascularization slightly improved the C-statistic to 0.72 in women (95% CI 0.68–0.75; C-statistic improvement for model 1: 0.11; NRI = 0.264; IDI = 0.038) and 0.73 in men (95% CI 0.69–0.77; C-statistic improvement for model 1: 0.12; NRI = 0.320; IDI = 0.040; Table 3).

Discussion

In this large registry of consecutive patients who underwent perfusion stress CMR for known or suspected CAD, the main findings are: (i) 11% of women and 27% of men had inducible ischaemia; (ii) inducible ischaemia and LGE were independent, long-term predictors of MACE regardless of patient sex; (iii) the presence of inducible ischaemia and LGE improved model discrimination for the prediction of MACE, after adjusting for traditional CV risk factors for both women and men, and (iv) irrespective of sex, coronary revascularization was associated with a lower incidence of MACE in patients with LVEF <50%, but there was no benefit in patients with LVEF ≥50%.

The prevalence of inducible ischaemia (20.1%) and LGE (38.9%) are consistent with prior studies in patients with similar levels of CV risk.^8,15 The rate of MACE (12.9%) is in-line with contemporary cohorts of patients referred for stress CMR^8,15 and the ISCHAEMIA trial.³ The long-term prognostic value of stress CMR is well established,^7,8,15 but sex-specific outcome data are scarce.^6,15 Some reports emphasized the challenge for detecting myocardial ischaemia in women.^5,6 Indeed, the pathophysiological features of myocardial ischaemia differ in women and men including the influence of sex hormones on the presentation of symptoms and ECG changes. In addition, the rate of microvascular angina is higher in women than in men, with a reported prevalence of 77% of women in a prospective study of consecutive patients with non-obstructive CAD and microvascular disease.^16,17 Interestingly, traditional hallmarks of myocardial ischaemia, such as stress-induced LV contractile and perfusion alterations are usually more difficult to detect in patients with microvascular disease.¹⁷ The current study demonstrates that inducible ischaemia and LGE are independently associated with MACE in women and men after a median follow-up of 8.8 years.

In line with previous cohorts,^15,18 the lack of inducible ischaemia indicated lower annual rates of MACE for both women and men (2.6%/year and 3.9%/year, respectively) with a similar warranty period in both genders of <5 years. Additionally, the extent of ischaemia was a strong predictor of MACE and CV mortality in both sex, concurring with the study of Coelho-Filho et al.,¹⁵ who assessed 405 patients at a median follow-up of 2.5 years. Importantly, inducible ischaemia and LGE assessed by stress CMR had incremental prognostic value in predicting MACE in both women and men, with better discrimination and reclassification over traditional risk factors. This finding is in agreement with previous myocardial perfusion single-photon emission computed tomography or echocardiographic studies, which have shown the incremental prognostic value of ischaemia to predict CV mortality irrespective of sex.^19,20 In agreement with a recent study, the prognostic value of unrecognized MI was similar to known MI for predicting MACE.²¹

The rate of early revascularization was 76% in patients with inducible ischaemia and was similar in women and men, which is consistent with recent studies.^7,20 While including procedural infarctions, the ISCHEMIA trial showed the lack of benefit of early coronary revascularization at median follow-up of 3.2 years.³ A sub-analysis that excluded procedural infarctions suggested better outcome at 5 years in the early revascularization strategy group.³ Previous large observational cohort studies also suggested a clinical benefit of coronary revascularization in patients with inducible ischaemia at mean follow-up of 4.6–5.5 years.^4,22 The COURAGE trial showed in patients with stable CAD that initial PCI did not reduce MACE when added to optimal medical therapy.²³ However, in the sub-set of patients who underwent serial functional testing with scintigraphy, PCI with a treatment target of ≥5% ischaemia reduction resulted in improved outcomes and a greater reduction in ischaemia compared with medical therapy alone.²⁴ In the current study, the potential benefit of CMR-related coronary revascularization for reducing the incidence of MACE and CV was driven by a greater impact on the sub-group of patients with altered LVEF. Indeed, coronary revascularization was associated with a lower incidence of MACE in patients with LVEF <50%, but there was no benefit in patients with LVEF ≥50%. These results are concordant with those of the ISCHEMIA trial, which included a large majority of patients with LVEF ≥50% [median IQR = 60 (55–65)%].³ Of note, a recent ancillary study of ISCHEMIA showed that revascularization was associated with better event-free survival in the sub-group of patients with LVEF <45%.²⁵ Although the results of the randomized STICH (comparison of surgical and medical treatment for congestive HF and CAD) and HEART (HF revascularization trial) studies, evaluating the prognostic value of CABG in patients with reduced LVEF, were negative at 5 years,^26,27 the extended follow-up of STICH (median 9.8 years) showed that surgical revascularization in addition to medical therapy resulted in a substantial benefit for all-cause and CV mortality, regardless of sex.²⁸ There are very few studies assessing PCI vs. medical treatment alone in patients with altered LVEF. The current guidelines state that coronary revascularization in patients with ischaemic cardiomyopathy and reduced LVEF may improve LV dysfunction and prognosis by reducing ischaemia in a viable hibernating myocardium.^2,29

Study Limitations

The study design was retrospective with a risk of referral bias and confounding factors regarding the association between management decisions after the stress CMR study and patient risks. Second, 7.3% patients were lost to follow-up, which can be explained by the relatively long follow-up period and the design of the study. Although all patients received optimal medical therapy according to current guidelines, baseline and post-revascularization pharmacological data were not collected. The analysis of the CMR perfusion scans was visual, but it represents the most widely accepted clinical method with optimal diagnostic accuracy. Although adenosine is commonly used for stress perfusion CMR, dipyridamole was used in our centre mainly because of medico-economic reasons and very close efficacy/safety profile compared to adenosine and regadenoson. As suggested by SCMR guidelines,³⁰ we believe that CMR protocols and workflow would be quite similar with adenosine. The use of regadenoson would result in a substantial gain of time and could facilitate the widespread applicability of perfusion stress CMR but at higher costs. The reasons for the absence of PCI in patients with ischaemia (non-significant lesions, technical difficulties, etc.) and the Syntax score were not collected. However, these limitations were related to patient care and reflect current clinical practice. Because of the lack of randomization, the prognostic impact of CMR-related revascularization could not be formally established.

Conclusion

In this large observational cohort of consecutive patients, inducible ischaemia and LGE on stress CMR were independently associated with MACE over a long-term follow-up, and offered incremental prognostic value over traditional risk factors irrespective of sex. CMR-related coronary revascularization was associated with a lower incidence of MACE in the sole subset of patients with LVEF <50%. Further prospective randomized studies are warranted to evaluate an initial revascularization strategy based on stress CMR, in particular in patients with altered LV function.

Supplementary data

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

Data Availability

All data generated or analysed during this study are included in this published article [and its supplementary information files].

Ethics Approval And Consent To Participate

The study was evaluated and approved by the local ethic committee of our institutions and conducted in accordance with the 1964 Declaration of Helsinki. All patients enrolled in this study were required to understand and give their consent for participation.

Conflict of interest: The authors declare that they have no competing interests. Solenn Toupin is an employee of Siemens Healthcare.

References