Prognostic value of dipyridamole stress perfusion cardiovascular magnetic resonance in elderly patients >75 years with suspected coronary artery disease

Abstract

Aims 

There are only very few data on the prognostic value of stress cardiovascular magnetic resonance (CMR) in elderly people, while life expectancy of the general population is steadily increasing. Therefore, this study aims to assess the prognostic value of vasodilator stress perfusion CMR in elderly >75 years.

Methods and results 

Between 2008 and 2017, we included consecutive elderly >75 years without known coronary artery disease (CAD) referred for dipyridamole stress CMR. They were followed for the occurrence of major adverse cardiovascular events (MACE) including cardiac death or non-fatal myocardial infarction. Univariate and multivariate analyses were performed to determine the prognostic value of ischaemia or late gadolinium enhancement. Of 754 elderly individuals (82.0 ± 3.9 years, 48.4% men), 659 (87.4%) completed the follow-up with median follow-up of 4.7 years. Using Kaplan–Meier analysis, the presence of myocardial ischaemia was associated with the occurrence of MACE [hazard ratio (HR) 5.38, 95% confidence interval (CI): 3.56–9.56; P < 0.001]. In a multivariable Cox regression including clinical characteristics and CMR indexes, inducible ischaemia was an independent predictor of a higher incidence of MACE (HR 4.44, 95% CI: 2.51–7.86; P < 0.001). In patients without ischaemia, the occurrence of MACE was lower in women when compared with men (P < 0.01).

Conclusion 

Stress CMR is safe and has discriminative prognostic value in elderly, with a significantly lower event rate of future cardiovascular event or death in subjects without ischaemia or infarction.

Introduction

The world’s ageing population with an increasing life expectancy raises the issue of screening for coronary artery disease (CAD) in elderly individuals at risk for CAD. Current guidelines do not recommend stress testing in individuals without known CAD or without chest pain.^1,2 Appropriate use criteria regard the utility of stress testing in asymptomatic individuals with multiple cardiovascular (CV) risk factors as ‘uncertain’ and without distinction of age.^3–6 However, many of these recommendations rely on studies involving younger population (35–60 years).⁷ Thus, we hypothesized that stress testing could be of prognostic interest in elderly people without known CAD for improved risk stratification.

Besides its operator-dependence and potential lack of echogenicity, stress echocardiography is often limited in older individuals by submaximal exercise or poor tolerance to high dose dobutamine-atropine tests.⁸ Single-photon emission computed tomography (SPECT) nuclear perfusion imaging may be hampered by artefacts associated with left bundle branch block (LBBB).^9–11 Coronary computed tomographic angiography (CCTA) has some limitations in elderly including the presence of coronary calcifications causing ‘blooming artefacts’.¹² Perfusion stress cardiovascular magnetic resonance (CMR) has emerged as an accurate technique that can assess ventricular function, stress myocardial perfusion, and viability within a single examination without radiation.¹³ Its ability to identify silent myocardial ischaemia has been established.^14,15 Moreover, a very recent randomized study shows that stress CMR was non-inferior to fractional flow reserve (FFR) with respect to major adverse cardiovascular events (MACE) and with a lower incidence of coronary revascularization than fractional flow reserve in patients with stable angina.¹⁶ However, there are only very few data on the tolerance and prognostic value of stress CMR in elderly people. We sought to assess the feasibility and prognostic value of vasodilator stress perfusion CMR in elderly individuals >75 years without known CAD.

Methods

Study population

Between December 2008 and 2017, we conducted a single-centre longitudinal study with retrospective enrolment in consecutive elderly individuals >75 years without known CAD referred for vasodilator stress perfusion CMR (CMR Laboratory in a tertiary Cardiovascular Centre) for the detection of myocardial ischaemia. Exclusion criteria were: (i) subjects ≤75 years old; (ii) history of CAD [percutaneous coronary intervention or coronary artery bypass graft or myocardial infarction (MI), defined by history of MI on the medical records or presence of significant Q wave on 12-lead ECG in a coronary territory]; (iii) contraindication to CMR (cerebral clips, metallic eye implant); (iv) contraindication to dipyridamole (severe asthma or chronic obstructive pulmonary disease, second- or third-degree atrioventricular block); (v) known cardiomyopathy (e.g. hypertrophic, dilated, or infiltrative) and acute or chronic myocarditis; (vi) known allergy to gadolinium-based contrast medium; and (vii) glomerular filtration rate <30 mL/min/1.73 m². Patients with atrial fibrillation at the time of CMR were included in the study. Clinical data were collected by medical history and clinical examination on the day of stress CMR by a trained physician using standardized criteria. All patients gave informed written consent and the study was approved by the local ethic committee of our institutions.

Patients follow-up and clinical outcome

Follow-up consisted of a clinical visit as a part of usual care (62%) or direct contact with the patient or the referring cardiologist (38%). A clinical questionnaire with a detailed description of clinical study endpoints was thoroughly filled out by two cardiologists (T.P. and M.K.). Clinical event adjudication was completely blinded to clinical and CMR data. The primary clinical endpoint was the occurrence of at least one of the combined MACE defined by cardiac death or non-fatal MI. Clinical event adjudication was based on the follow-up clinical visit or contact, with a consensus reached by two senior cardiologists. Cardiac death was defined as any death preceded by acute MI, acute or exacerbation of heart failure, documented fatal arrhythmias, or unexpected death without a non-cardiac cause, and confirmed by death certificates. Non-fatal MI was defined by symptoms and elevation of serum troponin to two-fold the upper normal limit. Patients who underwent coronary revascularization <90 days after the index examination were not censored.¹⁷

CMR protocol

CMR was performed in a dedicated CMR laboratory on a Siemens MAGNETOM Espree^® 1.5 T scanner (Erlangen, Germany) from December 2008 to August 2015, and on a Siemens MAGNETOM Aera^® 1.5 T scanner (Erlangen, Germany) from August 2015 to December 2017, with a 16- and 18-channel Torso/Cardiac surface coil of the same manufacturer. Long-axis (2-, 3-, and 4-chamber) and short-axis cine MR images encompassing the left ventricle (LV) from base to apex were obtained with a retrospectively gated balanced Steady State Free Precession (b-SSFP) sequence. Vasodilatation was induced with dipyridamole injected at 0.84 mg/kg over 3 min in all patients. At the end of dipyridamole infusion, a bolus of gadolinium contrast (Dotarem^®, Guerbet, Aulnay, France, 0.1 mmol/kg) was injected at a rate of 5.0 mL/s with an injector (Optistar Elite Mallinckrodt) with concomitant first-pass myocardial perfusion imaging. A saturation-recovery b-SSFP sequence was used with the following typical parameters: repetition time/echo time = 287/1.2 ms, GRAPPA acceleration factor = 2, field of view = 370 × 314 mm, matrix = 224 × 180, reconstructed pixel size = 1.7 × 1.7 × 8 mm. Six slices (four short-axis views, a two-chamber, and a four-chamber views) were acquired with a temporal resolution of two heartbeats. Cine acquisitions were then acquired at stress with the same geometry used at rest. Theophylline was intravenously injected (250 mg i.v. over 5 min) to null the effect of dipyridamole at the end of the stress test. Late gadolinium enhancement (LGE) imaging was acquired 10 min after contrast injection using a 3D segmented T1-weighted inversion-recovery gradient-echo sequence. Data processing and reporting were performed using two software packages: Syngo.via (Siemens Healthcare, Erlangen, Germany) for image display and reporting and Clinigrid (Hemolia Inc., Paris, France) for data processing. All patients were asked to avoid caffeine at least 12 h before CMR. A 12-lead ECG was performed before and after CMR.

CMR image analysis

LV volumes and function were quantified on the short-axis cine stack. An LV ejection fraction (LVEF) ≥50% was considered normal.¹⁸ Stress-perfusion CMR images were evaluated according to the 17-segment model of the American Heart Association (AHA).¹⁹ The analysis of perfusion images was done visually by two experienced cardiologists (J.G. with Euro-CMR level 3 certified and F.S.) blinded to clinical and follow-up data. Inducible ischaemia was defined as a subendocardial perfusion defect that: (i) occurred in at least one myocardial segment with at least 50% transmurality of the perfusion defect; (ii) persisted for at least three phases beyond peak contrast enhancement; (iii) followed a coronary distribution; and (iv) in the absence of LGE in the same segment.^20–23 Additional criteria indicative for true hypoperfusion vs. artefacts were subendocardial signal reduction persisting longer than the contrast medium first-pass through the LV cavity, signal reduction in several slices and neighbouring regions, and absence of breathing motion and triggering artefacts during contrast medium first-pass.^20–23 The total number of ischaemic and LGE segments was calculated using the AHA 17-segment model.¹⁹ The presence of viability was defined by the presence of LGE with <50% transmurality.²⁴

Statistical analysis

Continuous variables were expressed as mean ± SD, and categorical variables as frequency with percentage; follow-up is presented as median and interquartile range. Differences between patients with and without inducible ischaemia in terms of clinical baseline, and CMR characteristics, were compared using the Student’s t-test or Wilcoxon rank-sum test for continuous variables, and the χ² or Fisher’s exact test for categorical variables, as appropriate. Normal distribution was assessed by the Shapiro–Wilk test. Cumulative incidence rates of individual and composite outcomes were estimated using the Kaplan–Meier method and compared with the log-rank test. Data on patients who were lost to follow-up were censored at the time of the last contact. Cox proportional hazards methods were used to identify the predictors of MACE among patients with and without inducible ischaemia.

The multivariable model was built by stepwise variable selection with entry and exit criteria set at the P ≤ 0.2 level. Statistical analyses were performed using SPSS, Version 21.0 (IBM Corp., Armonk, NY, USA).

Results

Patients characteristics

During the inclusion period, 35 280 individuals referred for dipyridamole stress CMR were screened and among them 2681 were >75 years. The flowchart of study participants is given in Figure 1, resulting in 659 individuals who completed the clinical follow-up and constituted our study cohort. Baseline patients’ characteristics and baseline CMR data are shown in Tables 1 and2. The main primary indications for stress CMR included symptomatic angina and dyspnoea (>60% of patients were symptomatic), and high CV risk patients (∼40%) defined by Framingham risk score.²⁵ Median basic CAD Consortium Score²⁶ was 31% (mean 33 ± 18%), which is indicative of an average intermediate pre-test likelihood of CAD. The study cohort consisted of elderly individuals with a mean age of 82 years (women 54%). Most subjects were in sinus rhythm (93%). There was no significant difference between the two groups in baseline clinical characteristics. The overall study cohort had a median LVEF of 59 ± 9.1%. Individuals without evidence of inducible ischaemia on CMR had a significantly greater LVEF (60.2 ± 8.7% vs. 56.2 ± 10.8%, P < 0.001) and lower prevalence of LGE (4% vs. 36%, P < 0.0001).

Clinical outcome

CMR study

Out of 787 individuals without known CAD, 747 patients completed the stress CMR protocol. Reasons for failure were renal impairment (n = 16), declining participation (n = 8), intolerance to stress agent (n = 7), claustrophobia (n = 6), and poor gating (n = 3). No patient died during or shortly after CMR and without major adverse event in relation with the study. The main adverse events during or immediately after the study were chest discomfort due to dipyridamole, dyspnoea, headache, dizziness, nausea, blood pressure drop, tachycardia, and angina with ECG evidence of ischaemia. For patients with angina pectoris, symptoms rapidly resolved with 250 mg IV aminophylline over 5 min. In all other patients, additional sublingual nitrates in 89% and additional IV beta-blockers in 5% were added.

Prognostic value

Median (interquartile range) follow-up was 4.7(2.8–7.1) years. In the entire cohort of 659 patients, there were 64 MACE (9.7%) including 49 cardiac deaths (7.4%) and 15 non-fatal MI (2.3%). Moreover, other CV events were noted during study follow-up: 166 deaths of any cause (25.2%), 28 hospitalizations for HF (4.2%), 25 late coronary revascularizations (3.8%) (with 2 CABG), 20 strokes (3.0%), 17 patients underwent early coronary revascularization within 90 days of CMR (2.6%), and 5 sustained documented ventricular tachycardia (0.8%). For the whole-study cohort, annualized event rates were 4.5% for MACE and 7.1% for all-cause mortality.

The event rates for MACE are shown in Figure 2, based on the presence of inducible ischaemia and LGE on CMR. The 131 patients with positive ischaemia had a higher annualized rate of MACE than the 528 patients with negative ischaemia (13.3% and 2.3%, respectively; P < 0.001). Moreover, 72 patients with positive LGE had a higher annualized rate of MACE than the 587 patients with negative LGE (12.5% and 3.5%, respectively; P < 0.001). In addition, 155 patients with positive ischaemia or positive LGE had a higher annualized rate of MACE than the 504 patients with negative ischaemia and negative LGE (12.9% and 1.9%, respectively; P < 0.001).

The univariate analysis of baseline individuals and CMR characteristics for the prediction of MACE and all-cause mortality is shown in Table 3. Age, sex, LGE, absence of viability, and inducible ischaemia were all significantly associated with MACE in univariate Cox models. In univariate analysis, the number of segments with inducible ischaemia was associated with the occurrence of MACE [hazard ratio (HR) 2.31, 95% confidence interval (CI): 1.38–3.91; P < 0.001]. Kaplan–Meier event-free survival curves for MACE and all-cause mortality are shown in Figures 3 and4. Elderly patients with inducible ischaemia on stress CMR had a significantly lower event-free survival when compared with subjects without inducible ischaemia (P < 0.0001). In a multivariate stepwise Cox regression including clinical characteristics and CMR indexes, a positive ischaemia was independent predictor of a higher incidence of MACE (HR = 5.83, 95% CI 3.56–9.56; P < 0.001). Moreover, positive LGE was associated with higher incidence of all-cause mortality (HR = 2.35, 95% CI 1.04–5.29; P = 0.04) (Table 4). In individuals without inducible ischaemia on stress CMR, the occurrence of MACE during follow-up was lower in women when compared with men (P = 0.008). There was no significant difference in the occurrence of MACE between male and female who had inducible ischaemia on CMR (P = 0.47).

Discussion

In this specific cohort of elderly individuals >75 years without known CAD referred for stress CMR, the present report demonstrates that: (i) stress CMR was highly feasible and safe, (ii) inducible ischaemia was a strong and independent predictor of MACE, and (iii) the lack of inducible ischaemia or LGE on stress CMR is associated with significantly lower event rates of MACE and all-cause mortality.

Despite a rapidly growing elderly population, studies assessing the prognostic value of stress imaging techniques in elderly remain limited. Although a negative stress echocardiography or a negative SPECT carries an annualized rate of MACE of, respectively, 1.3% and 1.2% in the general referral population,^27,28 the prognostic impact of these techniques in elderly individuals have not been formally established. These techniques may be hampered by some technical limitations in elderly, especially the difficulty of exercising or the high prevalence of left bundle branch block. In the CE-MARC study that has established CMR’s high diagnostic accuracy and superiority over SPECT in 752 recruited patients with suspected CAD, the mean age was 60 ± 10 years without assessment in elderly.²⁹ A study with 513 patients demonstrated the incremental prognostic value of stress CMR over clinical risk factors and resting wall motion abnormalities,³⁰ with mean age of 61 ± 9 years and the age of patients who presented clinical events during follow-up was 38–76 years.³⁰ In a meta-analysis (19 studies, 11 636 patients) showing the prognostic value of stress CMR, the mean age of the population was 63 ± 12 years and there was no specific assessment of elderly >75 years.³¹ Similarly, Kelle et al.³² reported a bi-centre CMR outcome study in 3138 consecutive patients without focused data in elderly. Our data compare favourably with the report by Stacey et al.³³ emphasizing the prognostic value of stress CMR among asymptomatic middle-aged individuals for a sentinel CV event, with 327 participants with mean age of 68 ± 8 years but without specific analyse of elderly patients.³³

Our data are in agreement with large reports in the general population attesting the prognostic value of stress CMR,^31,32 and show that prognostic implications may extend to elderly individuals over a long follow-up. The annual event rate for cardiac death and non-fatal MI has been consistently reported as <1% in several studies for patients with suspected or known CAD and a normal stress-perfusion CMR.^34,35 Not surprisingly, MACE cumulative events rate is higher in our study due to the older age of the participants. In addition, ischaemia is an independent predictor of all-cause mortality and cardiac mortality. This finding acquires a particular importance in this specific population in which many other causes can be responsible for death. Moreover, there was a slight significant association between the presence of LGE and all-cause mortality (P = 0.04). This result must be interpreted clinically with caution in this very elderly population with a high mortality rate. Among individuals with negative stress CMR, we report lower rates of MACE in women than in men, which compares favourably with differences noted in a previous study with annual MACE rates of 0.3% in women vs. 1.1% in men.³⁶ Moreover, a recent study in elderly patients with asymptomatic myocardial ischaemia reported a higher rate of hospitalizations for CV event or death in men than women.³³ Finally, in a sex-specific analysis from CE-MARC study, CMR has greater sensitivity than SPECT in both sexes. Unlike SPECT, there are no significant differences in the diagnostic performance of CMR.²⁹

The present study extends the aggregate data on stress CMR by demonstrating its prognostic value in a specifically targeted population of elderly individuals. It also shows in an elderly population that inducible ischaemia and presence of LGE by stress perfusion CMR are robust markers of CV risk even in patients without a clinical history of prior infarction. Moreover, in our study LGE corresponded to silent or unrecognized MI (4.2% in group without ischaemia and 35.9% in group with ischaemia) in this population without known CAD. These results suggest that silent or unrecognized MI is a good predictor of MACE and that its diagnosis could be an important parameter of decision-making in the therapeutic management of the patient.

Besides, CCTA is a useful modality to assess obstructive CAD, with high diagnostic accuracy in the general referral population.^12,37 However, the prognostic impact of this anatomic approach has not been specifically addressed in a large cohort of elderly individuals, who often present with diffuse coronary calcifications.

Study limitations

The rate of patients lost to follow-up was 11.8%, which can be explained by a relatively long follow-up in this elderly population. Patients were declared lost to follow-up when they could not be reached by repeated phone calls, or by a direct contact with their referring cardiologists or general practitioner. The French National Registry of Death was carefully consulted for those patients and they were still alive at the end of the follow-up, which reinforces the data on all-cause mortality reported in this study.

Conclusions

Myocardial perfusion stress CMR is feasible and safe in elderly individuals >75 years. The absence of ischaemia or myocardial scar is associated with lower rates of MACE over a long-term follow-up. With a growing population of elderly often asymptomatic but at significant risk for CV disease, these results suggest that the use of stress CMR may be warranted for a better risk stratification in those individuals.

Data availability

Raw data of this study are available from the corresponding author upon reasonable request.

Conflict of interest: none declared.

References