Exercise electrocardiography stress testing in suspected angina and non-obstructive coronary arteries

In clinical guidelines, exercise electrocardiography stress testing (EST) is recommended for risk stratification in patients with angina, whereas coronary computed tomography angiography (CCTA) is recommended for the diagnosis of obstructive coronary artery disease.1 The exercise EST may be considered (class of recommendation IIb) as an alternative option when non-invasive imaging tests are unavailable. There is no recommendation on the use of exercise EST following CCTA.1 This gap in evidence is relevant since most individuals with stable angina who undergo CCTA do not have obstructive coronary artery disease2 and angina/ischaemia with non-obstructive coronary arteries (ANOCA/INOCA) may be causally relevant.1 A strategy of CCTA coupled with exercise EST may be diagnostically useful in suspected ANOCA/INOCA, but prior studies have involved a retrospective design3,4 and selected populations with a limited sample size.5–7 The aim of this study was to assess the diagnostic accuracy of exercise EST for identifying endotypes (microvascular angina, vasospastic angina, or both) in a community-dwelling population with suspected ANOCA/INOCA according to invasive functional testing [coronary flow reserve (CFR), index of microvascular resistance (IMR), and acetylcholine vasomotor responses] as a reference. We hypothesized that ischaemic group status by exercise EST criteria has at least moderate sensitivity and specificity for coronary endotypes, supporting adoption of the exercise EST in this population. The CorCTA trial involved 250 outpatients with suspected ANOCA/INOCA post-CCTA who underwent invasive endotyping for microvascular angina and/or vasospastic angina or non-cardiac chest pain.8 This study describes a predefined secondary analysis, and the methods have been described.8 Eligible patients provided informed consent and then completed the Rose Angina9 and Seattle Angina Questionnaires (SAQ).10 The exercise EST involved the Bruce treadmill protocol, with myocardial ischaemia defined as ≥.1 mV ST-segment depression 60 ms from the J-point (ischaemic group). Participants with suspected ANOCA/INOCA underwent functional coronary angiography involving guide-wire–based coronary thermodilution at rest and during adenosine hyperaemia (Abbott, Irvine, CA, USA; Coroventis, Uppsala/Sweden) followed by intracoronary infusion of acetylcholine (.182 μg/mL, 1.82 μg/mL, 18.2 μg/mL; 2 mL/min for 2 min). The analysis population was restricted to patients who performed an exercise EST and for those where ischaemic group status (ST-segment deviation) was definable. Categorical outcomes were compared between groups (ischaemic and non-ischaemic groups; final diagnosis endotype groups) using Fisher’s exact test, and continuous outcomes were compared between groups using the Wilcoxon–Mann–Whitney test (for data with a skewed distribution) and Student’s t-test (for normal distributed data). A two-tailed analysis was performed, and a P < .05 was considered as statistically significant. Measures of diagnostic accuracy (with 95% confidence intervals) of ischaemic group status for coronary microvascular dysfunction subgroups were assessed. Overall, 163 individuals [mean age, 54.6 years, body mass index 30.9 kg/m2, 102 (62.6%) women] with symptoms of suspected angina and ischaemic [n = 45 (27.6%)] or non-ischaemic [n = 118 (72.4%)] group status defined by ST-segment deviation on exercise EST were included. Of the 163 participants assessed using the Rose Angina Questionnaire, 91 (55.8%) had definite angina and 72 (44.2%) had probable angina. Compared with the non-ischaemic group, in the ischaemic group, angina stability measured by the SAQ was lower (worse) [48.3 (16.3) vs. 55.2 (22.2); P = .034]. Of the 163 participants, 68 (41.7%) had no evidence of coronary atherosclerosis [ischaemic (n = 15, 33.3%) vs. non-ischaemic groups (n = 53, 44.9%); P = .215]. Myocardial bridging was most prevalent in patients with vasospastic angina (29%) compared with other groups (normal coronary function 2.1%, microvascular angina 4.7%, mixed microvascular angina and vasospastic angina 14.3%; P = .002). Myocardial bridging did not occur in the ischaemic group [0 vs. 13 (11.0%); P = .020]. Eighty-three (50.9%) participants had microvascular angina, 20 (12.3%) had vasospastic angina, 14 (8.6%) had both, and 46 (28.2%) had no abnormality, and the distribution of these endotypes was similar in the ischaemic and non-ischaemic groups (P = .907). The IMR, CFR, microvascular spasm, and epicardial vasospasm in the left anterior descending coronary artery were not different between groups. Thirteen (28.9%) of the 45 patients in the ischaemic group had normal coronary function, indicating a probable false positive exercise EST result and no sex association (3 of 12 males vs. 10 of 33 females; P = 1.000). Ischaemic group status had moderate specificity (72.5%) to detect microvascular dysfunction (CFR < 2.5 and IMR ≥ 25 and microvascular spasm by acetylcholine) but low sensitivity (30.0%) (Table 1). The negative predictive value for the exercise EST was high for adenosine-CFR < 2.5, adenosine-CFR < 2.5, and microvascular spasm (acetylcholine response) and the combination of adenosine-CFR < 2.5 or microvascular spasm (acetylcholine response) or adenosine-IMR ≥ 25 (.814, .924, and .941, respectively). The exercise EST in ANOCA, ischaemic group status and endotypes of coronary microvascular dysfunction The bold highlight visually conveys that these two results are statistically significant (p-value), whereas the other values are not. The study population is restricted to patients that performed an exercise EST and for those where ischaemic group status was able to be defined. Between-group comparisons assessed using Student’s t-test (normal distributed data) or the Mann–Whitney U test (skewed data) for continuous variables and the Fisher’s exact test for categorical variables. CFR, coronary flow reserve; CI, confidence interval; EST, electrocardiography stress test; IMR, index of microvascular resistance; MV, microvascular; NPV, negative predictive value; PPV, positive predictive value. The exercise EST in ANOCA, ischaemic group status and endotypes of coronary microvascular dysfunction The bold highlight visually conveys that these two results are statistically significant (p-value), whereas the other values are not. The study population is restricted to patients that performed an exercise EST and for those where ischaemic group status was able to be defined. Between-group comparisons assessed using Student’s t-test (normal distributed data) or the Mann–Whitney U test (skewed data) for continuous variables and the Fisher’s exact test for categorical variables. CFR, coronary flow reserve; CI, confidence interval; EST, electrocardiography stress test; IMR, index of microvascular resistance; MV, microvascular; NPV, negative predictive value; PPV, positive predictive value. The main findings were, first, 27.6% of individuals with suspected ANOCA/INOCA fulfilled the exercise EST criteria for myocardial ischaemia, which is expectedly lower compared with populations downstream in the care pathway who have been selected for invasive management7; second, ischaemic group status was associated with worse angina stability (SAQ); third, ischaemic group status had high negative predictive value for coronary microvascular dysfunction. This novel design included prospective enrolment in multiple centres, minimal selection criteria, validated patient-reported outcome measures, central review of the exercise EST, a reference invasive centre, comprehensive invasive endotyping, and an independent clinical trials unit.7 Second, the population was larger than in prior studies5–7 and included patients with undifferentiated chest pain upstream in the outpatient setting, this being distinct from prior studies which included patients downstream in the care pathway following selection for invasive management.6 Finally, the exercise EST involved the Bruce protocol in outpatient clinics, implying the results are transferable to daily practice. The exercise EST is widely available, brief, inexpensive, repeatable, and free of ionizing radiation (noting most individuals with ANOCA/INOCA are female). Our study brings together non-invasive exercise EST and invasive endotyping, addressing an evidence gap on ANOCA/INOCA.1 We propose a clinical algorithm: in patients with suspected angina, CCTA should be considered initially.1 Then, in patients with suspected ANOCA/INOCA post-CCTA, the exercise EST represents a rule-out strategy for microvascular endotypes, exercise capacity, and prognosis. Alternative non-invasive stress test modalities are also useful.1,3 Non-invasive ischaemia testing with other diagnostic modalities was not included. No adjustment has been made for multiple testing. The anonymized data are available upon reasonable request. We thank the patients, staff, and the funders who supported this study. C.B. is employed by the University of Glasgow and holds research and consultancy agreements with Abbott Vascular, AstraZeneca, Auxilius Pharma, CorFlow, Coroventis, GSK, HeartFlow, Menarini, Neovasc, Novartis, Siemens Healthcare, XyloCor, and Valo Health. These companies were not involved in this manuscript.