By Michael H. Crawford, MD, Editor

SYNOPSIS: A systematic imaging protocol of coronary angiography, optical coherence tomography, and cardiac MRI in women clinically diagnosed with myocardial infarction with non-obstructive coronary artery disease revealed a cause in 84%, with three-quarters exhibiting an ischemic etiology.

SOURCE: Reynolds HR, Maehara A, Kwong RY, et al. Coronary optical coherence tomography and cardiac magnetic resonance imaging to determine underlying causes of MINOCA in women. Circulation 2020; Nov 14. doi: 10.1161/CIRCULATIONAHA.120.052008. [Online ahead of print].

Myocardial infarction with non-obstructive coronary artery disease (MINOCA) is an infrequent event largely confined to women, which can produce considerable subsequent adverse events and long-term mortality. Clinical and autopsy studies have revealed various vascular and non-vascular mechanisms for this syndrome, but early diagnosis of the underlying cause would greatly facilitate management.

Investigators from the Women’s Heart Attack Research Program (HARP) designed a comprehensive imaging protocol to identify the causes of MINOCA in patients referred for coronary angiography after the clinical diagnosis of acute MI. Those with < 50% stenosis of all coronary arteries were studied by optical coherence tomography (OCT) at the time of angiography. Later, they underwent cardiac magnetic resonance imaging (CMR) within one week of presentation. Exclusion criteria included alternative explanations for troponin elevation, prior history of obstructive CAD, recent use of vasoactive agents, impaired renal function, pregnancy, and treatment with thrombolytic therapy. Also, patients with MI caused by supply demand mismatch were excluded.

OCT and CMR findings were integrated to identify the most likely cause of MINOCA in each patient. Of 301 women enrolled at 16 sites, there was interpretable OCT imaging in at least one coronary artery for 145, and they were included in this analysis. Their median age was 60 years, and 50% were non-Hispanic white. Non-ST elevation MI was the working diagnosis at the time of admission in 97% of patients.

The ECG was abnormal in 65% of patients. In the 111 who underwent an acute echocardiogram, 44% exhibited a segmental wall motion abnormality. At angiography, the sites categorized 54% of the coronary arteries as normal. The core lab found no patient with zero stenosis (3% had 1-10%, 62% had 11-30%, and 35% had 31-50% stenosis).

By OCT, a definite or probable lesion was found in 46% of patients. The most frequent abnormality was plaque disruption (in 39%). In 3.4% of patients, thrombus without disruption of plaque was found, and 2.1% exhibited intimal bumping suggestive of spasm. One patient experienced a coronary dissection, and there was more than one finding for five patients. In a multivariate analysis, an OCT culprit lesion was independently associated with older age (OR, 1.05) and diabetes (OR, 5.41).

CMR, performed a median six days after MI onset, was interpretable in 116 of 145 patients with an interpretable OCT image. A late gadolinium enhancement distribution consistent with infarction was found in 33% of patients, almost all of whom showed evidence of myocardial edema. A regional injury pattern with more extracellular volume was found in 21% of subjects. Myocarditis was found in 15%. In 3%, Takotsubo cardiomyopathy was found, and 2% exhibited non-ischemic cardiomyopathy. The remaining 26% were normal on CMR.

Of the 116 women who underwent both OCT and CMR, 98 exhibited abnormalities on one or both studies, resulting in a final diagnosis of MI in 64%, myocarditis in 15%, Takotsubo in 3%, and non-ischemic cardiomyopathy in 2%. No cause could be identified for the remaining 16%. The authors concluded that among women with the clinical diagnosis of MINOCA, OCT plus CMR identified a cause in 84%, with 75% ischemic and 25% non-ischemic or alternative diagnoses to MI.


MINOCA is thought to represent 5-15% of MI patients; many of these are women. Post-MI adverse events occur in about 25% and mortality in about 11% over five years. The authors of clinical and autopsy studies have identified various potential mechanisms of MINOCA. Vascular causes include plaque rupture, erosion, and cavitation; coronary spasm, embolism, and dissection; and combinations of these. Non-vascular causes include myocarditis, Takotsubo syndrome, and cardiomyopathy. MINOCA is not thought to include microvascular disease despite its predominance in women.

In any given case, a precise mechanistic diagnosis likely would aid in acute and long-term treatment. For example, the treatment and secondary prevention of myocarditis differs greatly from those of myocardial ischemia/infarction. Thus, this analysis from the HARP study is of interest.

OCT is superior to angiography and intravascular ultrasound for identifying subtle coronary lesions and plaque characteristics, whereas CMR is superior for identifying myocardial injury caused by ischemia, inflammation, or other causes. Combining these two techniques makes sense for evaluating MINOCA patients. Unfortunately, even in this research study, there were significant challenges in successfully accomplishing both imaging modalities. Of 301 patients, there were interpretable OCT studies for only 145. Of those, only 59% studied all three main coronary arteries. Of the 145 with successful OCT, there were interpretable CMRs for 116.

Perhaps the most surprising finding is that of 145 patients who underwent OCT, the site investigators thought 53% of these patients had normal coronary arteries, whereas none of the 145 patients did according to OCT. Of course, the definition of normal coronary arteries may have been different between the two studies. OCT revealed about two-thirds of the patients had at least one stenosed artery at 11-30%. For a little more than another one-third of patients, there were 31-50% stenoses. Clearly, there was a disconnect between angiography and OCT.

On the other hand, not all showed evidence of MI or myocardial injury on CMR. Many of these coronary plaques were not the MINOCA culprits. Plaque disruption by OCT did correlate with CMR findings, which strengthens the conclusion that these were culprit lesions. These disruptions could have caused small thrombi that briefly occluded the artery (or embolized distally or provoked transient coronary spasm). The low incidence of spasm detection may have been caused by nitroglycerin administered during angiography or arterial vasodilators given in radial cases. No provocative tests for spasm were conducted. Also, 40% of patients with normal CMR studies exhibited culprit lesions on OCT, yet another disconnect. These sophisticated imaging techniques could aid cardiologists with their diagnosis and treatment of MINOCA. However, these techniques are challenging to accomplish in this clinical setting and may not always give the most helpful information.