Abnormalities in myocardial perfusion in the absence of coronary artery disease (CAD) occur in patients with hypertrophic cardiomyopathy (HCM) [
1‐
4]. While fixed perfusion defects can occur from extensive HCM-related myocardial fibrosis [
5], there is also a high prevalence of abnormal myocardial blood flow (MBF) reserve during vasodilator testing [
1,
3]. In some studies, reduced flow reserve has been attributed to low hyperemic perfusion rather than from the high resting flow that can occur from hypertrophy-related increased wall stress and oxygen demand [
1,
6]. The presence of a relationship between inducible ischemia and the degree of either myocardial fibrosis or left ventricular outflow tract gradients in HCM is uncertain based on mixed results [
1,
3,
4,
7]. Less controversial is the association between inducible ischemia and adverse clinical outcomes [
1,
3].
Ischemia in the absence of CAD in HCM has been attributed to structural abnormalities such as microvascular rarefaction and arteriolar medial hyperplasia [
8‐
10]. Functional abnormalities are also possible from exaggerated systolic flow reversal and delay in diastolic forward flow in the distal coronary arteries, or from epicardial-endocardial perfusion pressure loss [
11‐
13]. These mechanisms could contribute to ischemia in non-hypertrophied segments which has been reported in HCM and even in high-risk gene carriers [
14,
15]. In the current study, vasodilator stress myocardial contrast echocardiography (MCE) perfusion imaging was performed in patients with septal-variant HCM to assess the prevalence of reversible ischemia in both the hypertrophied and non-hypertrophied regions; and to spatially characterize ischemia as transmural diffuse, patchy, or subendocardial in distribution. Using parametric analysis, abnormalities in MBF at rest or during vasodilator stress were classified as being attributable to an impairment in microvascular flux or from loss of functional microvascular units. In a small subset of subjects undergoing septal myectomy for symptomatic obstruction, repeat MCE was performed to assess for changes in myocardial MBF reserve based on the potential impact of improving vascular function by reducing late intraventricular systolic pressures.
Methods
Subjects
The study was approved by the Investigational Review Board at Oregon Health & Sciences University and registered with ClinicalTrials.gov (NCT02560467). The study design was a prospective, non-blinded study of seventeen subjects between the ages of 19 and 70 with a diagnosis of septal variant HCM and fifteen age-matched subjects free of cardiac symptoms with no more than one CAD risk factor (lipid disorder, hypertension, diabetes, smoking) who were recruited to serve as normal controls. Subjects with HCM were recruited if they had a diagnosis made by echocardiography or cardiac magnetic resonance imaging (CMR) with a maximal septal thickness of 15 mm or greater, and also had undergone CMR with late gadolinium enhancement (LGE) imaging for quantification of fibrosis within the preceding 6 months. Subjects were excluded for known coronary or peripheral artery disease, significant valvular heart disease other than that caused by systolic anterior motion which could be no more than moderate in severity, history of resuscitated sudden cardiac death (SCD), left ventricular (LV) systolic dysfunction (ejection fraction [LVEF] < 50%), pregnancy, contraindications to regadenoson, allergy to ultrasound enhancing agents, or elite athlete status. Additional exclusion criteria for HCM subjects included prior septal reduction therapy (either surgical or alcohol ablation), pacemaker-dependent rhythm, presence of LV aneurysm, or treatment with cardiac myosin inhibitor.
Symptom status and SCD risk
Angina symptoms in subjects with HCM were determined by history. Risk for SCD was determined by an established risk prediction model based on subject age, echocardiographic indices, history of arrhythmia, symptoms, and family history [
16].
Vasodilator stress myocardial contrast echocardiography
Vasodilator stress MCE was performed in all subjects and was repeated at least 12 months after myectomy in those who were referred for surgical septal reduction. Subjects abstained from caffeine for 48 h. MCE perfusion imaging (iE33, Philips Ultrasound, Andover, MA) was performed at a centerline frequency of 2.0 MHz with multi-pulse amplitude-modulation imaging at a mechanical index of 0.12–0.16. Overall gain was adjusted to levels just under those that produced background myocardial speckle. Images were acquired in the apical 4-chamber, 2-chamber, and long-axis imaging planes. Lipid-shelled microbubbles with a gas core containing either sulfur hexafluoride (Lumason, Bracco Diagnostics, Monroe Township, NH) or octafluoropropane (Definity, Lantheus Medical Imaging, North Billerica, MA) were used. Lumason was reconstituted in 5 mL of normal saline while activated Definity was diluted to 30 mL total volume in normal saline. Infusion rates were kept constant for each individual at a rate of 1.0 to 1.5 ml/min. A 5-frame high-power (mechanical index > 0.9) sequence was applied to destroy microbubbles in the imaging sector through inertial cavitation, after which electrocardiographically-triggered end-systolic frames were acquired until visual replenishment had occurred. MCE was performed at rest and during vasodilator stress produced by intravenous administration of regadenoson (0.4 mg). Heart rate and blood pressure were recorded at baseline and three minutes after injection of regadenoson.
MCE analysis
Analysis was performed by a reader blinded to MRI and clinical data, other than diagnosis of HCM which is readily apparent on the echocardiogram. Perfusion was qualitatively defined as abnormal if there was lack of complete microvascular refill within 5 s at rest, or within 2 s during vasodilator stress [
17] (See Additional file
1) Perfusion abnormalities were categorized as being evenly transmural, subendocardial, or patchy in appearance. Quantitative perfusion analysis was performed using software developed for MCE perfusion imaging (iMCE, Narnar LLC, Portland, OR). For control subjects, data were averaged from transmural regions-of-interest placed over each perfusion territory of all three major coronary arteries. For HCM subjects, transmural regions-of-interest were separately drawn over the hypertrophic and non-hypertrophic control subjects in either the 4-chamber or 3-chamber view. Regions with obvious rib artifact or cavity attenuation were excluded. The first frame after inertial cavitation was digitally subtracted from all subsequent frames and background-subtracted time-intensity data were fit to the function:
$$\mathrm y=\mathrm A(1-\mathrm e^{-\mathrm{\beta}t})$$
where
y is signal intensity at time
t,
A is the plateau intensity reflecting relative microvascular blood volume (MBV), and
β is the rate constant reflecting microvascular blood flux rate. Myocardial MBF was quantified by the product of MBV and β [
18].
Echocardiography
Echocardiography (iE33, Philips Ultrasound, Andover, MA) was performed to assess chamber dimensions, wall thickness, left ventricular function, and peak LVOT gradient according to guidelines published by the American Society of Echocardiography [
19]. LV volumes, LVEF, and stroke volume in HCM subjects were calculated using the modified Simpson’s method. Stroke volume in controls was calculated by the product of LVOT area and time-velocity integral measured by pulsed-wave spectral Doppler. LV stroke work index was calculated by:
$$0.0136\times\mathrm{mean}\,\mathrm{arterial}\,\mathrm{pressure}\times\mathrm{stroke}\,\mathrm{volume}\,\mathrm{index}$$
For HCM subjects, LVOT gradient was added to mean arterial pressure, although this approach overestimates actual work based on the end-systolic nature of the gradient. Myocardial work index was calculated by the product of stroke work index and heart rate. The LVOT gradient in subjects with HCM were measured both at rest and during vasodilator stress upon completion of MCE perfusion imaging.
Assessment of fibrosis by CMR
A standardized CMR protocol was performed on a 1.5 Tesla (T) scanner (Philips 1.5 Achieva or Integra) with multi-channel channel phased-array chest coils and electrocardiographic gating. Cine steady-state free precession imaging was performed covering the whole heart in 8 mm thick slices, though these data were not used for analysis. For LGE, a phase-sensitive inversion-recovery sequence was acquired 12–15 min after intravenous gadolinium contrast administration (0.2 mmol/kg). Distribution and extent of LGE was assessed both visually and quantitatively using the six standard-deviation threshold according to the Society for Cardiovascular Magnetic Resonance standards [
20].
Statistical analysis
Data were analyzed using Prism (version 9.0, GraphPad, San Diego, CA). For data determined to be normally distributed by the D’Agostino and Pearson omnibus test, differences were assessed by one-way ANOVA with post-hoc comparisons made by paired or unpaired Student’s t-test with Tukey’s test to adjust for multiple comparisons. Unless otherwise described, normally-distributed data are expressed as mean ± standard deviation. Differences for non-normally distributed data were assessed with Friedman’s test with post-hoc individual comparisons by Mann–Whitney U test for non-paired data or Wilcoxon signed-rank test for paired data. Non-normally treated data are expressed as median with interquartile range (IQR). Differences in proportions were compared using χ2 analysis. Relationship between MCE perfusion data and other clinical data were determined using either Spearman’s rank correlation coefficient (ρ) or Pearson correlation coefficient. Differences were considered significant at p < 0.05.
Discussion
Hypertrophic cardiomyopathy is a disease with a wide degree of phenotypic variability. Clinical studies demonstrating inducible ischemia in patients obstructive HCM were published four decades ago using non-quantitative radionuclide imaging [
2,
22,
23]. Since then, quantitative perfusion imaging with positron emission tomography, CMR, and MCE in patients with HCM have confirmed that myocardial blood flow in the hypertrophied and non-hypertrophied regions is frequently reduced during exercise or vasodilator stress, and occasionally at rest [
1,
3,
4,
7]. In the current study, we demonstrated that vasodilator stress MCE with regadenoson can be used to identify abnormalities in perfusion at rest and during stress in patients with HCM, and that the spatial manifestations of perfusion defects are varied, with subendocardial or patchy abnormalities being more common than transmural diffuse defects in hypertrophied segments. Substantial reduction in hyperemic perfusion during vasodilator stress is commonly found in those with large amounts of fibrosis detected by LGE. We also demonstrated that improvement in hyperemic perfusion in both the hypertrophied and non-hypertrophied regions can occur late after septal myectomy.
In HCM, reduced perfusion reserve in the absence of atherosclerotic CAD has been attributed, in part, to structural abnormalities of the vasculature. On histopathology, medial hyperplasia and lumen narrowing of small coronary arteries and arterioles, and a reduction in myocardial capillary density in hypertrophic regions has been described in HCM [
8‐
10]. The latter feature indicates a failure in compensatory remodeling of the distal circulation to address the increased LV mass, cellular hypertrophy, and increased LV work and wall stress in HCM. Because of the compensatory reserve in the capacity of arterioles to dilate and capillary units to recruit, resting perfusion can be preserved in most patients with HCM. Yet, partial exhaustion of reserve and increased resistance from arteriolar narrowing and reduced capillary density is expected to produce myocardial ischemia during hyperemic stress or increased metabolic demand.
Functional abnormalities of the microcirculation in HCM have also been described. Extravascular compressive forces from high LV systolic and diastolic pressures in combination with the normal transmural pressure drop would be expected to reduce maximal flow in HCM, particularly in the endocardium. This mechanism has been proposed to explain reduced endocardial flow reserve in HCM, particularly in patients with high LV end-diastolic pressures or extreme septal hypertrophy [
4,
7]. Abnormalities in the phasic flow of coronary arteries can occur from altered hemodynamic forces in HCM. Studies using invasive coronary flow wires or non-invasive coronary wave intensity measurements have revealed a marked predominance of diastolic flow and more prominent retrograde systolic flow in distal coronary arteries in subjects with HCM, which can be further accentuated by inotropic stress [
12,
24]. Exaggerated retrograde flow combined with delayed or shortened diastolic relaxation can result in reduced antegrade discharge from small arteries or large arterioles that normally act as a type of “hydraulic capacitor” [
12,
25]. From a clinical perspective, this functional abnormality is likely to worsen as LV end-systolic pressure, myocardial diastolic pressure, and heart rate increase.
In the current study, the spatial distribution of perfusion abnormalities during vasodilator stress, whether from structural or functional causes, was assessed by MCE. This technique provides parametric information on whether abnormalities in perfusion are secondary to reduced MBV, which can occur from either capillary rarefaction or functional non-patency of microvascular units [
26]. It also measures microvascular flux rate which can be reduced from high resistance anywhere along the vascular network [
27]. At rest, MCE revealed very modest reductions in MBV in both the hypertrophic and non-hypertrophic regions of patients with HCM despite these subjects having higher systolic wall stress and work. There is reason to believe that this abnormality was from abnormalities in phasic flow based on results from previous studies showing a high degree of cyclic video intensity at the LV apex, primarily from low systolic intensity, during resting MCE in patients with apical HCM [
28]. Ordinarily, our finding of reduced perfusion at rest and increased work would be expected to result in ischemia. Yet these subjects were not symptomatic and LV systolic function was normal. This paradox could be related to compensatory mechanisms to increase oxygen delivery, even out of proportion to calculated work, based on studies using
11C-acetate PET indicating that myocardial oxygen consumption is not reduced in subjects with HCM who have normal to high LVEF [
29].
Perfusion abnormalities in those with HCM became much more prominent during vasodilator stress, primarily because of a deficit in the ability to appropriately augment microvascular flux rate. This finding is somewhat different from previous quantitative MCE studies that found that reduced MBF in HCM, both at rest and during stress, is attributable to abnormal MBV [
7]. We believe differences between the two studies can be explained by much less severe LVOT obstruction in the current study. We found that the dominant spatial pattern for hyperemic flow deficits was subendocardial or patchy in distribution. These patterns do not indicate any one mechanism since they could occur from pre-capillary drop in resistance from arteriolar narrowing, microvascular rarefaction, or phasic functional abnormalities of coronary flow. We observed a marked improvement in hyperemic flow, including in non-hypertrophied territories, after septal myectomy in two subjects who had very high resting LVOT gradients and low hyperemic flow (approximately one-third of control subject average) prior to surgical intervention. This finding suggests that abnormal flow from high systolic compressive forces in combination with delayed relaxation can affect global myocardial perfusion and is reversible late after correction of the high systolic gradient.
There are several important limitations of the study. The total number of subjects studied and the number of subjects undergoing myectomy was low because of strict entry criteria, including the need for recent CMR and exclusion for treatment with a myosin inhibitor which was being investigated concurrently with recruitment for this study. Yet data indicating a potential beneficial effect of myectomy on perfusion can be used to justify a larger prospective study in that narrow population of patients. Although MCE can be used to calculate absolute MBF in mL/min/g, this analysis was not performed because the requisite calculation of absolute MBV is valid only if blood pool microbubble signal is below the upper limit of the dynamic range which generally requires lower contrast infusion rates and appropriate scaling. Perfusion data were also not expressed as MBF normalized to work because of limitations in using end-systolic pressures to reflect total systolic load. Instead, we simply concluded that perfusion deficits in HCM occurred despite greater workload based on high systolic LV pressures. It should also be noted that vasodilator stress rather than exercise stress was used. The latter would provide a better test for stress-induced deficits in MBV, although the level of stress induced would be difficult based on difficulties in determining true afterload in those with dynamic gradients. Similarly, we have not tested other vasodilator agents, such as NO donors that could produce different results based on their prominent effects on pre-load which would affect wall stress and myocardial work, and based differences in the circulatory network where NO acts. Finally, ischemia from CAD was excluded by angiography in only about half of the HCM subjects, all of whom had anginal symptoms. One patient was excluded from analysis based on the presence of severe CAD on angiography performed for a typical coronary distribution of perfusion deficits on stress MCE.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.