Segmental cardiac strain assessment by two-dimensional speckle-tracking echocardiography in surviving MIS-c patients: Correlations with myocardial flow reserve (MFR) by 13 N-ammonia PET-CT.

Background: Multisystem inflammatory syndrome in children (MIS‐c) is associated with severe cardiovascular impairment and eventually death. Pathophysiological mechanisms involved in myocardial injury were scarcely investigated, and cardiovascular outcomes are uncertain. Autopsy studies suggested that microvascular dysfunction may be relevant to LV impairment. Objective: We aimed to evaluate segmental LV longitudinal strain by 2DST echocardiography and myocardial flow reserve (MFR) by 13 N‐ammonia PET‐CT, in six surviving MIS‐c patients. Methods: Each patient generated 34 LV segments for combined 2DST and MRF analysis. MFR was considered abnormal when <2, borderline when between 2 and 2.5 and normal when >2.5. Results: From July 2020 to February 2021, six patients were admitted with MIS‐c: three males, aged 9.3 (6.6–15.7) years. Time from admission to the follow‐up visit was 6.05 (2–10.3) months. Although all patients were asymptomatic and LV EF was ≥55%, 43/102 (42.1%) LV segments showed MFR <2.5. There was a modest positive correlation between segmental peak systolic longitudinal strain and MFR: r =.36, p =.03 for basal segments; r =.41, p =.022 for mid segments; r =.42, p =.021 for apical segments. Median peak systolic longitudinal strain was different among MRF categories: 18% (12%–24%) for abnormal, 18.5% (11%–35%) for borderline, and 21% (12%–32%) for normal MFR (p =.006). Conclusion: We provided preliminary evidence that surviving MIS‐c patients may present subclinical impairment of myocardial microcirculation. Segmental cardiac strain assessment 2DST seems useful for MIS‐c cardiovascular follow‐up, given its good correlation with 13 N‐ammonia PET‐CT derived MFR.

Keywords: children; COVID‐19; echocardiography; PET‐CT; speckle‐tracking

• Abbreviations
• 2DST two‐dimensional speckle‐tracking echocardiography
• LV left ventricle
• MFR myocardial flow reserve
• MIS‐c multisystem inflammatory syndrome
• PET‐CT Positron emision tomography/computed tomography

Multisystem inflammatory syndrome in children (MIS‐c), which is temporally related to severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) infection, is a rare and serious condition associated with cytokine storm, cardiovascular impairment, pediatric intensive care admission, and potentially death. Ventricular dysfunction, pericardial effusion, valvar regurgitation, and coronary artery inflammation in MIS‐c were documented worldwide.1

Although ventricular systolic dysfunction in MIS‐c patients has been extensively described, the pathophysiological mechanisms involved in myocardial injury were scarcely investigated. Moreover, long‐term cardiovascular outcomes in MIS‐c surviving patients are still subject of research.2

A recent autopsy study conducted in our institution revealed that endothelial cell injury in MIS‐c seems to play a critical role in the pathogenesis of the disease, leading to thrombotic events and tissue ischemia. Electron microscopy demonstrated not only the presence of viral particles in endothelial cells but also cellular and tissue structural changes, such as disrupted capillary walls associated with fibrin clot formation in myocardial microvasculature.3

We have also observed significant d‐dimer elevation among MIS‐c patients who developed ventricular systolic dysfunction in the acute phase of the disease, detected by standard echocardiogram. These findings point out that clot formation and microvascular dysfunction are relevant to the pathophysiology of myocardial impairment in MIS‐c.4

Microvascular dysfunction can be indirectly assessed by measuring myocardial blood flow (MBF), before and after vasodilation (during stress). Cardiac positron emission tomography–computed tomography (PET‐CT) is the gold standard noninvasive test for MBF. The ratio of the MBF measured at maximal vasodilation over the MBF at rest is referred to as myocardial flow reserve (MFR), which is a measure of the vasodilatory reserve of the myocardium. Reduced MFR, in the absence of flow‐limiting coronary artery disease, is believed to reflect dysfunction in the myocardial microvasculature.5,6 Although PET‐CT remains the most accurate noninvasive modality for MFR quantification and the modality to which others are compared, multiple factors have limited widespread clinical application, including low availability of scanners, radiation exposure, and increased cost.7

Segmental cardiac strain assessment by two‐dimensional speckle‐tracking echocardiography (2DST) has been used to identify subclinical myocardial dysfunction in adult and pediatric coronavirus disease 19 (COVID‐19) patients and has proven to be a more sensitive tool than left ventricle ejection fraction.8,9 Strain echocardiography is a highly reproducible, largely available, and cost‐effective technique. Nevertheless, we are unaware of previous studies comparing speckle‐tracking derived myocardial strain and MRF by PET‐CT, in MIS‐c patients.

The present study aimed to evaluate segmental cardiac strain by 2DST echocardiography, as well as MFR by 13 N‐ammonia PET‐CT, in a series of surviving MIS‐c patients. We hypothesized that myocardial microvascular dysfunction (reduced MFR) would be detectable in these patients and would be associated with left ventricular segmental strain reduction.

This was a cross‐sectional study that included all MIS‐c patients admitted to our tertiary referral institution from July 2020 to July 2021.

MIS‐c diagnosis was made according to Centers for Disease Control and Prevention (CDC) criteria, as follows10:

• An individual aged <21 years presenting with fever (body temperature >38.0°C or report of subjective body temperature elevation lasting ≥24 h), laboratory evidence of inflammation, evidence of clinically severe illness requiring hospitalization, with multisystem (>2) organ involvement (cardiac, renal, respiratory, hematologic, gastrointestinal, dermatologic, or neurological); AND
• No alternative plausible diagnoses; AND
• Positive for current or recent SARS‐CoV‐2 infection by real‐time polymerase chain reaction (RT‐PCR), serology, antigen test, or exposure to a suspected or confirmed COVID‐19 case within the 4 weeks prior to the onset of symptoms.

According to CDC, evidence of inflammation includes, but is not limited to, one or more of the following: an elevated C‐reactive protein (CRP), erythrocyte sedimentation rate (ESR), fibrinogen, procalcitonin, d‐dimer, ferritin, lactic acid dehydrogenase (LDH), or interleukin 6 (IL‐6), elevated neutrophils, reduced lymphocytes, and low albumin. RT‐PCR in respiratory specimens was performed to detect SARS‐CoV‐2 RNA.

Serologic tests included two different methods during the COVID‐19 pandemic at our institution: immunochromatography assay for SARS‐CoV‐2 specific IgM/ IgG antibody detection and anti‐SARS‐CoV‐2 enzyme‐linked immunosorbent assay (ELISA) for IgG antibodies detection.4

Exclusion criteria were death or refusal to participate in the study.

All surviving MIS‐c patients were recruited by the attending physician, during routine visits to our outpatients' clinics. First, clinical, laboratory, and echocardiographic data from MIS‐c admission were collected from medical records. Then, as part of the research protocol, echocardiograms with speckle‐tracking derived strain analysis and 13 N‐ammonia PET‐CT scans were ordered, as well as follow‐up blood tests: C‐reactive protein, d‐dimers, fibrinogen, troponin I, hemoglobin, lymphocytes count, platelets count, urea, creatinine, alanine transaminase, and aspartate transaminases.

The pediatric cardiologist performing all echocardiograms and the radiologist responsible for the PET‐CT scans evaluation were unaware of patients' clinical status by the time of the examinations.

The Institutional Research Ethics Committee approved the study and informed consents were obtained from parents or legal guardians.

Standard transthoracic echocardiography was performed according to the recommendations of the American Society of Echocardiography and included M‐mode, two‐dimensional imaging, conventional, and tissue Doppler evaluation at the septal and lateral mitral annulus.11 The equipment used was a Philips Affiniti 70 (Andover, MA 01810 USA), with multifrequency transducers (S 5–1 and S 8–3 MHz). Cardiac chamber dimensions were obtained using two‐dimensional mode, and left ventricle ejection fraction (LV EF) was calculated by Simpson's method (normal LV EF ≥55%).11 The z‐score values for the measures were calculated according to the "Pediatric Heart Network echo z‐score Project" data.12 LV mass (g) was estimated using Devereaux's formula according to the Penn convention and indexed for height (m) raised to an exponential power of 2.7.11 LV hypertrophy was diagnosed whenever LV mass index was greater than the 95th percentile for sex and age, according to Khoury et al.13 Evaluation of LV diastolic function included mitral E/e' ratio, with e' being the average of values obtained by tissue Doppler at the septal and lateral annulus (normal E/e' <14).11

The main principle of 2DST is that each segment of myocardial tissue displays a specific pattern of gray values in the ultrasound image, commonly referred to as a speckle pattern. Tracking this acoustic pattern during the cardiac cycle enables the observer to follow myocardial motion and to directly assess ventricular deformation.14 To evaluate segmental LV longitudinal systolic strain, two‐dimensional harmonic image cine‐loop recordings of apical four‐, three‐, and two‐chamber views were acquired and stored digitally for analysis (Figure 1A–C). A sector scan angle of 30–60° and frame rates of 60–90 Hz were chosen. A good‐quality electrocardiogram signal was obtained simultaneously. The endocardial and epicardial tracing was automatically generated by the computer algorithm and manually adjusted to cover the whole myocardium wall, when necessary (QLabTM software, Philips).14 The extent of myocardial strain in longitudinal direction throughout the cardiac cycle was computed as percentages (absolute values). A bull's eye plot was generated for each patient, demonstrating peak systolic longitudinal strain values of the 17 LV segments analyzed14: basal anterior, basal anteroseptal, basal inferoseptal, basal inferior, basal inferolateral, basal anterolateral, mid anterior, mid anteroseptal, mid inferoseptal, mid inferior, mid inferolateral, mid anterolateral, apical anterior, apical septal, apical inferior, apical lateral, and apex (Figure 1D).

Positron emission tomography (PET) images were obtained with a Discovery 710 PET/CT scanner (GE Healthcare, Milwaukee, WI, USA). Subjects were maintained in a fasting state for at least 6 h before the study and were told not to consume methylxanthine containing foods or beverages (coffee, chocolates, soft drinks, and tea) for at least 24 h before the PET scan.

Each acquisition was preceded by computed tomography (CT) transmission scan for photon attenuation correction. 13N‐Ammonia was produced by means of an on‐site cyclotron installed at our institution (PET trace TM 880, GE Healthcare), by 16O (p, α)13N nuclear reaction with water irradiation, followed by a reduction of the 13N compounds to ammonia with titanium hydroxide.15

For measurement of MBF at rest and at pharmacological stress (adenosine‐induced hyperemia), 13N‐Ammonia was administered intravenously (0,286 mCi/kg) over a 10‐sec period, the intravenous line was flushed with additional saline over a 10‐sec interval and standardized imaging protocols were performed according to the American Society of Nuclear Cardiology guidelines. A dynamic rest imaging was performed for 15 min (frames: 20 × 5 s, 30 × 10 s, 20 × 10 s, and 1 × 300 s). Stress imaging was performed identically, after adenosine infusion over 6 min (0.142 mg/min/kg). The myocardial perfusion radiopharmaceutical was injected about halfway into the adenosine infusion (at 3 min), when maximal vasodilatation and myocardial hyperemia occurred. MBF was expressed as ml/g/mine. The electrocardiogram was monitored continuously throughout the procedure, using a three‐channel electrocardiographic monitor. Blood pressure was monitored every 3 min at rest and every 1 min during stress testing, using an electronic sphygmomanometer.16

The quantitative PET datasets were fused with CT using commercially available software (CardIQ Fusion, GE Healthcare). Quantitative MBF and MFR were determined using the PMOD TM software package, version 3.4002 (PMOD Technologies LLC). Myocardial and blood‐pool time‐activity curves (TAC) were obtained from dynamic frames corrected for radioisotope decay. Segmental MBF was measured in each phase (rest and stress adenosine) by the model fitting of the blood‐pool and myocardial TACs, corrected for spill‐over and partial volume. MFR was calculated as the ratio of stress MBF over the rest MBF (the 17‐segment model according to the American Society of Nuclear Cardiology recommendations). For each LV segment, MFR <2 was considered abnormal, 2 ≤ MFR ≤2.5 borderline, and MRF >2.5 normal (Figure 2). These three categories were adopted according to the European procedure guidelines for PET‐CT quantitative myocardial perfusion imaging,17 as well as previously published studies on cardiac PET imaging for the detection and monitoring of coronary artery disease and microvascular health.18 The mentioned thresholds aim to stratify coronary artery disease risk among evaluated patients.

Statistical analysis was performed using SPSS version 13.0 (SPSS Inc.). Categorical data were reported as percentages and continuous data as mean ± standard deviation (sd) or median (range). Student's t‐test was used to assess normally distributed continuous data and Mann–Whitney test to assess non‐normally distributed continuous data. Fisher's exact test was used to compare categorical data. Pearson's correlation was used to test the relationship between LV segmental longitudinal systolic strain and segmental myocardial flow reserve. The level of significance was set at 5% (p < .05). The data of reproducibility of our Echo Lab for standard transthoracic parameters, as well as for 2DST, have been published elsewhere.4,19

From July 2020 to February 2021, six MIS‐c patients were admitted to our institution: three females; aged 9.3 (6.6–15.7) years. None of the patients had positive RT‐PCR. 5 had positive serologic tests and 1 was exposed to a confirmed COVID‐19 case within the 4 weeks prior to the onset of symptoms. Demographic and clinical data from hospitalization period are shown in Table 1.

1 TABLEDemographic and clinical data in the acute phase of MIS‐c. Categorical data were reported as percentages and continuous data were reported as mean ± standard deviation (SD) or median (rage)

2 Prophylactic anticoagulation.

All of them survived and agreed to participate in the present study. Before MIS‐c diagnosis, none of the six patients had evidence of cardiac disorders. By the time they were seen at the outpatients' clinic, their cardiovascular clinical examination was normal and they were classified as New York Heart Association functional class 1.

Time from MIS‐c admission to the follow‐up visit was 6.05 (2–10.3) months. The interval between follow‐up echocardiogram and 13N‐ammonia PET‐CT was 1 (0.20–1.67) month.

Standard echocardiograms performed within the first week of MIS‐c admission revealed normal left chambers dimensions, normal LV mass index (≤95th percentile for sex and age), preserved LV ejection fraction (≥55%) and normal LV filling pressure (average E/e' <14) in all patients. Three patients showed discrete pericardial effusion and two had mild coronary artery abnormalities (one with higher left coronary artery echogenicity and one with mild ectasia of right and left coronary arteries). Follow‐up standard echocardiograms were also considered normal, except for one patient with persistent discrete pericardial effusion (Table 2). Laboratory data revealed significant reduction of inflammation and cardiac injury biomarkers throughout the study (Table 2).

2 TABLEStandard echocardiogram and laboratory data of MIS‐c patients at admission and at the follow‐up visit.

1 Note

• 3 Categorical data were reported as percentages and continuous data as mean ± standard deviation (SD) or median (range).
• 4 Discrete pericardial effusion.
• 5 One patient showed higher left coronary artery echogenicity and 1 patient had discrete ectasia of right (z‐score = +2.7) and left (z‐score = +2.9) coronary arteries.
• 6 The Significance of Bold values represent significant p values.

Left ventricle strain assessment by 2DST included 102 segments (17 for each patient), all successfully tracked. The median value of LV basal segments peak systolic longitudinal strain was 18% (13%–25%), of mid segments was 18.5% (11%–26%) and of apical segments was 24% (15%–35%). The median LV global peak systolic longitudinal strain was 20.6% (17.4%–23.2%).

The same 102 LV segments were evaluated by 13 N‐ammonia PET‐CT and the median MFR of basal segments was 2.55 (0.93–3.98), of mid segments was 2.74 (0.92–3.92) and of apical segments was 2.53 (0.83–3.85).

LV segmental strain and MFR values obtained for each patient are shown in Table 3. Reduced LV global longitudinal strain (<18%)14 was detected in only two individuals (patients 1 and 4). Of note, even patients with preserved LV global longitudinal strain (≥18%)14 showed segments with abnormal or borderline MFR (≤2.5) (patients 5 and 6).

3 TABLELeft ventricle segmental peak systolic longitudinal strain (Long strain) and myocardial flow reserve (MFR) for each patient

7 Abbreviations: LAD, left anterior descending coronary artery; LCX, left circumflex coronary artery.

Distribution of LV segments according to MFR categories for each patient is presented in Table 4. LV segments with abnormal or borderline MFR were detected in all three patients that were followed for less than 6 months (patients 4,5 and 6). On the other hand, only one out of three individuals followed for more than 6 months showed abnormal or borderline MFR LV segments (patient 1) (Table 4).

4 TABLEDistribution of left ventricular segments according to the three different myocardial flow reserve (MFR) categories for each patient

The small number of evaluated patients precluded any association test between demographic/clinical/laboratorial data and segmental LV MFR impairment.

Figure 3 Shows the distribution of LV segments (basal, mid and apical) among the MRF categories. The proportion of basal, mid and apical segments was statistically similar among MRF categories (p > .05).

Comparing MFR categories, the median value of segmental peak systolic longitudinal strain was significantly different (p = .006), as shown in Figure 4 Median peak systolic longitudinal strain was 18% (12%–24%) for abnormal MFR segments, 18.5% (11%‐35%) for borderline MFR segments and 21% (12%–32%) for normal MFR segments.

There was a significant positive correlation between segmental peak systolic longitudinal strain obtained by 2DST and MFR obtained by 13 N‐ammonia PET‐CT: r = .36, p = .03 for basal segments;r = .41, p = .022 for mid segment; r = .42, p = .021 for apical segments (Figure 5A–C).

This study stands out for the identification of myocardial microvascular dysfunction in a series of asymptomatic surviving MIS‐c patients, using 13 N‐ammonia PET‐CT derived MFR. A significant correlation between LV segmental peak systolic longitudinal strain and MFR was observed, reinforcing speckle‐tracking echocardiography as a promising tool for the assessment of microvascular derangements in this particular population.

Standard echocardiograpy was unable to identify LV dysfunction in our patients, both at MIS‐c admission and at the follow‐up evaluation. Even though there was a significant reduction of cardiac injury biomarkers and no patient had symptoms of overt congestive heart failure at follow‐up visit, 43 (42.1%) out of the 102 analyzed LV segments showed abnormal or borderline MFR on 13 N‐ammonia PET‐CT scan. Considering that peak systolic longitudinal strain was significantly higher in LV segments with normal MFR than in segments with borderline or abnormal MFR, larger studies may establish a cut‐off value of myocardial strain capable of detecting segments with MFR ≤2.5 with reasonable sensitivity and specificity.

Since a physiologic apex‐to‐base gradient of strain exists for all ages (higher values at apex), we analyzed possible correlations between segmental MRF and correspondent peak systolic longitudinal strain for apical, mid, and basal segments separately.14 Interestingly, speckle‐tracking derived peak systolic longitudinal strain, evaluated at rest, correlated with segmental MFR, obtained after pharmacological stress with adenosine. These findings suggest that impaired vasodilation capacity may contribute to myocardial deformation abnormalities, which were also reported in the follow‐up of COVID‐19 pediatric patients by previous authors.9,20

Cumulative evidence has clarified the picture of COVID‐19 as a vascular disease.21 SARS‐CoV‐2 uses the angiotensin‐converting enzyme 2 (ACE2) to facilitate entry into target cells, and vascular endothelium is known to express abundant ACE2.22 In the context of acute myocarditis induced by SARS‐CoV‐2, different inflammatory cells invade the myocardial architecture, release cytokines, growth factors and impair coronary microcirculation. The vascular tone of coronary microvessels is physiologically regulated by myogenic, neural, metabolic, and endothelial mechanisms, according to local demands.23 Myocardial inflammation interferes with this precisely regulated system, induces endothelial dysfunction and microcirculatory disturbances, which finally results in myocardial ischemia. Moreover, hyper‐inflammatory state leads to activation of the coagulation cascade and suppression of fibrinolytic mechanisms, with a resulting prothrombotic milieu.21 Ultrasound‐guided minimally invasive autopsy studies in MIS‐c patients corroborate the above‐described pathophysiological mechanisms, demonstrating viral particles in different cell lineages of the heart (cardiomyocytes, endothelial, mesenchymal, and inflammatory cells), apart from microvascular thrombosis.4

The segmental MFR compromise detected by PET‐CT in our surviving MIS‐c patients, paralleled by segmental peak systolic longitudinal strain reduction, may represent residual damage to the coronary microcirculation, secondary to previous SARS‐CoV‐2 infection. It is valid to emphasize that we documented microcirculatory impairment even in the absence of coronary artery aneurisms, at standard echocardiography. Similarly, diminished MFR was also identified by 13 N‐ammonia PET‐CT in late follow‐up of Kawasaki disease patients, with angiographically normal coronary arteries.24 On the other hand, we have documented that only one out of three MIS‐c patients followed for more than 6 months showed abnormal or borderline MFR LV segments, which raises doubt about the possible transitory nature of microcirculatory damage.

Impaired vasodilation capacity, measured by PET‐CT MFR, was reported to be an independent predictor of subsequent cardiac events and associated with an increased risk of progression of congestive heart failure and death, among adult patients with idiopathic cardiomyopathy.25 That said, the presence and clinical significance of myocardial microvasculature compromise in surviving MIS‐c patients should be further investigated in the long term. It is of paramount importance to evaluate the effects of superimposed cardiovascular risk factors in this particular population, such as obesity, dyslipidemia, smoking, and systemic hypertension, which are known to impact myocardial microvascular functionality as well.23 Understanding the key dysregulated cascade triggered by COVID‐19 in the vascular system of MIS‐c patients provides information on mediators of the pathologic process, which may help tailor successful therapeutic interventions, reducing short‐ and long‐term morbidity and mortality.21

This is a single center cross‐sectional study that included a limited series of MIS‐c patients, which may preclude generalizations. Besides, the interval between MIS‐c admission and the follow‐up investigations with speckle‐tracking derived segmental strain combined with 13 N‐ammonia PET‐CT was not uniform among patients, varying from 2 to 10.3 months.

This is the first study to provide evidence that surviving MIS‐c patients may present with subclinical impairment of myocardial microcirculation. Dedicated pediatric follow‐up programs must be set, in order to assess future cardiovascular outcomes in this population. Segmental cardiac strain assessment by two‐dimensional speckle‐tracking echocardiography seems to be a useful technique for this purpose, given its large availability, cost effectiveness and good correlation with 13 N‐ammonia PET‐CT derived MFR.

Future longitudinal studies should be held in larger cohorts of survivors, aiming to determine the real incidence of myocardial microvascular compromise and its eventual reversibility.

The authors would like to thank Professor Magda Carneiro Sampaio for her invaluable support to our research team at University of São Paulo.

Gabriela Nunes Leal, Camila Astley, Marcos Santos Lima, and Maria de Fátima Rodrigues Diniz contributed to conceptualization/design, methodology, investigation, supervision, data curation, formal analysis resources, and writing. Alessandro Cavalcanti Lianza, Karen Saori Shiraishi Sawamura, Carolina Rocha Brito Menezes, and Camila Lino Martins Rodrigues da Silva contributed to conceptualization/design, methodology, and investigation. Vera Bain, Rodrigo Imada, William Chalela, Maria Fernanda Badue Pereira, Heloisa Helena de Sousa Marques, Carlos Alberto Buchpiguel, Bruno Gualano contributed to data curation, and formal analysis resources. Clovis Artur Silva contributed to conceptualization/design, supervision, funding acquisition, formal analysis resources, and writing. All authors gave final approval of the version to be published and agree to be accountable for all aspects of the work.