Background
Acute myocarditis is a significant cause of sudden death in young adults, and accurate screening for subclinical disease is needed. The aim of this study was to test the hypothesis that newer measures of tissue deformation and twist can detect ventricular dysfunction in patients with myocarditis and preserved left ventricular ejection fractions (LVEFs).
Methods
Twenty-eight consecutive patients (median age, 26.5 years; interquartile range, 19.3–33.8 years) with normal LVEFs and cardiovascular magnetic resonance features of myocarditis were prospectively recruited. Left ventricular tissue velocities, deformation, and twist were measured and compared with values in 64 healthy controls (median age, 25.1 years; interquartile range, 13.5–31.7 years).
Results
Patients with myocarditis had reduced annular e′ velocity and longitudinal and circumferential strain parameters ( P < .01) but similar LVEFs. Reduced lateral e′ velocity (odds ratio [OR], 1.77; 95% confidence interval [CI], 1.34–2.34), longitudinal strain (OR, 1.81; 95% CI, 1.38–2.38), circumferential early diastolic strain rate (OR, 1.31; 95% CI, 1.08–1.71), increased twist rate (OR, 1.02; 95% CI, 1.01–1.04), and earlier time to peak twist (OR, 0.80; 95% CI, 0.72–0.88) were identified as independent predictors of myocarditis, with abnormalities in any two of five predictors having 93% sensitivity and 91% specificity. Longitudinal strain parameters and lateral e′ velocity were improved at 1 year ( P ≤ .03) but remained reduced compared with controls ( P ≤ .02).
Conclusions
Patients with acute myocarditis and normal LVEFs had detectable left ventricular systolic and diastolic dysfunction on echocardiography. Tissue velocity, deformation, and twist parameters have the potential to improve the detection of patients with myocarditis and preserved LVEFs.
Acute myocarditis associated with a normal left ventricular (LV) ejection fraction (LVEF) is challenging to diagnose. Autopsy series of young adults with unexplained sudden death have shown myocardial inflammatory changes as the only abnormality in up to 12%, and as such, “subclinical myocarditis” is both underrecognized and potentially lethal. The importance of this entity is also emphasized by the observation that nearly half of adult patients with myocarditis develop symptomatic diastolic heart failure at medium term. Echocardiography is the initial imaging modality used, but diagnostic accuracy is limited, especially in patients presenting with chest pain or low-grade heart block, in whom LV function and size are almost always preserved. Newer echocardiographic measures of LV myocardial deformation and twist may have better sensitivity than conventional echocardiography for the detection of subclinical LV dysfunction. The current reference standard for noninvasive diagnosis of myocarditis is cardiovascular magnetic resonance (CMR). We hypothesized that changes in tissue deformation and twist are present and detectable by speckle-tracking echocardiography in children and young adults with normal LVEFs and CMR features of myocarditis.
Methods
We prospectively recruited patients with suspected viral myocarditis aged 10 to 45 years presenting with new-onset chest pain at rest lasting >30 min, with two consecutive elevated troponin I measurements and LVEFs > 50% on initial echocardiography. We excluded patients with ischemic heart disease defined by coronary angiography and/or CMR, patients with normal CMR findings, and patients with inadequate echocardiographic windows. A research echocardiographic examination was performed within 24 hours of CMR, with a follow-up study scheduled 6 to 12 months after enrollment. Echocardiographic findings of patients with CMR features of myocarditis as defined by the Lake Louise consensus were compared with those in 64 historical controls. The control population was recruited from volunteers aged 10 to 45 years with normal blood pressure and normal results on electrocardiography and transthoracic echocardiography in a previous study examining maturational changes of LV twist and untwist performance in healthy individuals. The present study received approval from the University of Alberta Research Ethics Board, and all adult patients and legal guardians of pediatric patients gave informed consent.
Echocardiography
Conventional LV Parameters
All studies were performed on a Vivid 7 ultrasound platform (GE Medical Systems, Milwaukee, WI) with electrocardiographic tracings. After exclusion of anatomic abnormalities, conventional parameters of LV end-diastolic and end-systolic volumes were measured using the biplane method of disks (modified Simpson’s rule). LVEF was calculated as (end-diastolic volume − end-systolic volume)/end-diastolic volume and expressed as a percentage. LV volumes were indexed to body surface area. Peak early and late mitral inflow Doppler velocities were measured and the velocities expressed as ratios. Isovolumic relaxation time was measured from mitral inflow and aortic outflow Doppler spectra. LV septal and lateral annular velocities were measured from the four-chamber (4CH) apical plane color tissue Doppler images optimized for frame rate (>150 Hz) and minimizing the angle of ultrasound beam to LV annular direction of travel (<20°). The Doppler sample size was set at 5 × 5 mm and placed within the myocardium <1 cm from the annulus at end-diastole, covering the longitudinal excursion of the mitral annulus in systole.
Speckle-Tracking Imaging Analysis
LV tissue deformation and twist were analyzed using commercially available speckle-tracking imaging software (EchoPAC version 7.1; GE Medical Systems). Research two-dimensional grayscale images were captured at the base and apex in short axis for circumferential strain and twist parameters measurements. The apical 4CH plane was used for longitudinal strain parameters. All research images were optimized for higher frame rates (mean, 87 ± 17 Hz). At least three consecutive cardiac cycles in each imaging plane were acquired, with the breath held at end-expiration. The best-quality cardiac loop was selected, and analysis was accomplished as follows: (1) The systolic period was manually defined, with “zero” set at the onset of the Q wave and ending at aortic valve closure. (2) The endocardium was manually traced, and the region of interest was adjusted to fit the myocardium thickness. (3) The software tracking algorithm was applied, and only cardiac loops passing the minimum tracking correlation in at least five of the six LV segments as well as a qualitative inspection of adequate tracking were included in the analysis.
LV Tissue Deformation
Longitudinal strain parameters assessed in the 4CH plane included peak systolic strain, peak systolic strain rate, and early diastolic strain rate. Likewise, peak circumferential systolic strain, systolic strain rate, and early diastolic strain rate were measured on basal and apical short-axis imaging ( Figures 1 and 2 ).
LV Twist Deformation
Speckle-tracking imaging analysis of the basal and apical circumferential plane also calculated the angle of rotation of the speckles within the plane about the LV centroid. A positive value by convention was assigned to counterclockwise rotation when viewed from apex to base. LV twist was calculated as the difference between the counterclockwise rotation at the apex (positive) and clockwise rotation at the base (negative) for each isochronal point. Peak twist rate and untwist rate were recorded from a discrete time derivative of the twist-time curve. The time to each of the recorded variables was normalized to systole and reported as 0% at onset of the Q wave on the electrocardiogram and 100% at aortic valve closure ( Figures 3 and 4 ). LV untwist performance was calculated as the percentage of completed untwisting at mitral valve opening and early diastolic filling at 130% of systole as (peak twist − twist at the cardiac cycle point of interest)/peak twist, expressed as a percentage. The reproducibility of strain and twist parameters in our laboratory has been established in previous publications.
CMR
Recruited patients underwent routine contrast-enhanced CMR scans as part of their clinical investigation using a 1.5-T magnet (Siemens Sonata; Siemens Medical Systems, Erlangen, Germany). All subjects were imaged using electrocardiographic gating and a dedicated cardiac coil. The imaging protocol consisted of long-axis and short-axis steady-state free precession cines to assess LV function, volumes, and mass; dark-blood T2-weighted short inversion time inversion recovery imaging to assess for myocardial edema; and phase-sensitive inversion recovery gradient-recalled echo imaging to assess for late gadolinium enhancement (myocardial necrosis) ( Figure 5 ). Postprocessing was performed using commercially available software (Syngo Argus; Siemens Medical Systems) by an expert CMR interpreter (I.P.) blinded to the echocardiographic results. Patients were considered to have CMR diagnoses of myocarditis if they had subepicardial or midwall necrosis on late gadolinium enhancement imaging and/or myocardial edema on T2-weighted imaging.
Statistical Analysis
Continuous variables were analyzed using Wilcoxon’s rank-sum test and are presented as medians with interquartile ranges (IQRs). Categorical variables were analyzed using Fisher’s exact test and are presented as counts with percentages. Nonindependent continuous variables were analyzed using Wilcoxon’s signed-rank test. Multivariate stepwise logistic regression was used to identify independent variables associated with the CMR diagnosis of myocarditis. The multivariate logistic regression model included a subset of the variables with P values < .05 on the corresponding univariate analysis that were selected a priori. Statistical significance was defined as P < .05. Data analysis was performed using SAS version 9.2 (SAS Institute Inc., Cary, NC).
Results
Thirty-nine consecutive patients who fulfilled the study criteria were recruited between January 2008 and July 2010. In addition to chest pain, 60% of patients presented with a viral prodrome of fever and malaise. The majority were devoid of other symptoms, except for two patients with associated dyspnea (one patient had biopsy-confirmed myocarditis, and the other had a more complex clinical course, receiving a diagnosis of associated atypical pneumonia) and one with syncope. There was no documented rhythm disturbance. After CMR and/or coronary angiography, nine patients were excluded, with four adult patients having evidence of small myocardial infarctions on CMR and five having normal results on both CMR and angiography. The remaining 30 patients received CMR diagnoses of myocarditis, with 70% having T2 edema and all having subepicardial or midwall late gadolinium enhancement, with most occurring in the inferolateral and lateral segments. Thirty-seven percent of patients with myocarditis underwent angiography at the discretion of their lead physicians, with none showing significant coronary artery disease (one patient had 40% mid diagonal stenosis). Two patients with myocarditis were subsequently excluded for inadequate echocardiographic image quality. The remaining 28 patients with CMR diagnoses of myocarditis and adequate echocardiographic imaging had their echocardiographic results compared with those of 64 controls. There were no differences in age and blood pressure between patients with myocarditis and controls ( Table 1 ). There was a predominance of male patients (87%) in the myocarditis group, consistent with previous cohorts of myocarditis. There were no gender differences in our control population for LV tissue deformation and LV twist parameters.
| Variable | Patients with myocarditis ( n = 28) | Controls ( n = 64) | P |
|---|---|---|---|
| Men | 24 (87%) | 34 (53%) | <.01 |
| Age (y) | 26.5 (19.3–33.8) | 25.1 (13.5–31.7) | NS |
| BMI (kg/m 2 ) | 26.7 (24.6–31.2) | 23.4 (19.7–26.3) | .0001 |
| BSA (m 2 ) | 2.02 (1.91–2.21) | 1.73 (1.55–1.97) | <.0001 |
| SBP (mm Hg) | 118 (110–123) | 120 (113–132) | NS |
| DBP (mm Hg) | 71 (65–80) | 71 (60–76) | NS |
| Diabetes | 0% | — | — |
| Hypertension | 0% | — | — |
| History of smoking | 27% | — | — |
| Dyslipidemia | 7% | — | — |
| Family history of IHD | 7% | — | — |
| Peak troponin I (μg/L) | 12 (4–21) | — | — |
| Interval peak troponin to CMR (d) | 3 (1–7) | — | — |
| Primary ECG abnormality | |||
| Normal | 30% | — | |
| ST-segment elevation | 38% | — | — |
| ST-segment depression | 3% | — | — |
| T-wave changes | 22% | — | — |
| RBBB | 5% | — | — |
| MRI myocarditis features | |||
| T2 edema | 70% | — | — |
| LGE | 100% | ||
| Subepicardium | 83% | — | — |
| Midwall | 33% | — | — |
| Location of LGE | |||
| Anteroseptal | 0% | — | — |
| Septal | 17% | — | — |
| Inferior | 27% | — | — |
| Inferolateral | 50% | — | — |
| Lateral | 60% | — | — |
| Anterior | 10% | — | — |
| Apex | 7% | — | — |
| Angiography | 37% | — | — |
Conventional LV Parameters
The conventional echocardiographic measure of LV size normalized to body surface area, LVEF, isovolumic relaxation time, LV inflow Doppler velocities and ratio, and Doppler tissue imaging s′ and a′ were not different between patients with myocarditis and controls. However, both Doppler tissue imaging septal and lateral e′ velocities were reduced, with concurrent increases in E/e′ ratios among patients with myocarditis ( Table 2 ).
| Variable | Patients with myocarditis ( n = 28) | Controls ( n = 64) | P |
|---|---|---|---|
| LVEF (%) | 59 (53–64) | 58 (55–62) | NS |
| LVEDV/BSA (mL/m 2 ) | 58 (50–69) | 62 (55–71) | NS |
| LVESV/BSA (mL/m 2 ) | 25 (19–29) | 26 (23–30) | NS |
| LV length (cm) | 8.8 (8.2–9.6) | 7.9 (7.3–8.5) | <.0001 |
| E wave (m/sec) | 0.73 (0.63–0.91) | 0.78 (0.70–0.86) | NS |
| A wave (m/sec) | 0.43 (0.38–0.52) | 0.41 (0.36–0.49) | NS |
| E/A ratio | 1.68 (1.40–2.00) | 1.84 (1.52–2.24) | NS |
| IVRT (msec) | 70 (62–83) | 70 (60–80) | NS |
| Septal e′ (cm/sec) | 9.1 (8.0–10.9) | 11.2 (9.9–12.6) | <.0001 |
| Septal E/e′ ratio | 8.4 (6.9–9.6) | 6.9 (5.8–8.0) | .002 |
| Septal a′ (cm/sec) | 6.0 (5.1–7.1) | 5.4 (4.8–6.0) | NS |
| Septal s′ (cm/sec) | 7.0 (6.3–7.8) | 6.9 (6.4–7.4) | NS |
| Lateral e′ (cm/sec) | 11.3 (9.5–12.9) | 13.6 (12.1–15.2) | <.0001 |
| Lateral E/e′ ratio | 6.9 (6.0–8.0) | 5.7 (4.9–6.6) | .0004 |
| Lateral a′ (cm/sec) | 4.7 (3.6–6.1) | 4.8 (3.8–5.3) | NS |
| Lateral s′ (cm/sec) | 7.9 (6.6–8.5) | 7.2 (6.3–8.1) | NS |
LV Twist and Tissue Deformation Parameters
Peak LV twist and untwisting rate was unchanged. Peak twisting rate was increased ( P = .001), with an earlier time to peak twist ( P < .0001), and LV untwist performance at mitral valve opening was greater ( P = .004) in patients with myocarditis. In early diastole at 130% of systole, untwist performance was no longer different ( Table 3 ). Basal circumferential and 4CH longitudinal strain, systolic strain rate, and early diastolic strain rate were reduced in patients with myocarditis. No differences were observed in apical circumferential strain parameters ( Table 4 ).
| Variable | Patients with myocarditis ( n = 28) | Controls ( n = 64) | P |
|---|---|---|---|
| Peak twist (°) | 12.6 (8.0 to 15.9) | 12.6 (9.4 to 16.1) | NS |
| Time to peak twist (%) | 95 (90 to 100) | 105 (102 to 108) | <.0001 |
| Peak twist rate (°/sec) | 90 (74 to 117) | 69 (52 to 92) | .0014 |
| Time to peak twist rate (%) | 59 (45 to 69) | 63 (58 to 73) | NS |
| Untwist rate (°/sec) | −95 (−116 to −60) | −87 (−114 to −69) | NS |
| Time to peak untwist rate (%) | 124 (114 to 141) | 128 (123 to 133) | NS |
| Untwist performance at MVO (%) | 22 (5 to 34) | 4 (−1 to 17) | .0044 |
| Untwist performance at 130% systole (%) | 44 (18 to 55) | 45 (32 to 63) | NS |
| Variable | Patients with myocarditis ( n = 28) | Controls ( n = 64) | P |
|---|---|---|---|
| Basal circumferential strain (%) | −15.1 (−18.3 to −11.3) | −18.9 (−20.3 to −15.9) | .0004 |
| Basal circumferential systolic strain rate (sec −1 ) | −0.97 (−1.18 to −0.83) | −1.16 (−1.26 to −1.03) | .0069 |
| Basal circumferential early diastolic strain rate (sec −1 ) | 1.18 (0.86 to 1.47) | 1.70 (1.40 to 2.05) | <.0001 |
| Apical circumferential strain (%) | −23.1 (−27.1 to −19.7) | −24.9 (−27.0 to −22.2) | NS |
| Apical circumferential systolic strain rate (sec −1 ) | −1.38 (−1.78 to −1.28) | −1.48 (−1.70 to −1.32) | NS |
| Apical circumferential early diastolic strain rate (sec −1 ) | 2.29 (1.84 to 2.88) | 2.49 (2.16 to 2.99) | NS |
| 4CH longitudinal strain (%) | −18.1 (−19.4 to −16.5) | −20.8 (−22.0 to −19.7) | <.0001 |
| 4CH longitudinal systolic strain rate (sec −1 ) | −1.01 (−1.09 to −0.92) | −1.11 (−1.23 to −1.04) | .0003 |
| 4CH longitudinal early diastolic strain rate (sec −1 ) | 1.45 (1.17 to 1.57) | 1.80 (1.52 to 2.04) | <.0001 |
Univariate and Multivariate Regression Analysis
Conventional LV Parameters
On univariate analysis, increase in LV end-diastolic volume indexed to body surface area, reduction in septal and lateral e′ velocities, and increased E/e′ ratio were found to be associated with a CMR diagnosis of myocarditis in this cohort. Changes in the newer tissue deformation and twist parameters, including earlier time to peak twist, increased peak twist rate, impaired basal circumferential and 4CH longitudinal strain, systolic strain rate, and early diastolic strain rate, were also found to be associated with a CMR diagnosis of myocarditis ( Table 5 ). On stepwise multivariate regression analysis, only reduced lateral e′, earlier time to peak twist, increased peak twist rate, basal circumferential early diastolic strain rate, and 4CH longitudinal strain were identified as independent predictors of a CMR diagnosis of myocarditis ( Table 6 ). The regression model was reanalyzed with the exclusion of patients with body mass indexes > 30 kg/m 2 and yielded the same independent predictors and odds ratios.
| Variable | Odds ratio | 95% confidence interval | P |
|---|---|---|---|
| Time to peak twist (%) ∗ | 0.80 | 0.72–0.88 | <.001 |
| Peak twist rate (°/sec) ∗ | 1.02 | 1.01–1.04 | .003 |
| Basal circumferential strain (%) ∗ | 1.31 | 1.13–1.51 | <.001 |
| Basal circumferential systolic strain rate (s −1 ) ∗ | 21.3 | 2.16–210 | .009 |
| Basal circumferential systolic strain rate (0.1 s −1 ) | 1.36 | 1.08–1.71 | .009 |
| Basal circumferential early diastolic strain rate (s −1 ) ∗ | 0.03 | 0.01–0.15 | <.001 |
| Basal circumferential early diastolic strain rate (0.1 s −1 ) | 0.71 | 0.60–0.83 | <.001 |
| 4CH longitudinal strain (%) ∗ | 1.81 | 1.38–2.38 | <.001 |
| 4CH longitudinal systolic strain rate (s −1 ) ∗ | 500 | 12.65–999 | <.001 |
| 4CH longitudinal systolic strain rate (0.1 s −1 ) | 1.86 | 1.29–2.69 | <.001 |
| 4CH longitudinal early diastolic strain rate (s −1 ) ∗ | 0.05 | 0.01–0.23 | <.001 |
| 4CH longitudinal early diastolic strain rate (0.1 s −1 ) | 0.74 | 0.63–0.83 | <.001 |
| LVEDV/BSA (mL/m 2 ) | 1.02 | 1.00–1.04 | .02 |
| Septal annular E′ (cm/sec) ∗ | 1.80 | 1.33–2.43 | <.001 |
| Septal annular E/E′ ratio ∗ | 1.58 | 1.19–2.10 | .002 |
| Lateral annular E′ (cm/sec) ∗ | 1.77 | 1.34–2.34 | <.001 |
| Lateral annular E/E′ ratio ∗ | 2.09 | 1.39–3.14 | <.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
