Abstract
To determine whether the extent of myocardial fibrosis by late-gadolinium enhancement cardiovascular magnetic resonance (LGE-CMR), and echocardiographic ventricular dyssynchrony are independently associated with response to medical therapy in patients with newly diagnosed idiopathic dilated cardiomyopathy (DCM). Myocardial fibrosis and ventricular dyssynchrony are frequent findings in DCM. Previous studies focused on patients with established cardiomyopathy; however, the degree of myocardial fibrosis and ventricular dyssynchrony at presentation and their role in perpetuating left ventricular (LV) dysfunction in DCM remains unclear. Those studies of individuals with long-standing DCM did not characterize patients early in the disease course, and may not have included those with significant improvement in LV function. Thus factors contributing to LV recovery are uncertain.
Consecutive patients with a new diagnosis of DCM [LV ejection fraction (EF) ≤45%] made within the preceding 2 weeks were recruited. Patients underwent LGE-CMR, echocardiography, 6-minute walk testing, cardiopulmonary exercise testing, and blood sampling for measurement of serum amino-terminal pro-brain natiuretic peptide (NT-pro-BNP) concentration at baseline. Baseline patient characteristics were compared with a cohort of healthy volunteers. Myocardial fibrosis by LGE-CMR was quantified, identified by experienced observers blinded to patient outcome. Left ventricular systolic function was reassessed after 5 months of optimal medical therapy. Sixty-eight patients with DCM and 19 healthy volunteers were studied. DCM patients were studied a median 12.5 days following diagnosis. Compared with healthy controls, DCM patients exhibited greater inter- and intra-ventricular dyssynchrony. Twenty-four per cent of DCM patients exhibited LGE at diagnosis. Among DCM patients with LGE, the mean fibrosis mass was 2.2 ± 1.3 g. On multivariate analysis, strain dyssynchrony index, and fibrosis mass were independently associated with change in the LVEF over time (P≤ 0.001). Late-gadolinium enhancement cardiovascular magnetic resonance conferred additive value for modelling change in the LVEF beyond clinical and echocardiographic dyssynchrony parameters.
The extent of myocardial fibrosis is independently associated with lack of response to medical therapy in new-presentation DCM, and LGE-CMR may thus be an important risk-stratifying investigation in these patients. Accurate risk stratification may permit more targeted pharmacological and device therapies for patients with newly diagnosed DCM.
Introduction
Idiopathic dilated cardiomyopathy (DCM) is a condition characterized by impairment in systolic left ventricular (LV) function following a poorly defined myocardial insult. A hallmark of DCM is myocardial fibrosis,1 whose presence [as detected by late-gadolinium enhancement cardiovascular magnetic resonance (LGE-CMR)] portends an adverse prognosis.2,3 Additionally, dyssynchronous ventricular contraction is a frequent finding in DCM patients4 whose presence has been associated with adverse cardiovascular events.5
Despite this emerging evidence, reports of the DCM disease course vary widely, with respect to patient outcomes and prognostic factors.2,3,5,6 Previous studies focus on patients with established cardiomyopathy; however, the extent of myocardial fibrosis at presentation of DCM and its association with the subsequent response to medical therapy is unknown. Furthermore, the prevalence of ventricular dyssynchrony at presentation in DCM is uncertain and its correlation with adverse LV remodelling remains unclear.
The LGE-CMR imaging technique has facilitated a paradigm shift in the evaluation of the failing heart through its ability to detect myocardial fibrosis.7 Novel echocardiographic analyses, using Doppler and speckle tracking, are able to quantify interventricular and intra-LV mechanical dyssynchrony. Using these techniques, we first sought to characterize myocardial fibrosis and ventricular dyssynchrony in patients with newly diagnosed DCM by comparison with healthy volunteers. Secondly, we determined whether the extent of myocardial fibrosis and ventricular dyssynchrony at first presentation is independently associated with the response to conventional medical therapy in these patients. Our primary hypotheses were that the degree of myocardial fibrosis as detected by LGE-CMR, and ventricular dyssynchrony by echocardiography would distinguish newly diagnosed DCM patients who are less likely to exhibit LV functional recovery despite optimal medical therapy.
Methods
Subject characteristics
Consecutive patients with a new diagnosis of DCM made within the preceding 2 weeks were recruited from three university hospitals from April 2008 to December 2009. The inclusion criterion was the presence of the LV ejection fraction (EF) ≤45% at baseline echocardiography or CMR. Exclusion criteria included the diagnosis of significant coronary artery disease (defined as the presence of >70% luminal stenosis in an epicardial coronary artery at angiography, non-invasive stress imaging suggestive of ischaemia, or prior myocardial infarction), severe valvular heart disease, thyroid dysfunction, infiltrative cardiomyopathy, or extra-cardiac systemic features to suggest sarcoidosis or amyloidosis, chemotherapy-induced cardiomyopathy, hypertrophic cardiomyopathy, and myocarditis. Myocarditis was excluded in potential DCM cases by the absence of classic clinical features, the presence of normal serum troponin I concentration at presentation, and by the lack of evidence of myocardial oedema on T2-weighted CMR.8 Exclusion from the CMR arm of the study was mandated by renal impairment (eGFR < 60 mL/min), or other conventional CMR contra-indications. Control subjects were recruited from hospital staff, who prior to enrolment were screened to ensure no history of cardiac disease, hypertension, or diabetes mellitus. All subjects provided written, informed consent to the study protocol, which complies with the Declaration of Helsinki and was approved by the relevant Human Research Ethics Committees.
Treatment of heart failure
Heart failure therapies were administered in accordance with current guidelines,9 using angiotensin-converting enzyme inhibitors or angiotensin receptor blockers (ACE-I/ARB), beta-blockers, and, when clinically indicated, aldosterone antagonists. These therapies were initiated and titrated to maximum tolerated doses by the treating physician.
Study protocol
Patients were studied within 2 weeks of presentation with new-onset DCM when clinically stable. Patients diagnosed as an outpatient were studied as soon as possible following diagnosis, but within 2 weeks following presentation. Any patient ineligible for CMR but consenting to the remainder of the protocol underwent these other study investigations. The study protocol consisted of the following baseline investigations: blood sampling, LGE-CMR, echocardiography, 6-minute walk testing, Minnesota Living with Heart Failure Questionnaire, and cardiopulmonary exercise testing. To minimize variability due to loading conditions, CMR and echocardiography were performed in immediate succession for each individual. Patients underwent repeat echocardiography, 6-minute walk testing, Minnesota Living with Heart Failure Questionnaire, and cardiopulmonary exercise testing a median 5 months [inter-quartile range (IQR) 4–7 months] later. Follow-up echocardiography was performed at the same time of day as baseline imaging.
Cardiac magnetic resonance protocol
Cardiac magnetic resonance was performed using commercially available 1.5T machines (Siemens Avanto, Erlangen, Germany or Philips Intera, Best, The Netherlands). Electrocardiographically gated steady-state free precession imaging of the left ventricle in the short-axis plane was undertaken (TE/TR, 1.5/3.0 ms; flip angle 60°) with slice thickness of 6 and inter-slice gap of 4 mm. A breath-hold T2-weighted, triple inversion recovery sequence was applied in three short-axis slices (basal, midventricular, and apical) and in three long-axis views using the body coil, to exclude myocardial oedema (TR = 2 RR, TE 65 ms, TI 140 ms, slice thickness 8 mm, matrix 256 × 256). T1-weighted contrast-enhanced imaging was then undertaken 10 min following intravenous administration of gadolinium-DTPA 0.1 mmol/kg using an inversion-recovery segmented gradient echo sequence. Inversion times were adjusted to null normal myocardium (260 400 ms) with voxel sizes of 1.9 × 1.4 × 7.0 mm. To exclude artefact, cross-cut imaging and imaging with another phase-encoding direction were performed.
Cardiac magnetic resonance data analysis
Left ventricular mass was measured as previously described,10 using commercially available software (CAAS MRV Version 3.3, PIE, Netherlands). The presence of myocardial oedema was excluded by visual inspection of the T2-weighted images. The presence and distribution of LGE in a 16-segment model were determined by two experienced, independent observers blinded to patient outcomes. Each segment was awarded 0 for LGE absence and 1 for LGE presence, yielding a fibrosis score out of 16. Regions in which fibrosis was identified by LGE were manually planimetered to measure the fibrosis mass, which was indexed to (LV mass × 100). Cases of disagreement were adjudicated by a third blinded observer.
Echocardiography protocol
Imaging was performed with patients in the left lateral decubitus position using a 3.5 MHz transducer (Vivid 7, General Electric-Vingmed, Milwaukee, WI, USA). All grey-scale images were acquired with frame rates exceeding 70 fps. Colour-coded tissue Doppler imaging (TDI) of the left ventricle was performed in the three apical views. Pulse wave Doppler recordings were taken with the sample volume in the right ventricular outflow tract in the parasternal short-axis view, and in the LV outflow tract in the apical five-chamber view. Transmitral blood flow velocities (E- and A-wave) were measured by pulse wave Doppler at the mitral valve leaflet tips. The left atrium was imaged in orthogonal planes for measurement of left atrial (LA) volume using the modified Simpson's rule. Tricuspid annular plane systolic excursion (TAPSE) was measured by M-mode echocardiography of the lateral tricuspid annulus. A minimum of three consecutive cardiac cycles was recorded for each image. Data were stored digitally for the subsequent analysis.
Echocardiographic analysis
Data were analysed using commercially available off-line software (EchoPac version 7.0.0 General Electric-Vingmed).
Left ventricle
The following analysis was performed of LV structure and function: (i) LV end-diastolic and end-systolic volumes were measured from apical four- and two-chamber images using Simpson's biplane method of discs, and indexed to body surface area. (ii) LV strain profiles were derived by speckle-tracking of grey-scale apical images as previously described, to determine the global LV longitudinal strain score (GLS, Figure 1E and F).11 (iii) E’ was measured off-line using colour-coded TDI as the peak septal mitral annular tissue velocity in early diastole. (iv) Intra-LV dyssynchrony was assessed off-line by speckle tracking as previously described.12 A strain dyssynchrony index (SDI) was defined as the standard deviation of values of time from QRS onset to peak longitudinal systolic strain on an 18-segment model as measured using speckle-tracking strain.12 (v) Intra-LV dyssynchrony was also assessed off-line using colour-coded TDI. A tissue Doppler dyssynchrony index (TD DI) was defined as the standard deviation of time from QRS onset to peak systolic tissue velocity during LV ejection on a 12-segment model (Figure 1C and D).13 (v) Using pulse wave Doppler measurements in the right ventricular and LV outflow tracts, interventricular mechanical delay (IVMD) was calculated as the difference in the time-to-onset of ejection from the right and left ventricles.
(A and B) Horizontal long-axis and short-axis T1-weighted late-gadolinium enhancement cardiovascular magnetic resonance images, respectively, both acquired from an individual 7 days after presentation with new onset idiopathic dilated cardiomyopathy. These demonstrate mid-wall septal enhancement (solid arrows). (C) Baseline colour-coded tissue Doppler imaging of the same individual in the apical four-chamber view: peak systolic ejection tissue velocity occurs simultaneously in the basal septal and basal lateral segments (dashed line). (D) Follow-up colour-coded TDI of the same individual in the apical four-chamber view: difference in time-to-peak systolic tissue velocity in the basal septal and basal lateral segments is shown to have increased from baseline (dashed arrows). (E) Baseline speckle-tracking strain imaging of the same individual in the apical four-chamber view (dotted line represents the mean four-chamber longitudinal strain). (F) Follow-up speckle-tracking strain imaging of the same individual in the apical four-chamber view demonstrating worsening in longitudinal strain from −7 to −6%.
Exercise testing protocol
Cardiopulmonary exercise testing was conducted on a stationary cycle ergometer to the measure peak oxygen uptake (VO2 PEAK) according to current recommendations.14 VO2 PEAK has previously been shown to be an important predictor of outcome in non-ischaemic systolic heart failure patients.15 The six-minute walk test was also performed to assess its prognostic value for the LVEF, given its recognized ability to predict heart failure hospitalization and mortality.16 Both these indices of functional capacity were evaluated for their ability at baseline to predict improvement in the LVEF.
Blood samples
Venous blood specimens were collected, immediately centrifuged for 10 min at 1000 g, and the serum aliquoted and stored at −80°C for batch analysis of amino-terminal pro-brain natiuretic peptide (NT-pro-BNP), whose concentration has recently been shown to have prognostic value in stable heart failure patients.17
Statistical analysis
Data from controls were compared with patients by Student's t-test for normally distributed continuous variables, the Wilcoxon–Mann–Whitney test for non-normally distributed continuous variables, and the χ2 test for categorical variables. We performed two multivariate regression analyses. Firstly, change in continuous variables from baseline to follow-up and the association of baseline covariates with changes in the LVEF from baseline to follow-up were evaluated by mixed effects modelling. For univariate analysis, the covariate of interest and subject visit and the covariate * visit interaction were modelled as fixed effects, with subject identity as a random effect to account for repeated measures within patients. For significant covariate * interaction terms, post hoc testing was performed at baseline and follow-up visits. For the multivariate model, any covariate with significant visit-dependent association with LVEF on univariate analysis, and any covariate whose main effect on the LVEF independent of visit exhibited a P-value < 0.2 was included. Backward elimination was used to identify factors independently associated with change in the LVEF.
In the second analysis, we determined whether incremental value for modelling change in the LVEF was conferred by the stepwise inclusion of echocardiographic markers of ventricular dyssynchrony and LGE-CMR to evaluation of DCM patients by routinely acquired information (age, QRS duration, NT-pro-BNP concentration), using the likelihood-ratio test.
Finally, the relationship between the presence of LGE and change in echocardiographic dyssynchrony indices from baseline to follow-up was examined by mixed effects modelling.
All statistical tests were two-sided and a P-value < 0.05 was considered significant. Statistical analysis was performed with STATA 11 (Stata Corp, College Station, TX, USA).
Inter-observer variability
Inter-observer agreement was assessed in 20 randomly selected cases, in which a second blinded observer measured strain and dyssynchrony parameters. Inter-observer agreement was evaluated by the method of Bland and Altman. In our laboratory the inter-observer Spearman correlation coefficient for the 16-segment quantification of LGE is 0.81 (P< 0.001).
Results
Eighty-two patients presenting with new-onset heart failure were screened for enrolment, of whom 14 were ineligible. The study cohort thus consisted of 68 patients, who were compared with 19 healthy volunteers (Figure 2). All of the 68 DCM patients underwent echocardiographic evaluation at both time points; however, 17 were unable to undergo baseline CMR. Reasons for patient ineligibility for CMR are illustrated in Figure 2. At follow-up, one DCM patient was intolerant of ACE-I/ARB due to hypotension. Three patients were intolerant of beta-blockers due to hypotension or bradycardia. All other patients were treated with both ACE-I/ARB (99% of cases) and beta-blockers (96% of cases) at maximum tolerated doses. In addition, at follow-up, 47 (69%) of patients received frusemide and 24 (35%) of patients received spironolactone.
Flow diagram illustrating the numbers of participants undergoing cardiovascular magnetic resonance during the study.
Baseline characterization of idiopathic dilated cardiomyopathy patients
Initial patient evaluation was performed a median of 12.5 days (IQR: 7–23 days) following their first presentation with symptoms of heart failure. The median duration of patient symptoms prior to presentation was 9 weeks (IQR: 2–26 weeks). Nine (13%) of the 68 DCM patients had AF at presentation.
There was no difference in the baseline age (59 ± 14 vs. 54 ± 7 years, P= 0.2) and gender between patients and controls. On CMR analysis, DCM patients displayed greater LV mass index (81 ± 22 vs. 49 ± 9 g/m2, P< 0.001). Twelve (24%) DCM patients exhibited LGE by CMR, whereas LGE was absent in all controls. The mean fibrosis score among DCM patients was 0.57 ± 1.3 and the mean fibrosis mass was 0.54 ± 1.1 g. Among patients with LGE, the mean fibrosis mass was 2.2 ± 1.3 g, and fibrosis mass indexed to LV mass was 1.4 ± 0.74%. Late-gadolinium enhancement was typically found as mid-wall striae or as foci at the junction points of the right ventricular free wall and the interventricular septum. No DCM patient exhibited LGE in a subendocardial or transmural distribution.
Echocardiographic analysis revealed significantly lower the LVEF (29 ± 8 vs. 67 ± 6%, P< 0.001) and less negative global longitudinal strain (GLS, −12 ± 2 vs. −21 ± 2%, P< 0.001) among DCM patients compared with controls, and greater indexed LV end-systolic volume (60 ± 22 vs. 16 ± 4 mL/m2, P< 0.001). DCM patients also demonstrated worse diastolic grade (2.3 ± 1.2 vs. 0.3 ± 0.5, P< 0.001) and adverse LA structural remodelling (LA volume index 48 ± 14 vs. 30 ± 8 mL/m2, P< 0.001). Assessment of ventricular synchrony displayed significantly greater inter- (IVMD: 31 ± 24 vs. 17 ± 12 ms, P= 0.03) and intra-ventricular (SDI: 79 ± 33 vs. 38 ± 14 ms and TD DI: 41 ± 14 vs. 25 ± 14 ms, P< 0.001) dyssynchrony than controls.
Response to heart failure therapy
Patient follow-up was performed a median of 5 months (IQR: 4–7 months) after first presentation. Changes over follow-up are displayed in Table 1. There was significant improvement in left and right ventricular function with favourable LV and LA remodelling (P≤ 0.001). Six-minute walk distance improved over follow-up (P< 0.001); however, despite these functional and structural changes, indices of ventricular dyssynchrony and VO2 PEAK failed to improve significantly. Minnesota Living with Heart Failure questionnaire scores significantly lessened with treatment (P< 0.001).
Comparison of baseline and follow-up parameters in DCM patients
| Characteristic (± SD) | Baseline | Follow-up | P-value |
|---|---|---|---|
| Minnesota living with heart failure score | 44 ± 24 | 29 ± 20 | <0.001 |
| QRS duration (ms) | 114 ± 29 | 115 ± 32 | 0.4 |
| Six-minute walk distance (m) | 420 ± 96 | 472 ± 95 | <0.001 |
| VO2PEAK (mL/kg/min) | 20 ± 6 | 22 ± 6 | 0.09 |
| Median N-terminal pro-brain natiuretic peptide (inter-quartile range, ng/L) | 1183 (720–3171) | 586 (276–1078) | 0.02 |
| Echocardiogram | |||
| LVESVI (mL/m2) | 60 ± 22 | 44 ± 22 | <0.001 |
| LVEDVI (mL/m2) | 83 ± 24 | 75 ± 21 | 0.04 |
| LV ejection fraction (%) | 29 ± 8 | 43 ± 13 | <0.001 |
| Global longitudinal strain score (%) | −12 ± 2 | −16 ± 4 | <0.001 |
| Diastolic grade | 2.3 ± 1.2 | 1.3 ± 0.9 | <0.001 |
| E/A | 1.5 ± 1.1 | 1.0 ± 0.7 | 0.02 |
| E/E’ | 25 ± 15 | 20 ± 17 | 0.1 |
| LAVI (mL/m2) | 48 ± 14 | 35 ± 11 | <0.001 |
| TAPSE (cm) | 1.8 ± 0.5 | 2.1 ± 0.4 | 0.004 |
| Strain dyssynchrony index (ms) | 79 ± 33 | 79 ± 28 | 0.5 |
| Tissue Doppler dyssynchrony index (ms) | 41 ± 14 | 38 ± 19 | 0.07 |
| Interventricular mechanical delay (ms) | 31 ± 24 | 39 ± 34 | 0.3 |
| Characteristic (± SD) | Baseline | Follow-up | P-value |
|---|---|---|---|
| Minnesota living with heart failure score | 44 ± 24 | 29 ± 20 | <0.001 |
| QRS duration (ms) | 114 ± 29 | 115 ± 32 | 0.4 |
| Six-minute walk distance (m) | 420 ± 96 | 472 ± 95 | <0.001 |
| VO2PEAK (mL/kg/min) | 20 ± 6 | 22 ± 6 | 0.09 |
| Median N-terminal pro-brain natiuretic peptide (inter-quartile range, ng/L) | 1183 (720–3171) | 586 (276–1078) | 0.02 |
| Echocardiogram | |||
| LVESVI (mL/m2) | 60 ± 22 | 44 ± 22 | <0.001 |
| LVEDVI (mL/m2) | 83 ± 24 | 75 ± 21 | 0.04 |
| LV ejection fraction (%) | 29 ± 8 | 43 ± 13 | <0.001 |
| Global longitudinal strain score (%) | −12 ± 2 | −16 ± 4 | <0.001 |
| Diastolic grade | 2.3 ± 1.2 | 1.3 ± 0.9 | <0.001 |
| E/A | 1.5 ± 1.1 | 1.0 ± 0.7 | 0.02 |
| E/E’ | 25 ± 15 | 20 ± 17 | 0.1 |
| LAVI (mL/m2) | 48 ± 14 | 35 ± 11 | <0.001 |
| TAPSE (cm) | 1.8 ± 0.5 | 2.1 ± 0.4 | 0.004 |
| Strain dyssynchrony index (ms) | 79 ± 33 | 79 ± 28 | 0.5 |
| Tissue Doppler dyssynchrony index (ms) | 41 ± 14 | 38 ± 19 | 0.07 |
| Interventricular mechanical delay (ms) | 31 ± 24 | 39 ± 34 | 0.3 |
LAVI, left atrial volume index; LVESVI, left ventricular end-systolic volume index; LVEDVI, left ventricular end-diastolic volume index; TAPSE, tricsupid annular plane systolic excursion.
Comparison of baseline and follow-up parameters in DCM patients
| Characteristic (± SD) | Baseline | Follow-up | P-value |
|---|---|---|---|
| Minnesota living with heart failure score | 44 ± 24 | 29 ± 20 | <0.001 |
| QRS duration (ms) | 114 ± 29 | 115 ± 32 | 0.4 |
| Six-minute walk distance (m) | 420 ± 96 | 472 ± 95 | <0.001 |
| VO2PEAK (mL/kg/min) | 20 ± 6 | 22 ± 6 | 0.09 |
| Median N-terminal pro-brain natiuretic peptide (inter-quartile range, ng/L) | 1183 (720–3171) | 586 (276–1078) | 0.02 |
| Echocardiogram | |||
| LVESVI (mL/m2) | 60 ± 22 | 44 ± 22 | <0.001 |
| LVEDVI (mL/m2) | 83 ± 24 | 75 ± 21 | 0.04 |
| LV ejection fraction (%) | 29 ± 8 | 43 ± 13 | <0.001 |
| Global longitudinal strain score (%) | −12 ± 2 | −16 ± 4 | <0.001 |
| Diastolic grade | 2.3 ± 1.2 | 1.3 ± 0.9 | <0.001 |
| E/A | 1.5 ± 1.1 | 1.0 ± 0.7 | 0.02 |
| E/E’ | 25 ± 15 | 20 ± 17 | 0.1 |
| LAVI (mL/m2) | 48 ± 14 | 35 ± 11 | <0.001 |
| TAPSE (cm) | 1.8 ± 0.5 | 2.1 ± 0.4 | 0.004 |
| Strain dyssynchrony index (ms) | 79 ± 33 | 79 ± 28 | 0.5 |
| Tissue Doppler dyssynchrony index (ms) | 41 ± 14 | 38 ± 19 | 0.07 |
| Interventricular mechanical delay (ms) | 31 ± 24 | 39 ± 34 | 0.3 |
| Characteristic (± SD) | Baseline | Follow-up | P-value |
|---|---|---|---|
| Minnesota living with heart failure score | 44 ± 24 | 29 ± 20 | <0.001 |
| QRS duration (ms) | 114 ± 29 | 115 ± 32 | 0.4 |
| Six-minute walk distance (m) | 420 ± 96 | 472 ± 95 | <0.001 |
| VO2PEAK (mL/kg/min) | 20 ± 6 | 22 ± 6 | 0.09 |
| Median N-terminal pro-brain natiuretic peptide (inter-quartile range, ng/L) | 1183 (720–3171) | 586 (276–1078) | 0.02 |
| Echocardiogram | |||
| LVESVI (mL/m2) | 60 ± 22 | 44 ± 22 | <0.001 |
| LVEDVI (mL/m2) | 83 ± 24 | 75 ± 21 | 0.04 |
| LV ejection fraction (%) | 29 ± 8 | 43 ± 13 | <0.001 |
| Global longitudinal strain score (%) | −12 ± 2 | −16 ± 4 | <0.001 |
| Diastolic grade | 2.3 ± 1.2 | 1.3 ± 0.9 | <0.001 |
| E/A | 1.5 ± 1.1 | 1.0 ± 0.7 | 0.02 |
| E/E’ | 25 ± 15 | 20 ± 17 | 0.1 |
| LAVI (mL/m2) | 48 ± 14 | 35 ± 11 | <0.001 |
| TAPSE (cm) | 1.8 ± 0.5 | 2.1 ± 0.4 | 0.004 |
| Strain dyssynchrony index (ms) | 79 ± 33 | 79 ± 28 | 0.5 |
| Tissue Doppler dyssynchrony index (ms) | 41 ± 14 | 38 ± 19 | 0.07 |
| Interventricular mechanical delay (ms) | 31 ± 24 | 39 ± 34 | 0.3 |
LAVI, left atrial volume index; LVESVI, left ventricular end-systolic volume index; LVEDVI, left ventricular end-diastolic volume index; TAPSE, tricsupid annular plane systolic excursion.
Correlates of response to medical therapy
We evaluated the following parameters to determine the association of their baseline value with subsequent improvement in the LV EF: age, gender, QRS duration, 6-minute walk distance, VO2 PEAK, markers of intra-LV (TD DI and SDI) and interventricular (IVMD) dyssynchrony, E/E’, LA volume, NT-pro-BNP concentration, and fibrosis mass by LGE-CMR. Of these, QRS duration, NT-pro-BNP concentration, SDI, and fibrosis mass at baseline were significant univariate correlates of improvement in the LVEF (Table 2). On multivariate analysis, SDI, and fibrosis mass were independently associated with change in the LVEF over time (P≤ 0.001) (Table 2). The influence of the presence of LGE on the LVEF and the GLS is displayed in Figure 3.
Univariate and multivariate correlates of change in the left ventricular ejection fraction following therapy
| Variable | Univariate model | Multivariate model | ||
|---|---|---|---|---|
| Coefficient (95% CI) | P-value | Coefficient (95% CI) | P-value | |
| Age | −0.06 (−0.3 to 0.2) | 0.8 | ||
| Gender | −4 (−10 to 3) | 0.8 | ||
| 6-minute walk distance (m) | 0.03 (−0.01 to 0.08) | 0.3 | ||
| VO2PEAK (mL/kg/min) | 0.3 (−0.7 to 1.3) | 0.6 | ||
| Tissue Doppler dyssynchrony index (ms) | −0.07 (−0.3 to 0.2) | 0.7 | ||
| E/E’ | 0.04 (−0.2 to 0.3) | 0.8 | ||
| LA volume index (mL/m2) | −0.07 (−0.3 to 0.2) | 0.8 | ||
| TAPSE (cm) | −1 (−10 to 7) | 0.7 | ||
| Interventricular mechanical delay (ms) | −0.1 (−0.3 to 0.05) | 0.3 | 0.3 (−0.06 to 0.6) | 0.1 |
| NT-pro BNP (μg/L) | −1 (−2 to −0.09) | 0.007 | −0.2 (−2 to 1) | 0.8 |
| QRS duration (ms) | −0.14 (−0.3 to −0.02) | 0.004 | −0.3 (−0.7 to 0.1) | 0.2 |
| Strain dyssynchrony index (ms) | −0.1 (−0.2 to −0.06) | 0.04 | −0.1 (−0.2 to −0.04) | <0.001 |
| Fibrosis mass (g) | −7 (−11 to −3) | <0.001 | −7 (−11 to −4) | <0.001 |
| Variable | Univariate model | Multivariate model | ||
|---|---|---|---|---|
| Coefficient (95% CI) | P-value | Coefficient (95% CI) | P-value | |
| Age | −0.06 (−0.3 to 0.2) | 0.8 | ||
| Gender | −4 (−10 to 3) | 0.8 | ||
| 6-minute walk distance (m) | 0.03 (−0.01 to 0.08) | 0.3 | ||
| VO2PEAK (mL/kg/min) | 0.3 (−0.7 to 1.3) | 0.6 | ||
| Tissue Doppler dyssynchrony index (ms) | −0.07 (−0.3 to 0.2) | 0.7 | ||
| E/E’ | 0.04 (−0.2 to 0.3) | 0.8 | ||
| LA volume index (mL/m2) | −0.07 (−0.3 to 0.2) | 0.8 | ||
| TAPSE (cm) | −1 (−10 to 7) | 0.7 | ||
| Interventricular mechanical delay (ms) | −0.1 (−0.3 to 0.05) | 0.3 | 0.3 (−0.06 to 0.6) | 0.1 |
| NT-pro BNP (μg/L) | −1 (−2 to −0.09) | 0.007 | −0.2 (−2 to 1) | 0.8 |
| QRS duration (ms) | −0.14 (−0.3 to −0.02) | 0.004 | −0.3 (−0.7 to 0.1) | 0.2 |
| Strain dyssynchrony index (ms) | −0.1 (−0.2 to −0.06) | 0.04 | −0.1 (−0.2 to −0.04) | <0.001 |
| Fibrosis mass (g) | −7 (−11 to −3) | <0.001 | −7 (−11 to −4) | <0.001 |
Univariate and multivariate correlates of change in the left ventricular ejection fraction following therapy
| Variable | Univariate model | Multivariate model | ||
|---|---|---|---|---|
| Coefficient (95% CI) | P-value | Coefficient (95% CI) | P-value | |
| Age | −0.06 (−0.3 to 0.2) | 0.8 | ||
| Gender | −4 (−10 to 3) | 0.8 | ||
| 6-minute walk distance (m) | 0.03 (−0.01 to 0.08) | 0.3 | ||
| VO2PEAK (mL/kg/min) | 0.3 (−0.7 to 1.3) | 0.6 | ||
| Tissue Doppler dyssynchrony index (ms) | −0.07 (−0.3 to 0.2) | 0.7 | ||
| E/E’ | 0.04 (−0.2 to 0.3) | 0.8 | ||
| LA volume index (mL/m2) | −0.07 (−0.3 to 0.2) | 0.8 | ||
| TAPSE (cm) | −1 (−10 to 7) | 0.7 | ||
| Interventricular mechanical delay (ms) | −0.1 (−0.3 to 0.05) | 0.3 | 0.3 (−0.06 to 0.6) | 0.1 |
| NT-pro BNP (μg/L) | −1 (−2 to −0.09) | 0.007 | −0.2 (−2 to 1) | 0.8 |
| QRS duration (ms) | −0.14 (−0.3 to −0.02) | 0.004 | −0.3 (−0.7 to 0.1) | 0.2 |
| Strain dyssynchrony index (ms) | −0.1 (−0.2 to −0.06) | 0.04 | −0.1 (−0.2 to −0.04) | <0.001 |
| Fibrosis mass (g) | −7 (−11 to −3) | <0.001 | −7 (−11 to −4) | <0.001 |
| Variable | Univariate model | Multivariate model | ||
|---|---|---|---|---|
| Coefficient (95% CI) | P-value | Coefficient (95% CI) | P-value | |
| Age | −0.06 (−0.3 to 0.2) | 0.8 | ||
| Gender | −4 (−10 to 3) | 0.8 | ||
| 6-minute walk distance (m) | 0.03 (−0.01 to 0.08) | 0.3 | ||
| VO2PEAK (mL/kg/min) | 0.3 (−0.7 to 1.3) | 0.6 | ||
| Tissue Doppler dyssynchrony index (ms) | −0.07 (−0.3 to 0.2) | 0.7 | ||
| E/E’ | 0.04 (−0.2 to 0.3) | 0.8 | ||
| LA volume index (mL/m2) | −0.07 (−0.3 to 0.2) | 0.8 | ||
| TAPSE (cm) | −1 (−10 to 7) | 0.7 | ||
| Interventricular mechanical delay (ms) | −0.1 (−0.3 to 0.05) | 0.3 | 0.3 (−0.06 to 0.6) | 0.1 |
| NT-pro BNP (μg/L) | −1 (−2 to −0.09) | 0.007 | −0.2 (−2 to 1) | 0.8 |
| QRS duration (ms) | −0.14 (−0.3 to −0.02) | 0.004 | −0.3 (−0.7 to 0.1) | 0.2 |
| Strain dyssynchrony index (ms) | −0.1 (−0.2 to −0.06) | 0.04 | −0.1 (−0.2 to −0.04) | <0.001 |
| Fibrosis mass (g) | −7 (−11 to −3) | <0.001 | −7 (−11 to −4) | <0.001 |
Graph illustrating the influence of baseline late-gadolinium enhancement status on the left ventricular ejection fraction.
In the second multivariate analysis, compared with a model containing patient age, QRS duration, and serum NT-pro-BNP concentration, the SDI did not add further value to the model; however, the fibrosis mass as determined by LGE-CMR did significantly add incremental prognostic value for follow-up LVEF (Figure 4).
Multivariate analysis illustrating the incremental value of late-gadolinium enhancement cardiovascular magnetic resonance in modelling improvement in the left ventricular ejection fraction.
Nineteen of the sixty-eight (28%) heart failure patients recruited failed to demonstrate improvement in the LVEF >5% and forty-nine (72%) did exhibit significant improvement by this criterion. The mean scar mass among non-responders was 0.9 ± 0.9 g, and was 0.2 ± 0.6 g among responders (P< 0.001). When indexed for LV mass, the baseline scar mass index among responders was 0.058 ± 0.25%, and among non-responders was 0.36 ± 0.64% (P= 0.02). The mean gadolinium score among non-responders was 1.1 ± 1.4 and among responders was 0.4 ± 0.9 (P= 0.01).
Comparison of DCM patients with and without baseline late-gadolinium enhancement
No significant differences in baseline characteristics were observed between DCM patients with and without LGE at index CMR (Table 3).
Characteristics of DCM Patients with and without baseline LGE
| Characteristic ( ± SD) | Patients with no baseline LGE (n= 39) | Patients with baseline LGE (n= 12) | P-value |
|---|---|---|---|
| Male gender, n (%) | 27 (70) | 10 (83) | 0.4 |
| Age (years) | 58 ± 15 | 58 ± 10 | 0.9 |
| Diabetes mellitus, n (%) | 7 (18) | 3 (25) | 0.3 |
| ACE-I/ARB, n (%) | 38 (97) | 12 (100) | 0.6 |
| Beta-blocker, n (%) | 37 (95) | 11 (92) | 0.6 |
| Spironolactone, n (%) | 15 (38) | 3 (25) | 0.3 |
| Frusemide, n (%) | 25 (64) | 9 (75) | 0.5 |
| QRS duration (ms) | 116 ± 32 | 111 ± 31 | 0.6 |
| 6-minute walk distance (m) | 420 ± 109 | 453 ± 73 | 0.4 |
| Minnesota living with heart failure score | 44 ± 26 | 42 ± 23 | 0.8 |
| LVESVI (mL/m2) | 62 ± 23 | 54 ± 12 | 0.3 |
| E/E’ | 23 ± 13 | 22 ± 8 | 0.9 |
| LAVI (mL/m2) | 50 ± 14 | 42 ± 12 | 0.2 |
| Characteristic ( ± SD) | Patients with no baseline LGE (n= 39) | Patients with baseline LGE (n= 12) | P-value |
|---|---|---|---|
| Male gender, n (%) | 27 (70) | 10 (83) | 0.4 |
| Age (years) | 58 ± 15 | 58 ± 10 | 0.9 |
| Diabetes mellitus, n (%) | 7 (18) | 3 (25) | 0.3 |
| ACE-I/ARB, n (%) | 38 (97) | 12 (100) | 0.6 |
| Beta-blocker, n (%) | 37 (95) | 11 (92) | 0.6 |
| Spironolactone, n (%) | 15 (38) | 3 (25) | 0.3 |
| Frusemide, n (%) | 25 (64) | 9 (75) | 0.5 |
| QRS duration (ms) | 116 ± 32 | 111 ± 31 | 0.6 |
| 6-minute walk distance (m) | 420 ± 109 | 453 ± 73 | 0.4 |
| Minnesota living with heart failure score | 44 ± 26 | 42 ± 23 | 0.8 |
| LVESVI (mL/m2) | 62 ± 23 | 54 ± 12 | 0.3 |
| E/E’ | 23 ± 13 | 22 ± 8 | 0.9 |
| LAVI (mL/m2) | 50 ± 14 | 42 ± 12 | 0.2 |
LGE, late-gadolinium enhancement.
Characteristics of DCM Patients with and without baseline LGE
| Characteristic ( ± SD) | Patients with no baseline LGE (n= 39) | Patients with baseline LGE (n= 12) | P-value |
|---|---|---|---|
| Male gender, n (%) | 27 (70) | 10 (83) | 0.4 |
| Age (years) | 58 ± 15 | 58 ± 10 | 0.9 |
| Diabetes mellitus, n (%) | 7 (18) | 3 (25) | 0.3 |
| ACE-I/ARB, n (%) | 38 (97) | 12 (100) | 0.6 |
| Beta-blocker, n (%) | 37 (95) | 11 (92) | 0.6 |
| Spironolactone, n (%) | 15 (38) | 3 (25) | 0.3 |
| Frusemide, n (%) | 25 (64) | 9 (75) | 0.5 |
| QRS duration (ms) | 116 ± 32 | 111 ± 31 | 0.6 |
| 6-minute walk distance (m) | 420 ± 109 | 453 ± 73 | 0.4 |
| Minnesota living with heart failure score | 44 ± 26 | 42 ± 23 | 0.8 |
| LVESVI (mL/m2) | 62 ± 23 | 54 ± 12 | 0.3 |
| E/E’ | 23 ± 13 | 22 ± 8 | 0.9 |
| LAVI (mL/m2) | 50 ± 14 | 42 ± 12 | 0.2 |
| Characteristic ( ± SD) | Patients with no baseline LGE (n= 39) | Patients with baseline LGE (n= 12) | P-value |
|---|---|---|---|
| Male gender, n (%) | 27 (70) | 10 (83) | 0.4 |
| Age (years) | 58 ± 15 | 58 ± 10 | 0.9 |
| Diabetes mellitus, n (%) | 7 (18) | 3 (25) | 0.3 |
| ACE-I/ARB, n (%) | 38 (97) | 12 (100) | 0.6 |
| Beta-blocker, n (%) | 37 (95) | 11 (92) | 0.6 |
| Spironolactone, n (%) | 15 (38) | 3 (25) | 0.3 |
| Frusemide, n (%) | 25 (64) | 9 (75) | 0.5 |
| QRS duration (ms) | 116 ± 32 | 111 ± 31 | 0.6 |
| 6-minute walk distance (m) | 420 ± 109 | 453 ± 73 | 0.4 |
| Minnesota living with heart failure score | 44 ± 26 | 42 ± 23 | 0.8 |
| LVESVI (mL/m2) | 62 ± 23 | 54 ± 12 | 0.3 |
| E/E’ | 23 ± 13 | 22 ± 8 | 0.9 |
| LAVI (mL/m2) | 50 ± 14 | 42 ± 12 | 0.2 |
LGE, late-gadolinium enhancement.
Baseline late-gadolinium enhancement and ventricular dyssynchrony
Absence of baseline LGE was significantly associated with improvement in both intra-LV and interventricular dyssynchrony. At baseline, the TD DI was similar between those without and with baseline LGE (41 ± 14 vs. 42 ± 15 ms, P= 0.9). At follow-up those without baseline LGE exhibited improvement in TD DI to 34 ± 18 ms; those with baseline LGE experienced progression in TD DI to 49 ± 21 ms (P= 0.01). There was no significant difference in IVMD between those without and with baseline LGE (31 ± 23 ms vs. 33 ± 28 ms, P= 0.9). At follow-up, those without baseline LGE had stable IVMD (29 ± 31 ms), whereas those with baseline LGE displayed worsening of IVMD (57 ± 39 ms, P= 0.01). Late-gadolinium enhancement presence was not associated with change in the SDI (P= 0.3).
Inter-observer agreement
Bland-Altman biases (95% limits of agreement) were 0.5 ms (−8 to 9 ms) for the TD DI, −6 ms (−11 to 23 ms) for the SDI, and 1.6% (−1.8 to 5.0%) for GLS.
Discussion
Major findings
In this prospective, multi-centre study of newly diagnosed DCM patients, we have demonstrated for the first time that: (i) myocardial fibrosis at initial presentation, as detected by LGE-CMR, and intra-LV dyssynchrony by speckle-tracking strain analysis were, among a broad range of physiologic covariates, independently associated with failure of improvement in the LV EF at the 5-month follow-up. The performance of LGE-CMR added incremental value to a multivariate model of follow-up LVEF. (ii) The presence of LGE at first presentation was associated with adverse outcomes with respect to global LV strain, and interventricular and intra-LV dyssynchrony. Thus, our findings have important implications for the clinical management of newly diagnosed DCM patients and for understanding the relationship between myocardial fibrosis, ventricular dyssynchrony, and lack of improvement in LV function.
Cardiovascular magnetic resonance late-gadolionium enhancement in DCM
In a study of patients with established DCM of at least 12 months duration, Assomull et al.2 showed that LGE was the sole independent predictor of death or cardiovascular hospitalization. Wu et al.3 reported similar findings among a cohort of patients with established non-ischaemic cardiomyopathy referred for implantable cardioverter-defibrillator insertion. Neither of these studies included QRS duration, echocardiographic indices of ventricular dyssynchrony, or NT-pro-BNP concentration, despite their reported prognostic value.5,18 In the present study, we included a variety of novel and established indices in a model of follow-up LVEF following optimal medical therapy. Late-gadolinium enhancement status at initial presentation was shown to be independently associated with improvement in LV function even accounting for these clinical and biochemical variables. Assomull and Wu reported a prevalence of LGE of 30 and 42%, respectively. The prevalence of LGE in the current study was slightly less, at 24%. This most likely reflects the exclusive recruitment of consecutive patients with newly diagnosed DCM in our study, in whom the prevalence of myocardial fibrosis may be less than when compared with the rates among patients with long-standing disease. Park et al.19 previously studied 46 patients with non-ischaemic cardiomyopathy, and found that LGE status predicted improvement in the LVEF. It was unclear, however, whether this study was performed in consecutive cardiomyopathy patients and most importantly, if CMR was performed at diagnosis and at the commencement of medical therapy. The reported 52% prevalence of LGE substantially exceeds that described in other series of non-ischaemic cardiomyopathy, including our own, reflecting that theirs was likely a cohort with long-standing dilated cardiomyopathy.
Cho et al.20 have recently shown that global two-dimensional strain may confer additive prognostic value to routine echocardiographic indices, including LVEF, in systolic heart failure patients. In the current study, we have demonstrated that the absence of myocardial fibrosis by LGE-CMR at presentation is associated with improvement in the global LV strain, whereas its presence is associated with failure to improve.
Ventricular dyssynchrony in DCM
We found that neither intra-LV nor interventricular dyssynchrony decreased with medical therapy despite significant improvement in all other major indices of cardiac structure and function. Our findings are in contrast to those of Takemoto et al.21 who reported that beta-blocker therapy improves intra-LV synchrony in a small cohort of patients with DCM. However, this study cohort did not comprise consecutive patients with DCM and follow-up dyssynchrony data were only obtained in 15 of 25 DCM patients. Furthermore, since assessment of ventricular dyssynchrony was not combined with direct assessment of myocardial fibrosis, the authors were unable to relate these two parameters in response to medical therapy. In contrast, our study indicates that the presence of myocardial fibrosis as detected by LGE-CMR at presentation in patients with newly diagnosed DCM is associated with subsequent deterioration in indices of ventricular synchrony.
This study found that among echocardiographic indices of ventricular dyssynchrony, only strain measures were independently associated with change in the LVEF. Despite its reported ability to identify responders to cardiac resynchronization therapy, tissue Doppler dyssynchrony was not associated with response to medical therapy in our cohort. Tethering of poorly contractile segments to adjacent viable segments may influence tissue velocities. This may account for the observed superior performance of strain imaging. Tigen et al.22 have recently reported an association between the extent of fibrosis in DCM patients and intra-LV dyssynchrony. The present study extends on these findings to demonstrate that the quantity of myocardial fibrosis by LGE-CMR is correlated with the magnitude of improvement in both LV function and ventricular synchrony.
Clinical implications
Our findings may have important implications for the clinical management of newly diagnosed DCM patients. Although implantable-defibrillator use has been shown to reduce cardiac mortality in patients with the LVEF ≤35%, uncertainty exists over the optimal timing of insertion following diagnosis with DCM. A substantial proportion of newly diagnosed DCM patients will experience improvement in the LVEF to >35% following medical therapy. Thus, routine device implantation at the time of diagnosis, purely based on the LVEF at presentation, is likely to result in insertion in many who would no longer qualify after 5 months of' medical therapy. The present study strongly suggests a role for LGE-CMR as a risk-stratifying investigation in cases of newly diagnosed DCM—an important direction for future research, in which the prognostic value of this imaging modality for hard clinical endpoints might be examined. Furthermore, the identification of individuals at high risk of non-response to heart failure treatment may permit more aggressive observation and more targeted medical therapy.
Study limitations
Patients were recruited and studied within 2 weeks of initial diagnosis with DCM. We were unable to determine the duration of LV dysfunction prior to enrolment, and hence it is likely that patients were presenting at different time points in thedisease course. We cannot be certain whether our findings are influenced by the chronicity of cardiomyopathy; however, our data within 2 weeks of clinical diagnosis are the earliest reported to date. Consequently, our study remains highly clinically relevant as it reflects real-life practice, in which identification of patients with subclinical DCM is not feasible. Moreover, those presenting with DCM do so at varying stages of the disease duration owing to inter-individual variation in symptom threshold and tolerance. Given the promptness with which patients were studied following diagnosis, there may have been heterogeneity in the degree of up-titration of medical therapy at the time of baseline imaging; however, all were early (<2 weeks) before significant remodelling would be anticipated to occur. Follow-up echocardiographic findings may also have been confounded by differences in fluid balance and loading conditions compared with the baseline study, although an attempt was made to control for this by performing imaging at the same time of day for each individual.
Cardiovascular magnetic resonance could not be performed in all patients at recruitment. This limits the generalizability of our findings. The LGE technique is limited in its ability to detect diffuse myocardial fibrosis because it relies on the contrast between more fibrotic tissue and relatively normal myocardium. T1 mapping is a CMR technique that has shown promise in the quantification of diffuse myocardial fibrosis, and may be of value in this setting.23 Myocardial biopsy was not undertaken as routine in the present cohort of patients. The practice of exclusion of myocarditis by clinical, laboratory, and CMR findings is standard of care in the participating institutions. The presence of an inflammatory myocardial infiltrate suggestive of myocarditis could not be excluded.
This clinical study is unable to ascribe a causal relationship between myocardial fibrosis and dyssynchrony. We found that both were independently associated with change in LV function. Further research on the influence of cardiac resynchronization in this cohort is likely to be informative, in those patients with baseline myocardial fibrosis. Lastly, owing to the limited sample size, these findings require confirmation in a larger study, perhaps comparing a strategy of routine baseline evaluation of myocardial fibrosis and ventricular dyssynchrony with routine care.
Conclusion
Using novel and complementary imaging techniques, we have shown that the presence of myocardial fibrosis and the degree of ventricular dyssynchrony are indepedently associated with a lack of response to medical therapy in new-presentation DCM. Identification of individuals at high risk of non-response to initial heart failure treatment may permit more individualized, timely, and targeted treatment.
Funding
Drs D.P.L. and P.M. are supported by a Medical Postgraduate Scholarship funded jointly by the National Health and Medical Research Council of Australia and the National Heart Foundation of Australia. Dr P.S. is supported by the National Heart Foundation of Australia
Conflict of interest: none declared.
Acknowledgements
Dr Hugh Greville, Dr Suchi Khurana.
Comments
We are grateful for the opportunity to engage in discussion over our research. We agree with Drs. Vest and Francis that it would be premature to come to a strong conclusion about the role of left ventricular (LV) dyssynchrony in perpetuating LV systolic dysfunction independent of myocardial fibrosis burden. While we consider our findings thought- provoking and hypothesis-generating, they require confirmation in larger numbers with more diverse heart failure aetiology. Given the observational nature of our research, we cannot be certain about the causative relationship between myocardial fibrosis and LV dyssynchrony, although the stronger predictive ability of myocardial fibrosis in the present study might suggest that it drives LV dyssynchrony, rather than the reverse. We presented data on the predictive value of myocardial fibrosis for LV dyssynchrony in the section entitled "Baseline late-gadolinium enhancement and ventricular dyssynchrony". In addition, late-gadolinium enhancement exhibited a trend towards a positive association with change in LV dyssynchrony by speckle-tracking strain. Lastly, Drs. Vest and Francis are correct in their assumption that LV ejection fraction was measured blinded to the numeric value of dyssynchrony and to late-gadolinium enhancement data In summary, our research is consistent with growing evidence of the prognostic importance of the presence and burden of myocardial fibrosis. In regards to LV dyssynchrony, we believe a distinction must be made between its measurement as proof of a physiologic concept and its evaluation to prognosticate for a particular patient in the consulting room. The instruments to assess LV dyssynchrony may not have the precision to provide a definitive answer in the latter situation. This imprecision on an individual basis should not mislead us to ignore a potential biological signal that is detectable in a cohort of individuals.
Conflict of Interest:
None declared
Dear Editor,
We read with interest the findings of Leong and colleagues(1) regarding the impact of myocardial fibrosis, as assessed by late- gadolinium enhancement cardiac magnetic resonance imaging (LGE-CMR), on the response to medical therapy in new-presentation idiopathic dilated cardiomyopathy. The investigators also reported the impact of strain dyssynchrony index (SDI). The latter part of the article draws parallels between the adverse prognosis conferred by greater baseline myocardial scar burden and the degree of ventricular dyssynchrony, and implies that they contribute independently to worse clinical outcomes.
We can see the conceptual attraction of such a conclusion, but can we be certain that the data point that way convincingly? Table 2 shows a significant association of SDI with change in left ventricular ejection fraction (LVEF) >5% at 5 months on multivariate analysis. However, it seems to us that Figure 4 illustrates no substantial incremental value of the addition of SDI to age, QRS duration and NT-pro BNP in predicting LVEF improvement (P=0.1).
The important finding demonstrated in Figure 4 is that the addition of LGE-CMR myocardial scar burden to age, QRS, BNP and SDI does significantly contribute to the LVEF prediction model (P<0.001). Could Leong and colleagues give an indication of their interpretation of the relative strength of LGE myocardial scar burden versus SDI for the prediction of cardiomyopathy outcomes? They are in a unique position to be able to do so. Figure 3 correlates baseline myocardial fibrosis to lower likelihood of LVEF improvement. However, the text indicates that the quantity of myocardial fibrosis by LGE-CMR is also associated with the magnitude of improvement in ventricular synchrony; is it possible to reveal those data too?
Although other investigators (2) have linked LGE myocardial fibrosis to global longitudinal strain derived from speckle-tracking echocardiography, the graphical data in the current paper by Leong do not appear, to us, to be supportive of a role for cardiac synchrony in the prediction of LVEF improvement. Instead it highlights the impressive utility of LGE-CMR quantification of myocardial scar, QRS duration and NT- pro BNP.
Finally are we correct to assume that LVEF was assessed echocardiographically with blinding to the numerical level of dyssynchrony but with the pattern of contraction visible, while the CMR information was unavailable? The relevance is that if observers had a prior belief that dyssynchrony was an important contributor to LVEF change they might have subconsciously tended to rate baseline LVEF lower in patients with eyeball -obvious dyssynchrony, which would cause dyssynchrony (but not CMR scar if they were blinded to it) to have an artefactual relation to LVEF response.
We ask this of the authors because we have faith in their neutral data analysis and respectfully wonder whether they could apply the same careful interpretation to the research methods for studies of this nature, of which there are several and with conflicting results. This strategy would require an elaborate protocol of mutually blinded measurements of baseline and pre/post markers, but not doing so could condemn our field to serial disappointment.
1. Leong DP, Chakrabarty A, Shipp N, Molaee P, Madsen PL, Joerg L, Sullivan T, Worthley SG, De Pasquale CG, Sanders P, Selvanayagam JB. Effects of myocardial fibrosis and ventricular dyssynchrony on response to therapy in new-presentation idiopathic dilated cardiomyopathy: insights from cardiovascular magnetic resonance and echocardiography. European Heart Journal. 2012; 33:640-648.
2. D'Andrea A, Caso P, Scarafile R, Riegler L, Salerno G, Castaldo F, Gravino R, Cocchia R, Del Viscovo L, Limongelli G, Di Salvo G, Ascione L, Iengo R, Cuomo S, Santangelo L, Calabr? R. Effects of global longitudinal strain and total scar burden on response to cardiac resynchronization therapy in patients with ischaemic dilated cardiomyopathy. Eur. J. Heart Fail. 2009; 11:58-67.
Conflict of Interest:
None declared