Early-Onset Late Gadolinium Enhancement is a Prognostic Factor for Duchenne Cardiomyopathy

Dilated cardiomyopathy (DCM) is an inevitable complication of Duchenne muscular dystrophy (DMD). Late gadolinium enhancement (LGE) demonstrated by cardiac MRI occurs in DMD-related DCM, indicating myocyte death and remodeling. We conducted a retrospective chart review identifying DMD patients in our center between January 2009 and July 2013. Subjects were cohorted by presence of LGE before age 14. We excluded patients in whom we could not determine LGE status prior to age 14. We reviewed comprehensive clinical data. Of the 41 subjects with complete data, 15 demonstrated LGE before age 14 (“early LGE”) and 26 had no LGE by age 14 (“controls”). Those with early LGE exhibited a more rapid decline in LV fractional shortening (p = 0.028). Patients with early LGE were younger at age of initiation of ACE inhibition (p = 0.025), mineralocorticoid receptor antagonism (p = 0.0024), and beta-blockade (p = 0.0017), suggesting aggressive clinical management in response to abnormal MRI findings. There were no significant differences in LV dilation between the two groups (p = 0.1547). Early LGE was not associated with obesity (p = 0.32), age at loss of ambulation (p = 0.31), or heart rate (p-value > 0.8). Early onset of myocardial fibrosis as indicated by LGE on cardiac MRI is associated with earlier progression of cardiomyopathic changes despite earlier medication therapy. Identifying this risk factor, observed in 34% of our cohort during preadolescence, may guide medical therapy and early counseling about cardiomyopathy progression. We advocate for obtaining at least one MRI in patients with DMD prior to age 14 to risk stratify patients.


Introduction
Dilated cardiomyopathy (DCM) is an inevitable complication of Duchenne muscular dystrophy (DMD) [1]. In these patients, both skeletal and cardiac muscle lack functional dystrophin and are thus mechanically fragile. Over time, loss of cell membrane integrity leads to pathologically increased calcium influx into muscle cells, culminating in apoptosis [2]. Cell death leads to replacement of myocardial tissue with fibrofatty tissue, which can be seen on cardiac MRI (cMRI) as late enhancement of the tissue, minutes after perfusion with gadolinium. Clinically, patients with DMD display tonic contraction of the myocardium and then progressive weakening of heart muscle, leading to ventricular dilation, dysfunction, and, ultimately, heart failure [3]. Detecting the earliest changes of this disease may help to guide timing and strategy of cardiac therapies.
Cardiac MRI provides detailed information about myocardial health and function and is not subject to the same limitations in ultrasound windows which plague older patients with DMD. Thus, this modality has been of particular interest in older patients as a replacement for older modalities, such as multigated acquisition (MUGA) nuclear scans. However, cMRI is time consuming, expensive, requires intravenous access to obtain contrast-enhanced images, and has limited availability.
In patients with DMD, LGE increases with age and correlates closely with progression of left ventricular dilation and dysfunction [4]. While the presence of LGE tends to increase over time, there has been little to no research examining the impact of timing of first appearance of LGE on how these patients with DMD fare in the long term [5]. The objective of this study was to examine the trajectory of cardiomyopathy in DMD patients who demonstrated LGE at an early age compared to those who demonstrated it later on in adolescence in order to describe disease progression between the two groups.

Patients
We conducted a retrospective chart review to identify patients with DMD in our center seen between January 2009 and July 2013. Forty-three patients between the ages of 0 and 18 years with a confirmed diagnosis of DMD and who underwent at least one cardiac MRI were identified. Subjects were divided by presence or absence of LGE before the age of 14 ("early LGE" and "control" groups). We excluded patients with insufficient MRI data to determine whether LGE was present before age 14. We also excluded from certain statistical analysis, patients under age 14 without LGE whose most recent MRI was before the age of 14, as these patients could not be safely assigned to the control group. If a patient had their first MRI after age 14 but was still free of LGE, they were assigned to the control group. Primary outcome was decline in left ventricular systolic function as assessed by fractional shortening (FS) on echocardiogram. Medical records were reviewed for patient age, demographic detail, confirmation of diagnosis, genetic testing, age at loss of ambulation, age at initiation of cardiac medications, and echocardiographic and cMRI information. Continuous objective data such as heart rate, pulse pressure, and body mass index were extracted from the chart at a date after yet closest to the patient's fourteenth birthday.

Imaging
We reviewed every lifetime echocardiogram and cMRI of the patients from our cohort. Transthoracic echocardiographic evaluation was performed using cardiac ultrasound imaging systems (Philips Medical Systems, Andover, MA) and was interpreted by board-certified pediatric cardiologists. Parameters assessed included left ventricular (LV) dimensions via M-mode measurements, systolic shortening, and contractility. M-mode measurements included LV end-diastolic and end-systolic dimensions (LVIDd and LVIDs), LV posterior wall thickness, and interventricular septal thickness at end diastole. The measurements allowed for calculation of fractional shortening and estimation of LV mass. Diastolic measurements, including tissue Doppler mitral velocities and longitudinal strain imaging, were inconsistently available and therefore not analyzed.
Cardiac MRI was performed without sedation on a Philips Achieva at 1.5 T, using a cardiac or torso phased-array coil. Gadolinium contrast with gadopentetate dimeglumine (0.2 mm/kg) or gadobutrol (0.1 mm/kg) was hand injected at infusion rates of one to two cc per second. Cardiac mass and function were acquired using short-axis steady-state precession images covering both ventricles, and it was assessed using semiautomatic planimetry (with manual refinement). After approximately 10 min following contrast injection, delayed hyperenhancement imaging was performed using 2D inversion recovery gradient echo imaging with cardiac triggering at two R-R intervals. A 10-inversion time Look-Locker sequence was used to estimate the optimal inversion time to maximize myocardial nulling. This sequence was repeated, as needed, if myocardial nulling was inadequate on the delayed hyperenhancement imaging. All cMRI images were interpreted by board-certified pediatric cardiologists with advanced imaging training and between 5 and 22 years of dedicated cMRI experience.

Statistical Analysis
To account for repeated measures (echocardiograms and MRIs), linear mixed models (LMM) were used to assess the effects of age between "early LGE" and "control" groups on FS, LVIDd, and LV end-diastolic volume (LVEDv) measures. To test effect modification of age on the "early LGE" group for each of the three outcome measures, main effects for age and LGE were included along with an age-by-LGE interaction term. Separate one-sided t tests were used to test whether mean FS or mean LV dimension after age 14 in the "early LGE" group was lower than in the "control" group. The correlation between echocardiographic LVIDd and cardiac MRI LVEDv values was measured and tested using a univariate linear regression model. Kaplan-Meier curves were used to estimate the time to ambulatory loss distribution using age as the time variable with a log-rank test to analyze differences in the two LGE groups. Cox proportional hazards regression was applied to model time to ambulatory loss and initiation of angiotensin converting enzyme inhibitors (ACEi), mineralocorticoid receptor antagonists (MRA), and beta-blockers. Unadjusted marginal logistic regression models were used to test the associations of continuous clinical measures of body mass index and heart rate with LGE groups. Fisher's exact test was used for tests of association between categorical clinical factors (e.g., obesity) and LGE group.

Results
Over the four-year period between January 2009 and July 2013, 43 subjects met inclusion criteria. Two were excluded due to inadequate MRI quality (n = 1) or lack of gadolinium contrast (n = 1). Of the remaining 41 subjects, 15 patients (34%) were found to have "early LGE" before age 14, and they were compared to those without LGE before age 14 (control group). Patients with early LGE had significantly lower fractional shortening (FS) by echocardiogram after age 14 when compared to the control group (24.89% vs 29.78%, p-value = 0.028) (Fig. 1). With each additional year of age, those with early LGE had an average reduction in FS by -1.11% compared to the average reduction of − 0.57% of the control group (p-value = 0.004) (Fig. 2).
When comparing LV dimension by echocardiogram, no significant differences over time were detected among the two groups. The LVIDd Z-score changed by approximately 0.04 per each additional year of life for the early LGE group compared to − 0.04 for the control group (p-value = 0.1547) (Fig. 3). For those who had echocardiograms after age 14, mean LVIDd Z-scores did not vary for the early LGE group compared to the control group (p-value = 0.3466) (Fig. 4). Similarly, three-dimensional measures of left ventricular end-diastolic volumes (LVEDv) by cardiac MRI were not significantly different between the early LGE group and the control group (p-value = 0.67). When correlating the echocardiographic measure of LVIDd with cardiac MRI measure of LVEDv for validation, a very positive significant correlation was observed between the two measures (R 2 = 0.48) (Fig. 5).   In patients with echocardiogram data taken after age 14 (n = 29), mean LVIDd Z-score in patients with LGE present before 14 years of age are not significantly higher than in patients without LGE present before age 14 (p-value = 0.3466) in a one-sided t test When evaluating other clinical factors that might modulate severity of cardiac disease in DMD patients, we did not find significant differences between early LGE patients and the control group. There was no significant difference observed between age of ambulation loss and development of LGE (10 vs 14 years, p-value = 0.31). Similarly, presence of obesity (p-value = 0.32) and heart rate (p-value > 0.8) were not associated with timing of LGE development in our small cohort.

Discussion
Duchenne muscular dystrophy is an X-linked neuromuscular disorder that affects 1 in 5000 boys [6]. Although these children present initially with skeletal muscle weakness, development of cardiomyopathy is inevitable and typically occurs in the second to third decade of life with more than 90% of adolescent males over 18 years of age demonstrating cardiac dysfunction [7]. With improvements in respiratory support for these patients, end-stage heart failure has become an increasing cause of death in this population. Thus, early and pre-emptive treatment of the cardiomyopathy are thought to be important, and identification of preclinical changes can allow for initiation of cardioprotective measures to slow the progression of adverse cardiomyopathic remodeling.
In recent years, cardiac MRI with late gadolinium enhancement (LGE) has been demonstrated to be a valuable tool in cardiac assessment of DMD patients. Because of comorbidities including body habitus, contractures, ventilatory disease, and scoliosis, echocardiography in this population is often limited with more severe disease. In addition to providing more detail regarding cardiac dimensions, systolic function, and diastolic function, the presence of LGE is a poor prognostic indicator for adverse cardiac events in adult patients with DMD [8,9]. While LGE is a known poor prognostic indicator in older DMD patients with overt cardiomyopathy, this study explores the significance of LGE detected in the pre-adolescent age range.
Our findings suggest that cardiac functional decline occurs much more rapidly in patients who demonstrate LGE at an early age, which can be observed in roughly one-third of pre-adolescents with DMD. Historically, symptomatic cardiomyopathy in these DMD patients has been identified later in the second decade of life or by age 20. This study shows there are preclinical changes detectable by cardiac MRI even if systolic function and cardiac dimensions measured by echocardiogram appear within normal range. This implies benefit to performing cardiac MRI on patients with DMD at least once as pre-adolescents as part of thorough risk stratification.
In this study, functional decline was determined via fractional shortening captured by echocardiography. While we highlight an important relationship between early-onset cardiac fibrosis with faster progression of systolic dysfunction, we are unable to draw further conclusion with regards to the evolution of LGE and its correlation with cardiac dysfunction due to the limited number of patients with serial cardiac MRIs in our study population. Of the forty-one patients who met inclusion criteria, only thirteen had more than one cMRI. In a similar study of pediatric and adult patients, Tandon et al. reported 98 patients with DMD who had at least four cMRI studies and showed that while LV ejection fraction remains stable in patients without LGE, LV ejection fraction declines significantly soon after development of LGE [9]. Our current study adds to this observation by denoting a specific age, 14 years, at which screening MRI can have an important future clinical implication. ACE inhibitors, beta-blockers, and MRAs are thought to delay progression of DMD cardiomyopathy by affecting neurohormonal regulation and attenuating myocardial fibrosis [10][11][12]. Research has demonstrated that early ACEi use is associated with delay in cardiac dysfunction and prolongation of life [13,14]. ACE inhibitors have now become standard of care for most patients with DMD, even prior to the onset of overt cardiomyopathy [15]. In our cohort, MRI findings likely led to earlier initiation of ACEi, MRA, and beta-blockade in the group of patients who developed LGE prior to age 14. Although it will take considerably more time to delineate whether programmatic personalization of medical therapy in response to early teen MRI data will prove important to patient longevity, this approach seems prudent compared with waiting for cardiac dysfunction to occur by echocardiogram before obtaining MRI data for the first time.
Early LGE patients trended toward earlier age of loss of ambulation, but this observation did not meet statistical significance. Larger studies have reported that loss of ambulation and heart function do not occur concomitantly, consistent with the observation that skeletal myocyte function and cardiac myocyte function are distinct [16].
Heart rate, however, has been previously reported to correlate with cardiac severity in DMD, including a case-control study where cardiomyopathy developed in 42% of patients with heart rate in the upper quartile but only in 14% of boys in the lower heart rate quartile (p < 0.05) [17]. Our study did not detect such a pattern, possibly due to the limited time points of which heart rate was captured in this study. Our study did not find a correlation between severity of obesity and earlier development of symptomatic cardiomyopathy, consistent with existing literature [18].
Importantly, if timing of loss of ambulation, heart rate trends, and development of obesity do not clinically correlate with development of cardiac dysfunction, this lends credence to use of cardiac MRI as a standard tool for early detection of myocardial fibrosis in patients before development of clinically symptomatic cardiomyopathy. As 35% of our cohort was found with early-onset fibrosis, utilizing cardiac MRI during this pre-adolescent period is clinically informative in a significant proportion of patients. Prognostic information is not only important for strategizing guideline-directed medical therapy for heart failure but also for counseling patients with DMD and their caregivers. Giving our patients knowledge about timeline of their hearts' functional decline is vital in a disease management that is ultimately palliative.
Frame shift mutations, such as deletions and duplications, are associated with more severe phenotypes of DMD. This is in contrast to point mutation/stop codon phenotypes [19,20]. Even large deletions of the region corresponding to the central triple helical repeats in the protein, if they do not alter the reading frame, can correspond with surprisingly mild phenotype, such as a Becker muscular dystrophy phenotype even if the mutation predicts Duchenne [21]. Allelic X-linked dilated cardiomyopathy can present as cardiac predominant, sparing the skeletal muscle which produces dystrophin by exon skipping or alternative splicing that the heart is not able to replicate [22,23]. Additionally, certain modifying genes such as LTBP4 have been shown to predict clinical outcomes, such as age at loss of ambulation [24,25]. The scope of our study did not include enough data to draw conclusions about genetics and severity of disease, so only descriptive data were included.
Our study is inherently limited by its retrospective nature and by it representing a single institution, whose patient population and clinician practices may vary from other institutions. We restricted our study to those with adequate follow-up after age 14, further curtailing our sample size. Additionally, this study era pre-dated protocolized age at first MRI at our institution. Finally, the age cutoff of 14 for determining early LGE was a clinically subjective decision, intended to optimize statistical analysis for this study.
One-third of pre-adolescent DMD patients exhibit cardiac myocardial fibrosis as indicated by LGE on cardiac MRI. This early finding is strongly associated with faster progression to cardiac systolic dysfunction. Aggressive augmentation of cardioprotective drugs in these patients who demonstrate early LGE does not prevent development of cardiac dysfunction but may potentially slow the course of cardiomyopathy. Early assessment by cMRI provides valuable prognostic information to help counsel families and guide medical decisions for young patients with DMD.
Author Contributions Dr. James collected the data, contributed to study design, carried out the initial analyses, drafted the initial manuscript, and reviewed and revised the manuscript. Dr. Menteer reviewed initial analysis, contributed to study design, and reviewed and significantly revised the manuscript. Drs. Lilith Moss and Durazo-Arvizu performed statistical analysis, wrote the section on statistics methodology, composed the figures and table design, and reviewed and revised the manuscript. Dr. Wood reviewed initial analysis, contributed to study design, wrote the section on imaging methodology, and reviewed and revised the manuscript. Drs. Ramos-Platt and Tiongson compiled genetic data and reviewed and revised the manuscript. Dr. Su conceptualized and designed the study, supervised data collection, reviewed initial analyses, coordinated statistical resources, and reviewed and significantly revised the manuscript.

Funding
The authors have no relevant financial or non-financial interests to disclose. Partial funding for statistics was obtained through the Division of Cardiology and the Pediatric Residency Program at Children's Hospital Los Angeles.

Conflict of interest
The authors have no conflict of interest to disclose.

Informed consent
The Institutional Review Board at Children's Hospital Los Angeles approved this study and waived the need for informed consent.
Ethical approval All authors approved the final manuscript as submitted and agreed to be accountable for all aspects of the work.