Patterns of late gadolinium enhancement in Duchenne muscular dystrophy carriers

We enrolled 30 consecutive female DMDc aged 11 to 63 years (mean age 36 ± 5 years) followed at the Center for Neuromuscular Diseases (Uildm) of Rome, Italy and 37 age-matched healthy female controls (mean age 34 ± 5 years, p = NS). Controls had no cardiac symptoms, history of cardiomyopathy or skeletal muscle disorders and presented with normal electrocardiography (ECG), 2D-echocardiography and CMR. No drug was given to DMDc and controls. Creatine kinase levels were screened in the control group and in all DMDc. The study was approved by the Ethical Board of the Catholic University of Rome. Written informed consent was obtained from all patients and controls. Exclusion criteria were left ventricular (LV) systolic and diastolic dysfunction, diabetes, hypertension, ECG changes suggestive of ischemic heart disease, atrial flutter/fibrillation, any degree of atrioventricular block, valvular heart disease, and LV hypertrophy.

DMDc and controls underwent physical examination and surface 12-lead ECG. DMDc, none with cardiac symptoms, were classified regarding muscle involvement as: (1) asymptomatic (asy), when characterized by the presence of high CK levels and/or minor myopathic signs like muscle cramps and myalgia, without muscle weakness and (2) symptomatic (sym), when presenting variable degrees of muscle weakness [14]. The degree of skeletal muscle involvement was assessed by a neurologist and all DMDc were divided into two functional groups: a) 21 asy (63%) and b) 9 sym (27%). Of the 9 sym-DMDc, 3 were mildly affected, showing some impairment in running and jumping, 5 moderately affected, with clear strength deficits in specific muscle districts and 1 severely affected and wheelchair-bound. In asy-DMDc, significant coronary artery disease was excluded by negative treadmill exercise testing. In sym-DMDc, myocardial perfusion was assessed by adenosine thallium-201 myocardial perfusion scintigraphy (ATl-201-MPS), according to the guidelines for the clinical use of cardiac radionuclide imaging [15]. ATl-201-MPS was performed in these patients since they could not perform treadmill exercise testing.

Conventional 2D-echocardiography was performed by the same operator (VG) in all patients, using a SONOS 5500 (Philips Andover, Mass, US). LV systolic function was assessed by ejection fraction (EF) from LV volumes (derived using the modified Simpson’s rule). LV systolic dysfunction was defined as LVEF ≤55%. LV diastolic function was assessed by Doppler analysis as previously reported [16].

Extensive molecular analysis was performed in all DMDc. Mutation detection was carried out using Multiplex-Ligation dependent Probe Amplification (MLPA) (deletions and duplications) or sequencing (small mutations), accordingly to the DMD guidelines [17]. Although muscle biopsy may be used for establishing the carrier condition, it is well known that approximately 60% of carriers do not show muscle abnormalities [18].

CMR was performed using a 1.5-T MR system (INTERA, Philips Medical Systems, Best, the Netherland) with a cardiac 5-element phased-array receiver coil . All images were acquired with ECG-gating, breath-hold steady-state free precession (SSFP) cine sequence for functional analysis, in contiguous short-axis view (10-mm intervals, interslice gap 2 mm, slice thickness 8 mm in plane-resolution 1.2 × 1.8 mm) from the mitral annulus to the apex and 3 long-axis planes, with the patient in a supine position. To assess the contribution of cardiac edema, we performed a T2-weighted segmented triple inversion recovery (T2-wSTIR) imaging module, in 3 short axis slices (8 mm, flip angle 90°, repetition times 2 RR intervals) at the base, mid, and apex and a single long axis-slice in a 4 chamber view, using for imaging a functional surface coil intensity correction. All patients underwent an LGE imaging protocol (repetition time 4.5 ms, echo time 1.7 ms, inversion time 200 to 300 ms) for myocardial scar using a segmented Inversion Recovery-Gradient Echo (IRGE) sequence, adjusting the inversion time and nulling the signal of normal myocardium. Contrast CMR images were acquired on average 10 to 15 min after injection of cumulative 0.1 mmol/kg gadolinium DTPA (Magnevist, Gd-DTPA, Shering AG). Images were obtained in 8 to 14 short-axis and 3 radial long-axis planes. Myocardial enhancement on LGE images was assessed visually and considered positive to a signal-intensity threshold of >2 SD above the mean intensity of a remote reference region [19] and interpreted as present or absent by the consensus of two cardiac CMR-trained physicians. LGE quantity was quantified using manual planimetry, summing the LGE positive areas yielding a total volume (ml). LGE percentage was obtained dividing the total LGE volume by the LV mass. The LGE location and wall motion abnormalities was classified according to AHA for heart imaging [20]. Height and weight were measured in DMDc and controls on the day of scanning; values for volumes and mass were indexed by body surface area (BSA). The Simpson’s method was applied to determine myocardial mass, end-diastolic volume, end-systolic volume, right ventricular (RV) and LV EF by a dedicated software, manually tracing the endocardial and epicardial borders in each short axis slice. Depressed RV and LV systolic function were defined according to the reference values for age and gender [21]. The ventricular volumes, RV and LV function and extent of contrast enhancement, were analysed off-line on a dedicated workstation (Extended MR WorkSpace Release 2.6.3.2, Philips Medical Systems, Best, the Netherland) by 2 CMR-experienced operators blinded to clinical DMDc status.

After testing for normal distribution based on standard parameters, intergroup differences and relations were compared by analysis of variance (ANOVA), correlation matrices and/or unpaired Wilcoxon t-test. NCSS software version 2007 (http://​www.​ncss.​com) was used. A value of p < 0.05 was considered statistically significant.

Table 2 shows LGE data in asy and sym carriers: 13 asy-DMDc were LGE negative and 8 were LGE positive. In one young sym-DMDc (ID 1, 20 year-old) who was moderately affected, subepicardial plus mid-miocardial LGE were seen. The average RV and LV EF values were normal, both in DMDc and controls (Table 3). No evidence of myocardial edema was seen in DMDc. LGE positive segments were as follows: 4 basal inferior, 5 basal infero-lateral, 6 basal antero-lateral, 10 mid inferior, 11 mid infero-lateral and 12 mid antero-lateral. Absolute LGE quantity was higher in sym-DMDc than asy-DMDc (1.77 ± 0.27 vs 0.76 ± 0.17 ml, F = 19.6, p < 0.0001). LGE percentage was also higher in sym-DMCc than in asy-DMDc (1.86 ± 0.26 vs 0.68 ± 0.17%, F = 22.1, p < 0.0001) (Table 4). There was no correlation between LGE presence and age. LGE-involved LV segments were most frequently segments 5 and 6. Segment 5 was involved in all LGE positive DMDc. In asy-DMDc (Figure 1) and sym-DMDc, subepicardial LGE was more prevalent than mid-myocardial one. In 5 DMDc (2 asy and 3 sym) (Table 2), LGE was subepicardial plus midmyocardial (Figure 2). There were statistically significant LGE distribution differences in sym-DMDc vs asy-DMDc (Table 4): in the formers, midmyocardial LGE was significantly (0.56 ± 0.09 vs 14 ± 0.05, F = 15.1, p < 0.0001) more frequent than subepicardial LGE. Subendocardial LGE was never seen. RV was LGE negative in all DMDc. All controls were LGE negative in RV and LV.

Dystrophin, genomic analysis and dystrophin gene mutations study. We identified 16 different exons deletions of the dystrophin gene, two duplications and a nonsense mutation (Table 1), in agreement with the known allelic heterogeneity occurring in distrophinopathies [22]. All deletions/duplications were out-of-frame, predicting protein truncations. In two DMDc (ID 4 and 14), a deletion removing the muscle promoter region was identified, nulling transcription of the in cis dystrophin allele. The majority of the identified mutations were located in the domain II spectrin-like repeats (exons 10–60, 24 patients), clustered in the proximal rod domain, (exons 10–44, 4 patients). The exon’s study, the mutation type analysis and the affected gene region and protein domain study, did not show correlation between the cardiac phenotype and genotype.

A major limitation of our CMR technique was that T1 mapping for measuring diffuse myocardial fibrosis [32] was not performed. Current CMR LGE does not detect diffuse microscopic fibrosis. We have assumed that LGE was related to fibrosis, although no histological validation has been performed in our population. Moreover, we did not report data on troponin T to rule out the possible correlation between acute myocarditis and LGE. The cross-sectional nature of this study limits its ability to determine the precise mechanisms or rate of progression of fibrosis, because cause-and-effect cannot be defined. A larger, prospectively followed-up cohort, may give further information on additional causes of fibrosis, its prognostic value, and its potential response to medical therapy if any. In refining LV diastolic and systolic function as causative of myocardial LGE presence, we did not perform Tissue Doppler Imaging (TDI). In this regard, both E/E’ ratio and Sa peak early velocity study, in the setting of normal diastolic and systolic functions, may not be useful in familial cardiomyopathy carriers, before they develop full phenotype [33]. Moreover, neither were data by CMR strain and strain rate measured which may have provided additive insight.

The myocardium in DMDc is a cell milieu where LV dysfunction might be virtually present. We have shown that LGE is frequently found, it is not correlated with age and gene deletion type, and in absolute and relative terms it is modest. However, subepicardial LGE is present yet at an early age. We speculate that when LGE extends from basal infero-lateral subepicardial level to midmyocardium, cardiac involvement might be advanced. Presently, we ignore whether subepicardial or midwall LGE may progress towards other myocardial segments, and how long is needed and/or whether this potential progression may in some way influence DCM development. The association between LGE presence and quantities and the development of LV dysfunction while time passes, remain to be investigated. Based on the present study, the largest conducted thus far, CMR may be useful in DMDc to define the starting LV segment, an information which may not be obtained by ECG, conventional Doppler or 2D-echocardiography.