Background
Duchenne muscular dystrophy (DMD) is a genetic neuromuscular disorder caused by mutations in the dystrophin gene. Clinical symptoms of DMD include progressive muscle weakness in early childhood resulting in the loss of independent ambulation (LOA) before the age of 12 [
1]. Often, in Becker muscular dystrophy (BMD), which is the mild allelic form with in-frame mutations of the dystrophin gene, LOA occurs after the age of 16 [
1]. With the increased availability of the standard of care and glucocorticoid (GC) treatments, the average age at LOA in DMD has been delayed until after the age of 14 [
2].
The recent discoveries of new therapeutic approaches to DMD have resulted in several clinical trials being conducted globally, participated by Chinese patients. We built a national registry in collaboration with the Treatment of Neuromuscular Diseases (TREAT-NMD) Network in 2012 (
www.dmd-registry.com) and reported the initial genetic characterization of dystrophinopathies in China [
3].
In this study, we recruited a large cohort and conducted more detailed genetic analyses on patients from nine neuromuscular centers in China. We performed detailed clinical studies of different genotypes, including those that are currently, or predicted to be, amenable to exon-skipping therapies. We compared the ages at LOA and the duration of survival to provide data for designing future clinical trials.
Discussion
Natural history plays a paramount role in rare disease clinical trials and drug development. The United States Food and Drug Administration recommended that in single-arm interventional trials for rare diseases, natural history studies should be used as an external control [
4]. In the current study, we described the genotype distribution of a cohort of Chinese patients with DMD and their responses to GC treatment. The clinical progression of the different genotype groups was measured by age at LOA.
Although detecting disease onset is an important part of the natural history of DMD, measuring the precise age at symptom onset is difficult because the delay in motor development is insidious and the symptoms progress slowly during the first few years of life. Approximately half the patients in this study who were diagnosed before the age of 3 years had high CK and/or elevated transaminases during their routine medical examination for kindergarten enrollment, which is a widespread practice in China. The average age at diagnosis of DMD is approximately 4 to 5 years worldwide [
5‐
7]. The mean age at diagnosis was 4.59 years in our study, which is similar to the age of 4.43 years reported by the Parent Project Muscular Dystrophy [
6]. The mean age at diagnosis in Italy was 3.4 years overall [
8] and 3.5 years in patients born in or after 2010 in our study. These data indicated that patients born more recently tend to be diagnosed earlier, possibly because of the increased awareness of DMD and the availability of diagnostic tools. Several provinces in China have now included DMD in newborn screening to provide earlier diagnosis [
9].
The distribution of DMD genotypes is similar worldwide. Our study showed that 75.9% of the probands had out-of-frame large deletions and duplications (multi-exon or single exon) and 24.1% had small mutations. These data corroborated the results of the TREAT-NMD DMD Global Database (80% large mutations and 20% small mutations) [
10], which agrees with our previously finding [
3]. The proportion of small mutations was slightly higher in this study, which could have been caused by the increased recruitment efforts to implement an ongoing clinical trial in China (PTC124-GD-041-DMD) studying patients with nonsense mutations in the dystrophin gene.
The benefits of GC treatment shown on the time function test, muscle strength, and forced vital capacity of patients with DMD were first reported in 1974 [
11]. Further studies have confirmed the advantage of this treatment [
12], and it has been recommended as a part of the standard care for DMD [
13,
14]. As previously reported, treatment with deflazacort delayed the age at LOA by 1 year compared with treatment with prednisone/prednisolone [
2,
15]. In the current study, we observed a nearly 2-year delay in LOA in the deflazacort group compared with the prednisone/prednisolone group. This is similar to the results of the NorthStar Database [
16,
17].
Correlations between age at LOA and DMD genotype subgroups have been reported by previous studies [
18,
19]. Exon 3–7 and single exon 45 deletions (amenable to exon 8 and exon 44 skipping) maintained longer ambulation than the other mutation subgroups. The milder phenotype was proposed to be due to endogenous exon skipping resulting in retaining partial dystrophin [
20‐
22]. In our study, the age at LOA for those with deletions eligible for exon 44 skipping and those with nonsense mutations was significantly greater than that of patients in the “other deletions” subgroup. No patient with exon 3–7 deletion lost independent ambulation during the study period. Patients in the “deletions amenable to exon 45, 51, and 53 skipping” subgroups were younger at LOA, as also previously reported [
23]. Therefore, genotype differences clearly influenced the clinical progression of DMD. Three mutation subgroups, i.e., the “nonsense mutations,” “exons 3–7 deletion,” and “deletions amenable to exon 44 skipping” had significantly milder clinical progression than the other subgroups. Some of these patients had such mild phenotype that they were excluded from the current study because of their “atypical” phenotype at the initial recruitment.
GC treatment has been proven beneficial for DMD. Our current study showed that all genotype subgroups benefited from GC treatment by significantly delaying LOA. Treatment with deflazacort appears to be more beneficial in delaying LOA compared with prednisone in all genotypes. However, it is possible that different mutation subgroups could show differential HR to GC treatment. Compared with other genotypes, deletions amenable to skipping exon 44 had a lower hazard ratio, which may indicate a stronger protective effect of GC treatments on this subgroup. Subgroups with severe phenotypes, such as the “deletions amenable to skipping of exon 45 and exon 51” subgroups, had a higher HR than the “deletions amenable to skipping of exon 44” regarding the response to GC treatments, which might be due to the rapid disease progression from an early age and the lack of sufficient time for the GC treatment to work. However, more evidence is needed to confirm this speculation.
With improvements in medical technology and the implementation of the standard of care, most DMD individuals are now living into their 30s and even 40s in Japan and some European countries [
14,
24,
25]. Multidisciplinary intervention, including pulmonary, orthopedic, cardiac, and rehabilitation care, is currently the most effective way to enable long-term survival for patients with DMD. Currently, cultural and economic barriers have prevented parents from accepting the use of a ventilator for patients with DMD in China. Among the 290 non-ambulatory patients in our study, only two patients used intermittent nocturnal assisted ventilation. This reluctance to use assisted ventilation, coupled with the systemic difficulty in implementing the standard of care, may have contributed to the shorter survival of our study cohort. Since 2015, we have initiated a “one city, one doctor” project to promote awareness and implement the standard of care for patients with DMD in China [
26]. We have seen improvements in the long-term care and survival of this group of patients.
Methods
Protocol approvals, registrations, and consent
This study was approved by our institutional ethics committee (No. 2015003). Nine major neuromuscular centers participated in this study (see author affiliations). Participants were recruited between March 1, 2015, and August 31, 2019. They were enrolled and followed up every year. Written or online informed consent was obtained from patients or their legal guardians. All data were entered and analyzed anonymously.
Recruitment criteria
Inclusion criteria were confirmed DMD diagnosis by genetic analysis and no participation in any clinical trial. The patients had to have shown phenotypic evidence of DMD before the age of 5, including progressive muscle weakness (proximal > distal), Gowers' sign, calf pseudohypertrophy, characteristic waddling gait, and elevated serum creatine kinase (CK).
Two groups of patients were excluded from this study since they had atypical clinical phenotypes, i.e., those who did not receive GC treatment and were able to maintain independent ambulation beyond the age of 12 and those who received GC treatment and were ambulatory beyond the age of 16. Participants who did not have valid follow-up data or could not be followed up for more than one year were also excluded.
Genotype assignments
Patients were categorized according to their genetic mutations: Large deletions (equal to or larger than one exon length), large duplications (equal to or larger than one exon length), and small mutations (including small deletions/insertions, single base-pair mutations such as missense/nonsense mutations and splice site mutations). For DMD mutation studies, single- or multi-exon out-of-frame deletions and duplications were confirmed using multiplex ligation-dependent probe amplification; then, second-generation sequencing was performed to identify small mutations. The genetic mutation results were confirmed by submission of a genetic report to the Duchenne Registry genetic counselors. Large deletions were further divided into deletions amenable to exon-skipping therapies at the exon 44, 45, 51, and 53 sites and to a group with exon 3–7 deletions. A patient with a single exon 52 deletion, amenable to skipping of both exon 51 and 53, was assigned to the “deletions amenable to exon 53 skipping” subgroup because of the emergence of golodirsen and viltolarsen treatments. Large deletions that could not be assigned to the aforementioned subgroups were included in the “other deletions” subgroup.
Monitoring the clinical progression of different genotypes
Participants were evaluated at outpatient visits every year. Those who had difficulty attending outpatient visits underwent follow-up by telephone or e-mail. We measured clinical progression by recording the age at LOA and the age at the time of death (length of survival). LOA was defined as a need for continuous wheelchair use. We collected these data from the patients' guardians during regular follow-ups.
Glucocorticoid treatments and clinical responses
In China, GC treatments for DMD usually start when the patient is between the age of 4 and 6 years before the rapid decline phase starts. Participants aged 5 years or older were assigned into three groups according to their exposure to GC: The GC-treated group, including those who received continuous GC treatment for 12 months or longer; GC-naïve group, including those who never received GC treatment or were only treated for less than 1 month; Intermittent group, including those with GC exposure that lasted longer than one month but less than 12 months. The patients in the GC group were further grouped into those who received prednisone/prednisolone and those who received deflazacort. Patients who switched between deflazacort and prednisone/prednisolone were assigned according to which treatment lasted longer. Clinical responses to these treatments among the different genotypes were measured by the age at LOA and length of survival.
Statistical analyses
Time-to-event analyses of LOA were performed to analyze the differences between groups with age (years) as the time variable and LOA as the event. The median age at LOA and corresponding 95% confidence intervals (CIs) were estimated by plotting empirical Kaplan–Meier curves for each group defined by the mutation type and by GC treatment administered while the patient was ambulatory. Cox proportional hazard models were used to estimate and compare the age-related risks of LOA. Covariates, including DMD mutations and GC drug (deflazacort or prednisone/prednisolone) treatments, were recorded. The significance level was set at p < 0.05. All analyses were conducted using the SPSS software package (SPSS 20 Inc., Chicago, IL, USA).
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.