Orthopedic outcomes of long-term daily corticosteroid treatment in Duchenne muscular dystrophy

Duchenne muscular dystrophy (DMD) is an X-linked recessive disorder that occurs in 1 in 3,500 live male births.¹ The disease presents in early childhood and is rapidly progressive, with most boys losing the ability to walk between ages 9 and 11.² Although the disease is known to be caused by deficiency of the protein dystrophin, there is no cure for this disorder. Despite multiple clinical trials since the 1960s, the only medications proven to alter the natural history of DMD in randomized controlled trials are corticosteroids.³ Multiple clinical trials have demonstrated that both prednisone and deflazacort, another corticosteroid, improve strength and function and prolong ambulation for up to 3 years.^4–15

The side effects of chronic steroid use are well documented, but few studies have examined orthopedic effects of long-term daily corticosteroid treatment in DMD.^{4,6,11,15–19} As a member of the Clinical Investigation of Duchenne Dystrophy (CIDD) group, the Ohio State University (OSU) has been involved in DMD steroid trials since 1983. OSU serves a large geographic area as a tertiary neuromuscular clinic and consequently has an extensive population of both steroid- and non-steroid-treated patients with DMD. In this study, we sought to document the clinical orthopedic effects of chronic daily corticosteroid treatment in boys with DMD.

METHODS

The charts of all patients with genetically confirmed dystrophinopathies (Duchenne or Becker dystrophy) followed at OSU from 2000 to 2003 were reviewed. Patients diagnosed as having Becker dystrophy by phenotype (independent ambulation after age 15)²⁰ or dystrophin gene analysis (i.e., in-frame deletion) were excluded from the analysis. Charts were reviewed with specific attention to demographic information, type and duration of steroid therapy (if any), ambulation status, arm and leg functional grades,²¹ and specific orthopedic complications including the frequency and location of long bone fractures, vertebral compression fractures, and the presence and degree of scoliosis. Patients were grouped according to whether or not they had ever been treated with corticosteroids for at least 1 year. For this analysis, patients on prednisone (n = 36; dose 0.75 mg/kg/day as single oral morning dose) or an equivalent dose of deflazacort^22,23 (0.9 mg/kg/day; n = 25) were analyzed together in the treated group; 14 patients had been on both drugs. Two boys who had received brief steroid trials at a subtherapeutic dose (less than 6 months of 5 mg total prednisone daily) were assigned to the nontreated group. Only those boys age 9 and older were included in the scoliosis and compression fracture analyses, as that is the age at which annual scoliosis series are performed in our clinic.

Height was obtained using a calibrated height meter on ambulatory subjects. If a subject could not stand, height was measured on the right side of the subject while he was lying on the examination table with the spine, hip joint, and knee joint straightened out as much as possible, in joint segments, with the median of three measures recorded.

Independent ambulation was defined as functional walking without orthoses or any assistive device. If a patient had ceased independent ambulation prior to beginning steroid therapy, he was assigned to the nontreated group for ambulation. Similarly, a boy who developed a scoliotic curve (defined as a structural curve read as greater than 10° by the interpreting radiologist) prior to steroid initiation was assigned to the nontreated group for scoliosis. Bone densities were also examined by the same evaluator on a subgroup of boys through dual energy x-ray absorptiometry (DEXA) using Lunar Prodigy in standard mode. The presence and type of fracture, including vertebral compression fractures, were determined by the interpreting bone radiologist using accepted criteria.

SAS JMP version 5 (SAS Institute, Cary, NC) was used for the summary statistics and statistical analyses. Summary statistics are reported as means ± SD. Two group comparisons were based on two-sample t tests for the means (with the assumption of equal or unequal variances as appropriate) and Fisher exact test for the proportions. The comparisons of the risk rates were based on likelihood ratio test applicable to Poisson processes. The reported p values correspond to two-tailed tests, and the significance level was set at 0.05.

RESULTS

Scoliosis.

The mean degrees of scoliosis for the nontreated group was 33.15 ± 29.98 vs 11.58 ± 15.65° for the treated group (p < 0.0001; figure). Ninety-one percent of the nontreated group age 9 and older had a structural scoliosis curve of 10° or more compared with only 31% in the treated group (p < 0.0001). Thirteen boys (29%) in the nontreated group age 9 or older had undergone scoliosis surgery vs 10 (15%) in the treated group. Five additional youths in the nontreated group had been referred for scoliosis surgery but their parents refused the procedure. Of the 10 young men in the treated group who had scoliosis surgery, 8 had ceased corticosteroid treatment by the time the spinal curve was discovered. The average time between the discontinuation of steroid use by these eight boys and the discovery of a scoliotic curve of at least 35° was 17.3 months (range 6 to 24 months). This subgroup of boys averaged 5.1 years (range 1 to 10 years) of steroid treatment. The average age at which they discontinued steroid treatment was 12.6 years (range 9 to 16 years).

• Download figure
• Open in new tab
• Download powerpoint

Figure Degrees of scoliosis by cohort

The mean degrees of scoliosis for the untreated group was 33.15 ± 29.98 vs 11.58 ± 15.65° for the treated group (p ≤ 0.0001). Ninety-one percent of the nontreated group age 9 or older had a scoliosis curve of 10° or more, whereas among the treated group the incidence was 31%.

DISCUSSION

Multiple studies have shown that corticosteroids improve strength and function in DMD.^{3,7,8,10,11,16} The analysis presented here supports a clinically meaningful correlate, that is, prolonged independent ambulation by 3.3 years in steroid-treated DMD boys vs patients not treated with steroids. The data are even more compelling considering that the treated group weighed an average of 13.9 kg more than the nontreated group. Another reflection in this study of the benefit of steroids is the fact that although the mean age of the treated group was 2.5 years older than the untreated group, the two groups had the same upper extremity functional grade. This grading reflects preserved gross muscle strength in the treated patients, who at a mean age of almost 17 years could still lift a 240-mL glass of water to their mouth.

This unequivocal benefit, however, comes at a cost of the side effects related to steroids. To make an informed decision, patients and parents have to balance the benefits and risks of steroid treatment. Most efficacy studies performed to date^3,7,8,10 have reported short-term results, and the cardiac and orthopedic effects have been unclear. We found differences between corticosteroid-treated and -nontreated DMD patients for prevalence and severity of scoliosis, frequency of vertebral compression fractures, and risk and distribution of long bone fractures.

In our population, long-term corticosteroid treatment also resulted in fewer boys with DMD with scoliosis and milder curves if present. Similar results have been reported in DMD boys without steroid treatment.^27,28 Furthermore, nearly three times as many young men were referred for scoliosis surgery in the nontreated group as in the treated group (40 vs 15%). Of the 10 boys in the treated group who were referred for scoliosis surgery, 8 had discontinued steroid treatment prior to developing scoliosis. These eight boys developed a scoliotic curve of at least 35° (a curve usually considered surgically remediable)^2,29 within 18 months after discontinuing corticosteroids. Taken together, these findings strongly suggest that steroid therapy protects against the development of scoliosis in DMD. This may simply reflect stronger axial and trunk muscles resulting from steroid treatment, although prolonged ambulation might also play a role. Whether corticosteroids truly prevent scoliosis or simply delay its onset remains to be determined.³⁰ Forced vital capacity is also improved with corticosteroids,^23,31 and this function is primarily dependent on the muscles of respiration within the trunk area.

Approximately one-third of corticosteroid-treated patients experienced a vertebral compression fracture in this cohort, virtually identical to that reported previously.³² No compression fractures were found in the untreated boys, again similar to prior statements.^33,34 Despite the prevalence of vertebral fractures in steroid-treated boys, it is worth noting that this complication was not a motive to discontinue treatment. At least 80% of these fractures were discovered incidentally during routine scoliosis screenings and not because of patient complaints.

A higher percentage of treated boys experienced long bone fractures with a risk 2.6 times greater than no treatment. Based on the fracture rate per year, a corticosteroid-treated boy would have to be followed for 12 years to observe a fracture, whereas the similar figure for the steroid-naive boy would be 28.5 years. Several factors are probably operative in this increased fracture rate in steroid-treated boys, including prolonged independent ambulation and increased body weight. Several authors have documented an increased risk of long bone fractures when DMD boys are still ambulatory.^32,33,35

Within the general pediatric population, the distal bones of both the upper and the lower extremities are the most frequently fractured, with the distal radius accounting for nearly 25% of all fractures in children^36–38 Our data suggest, and others have also noted,^32,33,39 that these figures are reversed in DMD, because the femur was the most likely to be fractured in the steroid-treated group (28.3%). The humerus was more vulnerable in the nontreated group (25.4%). A provocative finding in our subanalysis of DEXA data was that all boys had profoundly small (i.e., narrow) bone area for height. Although the retrospective nature of these data and the small sample size make it difficult to draw definitive conclusions concerning the effects of steroids, it may be that small bone size in these patients may account for some of the risk of long bone fracture in this population. This possibility will need to be explored through further studies.

One surprising finding in this analysis, given the known mechanism of action of steroids and its effects on growth hormone, was the fact that the average height of the nontreated patients (140.3 cm) was actually slightly less than the height of the steroid-treated patients (140.3 vs 143.7 cm). Although this is almost assuredly due to the fact that the nontreated group included younger boys, it is important to note that both groups were extremely short in stature.

In two prior series, 13 to 21% of subjects ceased ambulation because of a fracture.^32,33 In our cohort, only one fracture (3.1%) in the nontreated group led to a cessation of ambulation, whereas 11 of 27 (40.7%) independently ambulating treated boys had a fracture that precipitated the loss of ambulation (at an average age of 14.9 years). This disparity results from lower limb fractures occurring in steroid-treated boys.

Because of the orthopedic implications of steroid therapy in this population, aggressive prophylactic measures are indicated as soon as therapy is initiated. At the time of this study, all boys who began steroid treatment were also prescribed calcium supplements, either as calcium carbonate 350 mg TID or a calcium tablet with vitamin D supplement (750 to 1,200 mg) daily. Currently, all boys who begin corticosteroid treatment at OSU have dietary counseling and a consult to the bone mineral metabolism laboratory for baseline DEXA scanning. The DEXA scan is then repeated every 6 months to 2 years as recommended by the consultant and as dictated by clinical circumstances. The role of bisphosphonates and other bone-sparing agents in this population remains to be determined.

Further studies are needed to better assess the development of osteoporosis in steroid-treated and untreated boys with DMD and how this complication might best be prevented or treated. To define steroid bone complications in Duchenne, data should be comparative for age, calcium and vitamin D supplementation, age at initiation of treatment, and standardization of methodology, especially the inclusion of bone mineral content corrected for area, height for age, and bone area for height. Meeting this type of data collection head on requires cooperation between centers and collaborations with experts on bone mineralization. In the absence of such data, it is impossible to interpret the recommendations of dose regimens^40–42 attempting to combat the potential complications of steroid-induced osteoporosis in a disease with reduced muscle mass and small (narrow) bone area for height.

ACKNOWLEDGMENT

The authors thank the patients and their families for their participation. The authors also thank the Muscular Dystrophy Association.