Development of Contractures in DMD in Relation to MRI-Determined Muscle Quality and Ambulatory Function

Abstract

Background:

Joint contractures are common in boys and men with Duchenne muscular dystrophy (DMD), and management of contractures is an important part of care. The optimal methods to prevent and treat contractures are controversial, and the natural history of contracture development is understudied in glucocorticoid treated individuals at joints beyond the ankle.

Objective:

To describe the development of contractures over time in a large cohort of individuals with DMD in relation to ambulatory ability, functional performance, and muscle quality measured using magnetic resonance imaging (MRI) and spectroscopy (MRS).

Methods:

In this longitudinal study, range of motion (ROM) was measured annually at the hip, knee, and ankle, and at the elbow, forearm, and wrist at a subset of visits. Ambulatory function (10 meter walk/run and 6 minute walk test) and MR-determined muscle quality (transverse relaxation time (T₂) and fat fraction) were measured at each visit.

Results:

In 178 boys with DMD, contracture prevalence and severity increased with age. Among ambulatory participants, more severe contractures (defined as greater loss of ROM) were significantly associated with worse ambulatory function, and across all participants, more severe contractures significantly associated with higher MRI T₂ or MRS FF (ρ: 0.40–0.61 in the lower extremity; 0.20–0.47 in the upper extremity). Agonist/antagonist differences in MRI T₂ were not strong predictors of ROM.

Conclusions:

Contracture severity increases with disease progression (increasing age and muscle involvement and decreasing functional ability), but is only moderately predicted by muscle fatty infiltration and MRI T₂, suggesting that other changes in the muscle, tendon, or joint contribute meaningfully to contracture formation in DMD.

INTRODUCTION

Prior to and following loss of ambulation, boys and men with Duchenne muscular dystrophy (DMD) experience gradual loss of passive range of motion (ROM) related to muscle replacement with fat and connective tissue, prolonged static joint positioning, and muscle weakness [1]. While ankle plantar flexion contractures (reduced ROM) are the first to develop, other contractures including knee flexion, hip flexion, hip abduction, elbow flexion, forearm supination, and wrist flexion contractures are common in DMD [1, 2]. The development of contractures may contribute to impaired functional ability, particularly in the upper extremity, and positioning to accommodate contractures may lead to pain or discomfort in other areas of the body.

Several management techniques are used to combat contracture formation in DMD. These include daily home stretching at the major joints, splinting at night, standing devices, serial casting, and surgical interventions to release stress on tendons and ligaments holding joints in set positions [3, 4]. However, there is limited evidence for or against their efficacy, and optimal treatment of contractures is an area of controversy [5, 6]. A clear understanding of the pattern of contracture development is important for development and testing of contracture prevention and management strategies in DMD.

The development of contractures with disease progression in DMD has been well documented over several decades, [2, 5, 7–11] but many of these reports predate the widespread use of corticosteroids to treat DMD [12]. The ImagingDMD study includes a large cohort of predominantly corticosteroid-treated individuals with DMD who have participated in up to 10 years of annual ROM assessment as well as magnetic resonance imaging (MRI) of the limb muscles and tests of functional ability. Specifically, the ImagingDMD study has included measurement of fat fraction (FF) by magnetic resonance spectroscopy (MRS) and measurement of muscle transverse relaxation time (T₂) by MRI. MRI T₂ captures the behavior of water and fat molecules in the magnetic field. It is elevated by inflammation, edema, muscle damage, and cell necrosis, which increase the T2 of water within the muscle, as well as fatty infiltration, since intramuscular fat has a higher T₂ than muscle water [13–19]. In DMD, the trajectory of MRI T₂ closely mirrors the trajectory of muscle fat fraction as fat fraction increases, suggesting that progressive T₂ changes in DMD are largely driven by fatty infiltration [14]. We chose these measures due to their well-documented association with strength and function in DMD [20–22]. The objective of this investigation was to describe contracture formation in the ImagingDMD cohort and to examine the relationships between contracture formation, functional ability, and muscle quality measured by MRI and MRS. We hypothesized that contracture development would be associated with declining muscle quality and with decreasing ambulatory function.

MATERIALS AND METHODS

Study design

Males with DMD (ages 4–18 years at enrollment) participated in this longitudinal, observational study of MRI (MR) biomarkers and clinical natural history. Data included in this manuscript were collected between 2010 and 2018. Participants visited one of three study sites annually for functional assessment and MR examination, including evaluation of lower extremity ROM. In 2016, the evaluation protocol was expanded to include upper extremity ROM and MR measurements. Additionally, unaffected controls participated at a single timepoint for comparison. The study was approved by the institutional review boards at the University of Florida, Children’s Hospital of Philadelphia, Shriner’s Hospital for Children – Portland, and Oregon Health & Science University, and it complied with local, national, and international guidelines for human subjects research. Parents of each participant provided written informed consent, while the participants provided informed assent. Once a participant turned 18, he provided written informed consent himself. All participants with DMD in the ImagingDMD study have a genetically-confirmed dystrophinopathy with a clinical diagnosis of DMD, and onset of symptoms prior to the age of 5.

Measurement of ROM

Measurement of ROM and functional testing was performed by a total of nine trained functional evaluators. Seven functional evaluators were physical therapists, and two evaluators had education in exercise testing and measurement and were trained to perform the assessments in this study. Written manuals of procedures for the study tests were established, and all evaluators completed training prior to their first measurement for this study. Evaluators were certified based on satisfactorily demonstrating each test, using a checklist of criteria for conduct of study tests according to the standardized procedures of the study.

Passive ROM was measured at three lower and three upper extremity joints with values recorded to the nearest 5° [23]. ROM was evaluated in the right lower limb unless a prior fracture or injury had occurred, and it was evaluated in the upper limb the participant used for writing. Lower extremity ROM was measured with the participant in supine, and upper extremity ROM was measured with the participant seated. At all joints, the evaluator applied firm pressure to reach the end of the ROM that could be attained without pain. An appropriately sized standard plastic goniometer was aligned with the joint axis and segments as described below.

Hip extension was evaluated using the modified Thomas test position [24]. The subject was positioned in supine on a plinth with the table supporting the trunk to the level of the hip joint and both legs passively flexed by the evaluator to decrease the lordotic curvature of the lumbar spine. While maintaining left leg flexion, the right hip was allowed to extend passively. The axis of the goniometer was aligned with the greater trochanter of the femur, the stationary arm parallel to the trunk, and the moving arm in line with the lateral epicondyle of the femur. A measurement was made when resistance was felt or when the pelvis became unstable. The right leg was not allowed to externally rotate or abduct. A neutral position was considered 0°, with positive values for hip extension and negative values for hip flexion.

For passive knee extension, a pad was placed under the distal lower leg and the axis of the goniometer was placed on the femoral lateral epicondyle, with the stationary arm of the goniometer lateral and parallel to the mid-line of the femur aligning the greater trochanter and the moving arm of the goniometer lateral and parallel to the fibula aligning with the lateral malleolus. Gentle over pressure was applied, and full knee extension ROM was recorded as 0° with hyperextension indicated with a positive value and flexion indicated with a negative value.

Ankle dorsiflexion was evaluated in supine with the knee fully extended or extended to end range. The goniometer was placed with the stationary arm parallel to the longitudinal axis of the lower leg. The moving arm was held parallel to the plantar surface of the forefoot along the fifth metatarsal. The examiner’s hand was placed over the plantar surface of the foot with pressure applied to dorsiflex the ankle to end range while preventing inversion/eversion or supination/pronation of the foot. A neutral position at the ankle was recorded as 0°. Negative values corresponded to plantarflexion, while positive values corresponded to dorsiflexion.

Elbow extension was assessed in sitting with the humerus stabilized at the trunk and forearm in full supination. The axis of the goniometer was placed at the lateral epicondyle. The stationary arm of the goniometer was parallel with the humerus, and the moving arm of the goniometer was parallel with the radius. Full extension was recorded as 0° with positive values for elbow hyperextension and negative values for elbow flexion.

Forearm supination was also evaluated in sitting with the elbow flexed to 90° and the humerus stabilized at the trunk. The axis of the goniometer was placed just medial and posterior to the ulnar styloid process. The stationary arm of the goniometer was parallel with the humerus, and the moving arm of the goniometer was parallel to the anterior surface of the forearm. Supination, from a neutral position, was recorded as a positive value. Finally, wrist extension was evaluated with the elbow flexed to 90° and forearm pronated and supported by a table. The axis of the goniometer was placed over the triquetrum. The stationary arm of the goniometer was parallel with the ulna, and the moving arm of the goniometer was parallel with the 5th metacarpal. A neutral position was defined as 0°, with positive values for extension and negative values for flexion.

Functional measures

In addition to ROM assessment, at each visit, subjects participated in tests of ambulatory function. Training of functional evaluators, as well as detailed procedures for carrying out each functional test, have been described previously [25]. Reported here are results from the six minute walk test (6MWT) and 10 meter walk/run test. The 6MWT assesses the distance walked in six minutes on a 25-meter course, and the 10 m walk/run test measures how long it takes a participant to traverse 10 meters as fast as safely possible. Loss of ambulation was defined as the first visit at which the participant was unable to complete the 10 m walk/run unassisted and within 45 seconds.

Magnetic Resonance Imaging

Prior to ROM and functional testing, MRI of the lower and upper limb muscles was performed. In most cases this was performed on the same day, although in some cases the MRI was performed the day prior to functional evaluation. The leg and arm used for ROM measurement were also used for MRI acquisition. The MRI acquisition, analysis, and quality control procedures for this study have been extensively described in prior publications [26, 27]. Briefly, subjects were positioned in supine in the bore of a 3T MR scanner (Philips Achieva Quasar Dual 3T, Siemens Magnetom Verio 3T, and Siemens Magnetom TIM Trio 3T). Multiecho T₂ weighted images (TR = 3000 ms, TE = 20, 40, 60, 80, 100 ms) were collected at the fullest part of the calf, mid-thigh, mid-upper arm, and shoulder. T₂ maps were reconstructed excluding the 20 ms TE image, and muscle boundaries were manually defined on three slices at an internal anatomic landmark predefined for each segment. The mean T₂ value for the vastus lateralis (VL), biceps femoris long head (BFLH), medial gastrocnemius (MG), peroneal (PER), soleus (SOL), tibialis anterior (TA), biceps brachii (BB), triceps brachii (TB), and deltoid (DEL) was calculated as the average T₂ of all pixels across three slices for that muscle. Single-voxel ¹H-MRS was used to calculate relaxation-corrected fat fraction values in the SOL, VL, BB, and DEL as previously described [26, 28, 29]. The progression of these measurements have been previously reported [14, 21, 29, 30].

Data analysis

For participants with DMD, contractures were defined based on the interquartile range of the control values. Specifically, ROM values that were lower than the 1^st quartile value for ROM at each joint in the cohort of control participants were considered contractures. Bonferroni-corrected nonparametric Wilcoxon rank-sum tests were used to compare the control group ROM to the group ROM at each annual window relative to the first nonambulatory visit. Graphpad Prism software (version 8.0) was used for this analysis. For other analyses, all available data were utilized, including multiple observations per participant. Treating multiple data points from the same individual as independent typically leads to invalid estimates of uncertainty. Thus, we quantified uncertainty using a bootstrapping approach, in which participants (not data points) were randomly resampled; we used 1000 bootstrap samples [34]. Nonparametric Spearman correlations were used to examine relationships between ROM and age in boys with DMD and unaffected controls, MRI-determined muscle involvement (MRI T₂), and the ratio of MRI T₂ values in antagonistic muscles across a joint. Statistical significance was determined by the 95% bootstrap confidence interval for rho excluding zero. To compare ROM in ambulatory and nonambulatory individuals, statistical significance was based on the 95% bootstrap confidence interval for the difference in group means excluding zero. To compare six minute walk distance and 10 m walk/run velocity in groups with differing contracture severity, statistical significance was based on the 95% bootstrap confidence interval for the difference in group means excluding zero, with a Bonferroni correction applied for multi-group comparisons. To account for the larger number of comparisons, a greater number of bootstrap resamples was performed (20,000). Knime Analytics Platform (version 4.4.1) was used for the bootstrap resampling.

RESULTS

For this study, 178 individuals with DMD were enrolled between September 2010 and July 2018. Because enrollment occurred across many years (2010 to 2018), the duration of participation was varied. There was an average of 5 visits per subject with a minimum of one study visit and a maximum of 7 annual follow-up visits. In total, 909 study visits were included in this data set, including study visits that were not within 0.2 years of the target annual visit date (“out of window” visits). Lower extremity ROM was measured at all visits, while upper extremity ROM (initiated in 2016) was measured in 128 individuals over 318 visits. Age, ambulatory status, and corticosteroid status of the DMD cohort are reported in Table 1. Among corticosteroid users, most reported daily corticosteroid use (92%). Deflazacort was the most commonly used steroid (75%), followed by prednisone (22%).

Table 1.

Participant Characteristics at each study visit

	Baseline	1 yr	2 yrs	3 yrs	4 yrs	5 yrs	6 yrs	7 yrs	Out-of window
Number of participants	178	150	127	98	87	56	51	29	133
Age	8.9 ± 2.9	9.9 ± 2.9	10.6 ± 2.6	11.3 ± 2.2	12.4 ± 2.2	13.6 ± 2.0	14.6 ± 2.0	15.4 ± 2.1	10.2 ± 3.7
Steroid use	39/136/3	16/131/3	13/109/5	7/85/6	6/79/2	4/52/0	3/47/1	2/25/2	31/93/9
Ambulatory (% of group)	96%	92%	85%	83%	70%	57%	49%	57%	76%

Steroids = off/on/unknown.

ROM measurements were also made in the lower extremity of 65 unaffected control participants and the upper extremity of 40 unaffected control participants. Age was not significantly associated with ROM at any joint in control participants (p > 0.05). Unaffected boys had ROM values consistent with control ranges reported in the literature [31]. The 25th percentile of control ROM values were as follows: ankle = 10°; knee = 0°; hip = 0°; elbow = 0°; wrist extension = 80°; forearm supination = 80°. Based on this data, contractures for the DMD cohort were defined as ROM ≤ 5° at the ankle, ≤−5° at the knee, ≤ −5° at the hip, ≤ −5° at the elbow, ≤75° at the wrist, and ≤ 75° at the forearm.

Characterization of Contractures in DMD

Contractures were highly prevalent in the ImagingDMD cohort. Boys with DMD had at least 5° of ankle ROM limitations at 78% of study visits, at the forearm at 52% of study visits, at the hip at 45% of study visits, and at the knee at 29% of study visits. Elbow and wrist extension limitations were less prevalent, with contractures seen at only 27% and 25% of study visits, respectively. Severe contractures (at least 20° of limitation) were seen at the ankle at 27% of visits, forearm at 25% of visits, hip at 11% of visits, knee at 10% of visits, elbow at 15% of visits, and wrist at 7% of visits. In this large, heterogeneous cohort, average ROM was −4 ± 12° at the hip, −4 ± 12° at the hip, −3 ± 16° at the ankle, −4 ± 12° at the elbow, 71 ± 14° at the forearm, and 82 ± 10° at the wrist. The frequency of contractures increased with age (Fig. 1). ROM was significantly associated with age at the ankle (ρ = −0.53 (bootstrap CI: −0.44 to −0.60)), knee (ρ = −0.45 (bootstrap CI: −0.37 to −0.52)), hip (ρ = −0.27 (bootstrap CI: −0.17 to −0.36)), elbow (ρ = −0.33 (bootstrap CI: −0.18 to −0.45)), forearm (ρ = −0.36 (bootstrap CI: −0.21 to −0.50)), and wrist (ρ = −0.29 (bootstrap CI: −0.17 to −0.40)).

Fig. 1.

Frequency of contractures equal to or exceeding 5°, and of contractures equal to or exceeding 20°, in boys with DMD across age groups. Age groups are collapsed into 2 year bins in the upper extremity because fewer upper extremity measurements are available; upper extremity measurements began 5 years after lower extremity measurements. Contractures were most prevalent at the ankle (E) and hip (A) in younger boys, and at the hip (A), knee (B), ankle (C) and forearm (D) in the teenage years.

Relationship between lower extremity contractures and functional ability

In both the lower and upper extremity, contractures were more prevalent in nonambulatory individuals. Figure 2 presents lower extremity ROM by age in ambulatory (left) and nonambulatory (right) participants. ROM at each joint was significantly lower in nonambulatory than ambulatory individuals as a group, based on the bootstrap 95% CI excluding 0. Once individuals were nonambulatory, it was rare to have normal or near-normal ROM, while ambulatory individuals most commonly had near-normal ROM, particularly at the knee. Progression of contractures was slow in the ambulatory phase of DMD but rapid as participants entered the nonambulatory phase (Fig. 3). ROM fell below control values earlier at the ankle than at the hip or the knee. In ambulatory boys, greater loss of ROM was associated with greater impairment in six minute walking distance and 10 m walk/run time (Fig. 4), although there was a wide range of performances seen in individuals with full ROM. There was no difference in 6MWT performance in groups with no contractures and mild contractures at the ankle. In ambulatory individuals, knee contractures ≥ 20° were not seen, hip contractures ≥ 20° were uncommon, and ankle contractures ≥ 20° were quite common.

Fig. 2.

Contractures were more prevalent, and more severe, in nonambulatory boys and young men (B, D, F, H, J, L) compared with ambulatory boys and men (A, C, E, G, I, K). The dashed line represents the median control value. Age was weakly correlated with knee, elbow, wrist, and forearm ROM and moderately correlated with ankle ROM in ambulatory subjects but was only weakly correlated with wrist ROM in nonambulatory boys and men.

Fig. 3.

ROM was significantly decreased from normal from 5 years prior to the first nonambulatory visit at the ankle (C), 2 years prior to the first nonambulatory visit at the knee (B) and 4 years prior to the first nonambulatory visit at the hip (A). Contractures progressed rapidly following loss of ambulation. (*significantly different from control subjects, p < 0.05).

Fig. 4.

Boys with greater ROM had better performance on the 6MWT and 10 m walk/run across all 3 lower extremity joints measured. However, there was a wide range of function in boys with full ROM. Notably, no boys were able to walk with 20 degree or greater contractures at the knee, while many boys maintained fairly good functional performance with 20 degree or greater contractures at the ankle. *Bonferroni-corrected bootstrap confidence intervals for the mean difference between groups exclude 0.

Relationship between MR measures of muscle quality and contractures

MRI T₂ and MRS FF were significantly correlated with ROM (Table 2). However, while contractures were correlated with increasing muscle involvement on a population level, Fig. 5 shows 2 example images of the calves of individuals with severe contractures despite well-preserved muscle, and 2 images of the calves of individuals with preserved ROM despite high levels of fatty infiltration in the muscle. Finally, the ratio of MRI T₂ in agonist and antagonist muscles for a movement was not predictive of ROM at the knee and was only weakly predictive (value) of ROM at the ankle (Fig. 6).

Table 2.

Spearman’s rho values for the relationship between muscle MRI T₂ or FF and ROM at each joint measured. All correlations were negative, indicating that higher MRI T₂ and FF values were associated with loss of ROM. All relationships were significant, based on the 95% confidence intervals for the bootstrap rho value excluding 0. Most correlations were moderate (0.4 to 0.6), although wrist extension was only weakly correlated with MR. The strongest correlation (rho = 0.61) was found between ankle ROM and VL MRI T₂ Lower Extremity

Lower Extremity
	Hip ROM	Knee ROM	Ankle ROM
VL MRI T2	−0.42 (−0.51 to −0.33)	−0.58 (−0.63 to −0.50)	−0.61 (−0.67 to −0.55)
BFLH MRI T2	−0.40 (−0.49 to −0.31)	−0.52 (−0.59 to −0.45)	−0.59 (−0.65 to −0.52)
MG MRI T2	−0.40 (−0.49 to −0.29)	−0.45 (−0.51 to −0.35)	−0.48 (−0.56 to −0.38)
SOL MRI T2	−0.40 (−0.49 to −0.30)	−0.52 (−0.57 to −0.42)	−0.49 (−0.57 to −0.39)
VL MRS FF	−0.38 (−0.48 to −0.28)	−0.55 (−0.60 to −0.47)	−0.59 (−0.66 to −0.51)
SOL MRS FF	−0.38 (−0.48 to −0.29)	−0.50 (−0.56 to −0.41)	−0.49 (−0.57 to −0.41)
Upper Extremity
	Elbow ROM	Forearm ROM	Wrist ROM
BB MRI T2	−0.47 (−0.60 to −0.30)	−0.44 (−0.57 to −0.29)	−0.24 (−0.39 to −0.07)
DEL MRI T2	−0.43 (−0.60 to −0.23)	−0.41 (−0.56 to −0.25)	−0.28 (−0.46 to −0.11)
TB MRI T2	−0.46 (−0.60 to −0.29)	−0.41 (−0.56 to −0.21)	−0.20 (−0.35 to −0.02)

^*VL: vastus lateralis; BFLH: biceps femoris long head; MG: medial gastrocnemius; SOL: soleus; BB: biceps brachii; DEL: deltoid; TB: triceps brachii.

Fig. 5.

Example T₁-weighted gradient echo images of the calves in individuals with similar ROM and different muscle quality. MRI T₂ values: Subject A: TA = 39.8 ms, SOL = 45.4 ms, ratio = 0.88; Subject B: TA = 71.7 ms, SOL = 78.8 ms, ratio = 0.91; TA = 33.4 ms, SOL = 38.7 ms, ratio = 0.86; D: TA = 57.3 ms, SOL = 60.9 ms, ratio = 0.94). The presence of pronounced contractures in individuals with relatively healthy muscle, and absence of contractures in individuals with substantial fatty infiltration in the lower leg muscles, point to the important role factors other than muscle fibrofatty infiltration may play in contracture development.

Fig. 6.

ROM was moderately correlated with MRI T₂ in a flexor and an extensor muscle at both the knee and ankle (A-D). However, the ratio of the MRI T₂ from an agonist and an antagonist muscle was not strongly correlated with maximum ROM at either the knee (E: no significant correlation) or the ankle (F: SOL: rho = 0.18; bootstrap 95%CI = 0.05 to 0.30; G: MG: rho = 0.18, bootstrap 95%CI = 0.06 to 0.30). Bootstrap 95% CI for panels A-D can be found in Table 2. For panels E-F, higher ratios indicate greater involvement in the anterior compartment; lower ratios indicate greater involvement in the posterior compartment.

DISCUSSION

This study documents the natural history of contracture development in a large, contemporary, and predominantly corticosteroid treated longitudinal natural history cohort. Physical therapists, physicians, and parents are eager to understand the causality and consequences of contracture development, to better treat and prevent contractures in DMD. This study provides important information for further research. Specifically, we have found, in keeping with other contemporary and historical reports, that while contractures are present in ambulatory individuals with DMD, loss of ambulation precipitates loss of ROM almost universally [2, 5, 7]. ROM was significantly less than control values 5 years prior to loss of ambulation at the ankle, 2 years prior to loss of ambulation at the knee, and 4 years prior to loss of ambulation at the hip. During the ambulatory phase of DMD, better ambulatory function was associated with greater ROM at the hip, knee, and ankle, although this does not imply a causal relationship between ROM and ambulatory ability. Similarly, increased muscle involvement (increased MRI T₂ and FF) was associated with decreased ROM in the upper and lower extremity.

Recent work has established clear relationships between age and ankle dorsiflexion ROM in a large cohort of ambulatory boys with DMD [9]. The current study supports these findings, with a correlation coefficient of −0.44 for the relationship between ankle ROM and age in ambulatory individuals. However, age was only weakly to moderately correlated with hip and knee ROM in ambulatory individuals, and age was not significantly correlated with ROM in nonambulatory individuals. Another study compared ROM in ambulatory and nonambulatory individuals at the hip, knee, and ankle, and found, in keeping with the results of this study, that ambulatory status was closely related to ROM [5]. In the context of historical literature, these results confirm that while widespread use of glucocorticoids in DMD have increased the age at loss of ambulation, the overall pattern of relatively slow contracture evolution prior to loss of ambulation and rapid contracture evolution following loss of ambulation has persisted [2, 32].

The importance of maintaining ROM to improve ambulatory performance is an area of considerable discussion in the literature. Traditionally, it has been thought that contractures are detrimental to functional performance. Specifically, contractures have been shown to be associated with worse performance on the North Star Ambulatory Assessment [9], and in the current study we have shown that contractures were associated with poorer performance on the 10 m walk/run and 6MWT. It is not well-understood, however, if increased lower extremity contractures are a causative factor in ambulatory functional decline or if both ROM loss and functional decline are simply changing in parallel as the disease progresses. Knee contractures > 15° were not seen in ambulatory participants, indicating either that these contractures are incompatible with ambulation in DMD, or that they do not develop until individuals are nonambulatory. Some researchers have hypothesized that ankle plantarflexion contractures may help to maintain ambulation in individuals with DMD, helping to compensate for proximal muscle weakness [33]. Potentially supporting this hypothesis, 6MWD did not differ between groups with no ankle plantarflexion contractures and mild ankle plantarflexion contractures (Fig. 4). However, further evidence is necessary to confirm or refute this intriguing idea. Well-designed future studies may help clarify the role that contractures have in altering ambulatory function.

Muscle disease involvement is hypothesized to be a contributor to contracture development in DMD [34]. The MR data from this study provides evidence that while contractures are more common in individuals with greater muscle disease involvement, MRI T₂ or fat fraction alone does not fully explain individual ROM. This supports a previous report in a small number of individuals with LGMD 2A, in whom muscle involvement based on grading of T1 weighted MR images did not differ in individuals with vs. without a contracted phenotype [35]. Table 2 shows that correlations between MRI T₂ and fat fraction and ROM were generally moderate in the upper and lower extremities, explaining considerably less than half of the variance in ROM. Muscles crossing the affected joint were not more strongly correlated with ROM than other muscles of the body – MR measurements in the VL muscle, which crosses only the knee joint, was the strongest predictor of contractures at the hip, knee and ankle. The confounding role of overall disease progression is critical to recognize in a progressive disease such as DMD, where multiple manifestations of more severe disease are likely to be correlated despite there being no rationale for a direct causal relationship (as for the relationship between VL muscle quality and ankle plantarflexion ROM).). It is important to acknowledge that MRI T₂ elevation is nonspecific, capturing fatty infiltration as well as physiological changes to muscle water, including muscle damage, edema, an inflammation, while MRS fat fraction reflects muscle fatty infiltration [13–19].

It has been hypothesized that an imbalance in involvement of agonist and antagonist muscles across a joint contributes to contracture development in DMD [34], which was not supported by this data set; there was a very weak correlation (rho = 0.18) between ankle ROM and the ratio of plantarflexor (soleus or medial gastrocnemius) to dorsiflexor (tibialis anterior) muscle quality, with more severe contractures seen in individuals whose fatty infiltration is greater in the plantarflexors than the dorsiflexors. However, the known progression rates of each muscle indicate that this ratio is likely to be associated with disease progression [14], and no relationship was found between knee ROM and the ratio of quadriceps (vastus lateralis) to hamstrings (biceps femoris) muscle quality. Example images from individuals with different degrees of muscle fatty infiltration and ROM illustrate the hypothesis that factors beyond muscle fatty infiltration may be very important in contracture development in DMD (Fig. 5). Thus, it may be necessary to investigate muscle properties such as sarcomere length and non-muscular factors including the ligaments, tendons, and capsular structures surrounding each affected joint to fully understand and treat or prevent contracture formation in DMD [36].

A limitation of this study is the limited insight in to dystrophic muscle pathology provided by MRI T₂ and fat fraction. Fatty infiltration can be accurately quantified using MRS, and captured using MRI T₂. However, fibrotic involvement, often associated with contracture formation, isn’t easily quantified using standard MR techniques. Although MR-visible skeletal muscle changes in muscular dystrophy are sometimes described as fibrofatty replacement, [37] the precise relationship between fatty and fibrotic infiltration has not been well-established. Additionally, other potential muscle mediators of contracture development such as changes to muscle architecture were not quantified in this study. A second limitation is sparse sampling of nonambulatory individuals, particularly those who have been nonambulatory for multiple years. The study was designed to focus on ambulatory individuals and track them across the transition to full-time wheelchair use. Thus, most data collected in nonambulatory individuals were collected relatively soon after this transition, and the data set does not represent the full spectrum of nonambulatory people with DMD.

Recommended orthopedic care for DMD includes contracture prevention measures such as active and passive stretching, the use of orthoses or braces to maintain joint positioning, and tendon lengthening surgery for some individuals [4]. A majority of participants in this study reported using these methods – specifically, 87% report stretching, 70% report using night braces, and 65% report regular physical therapy. Despite their widespread use, evidence for or against these interventions is limited and contradictory [5, 8, 38–44]. In a recent conference report, contraction prevention measures such as stretching were reported as highly burdensome by families [8]. More research is needed to establish the most effective and least burdensome ways to prevent contractures in DMD. A recent investigation of changes in motor function following travel restrictions due to the COVID-19 pandemic suggested that maintaining lifestyle physical activity and physical therapy access may be important in maintaining ankle joint ROM [45]. The data presented in this manuscript represents an important resource for planning interventional research to identify optimal methods of contracture prevention in DMD.

In conclusion, this study examined the development of contractures in a large population of ambulatory and nonambulatory boys and young men with DMD in relation to age, ambulatory status, ambulatory performance, and MRI-determined muscle quality. Contractures increased with age in this cohort, but the relationship with age is weak in ambulatory individuals, and nonexistent in nonambulatory individuals, suggesting that loss of ambulation and consequent static positioning in older participants is a primary driver of contracture development. Muscle MRI T₂, which is increased by both muscle inflammation/damage/edema and fatty infiltration, and muscle fat fraction are significantly correlated with ROM at joints throughout the body. However, limited strength and specificity in these relationships suggests muscle fatty infiltration is only one of multiple factors driving contracture development in DMD. likely contribute substantially to contracture development in DMD.