Natural history of 10-meter walk/run test performance in Spinal Muscular Atrophy: A longitudinal analysis

Abstract

As trials and treatments for Spinal Muscular Atrophy (SMA) rapidly evolve, understanding the natural history and potential utility of the 10-meter walk/run test (10MWRT) in ambulant individuals is critical. Study aims were to: 1) establish change over time and across age for 10MWRT time in an untreated natural history cohort of young, ambulatory participants with SMA and 2) identify relations between 10MWRT time and age, SMA type, SMN2 copy number and anthropometrics. Untreated individuals (n=56) age 2 to 21 years who were enrolled in a long-term natural history study between 2005 and 2014 and met inclusion criteria were included. Linear mixed effects models were used to assess changes in 10MWRT time with age and associations with SMA type, SMN2 copy number, and body mass. SMA type 3b (versus 3a), SMN2 copy number 4 (versus 3) and lower body mass were associated with faster 10MWRT. 10MWRT performance improved between 3 and 8 years of age, was stable between 9 and 10, and gradually declined from 11 to 18. Findings provide the first longitudinal natural history report of 10MWRT time in young individuals with SMA and offer a critical foundation for interpreting childhood change in short distance walking speed with pharmacologic treatment.

1. Background

Spinal Muscular Atrophy (SMA) is a monogenetic disorder caused by mutations in the survival motor neuron 1 gene (SMN1) which results in progressive loss of anterior horn cells with subsequent, progressive muscular atrophy and weakness [1]. SMA presents across a broad phenotypic spectrum defined by age of symptom onset and maximum motor milestone achievement [2–4]. Children presenting earlier and at the weaker end of the spectrum (types 1 and 2) may or may not achieve sitting, whereas those presenting later and at the stronger end of the spectrum (types 3 and 4) achieve the ability to stand and ambulate, although these abilities may be lost over time [5].

The natural history of motor function in untreated, ambulatory children and adolescents with SMA has been characterized using the Hammersmith Functional Motor Scale Expanded (HFMSE), the Motor Function Measure, Revised Upper Limb Module (RULM) and the Six Minute Walk Test (6MWT) [5–7]. These motor tests of function have demonstrated ability to monitor disease progression in natural history studies and the impact of treatment on those receiving disease modifying pharmacotherapies [6,8–11]. However, the natural history of shorter functional capacity tests such as timed function tests (TFTs), including the 10-meter walk/run (10MWRT), time to climb 4 stairs and time to rise from supine to stand, which have high relevance to everyday life for those who are ambulant, remain underexplored [7,11–13].

Recent advances in pharmacotherapy intervention have made achievement of independent walking by infants and young children with SMA types 1 and 2 possible with early treatment [14, 15]. The changing motor profile and characteristics of this treated population necessitate a renewed examination of the TFTs, including the 10MWRT. The 10MWRT is a robust measure of short distance walking ability with minimal endurance demands that is quick and easy to administer, poses minimal use burden in both clinic and research settings, and has demonstrated sensitivity to change in those with neuromuscular disorders [11,12,16–18]. 10MWRT norms for healthy children have been established [19–21] and 10MWRT natural history in other neuromuscular disorders, including boys with Duchenne muscular dystrophy (DMD), has been well described [18,22]. However, to our knowledge there are no published reports that characterize observed 10MWRT trajectories in SMA across childhood age groups. A clearer natural history of 10MWRT times is critical for interpretation of performance over time, especially in young children whose walking abilities are evolving. In those with SMA, the 10MWRT appears to discriminate between older and younger ambulatory patients [12] and correlates with both longer distance walking endurance measured by 6MWT distance [11] and with knee flexor and extensor strength [12].

As the therapeutic pipeline for SMA continues to evolve and additional treatments receive FDA approval, the availability of untreated individuals with SMA to study the natural history of additional measures of motor function is rapidly decreasing. TFTs provide an additional metric of ambulatory motor function that is relevant and meaningful, may be sensitive to change over time and may more clearly define the natural history of SMA to ensure we continue to utilize outcomes that capture the complexity of the disease in ambulatory cohorts. As such, examining motor outcomes that reflect everyday function for children and adolescents with SMA, such as walking time over shorter distances and during transitions, is increasingly important. Understanding motor trajectories of the 10MWRT in untreated ambulant natural history cohorts may lead to more meaningful interpretation of everyday function in these individuals and better delineate 10MWRT performance as a comparator for future performance of treated cohorts.

The aims of this study were to characterize 10MWRT performance in a cohort of ambulatory individuals with SMA from early childhood through adolescence using a longitudinal SMA natural history database and to identify potential factors (age, type, SMN2 copy number, body mass) associated with 10MWRT time.

2. Methods

2.1. Ethics Statement:

The natural history study was approved by the Institutional Review Board (IRB) at all participating sites in the U.S. (n=5) and by the Ethics Committees at international sites n=2, Canada, Argentina). Before inclusion, all patients or their parent(s)/legal guardian(s) provided written informed consent. All research was done in accord with the Helsinki Declaration of 1975. De-identified data for analyses were obtained from the Project Cure SMA Longitudinal Pediatric Data Repository (LDPR), initially approved by the University of Utah and currently maintained under the Partners Healthcare IRB for Massachusetts General Hospital [23].

2.2. Participants:

The Project Cure SMA LPDR [23], was queried for all ambulatory participants (< 21 years of age) who completed the 10MWRT between the years of 2005 and 2014, participated in natural history studies through the Project Cure network, and had not received nusinersen/Spinraza, AVXS-101/Zolgensma, or R07034067/Risdiplam (Figure 1). Participants had confirmed homozygous deletion of the SMN1 gene and were classified as SMA type 3a, 3b, type 4, or pre-symptomatic by neurologists with expertise in SMA. By definition, individuals with SMA type 3 had disease onset after 18 months of age and gain the ability to walk; those with 3a had symptom onset prior to age 3, those with 3b had symptom onset after age 3[2,24]. Individuals with SMA type 4 had symptom onset in the second or third decade of life, mild motor impairment, and were without gastrointestinal or respiratory impairment. Pre-symptomatic patients had genetic confirmation of disease, but no symptoms at the time of testing. Inclusion criteria were availability of 10MWRT data, age between 18 months and less than 21 years, and the ability to complete the 10MWRT independently without physical assistance. Exclusion criteria were age ≥ 21 years, use of an assistive device and use of shoes or lower extremity orthotics during the 10MWRT, inability to complete the 10MWRT in ≤ 3 minutes (test termination point), and use of any known pharmacologic disease modifying treatment. Participants were not excluded for use of Valproic Acid (VPA) or albuterol/salbutamol, as no disease modifying therapeutic benefits have been reported [25].

Figure 1.

Consort Diagram

2.3. Demographic information:

Participant’s height, weight, age, type of SMA (3a, 3b, 4 or pre-symptomatic) and survival motor neuron (SMN2) copy number, and ongoing use of VPA were recorded at each visit, which typically occurred every 4–6 months.

2.4. Outcome measures:

Erect height was measured to the nearest centimeter using a stadiometer. Body weight was collected to the nearest 0.2 kilogram (kg) on a standard floor scale. Body mass index (BMI; weight in kg/height in m²) was calculated from measured weight and height. Age and gender-specific BMI reference values (CDC, NHANES) were used to assign BMI percentiles to study participants [26].

The 10MWRT was administered by trained physical therapists with SMA expertise as part of a comprehensive motor testing battery in standardized order. Frequency of testing was every 4–6 months. Children were tested barefoot. Use of lower extremity orthotics, shoes, and assistive devices was not allowed. A 10-meter distance was marked on an unobstructed, flat surface using tape. To limit the impact of acceleration and deceleration on gait speed, start and finish lines were placed 18 inches before and after the 10-meter distance. Participants were instructed to begin with toes on the start line and walk or run as fast as possible, without compromising safety, to the finish line. An additional rationale for the ramp down at the end of 10-meters was to minimize loss of balance and falls that can occur with rapid deceleration. The clinical evaluator walked behind the participant for safety. Time to complete 10-meters was measured using a stopwatch activated as the lead limb crossed the taped line marking the start of 10 meters and stopped when the trail limb crossed the taped line marking the end of 10 meters. Time required to walk/run the 10-meter distance was recorded in seconds to the nearest tenth of one second.

2.5. Statistical Methods:

Baseline demographic and clinical characteristics were summarized using frequency and percent for all categorical variables and mean, standard deviation, median and range for numeric variables. The association between age and the primary outcome of interest (10MWRT time) was estimated using linear mixed-effects models. Due to skewness, 10MWRT data were log-transformed. To flexibly model the relationship between age and 10MWRT time, linear and quadratic terms for age were included. A random intercept was used to account for correlation between repeated measurements on participants. BMI, SMA type, SMN2 copy number, and VPA treatment were considered potential confounders, and included as covariates in the model regardless of statistical significance. Since only one participant had a pre-symptomatic SMA type and the only one with type 4 was excluded secondary to being >21 years of age (Figure 1), analyses were limited to participants with SMA types 3a or 3b. Overall tests were followed by pairwise comparisons to evaluate the effect of categorical variables. As a sensitivity analysis, we conducted a post-hoc subgroup analysis, repeating the analysis specified above separately in 3a and 3b participants. All analyses were conducted using R (version 3.5.3, 2019, The R Foundation) and assumed a two-sided, 5% level of significance. No adjustment was made for multiple comparisons.

3. Results

Sixty-nine ambulant participants who completed the 10MWRT at least once were identified from the Project Cure SMA LPDR. After removing participant visits not meeting inclusion criteria (age ≥ 21 years; use of lower extremity orthotics or shoes; assistive device use) and one participant for whom outlier BMI values could not be validated, 56 participants with a total of 241 assessments met all remaining inclusion criteria (Figure 1). Table 1 summarizes number of visits across participants.

Table 1:

Distribution of Repeated Measures

Repeated Measures	Frequency (%) (n=56)
1	1 (1.8%)
2	8 (14.3 %)
3	9 (16.1%)
4	9 (16.1%)
5	24 (42.9%)
6	1 (1.8%)
7	1 (1.8%)
8	1 (1.8%)
9	0 (0.0%)
10	2 (3.6%)

3.1. Baseline characteristics of participants:

Median age at baseline (first 10MWRT) was 6.0 years and ranged from 1.8 to 17.5 years. Seventy-nine percent of children (n=44) were classified as SMA type 3a, 20% (n=11) as type 3b, and 2% (n=1) as untreated pre-symptomatic/not type classified. A majority (90%) had 3 (45%) or 4 (45%) copies of the SMN2 gene. Median baseline 10MWRT time was 8.0 seconds (Range 3.0 – 30.0). Participants’ baseline demographic characteristics are summarized by type and copy number in Table 2.

Table 2.

Summary of Baseline Demographics by SMA Type and SMN2 Copy Number

	SMA Type		Copy Number				Overall (N=56)
	3a (N=44)	3b (N=11)	2 (N=4)	3 (N=25)	4 (N=25)	5 (N=2)
Age (years)
Mean(SD)	6.47(3.62)	10.5(4.90)	8.96 (3.61)	7.76 (4.59)	6.67 (3.85)	4.24 (1.34)	7.23 (4.15)
Median (Q1–Q3)	5.19(3.72–8.93)	10.4(6.92–13.3)	8.58 (6.66–10.9)	6.70 (3.90–9.50)	5.02 (3.78–9.41)	4.24 (3.77–4.72)	5.96 (3.85–9.52)
[Min, Max]	[1.83,16.3]	[3.04,17.5]	[5.19, 13.5]	[1.83, 17.5]	[2.60, 17.2]	[3.29, 5.19]	[1.83, 17.5]
Height (cm)
Mean(SD)	115(22.8)	138(28.3)	133 (21.9)	122 (28.1)	115 (23.2)	101 (10.5)	119 (25.3)
Median (Q1–Q3)	110(94.7–129)	144(119–159)	132 (118–147)	119 (100–135)	106 (98.0–132)	101 (97.7–105)	111 (99.5–133)
[Min, Max]	[81.8,170]	[90.5,175]	[109, 158]	[81.8, 170]	[85.6, 175]	[94.0, 109]	[81.8, 175]
Weight (kg)
Mean(SD)	23.4(13.3)	36.0(17.5)	29.9 (13.5)	28.3 (17.7)	23.3 (12.0)	15.9 (3.32)	25.7 (14.9)
Median (Q1–Q3)	18.5(14.6–28.1)	33.7(24.5–43.4)	26.9 (19.1–37.7)	19.9 (14.6–36.0)	18.2 (15.2–29.8)	15.9 (14.7–17.0)	19.1 (14.9–33.3)
[Min, Max]	[9.40,70.0]	[14.6,64.9]	[19.0, 47.0]	[9.40, 70.0]	[11.9, 64.9]	[13.5, 18.2]	[9.40, 70.0]
BMI (kg/m 2 )
Mean(SD)	16.6(2.76)	17.6(2.55)	16.2 (2.40)	17.1 (3.21)	16.6 (2.34)	15.3(0.0483)	16.7 (2.72)
Median (Q1–Q3)	16.0(15.0–18.2)	16.8(16.3–18.4)	16.4 (15.3–17.3)	17.0 (15.0–19.3)	16.4 (15.0–17.3)	15.3 (15.3–15.3)	16.4 (15.0–18.2)
[Min, Max]	[11.0,24.4]	[14.3,22.8]	[13.2, 18.9]	[11.0, 24.4]	[12.1, 21.3]	[15.3, 15.3]	[11.0, 24.4]
SMA Type
3a	44(100%)	0(0%)	3 (75.0%)	21 (84.0%)	18 (72.0%)	2 (100%)	44 (78.6%)
3b	0(0%)	11(100%)	1 (25.0%)	4 (16.0%)	6 (24.0%)	0 (0%)	11 (19.6%)
pre-symptomatic	0(0%)	0(0%)	0 (0%)	0 (0%)	1 (4.0%)	0 (0%)	1 (1.8%)
VPA
No VPAEvent	39(88.6%)	11(100%)	4 (100%)	23 (92.0%)	22 (88.0%)	2 (100%)	51 (91.1%)
VPA Event	5(11.4%)	0(0%)	0 (0%)	2 (8.0%)	3 (12.0%)	0 (0%)	5 (8.9%)
Time to Walk/run 10m (sec)
Mean(SD)	9.48(5.50)	5.18(2.79)	7.50 (5.07)	10.3 (6.32)	6.68 (3.08)	11.5 (10.6)	8.54 (5.33)
Median (Q1–Q3)	9.00(5.00–12.0)	4.00(4.00–5.00)	6.50 (3.75–10.3)	9.00 (5.00–12.0)	5.00 (4.00–8.00)	11.5 (7.75–15.3)	8.00 (4.00–11.0)
[Min, Max]	[3.00,30.0]	[3.00,13.0]	[3.00, 14.0]	[3.00, 30.0]	[3.00, 14.0]	[4.00, 19.0]	[3.00, 30.0]

BMI= Body mass index; SMA= Spinal muscular atrophy; cm= centimeters; kg= kilograms; kg/m2= kilograms per meter squared; VPA= valproic acid; 10MWRT= 10-meter walk/run time; sec= seconds

3.2. Patterns of change in 10MWRT time from childhood through adolescence:

After adjusting for SMA type, SMN2 copy number, BMI, and VPA event, age was associated with log 10MWRT time, and the pattern of age-related change was quadratic in nature (quadratic β: 0.01, 95% CI: 0.01–0.01, p-value<0.001; linear β: −0.03, 95% CI −0.05 - −0.01, p-value<0.001; Figure 2). Table 3 provides the direction and magnitude of 1-year 10MWRT time change for each year across the age range studied in the full 3a/3b cohort. On average, 10MWRT performance of this cohort improved from 3 years of age through age 8, stabilized between ages 9 and 10, then progressively declined through age 18. For example, at 5 years of age, 10MWRT performance improved 7% in one year (e^β: 0.92, 95% CI 0.90 – 0.93). In contrast, at 13 years of age, 10MWRT performance declined 6% in one year (e^β: 1.06, 95% CI 1.04–1.08). The fastest predicted mean 10MWRT time was at 10.0 years of age during a plateau period.

Figure 2:

Predicted Values of 10-meter Walk/Run Test (10MWRT) Time by Age. Predictions for 3 SMN2 copies, SMA type 3a, mean BMI=16.80, and no VPA event.

Table 3.

Change in Time to Walk/Run 10m for each 1-Year Increase in Age (3a/3b cohort)

Age (years)	Number of Observations (n =256)	Delta 10MWRT (95% CI)	Percent Change for year increase in Age
1	1
2	8
3	24	0.90 (0.89–0.91)	10% Faster
4	33	0.92 (0.9–0.93)	8% Faster
5	28	0.93 (0.92–0.95)	7% Faster
6	19	0.95 (0.93–0.96)	5% Faster
7	17	0.96 (0.95–0.98)	4% Faster
8	16	0.98 (0.96–0.99)	2% Faster
9	26	0.99 (0.98–1.01)	Not Significant
10	19	1.01 (0.99–1.03)	Not Significant
11	6	1.03 (1.01–1.05)	3% Slower
12	7	1.04 (1.02–1.06)	4% Slower
13	7	1.06 (1.04–1.08)	6% Slower
14	6	1.08 (1.05–1.1)	8% Slower
15	5	1.09 (1.07–1.12)	9% Slower
16	7	1.11 (1.09–1.14)	11% Slower
17	5	1.13 (1.11–1.16)	13% Slower
18	2	1.15 (1.12–1.18)	15% Slower
19	5

After adjusting for confounders identified above (SMA type, SMN2 copy number, BMI, and VPA event), 10MWRT time in the SMA type 3a cohort was 74% longer on average than in the SMA type 3b cohort (e^β: 1.74, 95% CI: 1.22–2.48, p-value=0.002). Few participants (6 of 56) had copy number 2 (n=4) or 5 (n=2) and the association between 10MWRT time and SMN2 copy number was not significant overall (p-value=0.095). However, the 25 participants with SMN2 copy number 3 had a 45% slower 10MWRT time compared to the 25 participants with copy number 4 (e^β: 1.45, 95% CI: 1.08–1.94, p-value=0.013; Figure 3). An increase in BMI of 1.0 kg/m² was associated with an average 3% slower 10MWRT time (e^β: 1.03, 95% CI: 1.01–1.06, p-value=0.0084; Figure 4). There was no evidence to suggest VPA was associated with 10MWRT time (e^β: 1.00, 95% CI: 0.95–1.05, p-value=0.940). An increase in weight of 1.0 kg was also associated with an average 2% slower 10MWRT time after adjusting for age, SMA type, SMN2 copy number, and VPA event (weight estimate: 1.02, 95% CI (1.01, 1.03), p-value<0.001).

Figure 3:

Predicted Values of 10-meter Walk/Run Test (10MWRT) Time by Age and Copy Number. Predictions for SMA type 3a, mean BMI=16.80 and no VPA event.

Figure 4:

Predicted Values of 10-meter Walk/Run Test (10MWRT) Time by Age and BMI. Predictions for 3 SMN2 copies, SMA type 3a, and no VPA event

In post-hoc sensitivity analyses, model fits for the 3a sub-group were consistent with those from the full group analysis (Supplemental Table 1). This is as expected, since the 3a sub-group accounted for the majority of observations in the overall sample. In the underpowered 3b sub-group, no change in 10MWRT time across age was evident; neither BMI nor copy number was associated with 10MWRT as they were in the combined 3a/3b sample (Supplemental Table 1). The directions and significance of 1-year change across ages were comparable between the full 3a/3b sample and the type 3a sub-group, except the stabilization of walking speed began at age eight (rather than 9) in the 3a group. The sample size of the 3b sub-group was not sufficient to provide comparable 1-year change data across all ages.

4. Discussion

Outcome measures that capture changes in daily function in ambulatory individuals with SMA (historically classified as SMA types 3 and 4), as well as in newly evolving ambulatory SMA phenotypes, are much needed, particularly in light of the current treatments and continuing development of pharmacotherapies for SMA. 10MWRT time can be measured reliably in typically developing children [27] and in those with neuromuscular disorders [18,28]. Although the 10MWRT has been employed clinically and in SMA studies (e.g. Merlini 2004), only recently has it emerged as an exploratory endpoint in clinical trials [29]. Establishing expected 10MWRT performance changes across childhood in untreated individuals with SMA is a critical step to better characterization of disease progression across the ambulatory spectrum of SMA and will allow for more informed interpretation of 10MWRT change in children receiving pharmacological treatments.

Ambulatory, drug-naïve children with SMA demonstrate both impaired 10MWRT performance and a unique developmental trajectory compared to typically developing children. 10MWRT time for the SMA 3a cohort in this study was nearly two times slower than for same-age peers before age 6 [20] and was substantially slower than same-age peers after age 6 [19]. All SMA groups in our sample reached peak walking speed later in childhood (age 10) compared to previous reports of typically developing children (age 6). Maturation of gait mechanics in typically developing children evolves with walking experience from infancy through 5–6 years of age [30,31] when walking patterns become more adaptive and adult-like secondary to changes in gait-specific neural circuitry, sensory signaling, muscle activation demands and biomechanics [32–36]. Changes in gait speed in healthy children and in those with SMA in the early years are likely reflective of this maturation process and may be related to changes in height [20], but also to the evolution of muscle strength and power, coordination, and/or body composition.

The developmental trajectory of 10MWRT time in ambulatory, drug naïve children with SMA parallels those reported for typical development in young children, yet diverges from patterns reported for mid to late childhood [19,20]. Hoskens (2019) reported gradual improvements in 10MWRT performance with increases in age and height between 2.5 and 5 years of age in healthy boys. Improvements were statistically significant between age groups within the 2.5 to 4-year range. Similarly, Pereira (2016) reported improvement in both 10-meter walk and 10-meter run test times in healthy children from 2 to 6 years of age, with a plateau and relative stability from ages 6 through age 12. The pattern of improved performance in 10MWRT time in our SMA cohort initially parallels that of healthy children [19–20], with gradual improvements from 2.5 to 6 years of age. However, in contrast to Pereira (2016), our ambulatory, drug naive SMA cohort showed 1-year improvement in walking speed each year from age 3 through age 8, a brief plateau from 9 to 10 years, and then 1-year performance declines that escalated in magnitude each year through age 18.

Anthropometrics are associated with individual variability and stability affecting changes in walking function across development. Height and weight are both significantly correlated with 6MWT distance (6MWD) in healthy children [37,38]. Additionally, in normal weight healthy children and adolescents age and height, but not BMI, are independent predictors of 6MWD, whereas in those who were overweight, both height and BMI are independent predictors of 6MWD distance [20,39] and between age, height and the slow-paced 10MWT in typically developing children from 6–12 years old [27]. In this SMA study, both increased BMI and increased weight were associated with longer 10MWRT times across childhood. Capturing weight and BMI in SMA clinical trials and accounting for these variables in 10MWRT time statistical analyses will be important to address the influence of these potentially confounding variables.

The 10MWRT complements existing measures of walking function typically utilized as primary and secondary endpoints in SMA clinical trials. The 6MWT is a well validated, reliable and valuable measure for evaluating walking endurance and fatigue in SMA [5–9,40,41] and one that has been employed as a primary outcome in a number of trials [9,10]. The 6MWT measures long-distance, sustained walking speed and indexes functional community ambulation and endurance in children and young adults with neuromuscular conditions. However typical bouts of walking for children and adults, particularly those with SMA, are brief, intermittent household distances, using shorter duration bursts of muscle activity, rather than sustained walking over several minutes. Children 7–13 years of age most commonly walk in bouts of 20 second durations, with 80 percent of all walking bouts less than 1 minute in duration [42]. Only 7.7% of walking bouts are greater than 2 minutes in duration. During the 10MWRT, children with neuromuscular conditions may walk quickly and confidently for shorter distances, yet are unable to sustain the pace and pattern for longer distances [13]. Mean 10MWRT speed is 12% faster than reported speed for longer endurance walks. Therefore, the 10MWRT and 6MWT are complementary, with the 10MWRT adding a measure of power/strength capacity to the 6MWT’s endurance/fatigue metric [18]. Furthermore, the 10MWRT may more clearly reflect typical, daily ambulatory function in SMA and, as such, serve as a valuable addition to the toolbox of functionally relevant clinical trial endpoints indexing motor function. Use of the 10MWRT and 6MWT together in treatment trials may allow us to capture the effects of pharmacotherapies on unique aspects of walking function across the spectrum of ambulatory individuals, from lower functioning individuals able to walk only shorter household distances, to higher functioning individuals able walk community distance via sustained ambulation. Alternatively, one of the two walking measures may be a more appropriate endpoint than another in a given trial depending on the therapeutic target and mechanism of action of the drug being studied.

Analyzing motor trajectories of ambulatory SMA natural history cohorts can lead to more meaningful interpretation of motor function for patient care and clinical trials, particularly when the patient and participants are developing children. A U-shaped pattern of change in 10MWRT times across childhood characterized both the full sample and the SMA type 3a subgroup. However, the smaller size of the type 3b sample limits our ability to clearly characterize predicted patterns of age-related change in this subgroup. Additional research with a larger sample is needed to delineate the pattern of stability or change in 10MWRT time across childhood in untreated SMA type 3b. More clearly defined 10MWRT natural history trajectories by type and copy number have the potential to guide the selection of specific target age windows for treatment trials and interpret change over time in the context of expected developmental trajectories. For example, significant improvement in 10MWRT during an age range of expected plateau would be more compelling and straightforward to interpret than improvement during the early years before age 6, when annual improvement is expected. Similarly, maintenance of stable 10MWRT performance after age 12 in type 3a, when significant yearly downward decline is expected, would also support efficacy. The ability to stratify groups and compare 10MWRT trajectories of treated cohorts to one another, as well as to this untreated group, may help define characteristics of new motor phenotypes. This ability to stratify ambulatory groups may help identify early strong responders to new treatments, whilst avoiding overexposure of non-responders to longer term treatment.

4.1. Limitations and future research:

Limitations of this study include lack of a non-disease ambulatory comparator cohort and a limited number of study participants in the 3b cohort. Given the observational nature of this study, participant visits were not conducted at consistent intervals, not all participants contributed an equal number of visits, and not all age groups were equally represented. Candidate biomarkers such as compound muscle action potential, motor unit number estimation and neurofilaments were not collected and would add meaningful information for further interpretation[44,45]. Measures of muscle strength were not collected consistently enough to be included in analyses yet would be meaningful to include as potential covariates, given lack of clarity in the relations between strength and 10MWRT performance in previous SMA studies [7,12].

5. Conclusions

The 10MWRT is a timed functional test that is brief and straightforward to administer to ambulatory children as young as 2 years of age, can be conducted in clinical and research settings with limited space, shows good reliability and sensitivity in both typically developing and neuromuscular populations [27,43,46], and has high functional relevance and applicability to everyday mobility for ambulatory individuals with SMA. Although an optimal comparison group for clinical trials is an untreated, same age, similarly affected cohort, strong longitudinal natural history data for key clinical endpoints can also provide a useful benchmark, particularly when withholding treatment is unethical or when treatment with combination therapies is indicated. As the SMA drug pipeline, newborn screening and early/pre-symptomatic treatment evolve, functional ambulation is becoming more realistic for children and adults who may have never walked without pharmacotherapy, making the 10MWRT an even more salient clinical endpoint. The 10MWRT should be considered a complementary or alternative primary endpoint for use in SMA clinical trials with ambulatory patients as regulatory authorities require chosen endpoints to be related to clinically meaningful events. Future comparison of patterns of stability and change in 10MWRT time between untreated children with SMA and children with SMA undergoing treatment with disease modifying therapies will be critical to interpreting response to treatment across childhood and comparing effects of different pharmacological treatments and combination therapies.

Highlights.

• 10MWRT times are slower for those with SMA compared to typically developing peers.

• Direction and magnitude of 12-month change in 10MWRT varies by age in those with SMA.

• Age, SMA type, SMN2 copy number and BMI predict 10MWRT progression over time.

• SMA type 3b, four SMN2 copies, and lower BMI were associated with faster 10MWRT time.

• 10 MWRT times are 74 percent slower for SMA type 3a versus 3b across ages.

• In SMA a 1.0 kg/m² increase in BMI was associated with a mean 3% slower 10MWRT time.