One-Minute Walk Test in Children with Cerebral Palsy GMFCS Level 1 and 2: Reference Values to Identify Therapeutic Effects after Rehabilitation

ABSTRACT

Background: Children with cerebral palsy (CP) show age-driven development and individual fluctuations in walking capacity.

Aim: 1. To precisely quantify 1MWT changes in children with CP, GMFCS level 1 and 2, generating 1MWT percentiles, depicting expected development over 6 months; 2. to assess the effect of a 6-month rehabilitation using whole-body vibration (WBV).

Methods: Retrospective data analysis in 210 children with CP, GMFCS 1 and 2 who received standardized rehabilitation (DRKS00011331). 1MWT was assessed before (M0) and after treatment (M6), and at a 6-month follow-up (M12). Centiles were created using the lambda-mu-sigma method. Cohen’s d was used to assess effect size.

Results: We created 1MWT percentiles using data of all 210 children (M0 data). A small treatment effect size (d = 0.46) was found (M6 and M12 data).

Conclusions: Using the generated centiles clinicians may monitor 1MWT changes over 6 months. Combining WBV and conventional physiotherapy may improve 1MWT in children with CP.

Abbreviations: 1MWT: One-Minute Walk Test; 6MWT: Six-Minute Walk Test; CP: Cerebral palsy; ES: effect size; GMFCS: Gross Motor Function Classification System; GMFM-66: Gross Motor Function Measure 66; LOESS: Locally Estimated Scatterplot Smoothing; LMS: lambda-mu-sigma; SD: standard deviation; WBV: whole-body-vibration.

KEYWORDS:

• Cerebral palsy
• effect size
• one-minute-walk test
• walking capacity
• whole-body vibration

Introduction

Cerebral palsy (CP) is the leading cause of permanent motor impairment in childhood and adolescence, with a prevalence of 2–3 per 1000 live births in Europe.¹ Regardless of its complications such as spasticity, muscle loss and paresis, loss of functional selectivity, contractures, and skeletal deformities, the level of disability is not unchanging.^2,3 Thus, intensive multiprofessional care is needed to monitor and support the development of children and adolescents with CP.

Independent walkers with CP, with a Gross Motor Function Classification System (GMFCS) level of locomotion 1 or 2⁴, usually participate in activities next to typically developing children and adolescents, and thus experience difficulties related to insufficient walking capacity and speed.^5,6

Walking aids, splints, and physiotherapy as well as rehabilitation treatment are commonly used to improve the gross motor function and the walking capacity of children with cerebral CP. In neuropediatric practice, it is crucial to adequately assess the efficacy of such treatments, monitoring their effect using valid and diagnostic tools. The One-Minute Walk Test (1MWT) presents a valid, feasible and reliable tool, commonly used among children with CP.^5,7–10 A 1MWT test–retest agreement¹¹ has not been reported yet.

Age-related reference ranges are essential for the implementation of diagnostic tools assessing motor function, such as the Gross Motor Function Measure (GMFM).¹² Although the 1MWT is in use among children with CP¹³, no CP-specific reference ranges can be found in the literature.

Regarding the efficacy of interventions to improve gait speed in children with CP, an extensive review of the literature revealed that that gait training is very effective.¹⁴ Individual studies also suggest that training using whole-body vibration (WBV) can be effective towards the improvement of walking capacity¹⁵ and thus the approach warrants further investigation.¹⁴

The aim of this study was twofold: First, to present a method to precisely quantify changes in the 1MWT in children with CP, GMFCS level 1 and 2, generating and using age-related reference centiles to monitor changes in their walking capacity. These centiles would depict the expected progression under standard of care over a 6-month period. Second, to assess the effect of the neurorehabilitation program “Auf die Beine”¹⁶ on the walking capacity of these children, using this tool. The program combines conventional physiotherapy with WBV training in intervals of inpatient rehabilitation stay and home-training.

Methods

Study Population

We performed a retrospective analysis of prospectively collected data of independent walkers (GMFCS level 1 and 2)¹⁷, who received the rehabilitation treatment concept “Auf die Beine”, of the Centre of Prevention and Rehabilitation of the University of Cologne, Germany, from January 2006 to December 2017. For the generation of the reference centiles as well as the assessment of the effect of rehabilitation, we only included all children and adolescents with CP, GMFCS level 1 and 2, that were recruited between the third and the 12th year of age. Patients with additional syndromic, genetic conditions, i.e. collagen disorders, or further chronic conditions, such as bronchopulmonary dysplasia, that can significantly influence the walking capacity, were excluded from the study (Figure 1).

Figure 1. Flowchart of the study population.

The data had been prospectively registered in the database of the center after receiving written informed consent from the children’s guardians (www.drks.de, DRKS-ID: DRKS00011331). The study has been approved by the Ethics Committee of the Medical Faculty of the University of Cologne.

Intervention

“Auf die Beine”¹⁶ is an intensive, goal-oriented neuromuscular training program combining inpatient stays of in total 3 weeks, and a continuous 6-month WBV home-training, using a side-alternating WBV platform (System Galileo®, Novotec Medical, Pforzheim, Germany) to induce involuntary muscle stimulation. Each WBV session was 3 × 3 min long. The children received 3 WBV sessions per day during inpatient stays, and 10 WBV sessions per week. During the stays, caregivers were individually trained to conduct home-training with their children and document their activities in a log. Exercises and WBV frequency were adjusted to meet the individual goals set by the child, the parents, and the therapists. We should underline that WBV is a major, but not the only component of the rehabilitation program, which also includes for instance functional training, swimming, treadmill, and swimming. The rehabilitation program has already been described in detail.¹⁶

Investigation

The children’s walking capacity was assessed using the 1-min walk test (1MWT). The children were asked by a physiotherapist to walk as fast as possible around a circular 40-m track over a period of 1 min. Running was not allowed. The physiotherapist was continuously assuring that the track was consequently followed. If a patient fell or needed to take a break during the 1MWT, the measurement was regarded as incomplete and was excluded from the database. The test–retest reliability of the 1MWT to assess walking speed is regarded as good (ICC = 0.94)¹⁰ and validity between the 1MWT and the GMFM-66 and GMFM-88 children with CP⁹ is very good.⁵

The 1MWT were carried out at baseline (M0), at the end of the 6-month rehabilitation (M6), as well as in an outpatient follow-up 6 months later (M12). We used the 1MWT score at M0 to generate the centile curves. For the analysis of the treatment effect of the rehabilitation program, we used a double-baseline design, regarding each participating child as their own control, since the changes after the rehabilitation training (M6 vs. M0) were compared with their performance in the 6-month follow-up, and thus after a 6-month period of usual care (pharmacotherapy, splints, physiotherapy, orthopaedic interventions, etc.) (M12 vs. M6). The study design has already been reported.¹⁸ Only assessments performed using the same aids, such as splints, in all three investigations were included in the analysis.

Reference Centiles and Correlations

We used the LMS (lambda-mu-sigma) method to create the reference centiles for 1MWT.¹⁹ Only data of the M0 (210 children with CP) were used to create the reference curves. The demographics of this group are presented in Table 2 as “Participants in M0”. The methodology for the calculation of Z-scores, adjustment of skewness L(a), median value M(a) and the coefficient of variation S(a), as well as the calculation of differences between two z-scores and their correlation coefficients r, has been described in a previous publication.¹⁸ Correlation coefficients r of the Z-Scores at M6 and M12 were used to calculate the distribution of centile changes. We further applied LOESS regression to the generated correlations coefficients r. Calculation of Z-scores was performed by applying a modified Box-Cox transformation:

(1) $Z_{LMS}=\frac{1}{S\left(a\right)\ast L\left(a\right)}\ast \left[{\left(\frac{g}{M\left(a\right)}\right)}^{L\left(a\right)}-1\right]\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{S}(\mathrm{a}),\mathrm{L}(\mathrm{a})\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{M}(\mathrm{a})\ne 0$(1)

(2) $Z_{LMS}=\frac{\mathrm{l}\mathrm{n}\left(\frac{g}{M\left(a\right)}\right)}{S\left(a\right)}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{L}(\mathrm{a})=0\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{S}(\mathrm{a}),\mathrm{M}(\mathrm{a})\ne 0$(2)

Table 2. Tabulated 1MWT reference centiles for use with children having cerebral palsy (GMFCS 1 and 2), including the LMS values.

Age	C3	C10	C25	C50 = M	C75	C90	C97	L	S
3	9.1	18.7	29.8	42.1	53.8	63.9	73.5	1.306	0.462
3.5	10.1	20.5	32.6	45.9	58.7	69.8	80.3	1.262	0.453
4	11.1	22.5	35.4	49.7	63.6	75.6	87.2	1.219	0.444
4.5	12.3	24.5	38.3	53.5	68.4	81.4	93.9	1.176	0.435
5	13.6	26.6	41.1	57.2	73.0	86.9	100.4	1.132	0.427
5.5	15.0	28.8	43.8	60.6	77.2	92.0	106.4	1.089	0.418
6	16.6	30.9	46.4	63.8	81.1	96.6	111.8	1.045	0.410
6.5	18.4	33.0	48.7	66.5	84.4	100.6	116.5	1.002	0.402
7	20.3	35.0	50.9	68.9	87.3	103.9	120.5	0.958	0.394
7.5	22.2	37.0	52.8	71.0	89.6	106.7	123.8	0.915	0.387
8	24.3	38.8	54.5	72.8	91.6	109.0	126.6	0.871	0.379
8.5	26.3	40.6	56.1	74.3	93.3	111.0	128.9	0.828	0.372
9	28.4	42.3	57.6	75.7	94.7	112.6	130.9	0.784	0.364
9.5	30.4	43.9	58.9	76.9	95.9	114.0	132.6	0.741	0.357
10	32.2	45.4	60.1	77.8	96.8	114.9	133.7	0.698	0.350
10.5	33.9	46.7	61.1	78.5	97.4	115.5	134.5	0.654	0.343
11	35.5	48.0	62.0	79.1	97.9	116.0	135.1	0.611	0.337
11.5	37.1	49.1	62.8	79.7	98.3	116.4	135.6	0.567	0.330
12	38.6	50.3	63.6	80.2	98.6	116.8	136.0	0.524	0.323

Distance in meters, age in years.

We calculated r for age a between 3 and 12 years and for a 6-month interval, using the dataset of the follow-up assessments performed at M6 and M12.

We further defined a 3-year length section within the age interval from 3 to 12 years. We moved this section in steps of 0.1 year over the complete age interval. At each step, the correlation coefficient r was calculated with the 1MWT data covered in the moving section.

Finally, we calculated the standard deviation of ∆Z over the age interval from 3 to 12 years.¹⁸ According to Cole, the difference between two Z-scores

(3) $\mathrm{\Delta }\mathrm{Z}={\mathrm{Z}}_2-{\mathrm{Z}}_{1,}$(3)

determined at different time points t₂> t₁, are a measure for centile-crossing development. For the reference population, the following applies:

$\mathrm{e}\mathrm{x}\mathrm{p}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{u}\mathrm{e}\mathrm{E}\left[\mathrm{\Delta }\mathrm{Z}\right]=0$

(4) $\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{a}\mathrm{r}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{v}\mathrm{i}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\sigma \left(\mathrm{\Delta }\mathrm{Z}\right)=\sqrt{2\ast \left(1-\mathrm{r}\right)}$(4)

where r is the correlation coefficient between Z₂ and Z₁. The following applies to the standard deviation of ΔZ values:

(5) $\mathrm{Z}-\mathrm{s}\mathrm{c}\mathrm{o}\mathrm{r}\mathrm{e}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{i}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{g}\mathrm{e}=\frac{\mathrm{\Delta }\mathrm{Z}}{\sqrt{(2\ast \left(1-\mathrm{r}\right)}}$(5)

Effect Size Estimation

We used the effect size (ES) to quantify differences before and after the intervention.^18,20

‘Negligible’ effect is defined when ES<0.20, ‘small’ effect when ES ≥0.20 to <0.50, ‘medium’ effect when ES ≥0.50 to <0.80, and ‘large’ effect when ES≥0.80. For this assessment only data of children who were assessed thrice (M0, M6, and M12) were eligible. The demographics of this group are presented in Table 2 as “Participants in all three assessments”.

Statistical Analysis and Implementation

We analyzed the data using RStudio v. 1.0.136 in conjunction with R v. 3.2. (R Foundation for Statistical Computing, Vienna, Austria) and the statistical packages GAMLSS (v. 5.0.0), effsize (v. 0.7.1) and pROC (version 1.10.1).^21–23 The use of the resulting 1MWT reference centiles was finally exemplified on two cases, evaluating the effect of the rehabilitation treatment on the walking capacity in two children of our study cohort at M0 and M6.

We also analyzed the effect of the rehabilitation treatment on the 1MWT in our study cohort, following the methodology presented in Duran et al¹⁸, comparing Z-scores changes at M6 with Z-scores at M0. We compared the changes using the Wilcoxon signed rank test and the clinical effect of the intervention using Cohen’s d, since Shapiro test did not reveal normal distributions across the assessments.

Results

Study Population and Generation of Reference Centiles

In total, 210 1MWT assessments (97 girls, 113 boys) were eligible to generate percentile curves at M0. Follow-up assessments at M6 and M12 were available for 106 children with CP, GMFCS level 1 and 2 (45 girls, 61 boys).

The demographics of the study population are depicted in Table 1. Results are given as mean and standard deviation (SD), unless otherwise stated. The characteristics of the two groups (“Participants in M0” and “Participants in M0, M6, and M12”) did not differ significantly. The generated age-related percentiles for the 1MWT are depicted in Figure 2. The tabulated reference centiles are presented in Table 2. The tabulated, age-related values of the SD of both walking capacity assessments (6 months apart) shall be used in the tool, to correct for the age-driven character of the 1MWT development. They are depicted at the right y-axis of Figure 2, as well as in detail in Table 3.

Figure 2. Reference curves for 1MWT in children with CP GMFCS level 1 and 2. The centiles are labeled with the corresponding Z‐scores. The dashed line represents the standard deviation of the difference of two GMFM-66-Z-scores measured 6 months apart.

Table 1. Characteristics of the study population and evaluation of the effect of the rehabilitation program on the walking capacity assessed by the 1MWT.

Demographics	Participants in M0	Participants in all three assessments (M0, M6, and M12)
N	210	106
Females	97	45
Males	113	61
GMFCS level 1	61	36
GMFCS level 2	149	140
CP subtype
Bil. Spastic	139	61
Unil. Spastic	45	29
Dyskinetic	3	2
Ataxic	9	4
Mixed	14	10
Age (y) at M0
Mean	6.78	6.50
Standard deviation	2.51	2.17
1MWT at M0(m)
Mean	65.09	68.93
Standard deviation	27.70	27.01
1MWT at M6 (m)
Mean	N.A.	80.80
Standard deviation		23.21
1MWT at M12 (m)
Mean	N.A.	83.75
Standard deviation		21.41
Height at M0 (m)
Mean	117.05	115.98
Standard deviation	16.05	14.46
Weight at M0 (kg)
Mean	22.08	21.22
Standard deviation	8.32	6.76
BMI at M0 (kg/m^2)
Mean	15.60	15.40
Standard deviation	2.33	2.05
Evaluation of the effect of rehabilitation	1MWT M0–M6 changes	1MWT M6–M12 changes
Wilcoxon test V, p	V = 1173 p < .001	V = 2578 p = .418
Cohen’s d (95%CI)	d = 0.457 (small) (0.18, 0.73)	d = 0.007 (negligible) (−0.27, 0.26)
Mean value Cohen’s d of individual patients (95%CI)	d = 0.518 (moderate) (0.30, 0.73)	d = 0.006 (negligible) (−0.15, 0.16)

N.A. not applicable

Table 3. Tabulated values of the standard deviation of 1MWT Z-score differences (6 months apart).

Age	0.0	0.1	0.2	0.3	0.4	0.5	0.6	0.7	0.8	0.9
3	0.69046	0.69380	0.72645	0.72991	0.72991	0.73083	0.74537	0.76035	0.73423	0.73550
4	0.75013	0.73486	0.73436	0.75034	0.74949	0.74759	0.75390	0.75369	0.75373	0.75373
5	0.75373	0.75373	0.75373	0.75373	0.75161	0.75161	0.75161	0.76374	0.76067	0.76066
6	0.76104	0.76112	0.76168	0.76168	0.75480	0.75780	0.76797	0.76148	0.76740	0.76545
7	0.76278	0.80332	0.80368	0.79157	0.81129	0.80582	0.81216	0.73890	0.77310	0.77142
8	0.77307	0.76954	0.75260	0.75578	0.75768	0.75644	0.74636	0.75669	0.77949	0.77846
9	0.77428	0.79470	0.79593	0.77649	0.77723	0.77799	0.76045	0.75962	0.75821	0.75821
10	0.75821	0.75821	0.75821	0.75821	0.76088	0.76088	0.76088	0.75085	0.75922	0.75936
11	0.77068	0.77633	0.78235	0.77734	0.77734	0.77353	0.77342	0.78664	0.77525	0.78823

Age in years.

Correlation Coefficients R Estimation

The bold dotted line in Figure 2 depicts the standard deviation (SD) of the difference of two 1MWT Z-scores measured 6 months apart. For instance, in Figure 2, the 1MWT SD values, evaluated at the first measure, show a very slow decrease with increasing age until the age of seven, followed by a fast decrease within the next year of life, a plateau until 10 years and a consequent fast increase until 12 years, as can be read from the y-axis at the right.

Assessing 1MWT Changes in the Participants of “Auf Die Beine”

Children with CP, who participated in the rehabilitation program “Auf die Beine” showed a significant improvement in the 1MWT at M6 (Wilcoxon signed rank test, p < .001), while no statistically significantly changes were revealed at M12 (p = .418). The analysis of the efficacy of the intervention on the 1MWT showed a small (d 0.457 for the whole cohort) to moderate effect size (mean d of individual patients 0.518). The values of Cohen’s d were negligible for the follow-up phase. The detailed results are depicted in Table 1.

Assessing 1MWT Changes over a 6-Month Period in an Individual Patient

We have depicted the use of the reference centile using a case from the study group in Figure 3. Patient X is a girl with bilateral spastic CP, GMFCS level 2, with a 1MWT score of 34 m (at age: 5 years 11 months) at the beginning (M0) of the rehabilitation. After 6 months (M6) she scored 46 m (at age: 6 years 5 months) and after 12 months (M12) 51 m (at age 6 years 11 months).

Figure 3. Case study. The figure illustrates the use of the reference centiles. First, the two 1MWT scores are entered (three crosses, 34 m at M0, age₁ = 5 years 11 months, 46 m at M6, age₂ = 6 years 5 months, and 51 m at M12, age₃ = 6 years 11 months). The corresponding Z-scores are Z₁ = −1.146 for M0, Z₂ = −0.773 for M6 and Z₃ = −0.652 for M12. The age-dependant standard deviation (SD₁) of the centile change, is determined by using the dashed curve and the age of the child at the first 1MWT assessment at M0 (SD₁ = 0.761) and at M6 (SD₂ = 0.758). Using the formula (Z₂-Z₁)/SD₁, the Z-score for the centile change can be calculated. It corresponds to an effect size (−0.773 + 1.146)/0.761 = 0.49 for the first 6 months, during the active training in the rehabilitation program ‘Auf die Beine’. During the observational 6-month period (Z₃-Z₂)/SD₂ = 0.16 corresponding to a negligible effect size.

Is the improvement of 1MWT between M0 and M6 to be interpreted as therapeutic effect of the rehabilitation or was it to be expected due to the motor development under standard of care (e.g., weekly outpatient physiotherapy treatment)?

Interpretation: As seen in Figure 3, the Z-scores for the 1MWT at M0, M6, and M12 are estimated as:

Z₁ = −1.146 for M0, Z₂ = −0.773 for M6 and Z₃ = −0.652 for M12.

The age-dependent standard deviation (SD) of the centile change is determined by using the dashed curve, or for better accuracy by using the values in Table 3, and the age of the child at the first 1MWT assessment at M0 (SD₁ = 0.76) in order to assess the period of the first 6 months; and at M6 (SD₂ = 0.76) to assess the follow-up period.

Thus,

$\begin{array}{rl}& \mathrm{Z}-\mathrm{s}\mathrm{c}\mathrm{o}\mathrm{r}\mathrm{e}\mathrm{o}\mathrm{f}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{i}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{g}\mathrm{e}=\left({\mathrm{Z}}_2-{\mathrm{Z}}_1\right)/\mathrm{S}{\mathrm{D}}_1\\ & =\left(-0.773+1.146\right)/0.76=0.49,\end{array}$

indicating a small, but almost medium, effect size, and

$\begin{array}{rl}& \mathrm{Z}-\mathrm{s}\mathrm{c}\mathrm{o}\mathrm{r}\mathrm{e}\mathrm{o}\mathrm{f}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{i}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{g}\mathrm{e}=\left({\mathrm{Z}}_3-{\mathrm{Z}}_2\right)/\mathrm{S}{\mathrm{D}}_2\\ & =\left(-0.652+0.773\right)/0.76=0.16,\end{array}$

indicating a negligible effect size.

Thus, next to the expected development, the child experienced an additional effect associated with the rehabilitation program.

Discussion

We quantified changes in the walking capacity of ambulant individuals with CP, GMFCS level I and II, generating age-related centile curves for the 1MWT, considering the age-expected development. To our best knowledge, no method has been reported to calculate 1MWT changes in children with CP, as well as among typically developing children, taking into consideration expected development under standard of care.

In general, in our study cohort, we found a statistically significant improvement in the walking capacity assessed by the 1MWT in children with CP, GMFCS level 1 and 2, under treatment with an intensive neuromuscular rehabilitation program, including vibration-assisted home-training (p < .001) (Table 1). The treatment effect size of the whole cohort was small but almost moderate (Cohen’s d = 0.46; 95%CI 0.18–0.73), and thus potentially relevant for clinical decisions. The estimation of the mean value of the individual effect size that each patient experienced was very similar, with a Cohen’s d of 0.52 (95%CI 0.30–0.74, moderate effect) proving once more the efficacy of the treatment. Out of 103 children, 43 children had an individual effect size of ≥0.5. Thus, by applying the proposed method, it is not only possible to evaluate the overall effect of a therapeutic intervention in a cohort of patients, but also to identify from which treatment an individual patient can benefit the most. In the rise of the era of personalized medicine, this challenge will play an increasingly significant future role in rehabilitation medicine.²⁴

The 1MWT has been validated for use for children and adolescents with CP, utilizing the GMFM-66 and GMFM-88 as “gold standard”.⁹ Regardless of the small size of the studied patient cohort, the sample was representative, including all different types of CP and different GMFCS levels of locomotion.⁹ The same working group has also studied and confirmed the reliability of the tool.¹⁰ Although the tool is easy to be used and implemented in the daily rehabilitative and neuropediatric practice, it is still less often used in research settings among children with CP. We suppose that the lack of reference values for the patient’s age or height, even among typically developing children, as well as the significance of the expected 1MWT changes under standard of care, due to increasing age or growth in childhood and adolescence, including patients with CP, could restrain researchers from using this easy tool. This paper aims at bridging this gap, providing values regarding the expected development under standard of care in independent walkers with CP, to monitor the effect of a 6-month treatment, e.g., a newly prescribed foot-ankle splint, on the child’s walking capacity.

The situation differs when it comes to the more complicated 6-min walk test (6MWT), a tool that is often used to assess the walking capacity of typically developing children or children with CP in research settings. Of course, one may argue that children may perform differently after the first minute of walk, and that a diagnostic tool with a longer duration could be more relevant in a rehabilitation setting. However, at least six-times longer duration in comparison to the 1MWT, especially considering the fact that the test has to be restarted, if the patient falls or needs a break, makes the 6MWT less useful for the pediatric or neuropediatric practice. Further, we regard that the 1MWT may address the children’s everyday walking capacity more precisely, especially in CP patients, GMFCS level 2, assessing capacities related to their short distance performance, for instance how fast they can move from a classroom to another.

The children reported in this study received 6-month training, including WBV, and showed a statistically significant, and potentially relevant effect on their walking capacity assessed by the 1MWT at M6. A similar intervention was reported being offered, as a controlled trial, to children with cerebral palsy in Korea.¹⁵ The control group received conventional physiotherapy, while the intervention group received also additional WBV. Although the 1MWT score was not a study outcome parameter, Lee et al. reported a significant improvement in stride length and gait speed, when compared with normally developing children. This depicts the positive effect of additional WBV, combined with conventional physiotherapy, in children with CP.¹⁵

Limitations

We generated the reference centiles for independent walkers with CP (GMFCS level 1 and 2), with an age from 3 to 12 years, using data of patients, who have been treated with the rehabilitation program ‘Auf die Beine’. The treatment includes intervals of inpatient rehabilitation, as well as an intensive, 6-month home-training utilizing vibration-assisted physiotherapy. CP patients with GMFCS level 1 often fail to be recruited in the treatment. Therefore, selection bias is a relevant limitation of this analysis. For the generation of the reference ranges of the independent walkers, we did not differentiate the patients according to the GMFCS level.

We estimated the correlation coefficient r, regarding the expected development under standard of care of walking capacity in the 6-month follow-up, under the standard of care provided to children with CP in Germany. This is a relevant limitation of the generalisability of our findings, since the standard of care is not internationally identical.

A double baseline design, including an initial assessment of changes in 1MWT under standard of care, followed by the intervention, would be the ideal study design to answer our second research question. In the present study, the follow-up period under standard of care followed the intervention. This is the main limitation in our retrospective study. Thus, long-term effects of the 6-month rehabilitation, that could positively influence the effect size of the standard of care in Germany at M12 cannot be ruled out. Indeed, the patients did not show a deterioration of the 1MWT at M12, which could also be interpreted as long-term therapy effect. On the other hand, since the effect size of the treatment provided in the observational phase was statistically insignificant and clinically negligible, we regard that the significant 1MWT improvement achieved at M6 has a causal association with the rehabilitation training.²⁵

Further, the presented tool can evaluate Z-scores changes only for 6-month time intervals.¹⁸ For other intervals, we can only estimate whether the 1MWT score is equal or lies below or above the expected development. Further research could investigate changes in the walking capacity using other time periods.

Conclusions

The 1MWT is a valid and simple assessment to monitor changes in the walking capacity of children with chronic conditions. We have reported a method to monitor and estimate changes in the walking capacity, measured using the 1MWT, of individual children, as well as of cohorts of patients with CP, GMFCS level 1 and 2, over a period of 6 months. To assess these changes, we generated 1MWT reference centiles, depicting the expected development under standard of care among these patients, over a period of 6 months. We further assessed the effect size of the rehabilitation concept “Auf die Beine” on the walking capacity of children with CP, taking into consideration changes related to expected development, and identified a small, but relevant effect for potential clinical decisions.

Declaration of interest

We declare no competing interests. No funding to declare. ID is working at the Centre of Prevention and Rehabilitation of the University of Cologne and ES is directing the center.

Acknowledgments

We thank the physical therapists at the UniReha GmbH, Center of Prevention and Rehabilitation of the University of Cologne for their dedicated and qualitative work. We acknowledge Ida Alperstedt and Angelika Stabrey for database management.