Sage Journals: Discover world-class research

Abstract

BACKGROUND:

The high variability in muscle performance, especially in muscle strength, typical of people with multiple sclerosis (PwMS) implies that in order for strength variations to be clinically interpretable, the extent of variability should be properly quantified.

OBJECTIVE:

To establish the responsiveness of maximal isokinetic strength measurements obtained from the ankle dorsiflexors of PwMS and of a cohort of matched healthy controls.

METHODS:

Twenty PwMS and 20 healthy controls participated in this study. All subjects underwent a dedicated familiarization session and then three isokinetic testing sessions (Test 1; 1-day retest; 1-week retest) in which the peak moment (PM) and maximal work (MW) were recorded in the more affected side.

RESULTS:

PwMS proved significantly weaker than controls ( $p<$ 0.00005). The test-retest consistency of the measurements proved high for both groups (all intraclass correlation coefficients $>$ 0.9). Importantly, the smallest real difference (SRD) scores ranged 11.4–19.8% for PwMS and 9.2–14.4% for controls.

CONCLUSIONS:

The responsiveness of the strength and work findings in this cohort indicates that isokinetic measurements of dorsiflexion performance in PwMS may effectively be employed for detecting the efficiency of an intervention in spite of the fluctuating nature of this disease.

Keywords

Isokinetics measurement error multiple sclerosis test-retest reliability smallest real difference

1. Introduction

Reproducibility and responsiveness are two closely related methodological properties of clinical measurements [2]. Reproducibility, also known as test-retest reliability, is the ability to measure attributes in a consistent manner when administered on several occasions [2]. Responsiveness is the ability to detect relevant changes over time [13]. The recent decade saw a significant shift towards the preferential employment of the latter in rehabilitation research. This trend reflects the growing understanding that more than assessing the stability of a trait or a test measure, what is needed is a decision rule for assessing the efficiency of an intervention. Such a decision rule, which inevitably incorporates an element of reproducibility, is expressed by a number, termed the ‘cutoff score’. This score differentiates a meaningful change in performance from a non-meaningful one, even if statistically significant, which falls within the so-called error zone. The error (or measurement error) reflects the cumulative effect of of the ‘subject/patient’, ‘examiner’, ‘test instrument’, ‘protocol’, and the ‘test environment’ factors [27, 35]. One of the most common indicators of responsiveness is the smallest real difference (SRD). Significantly, if following an intervention, a change in an outcome parameter (OP) relative to the baseline level exceeds the SRD, this change is considered clinically important, otherwise the apparent ‘change’ carries no clinically meaningful value. Obviously, a higher reproducibility (e.g. higher intraclass correlation coefficients, ICCs) contributes to better responsiveness (e.g. lower SRD) but the relationship is not 1:1. Furthermore, the analysis of the responsiveness of relevant OPs is essential for assessing the efficiency of clinical interventions and for developing new evaluation methods.

These considerations are particularly pertinent for evaluating muscle performance in conditions characterized by high variability and inconsistent responses to an intervention such as those occurring in patients with multiple sclerosis (PwMS). These patients are characterized by fluctuations in their compromised muscle mechanical function (muscle strength, muscle power, explosive muscle strength and increased fatigability) [15, 19, 30], due to their impaired ability to fully activate the tested muscles [17, 32]. It has been proposed that quantitative assessment of muscle performance in PwMS may be biased by daily variability in strength or fatigue [18, 31].

The reproducibility of muscle performance findings in PwMS has been the topic of a few studies which reported excellent test-retest reproducibility of strength and fatigue parameters [3, 18, 27, 32]. However, all these studies: i) related mainly to ICCs, which are known to be insensitive to changes that occur within a group in a uniform mode [29]; ii) limited the basis for estimating reproducibility to two time points (test and retest) only, although the inclusion of a third time point in a non-consecutive day has been shown in different patient populations to portray this trait in a more clinically meaningful way [39]; iii) did not perform responsiveness analysis.

These shortcomings have motivated the present investigation, which focused on the responsiveness of isokinetic strength and work parameters of the ankle dorsiflexors in PwMS and in a matched cohort of healthy subjects, at 3 different time points. The ankle dorsiflexors were chosen as they are one of the most common sites of weakness in neurological conditions, particularly, in MS [40]. Therefore, the primary aim of the study was to establish cut-off values for clinically meaningful changes relating to the performance of the dorsiflexors in PwMS and healthy controls. In addition, the study characterized the dynamometric profile of the participants and explored any pattern of practice-based improvements [4].

2. Materials and methods

2.1 Subjects

This study is a continuation of a previously published single-system interventional case series of 8 PwMS [22], whose baseline assessments of isokinetic strength have been considered also for the current investigation. Twenty PwMS with predominantly unilateral muscle weakness of the dorsiflexors were recruited from those participating in a larger interventional trial taking place in the University Hospital facilities. Inclusion criteria were: diagnosis of relapsing-remittent MS according to 2010 revision of diagnostic criteria [28]; patient-reported evidence of strength asymmetry between dorsiflexor muscles; age $\geqslant$ 18 years; independent walking with or without unilateral assistance; Expanded Disability Status Scale score $\leqslant$ 6.5 with Pyramidal Functional System score $\geqslant$ 3. Exclusion criteria consisted of any medical condition contra-indicating participation in strength testing exercises; occurrence of relapses, treatment with corticosteroid and/or botulinum toxin, variations in disease-modifying drugs or symptomatic treatment in the previous 6 months; severe ataxia and postural instability; major depression; clinically relevant cognitive deficits; participation in rehabilitative or training programs within 6 months prior to the study. Participants deemed eligible after clinical examination were asked to refrain from any other exercise activity during the study. Twenty healthy individuals of comparable age, gender and anthropometric characteristics served as control subjects. The study was approved by the institutional Bioethics Committee of the Local Health Authority and conducted in accordance with the Declaration of Helsinki. A written informed consent was obtained by all the participants before enrollment.

2.2 Measurements of ankle dorsiflexion muscles’ strength

Bilateral concentric assessment of ankle dorsiflexors was conducted by the same examiner at a consistent time of the day, using an isokinetic dynamometer (Biodex System 3, Biodex, Shirley, NY, USA). The test was conducted in the sitting position with the trunk inclined at 85 ${}^{\circ}$ , the tested knee flexed at 30 ${}^{\circ}$ and the ankle in full plantar flexion as starting position [1]. The dynamometer shaft was aligned with the lateral malleolus. As all participants in this study were able to move the foot from full plantarflexion through an arc of 45–50 ${}^{\circ}$ the test range of motion was adjusted to 45 ${}^{\circ}$ . Participants in both groups had no prior experience with isokinetic testing. In the 2 weeks preceding the test, all participants attended a first orientation session where all the testing procedures were explained. During this session, each participant was positioned onto the device and the chair settings were recorded. Additionally, participants were instructed to perform a single concentric ankle dorsiflexion followed by a passive return to plantarflexion and then required to perform 4 to 6 very light movements. Visual feedback was provided at this stage by showing the real-time force traces during the motor task to ensure it was clearly understood.

In the next formal familiarization session, which was aimed at controlling the potential learning effect of practice-based improvements [4], the position settings previously recorded were verified and the subject was required to perform 2 sets of 4 submaximal single concentric dorsiflexions, with the first set carried out at a comfortable, self-selected level of exertion, and the second set performed at a moderate effort. These procedures resembled those employed for the warm-up phase preceding the criterion strength test. As in the orientation session, participants were requested to monitor the force traces on the PC screen located in front of them. This session was aimed at controlling potential practice-based improvements [4].

Prior to the formal testing sessions, all subjects underwent a 5-minute warm-up by performing two sets of 2-to-4 concentric submaximal repetitions at 45 ${}^{\circ}$ /s with a 3-minute rest in between. Following a 5-min rest, the criterion test consisted of 4 maximal attempts at 45 ${}^{\circ}$ /s. Sixty seconds of recovery were allowed among the attempts. Verbal communications were kept to the minimum and only the standardized instruction to “pull the foot and ankle upwards as fast and as hard as possible” was consistently delivered before each attempt. In the cohort of PwMS the less-affected leg was tested first. A 6-min rest elapsed between the testing of the less-affected and the more-affected leg in PwMS, and between the dominant and non-dominant leg in the control group. The entire procedure was repeated in 3 different sessions: 1) criterion test; 2) 1-day retest and 3) 1-week retest. Neither visual feedbacks nor verbal encouragements were provided to the participant during the testing procedure.

Table 1
Demographic and clinical features of participants

	PwMS	Controls	P values
	( $n=$ 20)	( $n=$ 20)
Gender (M:F)	4:16	6:14	0.60
Age (years)	42 $\pm$ 10.2	37 $\pm$ 11.1	0.14
Weight (kg)	60.1 $\pm$ 13.3	65.7 $\pm$ 10.8	0.16
Disease duration	14.4 $\pm$ 8.4	/	/
(years)
EDSS ${}^{*}$	3 (1.9–4.1)	/	/

PwMS, patients with multiple sclerosis; Controls, healthy subjects; M, males; F, females; values are expressed as mean $\pm$ standard deviation. EDSS, Expanded Disability Status Scale; ${}^{*}$ score calculated by median and interquartile range.

Table 2

Dynamometric features of the ankle dorsiflexion muscles by group, limb and testing session

	Angular	Measures	Session 1	Session 2	Session 3	Session	Session 1	Session 1	Session 2
	velocity		Mean $\pm$ SD	Mean $\pm$ SD	Mean $\pm$ SD	effect ${}^{\#}$	vs. 2	vs. 3	vs. 3
			(95% C.I.)	(95% C.I.)	(95% C.I.)	$p$ value ${}^{*}$	$p$ value ${}^{*}$	$p$ value ${}^{*}$	$p$ value ${}^{*}$
PwMS	LA	PM	24.4 $\pm$ 6.9 (21.2–27.6)	23.8 $\pm$ 7.0 (20.5–27.1)	22.5 $\pm$ 6.2 (19.5–25.4)	F ${}_{2,19}=$ 6.03 $p=$ 0.006	0.32	0.009	0.04
		MW	10.0 $\pm$ 4.7 (7.8–12.1)	9.6 $\pm$ 4.5 (7.4–11.7)	8.7 $\pm$ 3.6 (7–10.4)	F ${}_{2,19}=$ 7.65 $p<$ 0.002	0.23	0.009	0.04
	MA	PM	20.9 $\pm$ 6.6 (17.8–24.0)	19.7 $\pm$ 6.4 (16.7–22.3)	19.4 $\pm$ 6.2 (16.4–22.2)	F ${}_{2,19}=$ 5.13 $p=$ 0.01	0.02	0.03	0.98
		MW	7.9 $\pm$ 3.8 (6.1–9.6)	7.5 $\pm$ 3.2 (6.0–8.9)	7.0 $\pm$ 3.4 (5.4–8.6)	F ${}_{2,19}=$ 3.93 $p=$ 0.03	0.43	0.07	0.45
Control	DOM	PM	31.7 $\pm$ 11.6 (26.2–37.1)	33.7 $\pm$ 14.2 (27.1–40.4)	34.0 $\pm$ 14.2 (27.4–40.7)	F ${}_{2,19}=$ 4.64 $p=$ 0.02	0.06	0.09	0.70
		MW	16.2 $\pm$ 4.0 (14.3–18.1)	17.4 $\pm$ 4.9 (15.1–19.8)	17.4 $\pm$ 4.5 (15.3–19.5)	F ${}_{2,19}=$ 3.07 $p=$ 0.06	0.35	0.13	0.99
	nDOM	PM	30.7 $\pm$ 11.9 (25.1–36.2)	33.1 $\pm$ 15.1 (26.0–40.2)	33.2 $\pm$ 14.0 (26.6–39.7)	F ${}_{2,19}=$ 4.41 $p=$ 0.02	0.17	0.02	0.98
		MW	16.1 $\pm$ 6.1 (13.2–18.9)	17.4 $\pm$ 6.3 (14.4–20.3)	17.4 $\pm$ 7.1 (14.1–20.7)	F ${}_{2,19}=$ 5.50 $p=$ 0.008	0.03	0.03	1.00

LA, Less-affected limb; MA, Most-affected limb; DOM, Dominant; nDOM, non-dominant; PM, Peak Moment in Newton meter; MW, maximal Work in Joules; SD, Standard Deviation; C.I., Confidence Interval; ${}^{\#}$ Session effect indicating the main effect of time across the 3 testing sessions and calculated by repeated measures ANOVA, which were run separately for each group; ${}^{*}p$ value calculated by Bonferroni-adjusted pairwise comparisons. All tests were conducted at 45 ${}^{\circ}$ /s of isokinetic angular velocity.

2.3 Data analysis

Statistical analysis was performed using the SPSS software (version 18) for Windows (SPSS Inc, Chicago, IL, USA). The highest PM value (in Nm) and MW value (in J) were recorded and used in the analysis. The MW/PM ratio was also calculated (in J/Nm). One-way ANOVA was employed to compare age, weight and the strength levels in the two groups. The Chi-square test was used to compare groups by gender. Means, standard deviations (SD) and 95% confidence intervals (CI) were calculated by side for PM and MW across the 3 testing sessions. Each OP was determined as the maximum value recorded among the tests. The equality of variances was assessed by the Levene’s test. A repeated-measures analysis of variance (ANOVA) was conducted to compare measurements among the sessions in the 2 groups to verify the repeatability of maximal isokinetic strength measurements in PwMS in comparison with healthy subjects. To control for any bias deriving from gender pooling, the strength levels were compared by two-way ANOVA between the groups with and without the male subjects. Significance was set at $p<$ 0.05. When significance was detected, Bonferroni-adjusted pairwise comparisons were performed to locate the difference. Cohen’s $d$ effect size magnitudes (small $\leqslant$ 0.5; moderate 0.51–0.79; large $\geqslant$ 0.8) were also used to estimate the clinical relevance of the inter-session differences and of the inter-group differences [6]. Linear correlation coefficients (Pearson’s $r$ ) were used to evaluate the degree of linear relationship between PM and MW. Test-retest reproducibility and responsiveness were estimated in both groups by calculating the following indices: 1) 2-way random ICC (ICC ${}_{2,1}$ ) for single measures [13, 33]; 2) SRD at the individual level (SRDi) $=$ 2.77*SEM where SEM (standard error of measurement) is equal to SD $\times$ $\surd$ (1-ICC ${}_{2,1}$ ) while SEM% $=$ (SEM/mean) $\times$ 100 [21].

Table 3
Reproducibility and responsiveness of maximal strength measurements from the ankle dorsiflexion muscles by group, limb testing session

Groups	Leg	Outcome	Reproducibility			Responsiveness
		measures	ICC ${}_{2,1}$			SRD
			Session 1 vs. 2 (95% C.I.)	Session 1 vs. 3 (95% C.I.)	Session 1 vs. 2 vs. 3 (95% C.I.)	SRDi (Range)	SRDi (%)
PwMS	LA	PM	0.98 (0.94–0.99)	0.94 (0.82–0.97)	0.93 (0.81–0.97)	2.44–2.76	11.4
		MW	0.98 (0.95–0.99)	0.92 (0.81–0.97)	0.95 (0.89–0.98)	1.40–1.82	19.4
	MA	PM	0.94 (0.88–0.98)	0.93 (0.83–0.97)	0.94 (0.87–0.97)	2.45–2.60	13.0
		MW	0.94 (0.86–0.98)	0.91 (0.78–0.96)	0.92 (0.82–0.97)	1.25–1.47	19.8
Control	DOM	PM	0.96 (0.90–0.98)	0.94 (0.85–0.98)	0.96 (0.92–0.98)	4.57–5.57	13.8
		MW	0.82 (0.61–0.91)	0.83 (0.61–0.93)	0.84 (0.70–0.93)	1.56–1.77	9.2
	nDOM	PM	0.92 (0.81–0.97)	0.96 (0.90–0.98)	0.95 (0.90–0.98)	4.67–5.50	14.4
		MW	0.94 (0.86–0.98)	0.94 (0.86–0.98)	0.95 (0.89–0.98)	2.40–2.77	14.1

LA, Less-affected limb; MA, Most-affected limb; DOM, Dominant; nDOM, non-dominant; PM, Peak Moment in Newton meter; MW, maximal Work in Joule; ICC, Intraclass Correlation Coefficient (two-way random); C.I., Confidence Interval; SRDi, Individual Smallest Real Difference; SRDi%, Individual Smallest Real Difference in percentage. SRDi absolute values follow the same unit of measurement of the relative outcome measure. All tests were conducted at 45 ${}^{\circ}$ /s of isokinetic angular velocity.

3. Results

All participants completed the 3 testing sessions. All tests were well tolerated by the subjects and no exacerbation and/or injuries occurred in the PwMS group throughout the entire duration of the study. Demographic and clinical characteristics of the participants are detailed in Table 1.

No differences emerged between groups for age, weight, height and gender. Since no difference in isokinetic strength was detected between males and females within each group, pooled data are reported.

People with MS were significantly weaker than healthy controls in all OP, with strength deficits ranging 37–38% in PM and 42–55% in MW, at large to very large effect sizes (Cohen’s $d$ 0.86–1.77). Linear correlation revealed a moderately positive relationship between PM and MW both in the control (dominant limb: $r=$ 0.62; non-dominant limb: $r=$ 0.81) and PwMS groups (less-affected and more-affected limb: $r=$ 0.73).

No inter-side difference in strength was found in controls. In PwMS the more-affected limb was significantly weaker than the less-affected one both in PM (14.5%, $p<$ 0.0005) and MW (24.3%, $p<$ 0.0005), with the deficit in MW within the more-affected limb greater than that in PM ( $p<$ 0.0005).

The MW/PM ratio was significantly higher ( $p=$ 0.0004) in the controls (non-dominant: 0.56 J/Nm; dominant: 0.58 J/Nm) than in PwMS (less-affected: 0.41 J/Nm; more-affected: 0.38 J/Nm), as shown in two representative subjects in Fig. 1.

Figure 1.

Representative concentric ankle dorsiflexion moment-angular position (‘isokinetic strength’) curves obtained at an angular velocity of 45 ${}^{\circ}$ /s. A – Control subject, B – patient with MS.

3.1 Inter-session differences by means

Descriptive statistics for the isokinetic findings of dorsiflexors’ strength are presented in Table 2.

The repeated-measures ANOVA, which was run separately for each group, revealed a main effect of the testing session in the PwMS group in both PM and MW. Bonferroni-adjusted pairwise comparisons detected significant reductions in both OP from the less-affected limb between sessions 1 and 2 (PM: $-$ 2.5%; MW: $-$ 4.2%) and sessions 1 and 3 (PM: $-$ 8.4%; MW: $-$ 14.9%) whilst in the more-affected limb significant reductions were found only in PM between sessions 1 and 2 (PM: $-$ 6.1%; MW: $-$ 5.3%) and between sessions 1 and 3 (PM: $-$ 7.7%; MW: $-$ 12.9%). Overall, the effect sizes of such differences between sessions were small for both PM and MW (Cohen’s $d$ ranging 0.1 to 0.3).

A main effect of the testing session was detected also in the control group. Pairwise comparisons revealed significant increases (8.4%) only in the non-dominant limb in PM between sessions 1 and 3 and in MW between sessions 1 and 2 and sessions 1 and 3, at small effect sizes (Cohen’s $d$ values $<$ 0.3).

The significant Session*Group interactions revealed that the temporal pattern of the changes over the 3 testing sessions was statistically different between the more-affected limb of PwMS and the non-dominant limb of controls (PM: F $=$ 7.837; $p=$ 0.001; MW: F $=$ 7.715; $p=$ 0.002) as well as between the less-affected limb of PwMS and the dominant limb of controls (PM: F $=$ 6.987; $p=$ 0.003; MW: F $=$ 7.818; $p=$ 0.001).

3.2 Reproducibility and responsiveness

The values of the ICC (reproducibility) and responsiveness (SRDi and SRDi%) of the PW and MW measurements are detailed by group, limb and type of comparison in Table 3. Relative reproducibility analyses performed separately for females and males within each group revealed that the scores calculated for females only were very similar to the pooled data reported in Table 3. In detail, in female PwMS ( $n=$ 16/20) ICC was 0.93–0.94 in the less-affected limb and 0.91–0.94 in the more-affected limb. In the female controls ( $n=$ 14/20) the coefficients ranged 0.82–0.95 in the dominant limb and 0.91–0.95 in the non-dominant limb.

The SRDi values relating to the PM of both legs in PwMS were appreciably lower than their MW-related counterparts: 11.4 vs 13% and 19.4 vs 19.8%, respectively and even lower than the parallel scores calculated for the controls (13.8 vs 14.4%).

4. Discussion

The main finding of this study is that the responsiveness of maximal isokinetic performance of the ankle dorsiflexors in PwMS is clinically acceptable and hence may serve for the interpretation of the changes in strength observed following conservative interventions in this population that is characterized by high fluctuations in muscle performance [18].

4.1 General aspects

As expected, PwMS proved significantly weaker than controls, strongly supporting previous findings pointing to some 30–70% reduction in strength [26, 30, 32, 37]. On the other hand, the extent of asymmetry between the stronger and the weaker limb was differently portrayed by MW (significant inter-limb deficit of 20–24%) than by PM (non-significant inter-limb deficit of 12–14%). Linear correlation analysis showed that only 36–64% of the variability in one of these OPs is directly influenced by the other, suggesting that PM and MW relate differently to muscle function, echoing recent studies of the ankle dorsiflexors in healthy subjects both at baseline and after training [23, 24]. In particular, since MW represents the dissipated muscular output over the entire range of joint motion [25], adding it to the PM may complement the assessment of muscle performance in PwMS, who experience difficulties in maintaining the strength levels over distance and over time [30].

Furthermore, inspection of the morphology of the dynamometric curves and the calculation of the MW/ PM ratio revealed a significant difference between healthy subjects and PwMS. In the former, the MW obtained during a maximal dorsiflexion effort was 58% of the PM, while in the latter it was only 38%. These findings indicate that in PwMS the capacity of the dorsiflexors to sustain muscle tension over a given ankle range of motion is likely to be more affected than the ability to exert an instantaneous PM. It is worth noting that the analysis was based on the use of a relatively low angular velocity (45 ${}^{\circ}$ /s) in view of the difficulties experienced by neurological patients in reaching high angular speeds [36]. Whether these results are generalizable to other muscles, velocities and contraction type, requires further exploration. However, we argue that both the MW and the MW/PM ratio may be considered an adjunct to the PM in the interpretation of findings pertaining to isokinetic muscle performance in PwMS, both at baseline and after rehabilitation.

4.2 Reproducibility over sessions

Reproducibility of PM and MW measurements, as assessed by ICC, proved excellent in PwMS and healthy subjects, regardless of gender. This general finding supports previous research relating to the ability of isokinetic dynamometry to yield clinically acceptable consistency of strength measurements in healthy subjects [12] as well as in PwMS under dynamic [17] and isometric conditions [3, 27, 32]. However, despite the high ICCs, significant inter-session differences were observed within each group, indicating that both PwMS and healthy controls performed differently depending on the time point of assessment. Despite the dedicated familiarization procedures, the two groups showed a different pattern of change across sessions, with strength increasing from session 1 to 2 in controls, while declining from session 1 to 3 in PwMS, in line with a recent single-system case series [22]. Even though a certain amount of practice-based improvement ( $+$ 8.1%) may have occurred in the control group from session 1 to session 2, the decline in performance (from $-$ 2.5% to $-$ 14.9%) exhibited by PwMS raises questions regarding the temporal stability of maximal isokinetic strength measurements in this population. This is a crucial issue in the context of intervention since an unreliable baseline would not allow tracking of changes induced by training [2]. The onset of fatigue, which could account for the high day-to-day variability in performance exhibited by PwMS [18], may have played a role. However, it is worth noting that in the present protocol, changes occurred between sessions 1 and 2, which were separated by only 1 day, whereas significant changes were detected from sessions 1–3 and 2–3, when at least a 6-day rest period was allowed. Therefore, it is likely that, besides fatigue, other factors contributed to the instability of measurements. We argue that these are to be sought within the pathophysiology and clinical features of MS. In particular, in PwMS the level of muscle performance is affected by the impaired recruitment of motor units, which may result in reduced muscle strength, power and rate of force development leading to the physical deconditioning commonly observed in this population [15, 17, 32]. This feature has been linked to a decreased muscle cross-sectional area, reduced aerobic-oxidative energy supply [16], increased muscular fatigue [19], decreased motoneuron firing rates [30] and reduced cortico-spinal neural drive [10].

From a statistical standpoint, the results provided by the repeated-measures ANOVA and ICC analyses seem to be contradictory. Weir [38] suggests that an ICC can be large (indicating good reliability) while the ANOVA shows a significant trials effect. Indeed, as long as they are consistent across all the subjects (as indicated by all ICCs $>$ 0.9) even small differences among testing sessions may result in a statistically significant effect for trials. The finding of excellent reliability along with significant differences in the mean may raise some concerns regarding the exclusive employment of ICC without consideration of the trials effect. However, the analysis of the effect sizes revealed low scores (Cohen’s $d$ 0.1–0.3) indicating that, although significant, the inter-session differences were not clinically relevant. Supporting this position is the observation that in the current cohort of PwMS even the largest change in PM, although flagged as statistically significant, was equal to less than 2 Nm.

Taken as a whole, the findings seem to confirm that, despite the employment of methodological procedures to optimize the consistency of the measurements, the assessment of maximal strength in PwMS is particularly challenging. Our cohort of PwMS achieved their peak performances in the first session, thus suggesting that one test may be sufficient to obtain valid findings. However, implementing 3 testing sessions revealed a trend of declining strength across sessions, which otherwise would have not been captured by a single test. Although the present study remains inconclusive as to how many tests are advisable for assessing PwMS, defining such variability across at least 3 sessions is a valuable information to adequately characterize the dynamometric profile of this population.

4.3 Responsiveness

Based on the results of the present study, any change in muscle performance falling below the individual cut-off value, expressed by SRDi, should be discarded even if statistically significant since it can be attributed to the measurement noise. This concept is particularly meaningful in terms of an intervention, since not every observed difference is clinically meaningful. Indeed, clinically meaningful improvements are at best no less than 10% and typically around 20–30% in orthopedic conditions [7], 26–55% in stroke survivors [20] and 11–39% in individuals with post-polio syndrome [10]. Importantly, the PM-based SRDi scores ranged 11.4–13% in PwMS and 13.8–14.4% in the controls, with the value for the more affected side in PwMS (13%) lower than the corresponding values for both the dominant and non-dominant sides in the controls. Noteworthy, in PwMS the PM-based SRDi scores were 40–50% lower than their MW counterparts, indicating that for these patients, a meaningful improvement in isokinetic strength could be detected relatively earlier by using the PM rather than the MW as the OP of choice. That said, the SRDi for MW for the ankle dorsiflexors in PwMS is coincident with what has been indicated as a clinically meaningful margin for real improvement in muscle function [8]. The findings thus indicate that in spite of inter-session differences in isokinetic strength and work capacity, the responsiveness of these measures is very well within clinically acceptable limits and may therefore serve for determination of clinically meaningful improvement or deterioration of these isokinetic markers, due to intervention or naturally occurring variations. Using the specific SRDi figures obtained in the ankle dorsiflexors of PwMS with same characteristics and measured isokinetically at relatively low velocities may practically translate into proclaiming that a change of slightly over 13% in the PM should be interpreted as clinically meaningful, thus informing a decision relating to the nature of the intervention.

In conclusion, measurements of maximal isokinetic concentric dorsiflexion strength obtained from PwMS proved to be highly reproducible, temporally unstable albeit in a clinically irrelevant manner but significantly well responsive to change. If the findings of this study are to be generalized to other muscle groups, contraction types and angular velocities, appropriate procedures should be instituted in PwMS when assessing physical capacities and/or the occurrence of meaningful changes that may be induced by rehabilitative interventions. These procedures should include methodological steps such as the inclusion of a proper familiarization protocol. Moreover, a comprehensive statistical approach should be employed in order to establish reproducibility and responsiveness of the measurements. Finally, the employment of the ANOVA for checking measurements’ stability over time and of the effect size for clinical relevance analyses may help to optimize the quantification and interpretation of test-retest results, which is particularly challenging in “unstable” individuals such as PwMS.

4.4 Study limitations

This study focused only on concentric testing. Eccentric measurements would have provided important information as lengthening contractions have a functional role in gait as well as in other functional activities. However, they were not included in order to keep the patients as mechanically unloaded as possible.

Another limitation was the employment of a single test velocity. However, the addition of more test velocities would not have necessarily changed the tenor of the main findings, as revealed by a study of orthopedic patients [41]. Admittedly, the efficiency of multiple velocity testing in MS requires further exploration, although the difficulty that these patients exhibit when exercising at high angular velocities should be taken into proper account [36].

In terms of sample size, although the present study is in line with the recommendation that reliability studies should enroll at least 15–20 patients [11], the present results need to be confirmed over a larger sample.

Finally, and as a precaution, the present findings should, at this point, not be generalized to muscle groups other than the ankle dorsiflexors considering their neurophysiological and biomechanical peculiarities [34].

4.5 Practical applications

In general, the massive increase of the literature dealing with new strategies to characterize muscle strength in PwMS is a decisive proof for the growing interest in new competencies and methodologies which are derived from exercise physiology and sports sciences. This interest has recently stimulated a translational process mandating the inclusion of exercise and strength conditioning experts in the multidisciplinary clinical team of neurologists, physiatrists and physiotherapists.

From the more focused viewpoint, strength and conditioning professionals involved in testing and rehabilitation of muscle strength in PwMS should be aware of the unique importance of the concept of responsiveness when adjusting or designing an intervention program. They should also take note of the additive value of using the MW when assessing the performance of people afflicted with this disease.

Footnotes

Acknowledgments

We wish to thank Paul Gottlieb, MA (Cantab), MPhil, for editing the manuscript. This work was supported by Fondazione Italiana Sclerosi Multipla (FISM Grant 2016/R/11).

Conflict of interest

The authors declare no conflict of interest.

References

Backman

Oberg

. Isokinetic muscle torque in the dorsiflexors of the ankle in children 6–15 years of age. Normal values and evaluation of the method. Scand J Rehabil Med 1989; 21(2): 97-103.

Beckerman

Roebroeck

Lankhorst

Becher

Bezemer

Verbeek

ALM

. Smallest real difference, a link between reproducibility and responsiveness. Qual Life Res 2001; 10(7): 571-578.

Bever

Johnson

Panitch

Anderson

Smith

. Isometric Measurement of Hamstrings and Quadriceps Strength in Multiple Sclerosis Patients: Sensitivity and Variability. Neurorehabil Neural Repair 1995; 9(4): 221-227.

Carroll

Herbert

Munn

Lee

Gandevia

. Contralateral effects of unilateral strength training: Evidence and possible mechanisms. J Appl Physiol 2006; 101(5): 1514-1522.

Caruso

Brown

Tufano

. The reproducibility of isokinetic dynamometry data. Isokinetics Exerc Sci 2012; 20(4): 239-253.

Cohen

. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Earlbaum Associates, 1988.

Dvir

. Difference, significant difference and clinically meaningful difference: The meaning of change in rehabilitation. J Rehabil Med 2015; 11(2): 67.

Dvir

. How much is necessary to indicate a real improvement in muscle function? A review of modern methods of reproducibility analysis. Isokinet Exerc Sci 2003; 11: 49-52.

Fimland

Helgerud

Gruber

Leivseth

Hoff

. Enhanced neural drive after maximal strength training in multiple sclerosis patients. Eur J Appl Physiol 2010; 110(2): 435-443.

10.

Flansbjer

Lexell

. Reliability of knee extensor and flexor muscle strength measurements in persons with late effects of polio. J Rehabil Med 2010; 42(6): 588-592.

11.

Fleiss

. The Design and Analysis of Clinical Experiments. New York: John Wiley Sons, 1986.

12.

Gleeson

Mercer

. The utility of isokinetic dynamometry in the assessment of human muscle function. Sports Med 1996; 1; 21(1): 18-34.

13.

Guyatt

Walter

Norman

. Measuring change over time: assessing the usefulness of evaluative instruments. J Chron Dis 1987; 40(2): 171-178.

14.

Harvill

. Standard error of measurement. Educ Meas Issues Pract 1991; 10(2): 33-41.

15.

Jørgensen

MLK

Dalgas

Wens

Hvid

. Muscle strength and power in persons with multiple sclerosis – A systematic review and meta-analysis. J Neurol Sci 2017; 376: 225-241.

16.

Kent-Braun

Castro

Weiner

Gelinas

Dudley

, et al. Strength, skeletal muscle composition, and enzyme activity in multiple sclerosis. J Appl Physiol 1997; 83(6): 1998-2004.

17.

Kjølhede

Vissing

Dalgas

. Multiple sclerosis and progressive resistance training: A systematic review. Mult Scler J 2012; 18(9): 1215-1228.

18.

Lambert

Archer

Evans

. Muscle strength and fatigue during isokinetic exercise in individuals with multiple sclerosis. Med Sci Sports Exerc 2001; 33(10): 1613-1619.

19.

Lenman

AJR

Tulley

Vrbova

Dimitrijevic

Towle

. Muscle fatigue in some neurological disorders. Muscle Nerve 1989; 12(11): 938-942.

20.

Lexell

Flansbjer

U-B

Brogårdh

. Isokinetic assessment of muscle function: Our experience with patients afflicted with selected diseases of the nervous system. Isokinetics Exerc Sci 2012; 20(4): 267-273.

21.

Lexell

Downham

. How to assess the reliability of measurements in rehabilitation. Am J Phys Med Rehabil 2005; 84(9): 719-723.

22.

Manca

Cabboi

Ortu

Ginatempo

Dragone

Zarbo

, et al. Effect of contralateral strength training on muscle weakness in people with multiple sclerosis: Proof-of-concept case series. Phys Ther 2016; 96(6): 828-838.

23.

Manca

Pisanu

Ortu

De Natale

Ginatempo

Dragone

, et al. A comprehensive assessment of the cross-training effect in ankle dorsiflexors of healthy subjects: A randomized controlled study. Gait Posture 2015; 42(1): 1-6.

24.

Manca

Solinas

Dragone

Deriu

. Isokinetic testing of muscle performance: new concepts for strength assessment. Isokinetics Exerc Sci 2015; 23(2): 69-75.

25.

Morrissey

. The relationship between peak torque and work of the quadriceps and hamstrings after meniscectomy. J Orthop Sports Phys Ther 1987; 8(8): 405-408.

26.

Miller

Gelinas

Kent-Braun

. Functional relationships of central and peripheral muscle alterations in multiple sclerosis. Muscle Nerve 2004; 29(6): 843-852.

27.

Paltamaa

West

Sarasoja

Wikström

Mälkiä

. Reliability of physical functioning measures in ambulatory subjects with MS. Physiother Res Int 2005; 10(2): 93-109.

28.

Polman

Reingold

Banwell

Clanet

Cohen

Filippi

, et al. Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria. Ann Neurol 2011; 69(2): 292-302.

29.

Portney

Watkins

. Foundations of Clinical Research: Applications to Practice. 3

{}^{\text{rd}}

Edition. F.A. Davis Company, 2008.

30.

Rice

Vollmer

Bigland-Ritchie

. Neuromuscular responses of patients with multiple sclerosis. Muscle Nerve 1992; 15(10): 1123-1132.

31.

Sandroff

Sosnoff

Motl

. Physical fitness, walking performance, and gait in multiple sclerosis. J Neurol Sci 2013; 328(1-2): 70-76.

32.

Schwid

Thornton

Pandya

Manzur

Sanjak

Petrie

, et al. Quantitative assessment of motor fatigue and strength in MS. Neurology 1999; 53(4): 743-750.

33.

Shrout

Fleiss

. Intraclass correlations: Uses in assessing rater reliability. Psychol Bull 1979; 86(2): 420-428.

34.

Tallent

Goodall

Hortobágyi

St Clair Gibson

Howatson

. Corticospinal responses of resistance-trained and un-trained males during dynamic muscle contractions. J Electromyogr Kinesiology 2013; 23(5): 1075-1081.

35.

Thomas

Nelson

. Research methods in physical activity. Champagne, IL: Human Kinetics Books, 1990.

36.

Tiffreau

Ledoux

Eymard

Thévenon

Hogrel

. Isokinetic muscle testing for weak patients suffering from neuromuscular disorders: a reliability study. Neuromuscul Disord 2007; 17(7): 524-531.

37.

Van der Kamp

de Noordhout

Thompson

Rothwell

Day

Marsden

. Correlation of phasic muscle strength and corticomotoneuron conduction time in multiple sclerosis. Ann Neurol 1991; 29(1): 6-12.

38.

Weir

. Quantifying test-retest reliability using the intraclass correlation coefficient and the SEM. J Strength Cond Res 2005; 19(1): 231-240.

39.

Hart

Draper

Olsen

. Peer and teacher sociometrics for preschool children: Cross-informant concordance, temporal stability, and reliability. Merrill-Palmer Q 2001; 47(3): 416-443.

40.

Zackowski

Smith

Reich

Gordon-Lipkin

Chodkowski

Sambandan

, et al. Sensorimotor dysfunction in multiple sclerosis and column-specific magnetization transfer-imaging abnormalities in the spinal cord. Brain 2009; 132(5): 1200-1209.

41.

Zemach

Almoznino

Barak

Dvir

. Quadriceps insufficiency in patients with knee compromise: how many velocities should an isokinetic test protocol consist of? Isokinetics Exerc Sci 2009; 17(3): 129-33.

Strength and work parameters in people with multiple sclerosis and in healthy individuals: A responsiveness study of the ankle dorsiflexors

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Materials and methods

2.1 Subjects

2.2 Measurements of ankle dorsiflexion muscles’ strength

Table 1 Demographic and clinical features of participants

Table 3 Reproducibility and responsiveness of maximal strength measurements from the ankle dorsiflexion muscles by group, limb testing session

3.2 Reproducibility and responsiveness

4. Discussion

4.1 General aspects

4.2 Reproducibility over sessions

4.3 Responsiveness

4.4 Study limitations

4.5 Practical applications

Footnotes

Acknowledgments

Conflict of interest

References

Table 1
Demographic and clinical features of participants

Table 3
Reproducibility and responsiveness of maximal strength measurements from the ankle dorsiflexion muscles by group, limb testing session