SEARCH

SEARCH BY CITATION

Keywords:

  • cerebrovascular accident;
  • lower extremity;
  • mobility limitation;
  • quadriceps, knee flexors;
  • weakness

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Functional performance
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgments
  10. References

The objective of this study was to assess the nature of muscle weakness in both legs after stroke compared with able-bodied control individuals and to examine whether there is a relationship between the degree of muscle weakness and coactivation of knee extensors and flexors as well as voluntary activation capacity of knee extensors of both paretic and non-paretic legs and indices of functional performance. Maximal voluntary isometric torques of knee extensors (MVCe) and flexors (MVCf) were determined in 14 patients (bilaterally) and 12 able-bodied controls. Simultaneous measurements were made of torque and surface EMG from agonist and antagonist muscles. Coactivation was calculated. Supramaximal triplets were evoked with electrical stimulation to estimate maximal torque capacity and degree of voluntary activation of knee extensors. MVCs, activation and coactivation parameters were correlated to scores of seven functional performance tests. MVCe, MVCf and voluntary activation were lower in paretic lower limb (PL) compared with both non-paretic lower limb (NL) and control. Besides, all these parameters of NL were also lower than control. Electrically evoked torque capacity of knee extensors of PL was about 60% of both NL and control, which were not significantly different from each other. Strong significant correlations between strength, as well as voluntary activation, and functional performance were found. Coactivation did not correlate well with functional performance. Thus, whereas for NL activation failure can explain weakness, for PL both activation failure and reduced intrinsic torque capacity are responsible for the severe weakness. Activation capacity and muscle strength correlated strongly to functional performance, while coactivation did not.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Functional performance
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgments
  10. References

About 80–90% of patients has a spastic paresis in at least one of the limbs shortly after the stroke (Bonita & Beaglehole, 1988). After 6 months, half of the patient population still suffer from motor impairments revealed by hemiparesis (partial paralysis that is manifested by decreased muscle strength) or hemiplegy (taking into account both spastic and paretic components contribute to motor functional loss) (Bourbonnais & Vanden Noven, 1989). Furthermore, Bohannon (1989, 2007) showed that the severity of the paresis strongly relates to the loss of functional performance in stroke.

Given the importance of the locomotor muscles in daily life, this study will focus on the muscles of the thigh. Muscle weakness in hamstrings and quadriceps on the paretic side (contralateral to the lesion) is a well-known phenomenon (Davies et al., 1996; Newham & Hsiao, 2001). In addition, reduced muscle strength was also reported for the non-paretic limbs (ipsilateral to the lesion) (Bohannon & Walsh, 1992; Davies et al., 1996; Harris et al., 2001) and this weakness develops already in the first week following acute ischaemic hemiplegic stroke (Harris et al., 2001). Moreover, several studies reported correlations between muscle strength of knee extensors and knee flexors of both paretic and non-paretic lower limbs and functional performance (Bohannon, 2007), such as gait and stair-climbing speeds (Kim & Eng, 2003), short- and long-distance walking (Patterson et al., 2007), transfer capacity (Bohannon, 1988) and standing-up performance (Corrigan & Bohannon, 2001).

The common goal of rehabilitation of stroke patients is to retrain functional performance skills (e.g. locomotion, balance, stair climbing ability and transfer capacity). Clearly, muscle strength seems an important determinant for improving locomotor performance in stroke patients and Ouellette et al. (2004), Ada et al. (2006) and Yang et al. (2006) suggested that strength training should be an important part of the rehabilitation programme. To do this effectively and to optimize interventions or adaptation in rehabilitation programmes of stroke patients, it is important to elucidate which parameters may underlie muscle weakness (and impaired functional performance) and therefore are of most interest to improve.

Previous studies have shown an overall loss of muscle mass, type II fibre atrophy and predominance of type I fibres in paretic muscles after stroke (Scelsi et al., 1984; Dietz et al., 1986; Dattola et al., 1993; Hachisuka et al., 1997). It is therefore expected that at least part of the muscle weakness relates to adaptations in the intrinsic muscle properties. However, Bourbonnais & Vanden Noven (1989) and Newham & Hsiao (2001) showed that stroke patients have a disturbed central activation, which would lead to an impairment of the ability to maximally drive their muscles and most probably also to altered coordination of muscles involved. A disturbed neural control in the sense of abnormal coactivation of antagonist muscles during maximal voluntary extension and flexion contractions might be indicative for a decreased neural control during locomotor tasks such as walking or stair climbing. Moreover, excessive coactivation of antagonist muscles, in addition to reduced voluntary activation of agonist muscles might further reduce the net torques around the joints and are therefore also expected to contribute significantly to the impaired use of the muscles of the thigh. Some of these parameters indeed have been associated with functional performance. For example, Chae et al. (2002) found significant correlations between coactivation and motor impairment in upper limb hemiplegia. To our best knowledge, it is not known if such relationship also holds for the lower limbs in stroke patients.

Thus, there is clear evidence suggesting that both impaired neural control as a direct consequence of the stroke and adaptations in intrinsic muscle characteristics underlie the weakness and therefore impaired functional performance in patients after stroke. Information about the relative contribution of these factors is, however, lacking. Therefore, to provide scientific background information for developing specific intervention or rehabilitation programmes, the purpose of this study was to assess the nature of muscle weakness in both legs after stroke compared with able-bodied control individuals. Furthermore, we wanted to examine whether there is a relationship between the degree of muscle weakness and coactivation of the knee extensors and flexors as well as voluntary activation capacity of knee extensors of both paretic and non-paretic lower limbs and indices of functional performance.

Our hypothesis is that both limbs show reduced intrinsic muscle strength, impaired voluntary activation capacity and disturbed coactivation, which correlate significantly with indices of functional performance after stroke.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Functional performance
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgments
  10. References

Subjects

Fourteen stroke patients and 12 able-bodied control subjects volunteered in this study. Patients characteristics (see Table 1) with a wide range of motor impairments from very severe to very mild, based on the scores at the tests of functional performance, were recruited for this study. They entered the study on average 3·5 months after stroke and 2 months after admission in the rehabilitation centre. Controls were matched as much as possible for age, length and weight (mean ± standard deviation: see Table 1). Before participation, each subject was thoroughly informed about the procedures, filled out a health questionnaire and signed an informed consent. Inclusion criteria for patients were a minimum age of 18 years and a hemiparesis of the lower extremity as a result of first ever-stroke. The exclusion criteria were medical complications (such as unstable cardiovascular problems), severe cognitive and/or communicative problems preventing the ability to follow verbal instructions or limiting the ability to perform the requested tasks (e.g. aphasia, hemi-neglect) and contra-indications for electrical stimulation (unstable epilepsy, cancer, skin abnormalities, pacemaker). The project carried the approval of the institutional review board (Medical Ethical Committee) of the VU University Medical Centre, Amsterdam, the Netherlands.

Table 1.   Characteristics of the subjects.
 Sex (men/ women)Age (year)Weight (kg)Length (cm)Lesion side (left/right)Stroke type (haemorrhagic/ ischaemic)Time after stroke (days)Time after admission (days)
Mean ± SD(10M 4W)55·9 ± 10·474·9 ± 14·3174·2 ± 10·1(6L 8R)(5H 9I)109 ± 4667 ± 33
Control Mean ± SD(7M 5W)58·1 ± 12·275·7 ± 11·9175·6 ± 6·5    

Experimental set-up

Contractile properties and functional performance were measured. For the present study, only a selection of muscle function characteristics (strength, voluntary activation and coactivation) was used. The experiments were spread over 4 different days (sessions) with at least 1 day of rest in between.

Force measurements

Maximal voluntary and electrically evoked forces of the knee extensors and knee flexors were measured on a custom built Lower EXtremity System (LEXS). This set-up can be used in both supine and sitting position. Subjects were placed on the LEXS in a supine position (Fig. 1, upper) and were tilted to the measuring position (Fig. 1, lower): they were seated with their back 10° backwards and a hip angle of 100° (180º being fully stretched hips) and knee angle 60° (0º= full extension). This semi-supine position is comfortable for patients either with or without balance problems. The position of the LEXS was adjusted to establish these angles and to enable alignment of the knee axis (during maximal voluntary contraction) with the rotation-axis of LEXS. A hip belt (Fig. 1C, lower) was fixed tightly to avoid changes in hip and knee angle during (isometric) contractions and a trunk belt (Fig. 1A, lower) was fixed for stabilization. The lower leg was strapped tightly to a force transducer (KAP, E/200 Hz, Bienfait B.V Haarlem, The Netherlands, range: 0–2 kN) just above the ankle by means of a cuff (Fig. 1D). Upper–lower translation of the force transducer was adapted. Active knee angle was determined with a handheld goniometer (model G300; Whitehall Manufacturing, CA, USA) using the greater trochanter, the lateral epicondyl of the femur and the lateral malleolus of the fibula as references (Kooistra et al., 2005). The distance between the lateral femur epicondyl and a fixed point at the force transducer was measured representing the external moment arm. For the maximal voluntary contractions (MVCs) of the knee flexors, the leg was fixed tightly by a top restraining bar that was secured on the thigh, just proximal to the knee joint to minimize the movement of the leg and to avoid changes in hip and knee angles during force generation (Fig. 1B, lower). An ankle brace was placed to keep the ankle in plantar flexion position, reducing the contribution of the gastrocnemius medialis muscle to the knee flexion. During voluntary contractions the skin was shaved and cleaned with alcohol for EMG measurements (Biotel 99, The Netherlands). After that, surface Ag-AgCl electrodes were placed in a bipolar figuration, in line with the muscle fibre direction, with a centre to centre inter-electrode distance of 25 mm on the muscle bellies of the vastus lateralis muscle (VL), rectus femoris muscle (RF), vastus medialis muscle (VM), biceps femoris muscle (BF) and gastrocnemius medialis muscle (GM). One reference electrode was placed on the patella of the measured lower limb.

image

Figure 1.  Lower extremity system (LEXS). Upper: LEXS as used in supine position. Lower: LEXS with subject in sitting position. A, Shoulder belt. B, Upper leg fixation. C, Hip belt. D, Force transducer.

Download figure to PowerPoint

Experimental procedures

Familiarization session

The measurements of the familiarization session were performed with the non-paretic lower limb to check whether the instructions were understood by the subject. After a warming-up (existing of five submaximal contractions) subjects were trained to perform maximally isometric knee flexion and extension contractions. Subsequently, the subjects were familiarized with electrical stimulation.

Muscle strength

Subjects were asked to maximally generate isometric knee extensions for 3–4 s to determine maximal voluntary knee extension torque (MVCe). Alternately, MVCs with the knee flexors (MVCf) were performed. Two to four attempts were allowed for both MVCfs and MVCes, separated by 2 min of rest. MVC was taken as the highest value of these attempts, which did not exceed preceding attempts by >10%, allowing a maximum of four attempts. Real-time force was visible on a computer screen. Subjects were vigorously encouraged to exceed their previous maximal value, which was also displayed to confirm the subject’s achievement throughout the test. EMG was measured as described above (see section Force measurements). The same measurements performed on the paretic lower limb were repeated with the non-paretic lower limb, carried out on a separate day. Control subjects only performed one of these sessions, with the right leg.

Voluntary activation

Volitional tests rely heavily on the patient’s motivation and the ability to maximally recruit their muscles and are often not an accurate reflection of the maximal torque generating capacity of the muscle. Electrically evoked contractions are independent of the patient’s effort. Therefore, a modified superimposed stimulation technique was used in which electrically evoked triplets (pulse train of three rectangular 200 μs pulses applied at 300 Hz) were used to establish the subjects’ capacity to voluntarily activate their muscles (Kooistra et al., 2005).

After explanation of the procedure, the skin of the thigh of the subject was shaved (when necessary) and a pair of self-adhesive surface electrodes (13 × 8 cm; Model 283100; Schwa-Medico, Nieuw Leusden, the Netherlands) was placed over the proximal and distal part of the anterior thigh and moment arm was measured as described above. The knee extensors were electrically stimulated using a computer-controlled constant current stimulator (Digitimer DSH7; Digitimer Ltd., Welwyn Garden City, UK). Measurements started with the paretic lower limb, knee angle 60º. First, stimulation current was increased until torque measured in response to a triplet levelled off. The current (in mA) was then increased by a further 20 mA to ensure supramaximal stimulation. It was assumed that at this point all muscle fibres of the knee extensors were activated. These high frequency stimulations (triplets) produce maximal responses in terms of torque production (de Haan, 1998), thereby limiting the sensitivity to, for instance, length-dependent changes in calcium sensitivity and post-tetanic potentiation and improving the signal-to-noise ratio. Subjects underwent measurements consisting of a triplet superimposed on the plateau of the force signal of the MVC. Subsequently, these measurements were performed with the non-paretic lower limb. Control subjects performed this session with the right leg only.

Functional performance

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Functional performance
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgments
  10. References

To determine the relationship of changes in motor control and muscle function with functional performance, the following tests were performed by the subjects under supervision of a physiotherapist (except for the Rivermead mobility index, which was carried out by one of the researchers):

  • Timed ‘Up and Go’ test (TUG) requires patients to stand up from a chair, walk 3 m, turn around, return, and sit down again. Time to fulfil this test is measured (Podsiadlo & Richardson, 1991).
  • 10 meter walk test (10 m) is performed at comfortable (self selected) walking speed by patients who are able to walk independent with or without mobility aid and/or orthesis. Time to walk 10 m is measured and averaged over three trials (Smith & Baer, 1999).
  • Berg balance scale (BBS) assesses sitting and standing balance and exists of 14 test-items, scored on an ordinal five-point scale (0–4). It gives an estimation of the chance that patients with stroke will fall (Berg et al., 1989, 1992, 1995).
  • Motricity index (MI) evaluates the arbitrary movement activity and maximum isometric muscle force. Possible scores are 0–9–14–19–25–33 at each of the three parts of the test for lower extremities (Demeurisse et al., 1980; Collin & Wade, 1990; Cameron & Bohannon, 2000).
  • Functional ambulation categories score (FAC) evaluates the measure of independence of walking of the patient. Categories are scored on a six-point scale (0–5) (Holden et al., 1984, 1986).
  • Brunnstrom–Fugl-Meyer (FM), lower extremity, is a test for evaluation of patellar, knee flexor and Achilles reflexes, flexor and extensor synergies, isolated movements of knee flexor and ankle dorsal flexor function and normal reflex activity of the quadriceps and triceps surae muscles in hemiplegic patients (Fugl-Meyer et al., 1975).
  • Rivermead mobility index (RMI) comprises a series of 14 questions and one direct observation, and covers a range of activities from turning over in bed to running. It is a measure of mobility disability, which concentrates on body mobility (Collen et al., 1991).

Data analysis

Real-time force applied to the force transducer was displayed online on a computer monitor and digitally stored (1 kHz) on computer disc. The force signals were automatically corrected for gravity of the leg. All force signals were low-pass filtered (fourth order, 50 Hz, Butterworth).

Maximal voluntary extension and flexion torque (Nm) was determined as the peak force from the force plateau multiplied by the external moment arm. Ratios of MVCf and MVCe are used to indicate which muscle group is most affected.

Voluntary activation is defined as the completeness of skeletal muscle activity during voluntary contractions and was calculated by means of a modified interpolated twitch technique (Kooistra et al., 2005).

  • image

Here the superimposed triplet is the force increment during a maximal contraction at the time of stimulation and the control triplet is that evoked in the relaxed muscle (Shield & Zhou, 2004).

EMG signals of the voluntary maximal knee extensions and flexions were amplified (×100), digitized (1 kHz) and stored with the synchronized force signal on computer disc. All EMG signals were band-pass filtered (10–400 Hz) and rectified. Rectified surface EMG amplitude (rsEMG) was calculated for the RF, VL, VM, BF and GM for 1000 ms of the plateau of the force signal. rsEMG served to assess activation levels of the knee extensors and flexors during MVC as well as to determine the coactivation of the thigh muscles during MVC. To assess the level of coactivation, rsEMG of the antagonist muscles was normalized to the maximum rsEMG during their primary function (working as agonists). For instance, the level of coactivation of knee flexors during knee extension was expressed as a percentage of the maximal knee flexion rsEMG (rsEMG/rsEMGmax) (Macaluso et al., 2002).

Correlations were calculated between the coactivation parameters and the tests of functional performance, between MVCs and the tests and between voluntary activation and the tests.

Statistics

All results were presented as means ± standard deviations. For the comparison of MVC, MVCf:MVCe ratio, triplet torque, activation level and rsEMG/rsEMGmax between non-paretic, paretic and control lower limbs, one-way analysis of variance (ANOVA) was used. In case of significance, Bonferroni post hoc tests were used. To assess the relation between coactivation, activation level and strength and outcome measurements of the tests of functional performance, non-parametric Spearman correlation was used because of the ordinal scale of most of the tests. For each statistical analysis, the level of significance was set at P<0·05.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Functional performance
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgments
  10. References

All subjects of whom the data were included for analysis could tolerate the electrical stimulation. Subjects performed some MVCs without stimulation during the familiarization session. Only when MVCs with expected superimposed stimulation matched those without expected stimulation, data were used for analysis. Data were not complete for some of the patients due to discomfort (duration of the experiments, electrical stimulation) or unreliable data. The number of patients analysed are indicated in the text.

Muscle strength

Both MVCe and MVCf torques (Table 2) were significantly lower in the paretic lower limb as well as the non-paretic lower limb compared with control. Furthermore, MVCe and MVCf of the paretic limb were also significantly lower than the non-paretic limb. The ratio MVCf:MVCe (Table 2) was significantly lower for the paretic limb compared with both non-paretic limb and control, which in turn were not different from each other.

Table 2.   MVC during extension (MVCe) and flexion (MVCf), voluntary activation and triplet torque of knee extensors of the paretic lower limb (PL), non-paretic lower limb (NL) and control (mean ± standard deviation).
 ControlNLPL
  1. *Significantly lower than control.

  2. †Significantly lower than NL.

MVCe (Nm)223·2 ± 47·8 (n = 12)151·8 ± 54·8* (n = 13)62·1 ± 47·9*,† (n = 14)
MVCf (Nm)89·5 ± 29·0 (n = 12)56·3 ± 26·9* (n = 13)10·3 ± 16·1*,† (n = 14)
Ratio MVCf:MVCe0·40 ± 0·07 (n = 12)0·36 ± 0·09 (n = 13)0·14 ± 0·19*,† (n = 12)
Voluntary activation (%)93·6 ± 4·1 (n = 12)75·1 ± 7·3* (n = 10)57·8 ± 24·6*,† (n = 6)
Triplet torque (Nm)99·1 ± 12·6 (n = 12)93·7 ± 25·6 (n = 12)55·8 ± 26·8*,† (n = 10)

The extension torque evoked by a triplet for the paretic lower limb was about 60% (P<0·05) of both non-paretic lower limb and control, which were not different (Table 2). Maximal voluntary activation of both paretic and non-paretic lower limb was significantly lower than control and that of the paretic lower limb was significantly lower than the non-paretic lower limb (Table 2).

rsEMG for activation and coactivation

During maximal knee extension, rsEMG of VL, RF and VM was significantly lower for the paretic limb (n = 13) compared with both non-paretic limb (n = 12) and control (n = 12) (Fig. 2a), whereas no differences were observed across groups for rsEMG of the antagonist muscles (BF and GM, Fig. 2d).

image

Figure 2.  rsEMG values of knee extensors vastus lateralis (VL), rectus femoris (RF) and vastus medialis (VM) during maximal voluntary knee extension (a) and flexion (c) and absolute EMG values of knee flexors biceps femoris (BF) and gastrocnemius medialis (GM) during maximal voluntary knee flexion (b) and knee extension (d) for controls, non-paretic lower limb and paretic lower limb. *Significantly different from control + significantly different from non-paretic lower limb.

Download figure to PowerPoint

During maximal knee flexion, the paretic lower limb (n = 12) showed reduced rsEMG (P<0·05) of all five muscles studied (agonist and antagonists, Fig. 2b and c) compared with control (n = 12) and when compared with the non-paretic lower limb (n = 12) this was true for all muscles except for VL. No differences in rsEMG were found between the non-paretic limb and control, neither during MVCe nor during MVCf.

The ratio rsEMG/rsEMGmax was used as a measure of coactivation. This ratio (the level of coactivation) of the knee extensors (VL, RF and VM) during knee-flexion was not significantly different across groups (Fig. 3a). However, as can be observed in Fig 3, significantly higher ratio for BF and GM during maximal voluntary knee extension was found in the paretic compared with the non-paretic lower limb and controls. The ratio in the non-paretic lower limb did not differ from controls.

image

Figure 3.  (a) Coactivation ratio rsEMG/reEMGmax of vastus lateralis (VL), rectus femoris (RF) and vastus medialis (VM) during maximal voluntary knee flexion and (b) coactivation of biceps femoris (BF) and gastrocnemius medialis (GM) during maximal voluntary knee extension for controls, non-paretic lower limb and paretic lower limb. Note the unrealistic coactivation values above 1 for the paretic lower limb during extension (b), which is mainly due to problems with activation of muscles in this group. *Significantly different from control + significantly different from non-paretic lower limb.

Download figure to PowerPoint

Functional performance

Results of the tests of functional performance are shown in Table 3. Three subjects could not perform the 10-m walk test (10 m) and the Timed ‘Up and Go’ test (TUG). They were not used for data-analysis for these two tests. One of them scored zero at the Fugl Meyer (FM) and Functional ambulation categories (FAC). Note that significant correlations or trends were found between MVCe and MVCf of the paretic lower limb and all tests of functional performance (Table 4). Other significant correlations and trends are also shown in Table 4.

Table 3.   Maximal possible scores (if applicable) and the mean scores and standard deviations (SD) of stroke patients of the present study at the tests of functional performance.
 RMI10 mFACBBSTUGFMMI
  1. RMI, Rivermead mobility index; 10 m, 10-m walk test; FAC, Functional ambulation categories; BBS, Berg balance scale; TUG, Timed Up and Go; FM, Fugl Meyer and MI, Motricity index.

Maximum possible scores1555634100
Mean8·133·8s3·337·744·4s16·546·0
SD3·819·1s1·413·322·6s8·520·9
n14111414111411
Table 4.   Significant correlations (bold) and trends were found between the tests of functional performance and MVCes, MVCfs, some of the rsEMG/rsEMGmax parameters of the measured muscles (RF, rectus femoris; BF, biceps femoris and GM, gastrocnemius medialis) and activation measured in the paretic lower limb (PL) and non-paretic lower limb (NL).
 RMI10 mFACBBSTUGFMMI
  1. See details of abbreviation in Table 4. X represents no significant correlation or trend.

  2. †Trend 0·054 < P < 0·096.

  3. ‡Correlation is significant at the 0·05 level (two tailed).

  4. §Correlation is significant at the 0·01 level (two tailed).

  5. ††Correlation is significant at the 0·001 level (two tailed).

MVCe PL0·744§(n = 14)−0·545† (n = 11)0·780††(n = 14)0·764††(n = 14)−0·573† (n = 11)0·746§ (n = 14)0·761§ (n = 14)
MVCf PL0·643‡(n = 14)−0·763§(n = 11)0·463† (n = 14)0·527† (n = 14)−0·858††(n = 11)0·687§ (n = 14)0·727§ (n = 14)
MVCe NL0·510† (n = 13)−0·699‡(n = 10)X0·504† (n = 13)X0·647‡ (n = 13)0·615‡ (n = 13)
MVCf NLX−0·634‡(n = 10)0·485† (n = 13)0·516† (n = 13)−0·622† (n = 10)0·597‡ (n = 13)0·570‡ (n = 13)
rsEMG/rsEMGmax RF PLXX−0·636‡(n = 11)−0·543† (n = 11)XXX
rsEMG/rsEMGmax BF PLX0·738‡(n = 8)XX0·857§ (n = 8)XX
rsEMG/rsEMGmax GM PLX0·786‡(n = 8)XX0·714‡ (n = 8)XX
rsEMG/rsEMGmax GM NL0·579† (n = 11)X0·622(n = 11)0·556† (n = 11)XXX
Activation PL0·841‡(n = 6)X0·845‡ (n = 6)0·829‡ (n = 6)X0·883‡ (n = 6)X
Activation NL0·665‡(n = 10)−0·964††(n = 7)XX−0·893§ (n = 7)X0·828§ (n = 10)

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Functional performance
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgments
  10. References

The present report shows significant adaptations in both central neural activation and more intrinsic muscle properties in individuals 3.5 months after stroke. Weakness was observed in both paretic and non-paretic knee extensors and flexors. However, in the non-paretic lower limb this reduced strength seemed primarily due to impaired voluntary activation whereas in the paretic lower limb both impaired voluntary activation as well as reduced intrinsic torque capacity seemed responsible.

An interesting and new observation is the strong significant correlation between functional performance and voluntary activation capacity of both lower limbs in stroke patients.

Strength

Maximal voluntary contractions

The results of the present study demonstrate that muscle strength of knee flexors and extensors is reduced after stroke in both legs with the greatest reduction in voluntary strength observed in the paretic lower limb muscles (Table 2), which is in accordance with previous studies (Bohannon & Walsh, 1992; Davies et al., 1996; Harris et al., 2001; Newham & Hsiao, 2001). The finding by Harris et al. (2001) that muscle strength of the non-paretic lower limb was already reduced in the first week after stroke, suggests that the direct consequences of stroke are the most likely cause of this muscle weakness.

Furthermore, the lower MVCf:MVCe ratios in the paretic lower limb compared with control (Table 2) indicate a more pronounced weakness of the flexor muscles compared with extensors after stroke. This was also reported by Newham & Hsiao (2001), although they found a higher ratio for the paretic lower limb (0·31) compared with our results (0·14), which might be explained by possible differences in severity of stroke. Overall, the results suggest that the knee flexors are more affected than knee extensors after stroke. Alternatively, it may be possible that both muscle groups are equally affected by the stroke itself, but that the difference appeared because during rehabilitation, training of knee extensors is more emphasized than that of knee flexors. Moreover, Twitchell (1951) described predictable spontaneous motor recovery with a fixed sequence after (initial paralysis after) stroke. It was stated that extensor movement is the first movement to recover. Accordingly, as our subjects were studied within 4 months after stroke such sequential spontaneous recovery could explain why knee extensors of the paretic lower limb seem less affected than knee flexors. Also the presence of synergies might be a possible explanation for the lower MVCf:MVCe ratio in the paretic limb compared with the non-paretic limb and control. Some patients have lost independent control over certain muscle groups after stroke, which results in coupled joint movements which are often inappropriate for daily tasks. These coupled movements or inability to master individual joint movements are known as synergies. Although Brunnström (1966) and Welmer et al. (2006) demonstrated that the presence of synergies can negatively affect functional performance, Neckel et al. (2006) showed little evidence of abnormal synergy patterns in lower limbs in isometric conditions. In the present study, scores at the (Brunnström) Fugl Meyer test show that as much flexor as extensor synergies are present in the paretic lower limb. Thus, it seems that, in line with the results of Neckel et al. (2006), the presence of synergies are not a likely explanation for a lower MVCf:MVCe ratio in the paretic lower limb.

Voluntary activation

The present results show that maximal voluntary activation capacity of the knee extensors is reduced in both lower limbs after stroke (Table 2), which can be attributed to damage of the descending motor tracts following stroke (Ward, 2004). Voluntary activation of 58%, 75% and 94% for respectively paretic, non-paretic and control lower limb was found. This means activation failure (defined as 100-voluntary activation (%)) of 42% for the paretic lower limb, 25% for the non-paretic lower limb and 6% for control which is very comparable with the values of Newham & Hsiao (2001). It was shown before that the ability to maximally drive the knee extensor muscles is usually overestimated and that the level of overestimation increases with lower activation capacity (Kooistra et al., 2007) . This would suggest that as the stroke patients have a lower activation capacity compared with controls, the difference in the ability to access the muscles’ potential between limbs is most likely underestimated. In other words, the real differences in maximal voluntary activation between paretic lower limb and control in the present study will be even more pronounced than calculated here.

Unfortunately, from pilot experiments we concluded that presently it is difficult to obtain reliable maximal voluntary activation values in the knee flexors and therefore we are unable to estimate the possible influence of voluntary activation capacity on strength of the knee flexors. On the other hand, interestingly, the higher values found for EMG of flexor muscles during extension (Fig. 2d, paretic lower limb, PL) than during flexion (Fig. 2b, PL) suggest that the voluntary activation capacity of the knee flexors is possibly severely affected.

Maximal torque capacity

Triplet torque is a measure for the maximal (intrinsic) torque capacity of the muscle, independent of voluntary activation. While all muscle fibres were assumed to be recruited with the used supramaximal stimulation, the muscle fibres of the paretic knee extensors only generated about half of the torque of controls (Table 2). Thus, besides reduced voluntary activation there is a reduction in the intrinsic torque capacity of the paretic knee extensors. In contrast, torque capacity of the non-paretic lower limb did not differ from control (Table 2). This is different from Harris et al. (2001) who demonstrated that apart from reduced central activation there is an additional reduction in the twitch torque of the non-paretic quadriceps muscles.

Part of the reduction of the torque capacity in the paretic lower limb might be explained by muscle atrophy. For instance, Metoki et al. (2003) and Ryan et al. (2002) found hemiparetic skeletal thigh muscle atrophy more than 6 months after stroke that might contribute to functional disability in chronic hemiparetic stroke patients. Atrophy is not induced by upper neuromotor lesion itself, but may be attributed to the sedentary lifestyle as a result of functional disability (e.g. hemiparesis), immobilization (e.g. lying in bed or sitting in a wheelchair) and malnutrition after stroke (Harris et al., 2001; Metoki et al., 2003). Therefore, although muscle size or volume was not directly measured in the present study, the reduced torque capacity in the paretic limb is likely due to muscle atrophy, especially in view of the period of 4 months of relative disuse after stroke before patients were measured. Nevertheless, the found 50% reduction in torque capacity reported in our study would indicate a similar degree of atrophy, which seems very excessive if attributed solely to changes in muscle volume. Therefore, it seems reasonable to expect that other factors, like for example a disturbed excitation contraction coupling mechanism, may additionally contribute to the specific reduction of torque capacity (D’Antona et al., 2007).

Coactivation

Excessive coactivation, for instance as a result of impaired reciprocal inhibition (Ivanhoe & Reistetter, 2004), may underlie the lower torques as measured in the paretic and non-paretic lower limb compared with control.

Conflicting results are found between and within several muscle groups regarding coactivation after stroke. For instance, Knutsson & Martensson (1980) and Neckel et al. (2006) have indeed found increased coactivation of thigh muscles in hemiplegia following stroke, while Davies et al. (1996) and Newham & Hsiao (2001) did not. Some studies (Hammond et al., 1988; Dewald et al., 1995; Lum et al., 2003), but not all (Colebatch et al., 1986; Fellows et al., 1994) showed excessive coactivation in the hemiparetic elbow and shoulder muscles. Chae et al. (2002) even found correlation between coactivation in upper limb hemiparesis and functional measures. Levin & Hui-Chan (1994) found high coactivation ratios for the paretic dorsiflexors, while Becher et al. (1998) found a low level of coactivation in gastrocnemius muscle in spastic hemiplegia. The results of the present study show a higher rsEMG/rsEMGmax ratio of the knee flexors during extension in the paretic compared with the non-paretic lower limb. These findings may suggest that enhanced coactivation of knee flexor muscles might (partly) explain the lower MVCe in the paretic lower limb compared with the non-paretic lower limb and control, which is in agreement with Neckel et al. (2006). Remarkably, however, rsEMG of the flexors (BF and GM) during extension was the same for the three groups, while maximal rsEMG of these muscles during flexion (i.e. working as agonists), were significantly lower for the paretic lower limb compared with control and the non-paretic lower limb. Consequently, the ratio might be exaggerated by reduced agonist activity in the presence of normal antagonist activity (Clark et al., 2006) since the antagonist activity of the flexors is scaled to a lower ‘maximum’ agonist value in the paretic lower limb. Thus, although it seems tempting to interpret our data such that the reduced torques around the knee joint result from enhanced coactivation of antagonist muscles, the high rsEMG/rsEMGmax ratios most likely result from the low rsEMG for BF and GM during MVCf, which would argue against the existence of real excessive coactivation.

Functional performance

Strength and functional performance

The results show that isometric extensor as well as flexor torque of especially the paretic lower limb but also of the non-paretic lower limb is strongly related to a broad spectrum of functional performance (Table 4), which is in accordance with reported findings (Bohannon & Walsh, 1992; Patterson et al., 2007).

Thus, deteriorated functional performance is related to limitations in muscle strength of both extensors and flexors of both lower limbs, but the exact nature of that reduced functional performance is not completely clear yet. Given the nature of the disorder (cerebro vascular accident), a disturbed neural control is also a very likely explanation for impaired functional performance.

Neural control and functional performance

Most interestingly and to our best knowledge never investigated before, maximal voluntary activation of the paretic lower limb correlated significantly with RMI, FAC, BBS and FM, while activation of the non-paretic lower limb correlated significantly with RMI, MI, 10 m walk test and TUG (Table 4). The lower the activation, the longer the time to walk 10 m or perform the TUG and the lower the score at the RMI and MI. The very high correlation coefficient for the relationship between activation of the knee extensor muscles of the non-paretic lower limb and the 10 m walk time means that about 93% (r2) of the time of the 10 m walk test could be explained by the voluntary activation capacity of these muscles. In other words, activation capacity of the non-paretic knee extensor muscles seems to be very important in predicting functional performance. A possible cause for this strong correlation might be behavioural compensation strategies, for instance enhanced use of the non-paretic lower limb, to counterbalance the motor impairment of the paretic lower limb (Roby-Brami et al., 2003; Kwakkel et al., 2004).

The underlying nature of the reduced maximal voluntary activation could be impairments in either motor unit recruitment and/or rate of motor unit firing (Kamen et al., 1995; Harridge et al., 1999). This disturbance in central drive might not only cause a limited voluntary activation capacity during maximal isometric contractions, but might also lead to a disturbed neuromuscular control and coordination during more submaximal contractions while performing the functional performance tests. Hence, although the sample size is limited, the findings tend to reinforce measures as strength and activation as indicative of function and identify intrinsic muscle strength and voluntary activation as important targets of intervention for stroke patients. Moreover, we believe that improving voluntary activation might result in better functional performance, like increasing strength results in gains in measures of impairments and disabilities (Teixeira-Salmela et al., 1999), gait performance (Teixeira-Salmela et al., 2001) and velocity (Sharp & Brouwer, 1997). Besides, it is known that voluntary activation can be improved by strength training in healthy subjects (Duchateau & Hainaut, 1988; Harridge et al., 1999; Kamen & Knight, 2004). Therefore, involving strength training in rehabilitation programmes for stroke patients is recommended for both increasing muscle strength (Ouellette et al., 2004; Ada et al., 2006; Yang et al., 2006; Bohannon, 2007) and improving voluntary activation in order to improve functional performance after stroke.

The question remains whether our measure for the degree of coactivation correlates with functional performance. In the present study, indeed some significant correlations were found between the measured ratio during extension and functional performance. However, hardly any correlation between functional performance and the ratio during flexion was found. This supports our idea that high ratios were obtained due to low maximal rsEMG instead of high coactivation. Another possible explanation might be that the degree of coactivation during an isolated task such as maximal voluntary contraction is different from that during more complex functional tasks, thereby possibly not being representative to assess the influence of coactivation during functional performance.

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Functional performance
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgments
  10. References

Both paretic and non-paretic knee extensor and knee flexor muscle torques are reduced after stroke. Whereas for the non-paretic lower limb failure of activation can fully explain the weakness, for the paretic lower limb both reduced activation and reduced torque capacity are responsible for the severe weakness. There are no indications that coactivation during maximal contractions play a significant role in the observed weakness. The high correlations found between strength, as well as voluntary activation, and functional performance indicate that these parameters have a determinant influence on performance. These results clearly indicate that both muscle strength and voluntary activation are important parameters which can potentially be fruitful objects for rehabilitation programmes after stroke.

Acknowledgments

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Functional performance
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgments
  10. References

The authors thank the stroke patients for their participation in this study. In addition, we wish to thank Micha Paalman for the development of the lower extremity system and Peter Verdijk for the development of the software for collecting and analysing the data.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Functional performance
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgments
  10. References