**4. Discussion**

A considerable improvement was found in de scores at the tests of functional performance, especially up to F2. In general also the muscle variables improved over time. Changes in muscle variables of the 5 subjects that were measured at t=0, F1 and F2 were shown to correlate with improvements in tests of functional performance during the first 9 months after stroke.

**PL** 

**NL** 

\* significant (p<0.05) correlation, ^ trend (p<0.1)

**4.1 Functional performance** 

(MRTDstim), triplet torque and voluntary activation (VA).

the five subjects who came for these 2 follow-ups.

Functional Recovery and Muscle Properties After Stroke: A Preliminary Longitudinal Study 75

**MVCe** - 0.843^ 0.861^ - - - 0.832^

**MRTDstim** 0.736 0.990\* 0.860 - - - 0.815

**MVCf** - - - 0.985\* - - - **MRTDvol** - 0.750 - - - - -

**Triplet** 0.827^ 0.897\* 0.923\* - - - - **VA** 0.710 - - - - - -

**MVCe** 0.745 N.A. 0.833^ - - - - **MVCf** 0.922\* N.A. - - - - - **MRTDvol** - N.A. - - - -0.881 -

**MRTDstim** - N.A. 0.719 - - - 0.732

**Triplet** - N.A. - - - - - **VA** - N.A. 0.742 - - - -

Table 4. Correlation coefficients between changes (between t=0 and F2) in scores at tests of functional performance and changes in muscle variables for the non-paretic (NL) and paretic lower limb (PL) of 5 subjects. Berg Balance Scale (BBS), Motricity Index (MI), Functional Ambulation Categories-score (FAC), Rivermead Mobility Index (RMI) and Brunnstrom Fugl-Meyer (FM), Timed ''get-up-and-go'' test (TUG) and 10 meter walk test (10m). Maximal voluntary extension (MVCe) and flexion (MVCf) torque, maximal rate of torque development during voluntary (MRTDvol) and electrically evoked contractions

From the 5 subjects we followed until F2, subject 1 scored the lowest values of all subjects (n=14) that took part in the study at t=0 at all variables of the paretic lower limb (Table 3), whereas subject 3 scored the best. The actual improvement in the Timed ''get-up-and-go'' test (TUG) between t=0 and F2 may be even greater than presented in Table 2. Subject 1 was not able to perform the TUG at t=0, but during F1 that subject was able to do the TUG in 81 s. Nevertheless, the score of this subject could not be used, since the improvement could not be calculated due to the missing value at t=0. This same subject 1 scored 9 on the Motricity Index (MI) at t=0 but 0 at F1 and F2. Moreover, subject 3 achieved the maximal score of 100 at MI for all three measurement points and only one subject improved between F1 and F2 at this test. This may explain why we found no significant improvement over time for MI in

(m/s)

**FAC TUG**  (s)

**MI** 

**BBS FM RMI 10m** 


Table 3. Data on muscle variables for 5 subjects who were measured during the first measurement (t=0) and 3 and 6 months thereafter (F1, F2 respectively) for the non-paretic (NL) and paretic lower limb (PL). Maximal voluntary extension (MVCe) and flexion (MVCf) torque, maximal rate of torque development during voluntary (MRTDvol) and electrically evoked contractions (MRTDstim), triplet torque and voluntary activation (VA).


\* significant (p<0.05) correlation, ^ trend (p<0.1)

Table 4. Correlation coefficients between changes (between t=0 and F2) in scores at tests of functional performance and changes in muscle variables for the non-paretic (NL) and paretic lower limb (PL) of 5 subjects. Berg Balance Scale (BBS), Motricity Index (MI), Functional Ambulation Categories-score (FAC), Rivermead Mobility Index (RMI) and Brunnstrom Fugl-Meyer (FM), Timed ''get-up-and-go'' test (TUG) and 10 meter walk test (10m). Maximal voluntary extension (MVCe) and flexion (MVCf) torque, maximal rate of torque development during voluntary (MRTDvol) and electrically evoked contractions (MRTDstim), triplet torque and voluntary activation (VA).

#### **4.1 Functional performance**

74 Rehabilitation Medicine

Table 3. Data on muscle variables for 5 subjects who were measured during the first measurement (t=0) and 3 and 6 months thereafter (F1, F2 respectively) for the non-paretic (NL) and paretic lower limb (PL). Maximal voluntary extension (MVCe) and flexion (MVCf) torque, maximal rate of torque development during voluntary (MRTDvol) and electrically

evoked contractions (MRTDstim), triplet torque and voluntary activation (VA).

**PL NL Subject t=0 F1 F2 t=0 F1 F2** 

> From the 5 subjects we followed until F2, subject 1 scored the lowest values of all subjects (n=14) that took part in the study at t=0 at all variables of the paretic lower limb (Table 3), whereas subject 3 scored the best. The actual improvement in the Timed ''get-up-and-go'' test (TUG) between t=0 and F2 may be even greater than presented in Table 2. Subject 1 was not able to perform the TUG at t=0, but during F1 that subject was able to do the TUG in 81 s. Nevertheless, the score of this subject could not be used, since the improvement could not be calculated due to the missing value at t=0. This same subject 1 scored 9 on the Motricity Index (MI) at t=0 but 0 at F1 and F2. Moreover, subject 3 achieved the maximal score of 100 at MI for all three measurement points and only one subject improved between F1 and F2 at this test. This may explain why we found no significant improvement over time for MI in the five subjects who came for these 2 follow-ups.

Functional Recovery and Muscle Properties After Stroke: A Preliminary Longitudinal Study 77

Fig. 2. Correlations between ΔMRTDstim of PL and ΔFM (ρ=0.99, p=0.01, n=4), between Δtriplet of PL and ΔFM (ρ=0.90, p=0.04, n=5), between Δtriplet of PL and ΔRMI (ρ=0.92, p=0.03, n=5), between ΔMVCf of NL and ΔBBS (ρ=0.92, p=0.03, n=5) and between ΔMVCf of

Δ: delta, changes between t=0 and F2; MRTDstim, electrically evoked maximal rate of torque development; MVCf, Maximal Voluntary Contraction torque of the knee flexors; PL, paretic lower limb; NL, non-paretic lower limb; FM, Fugl-Meyer assessment; BBS, Berg Balance Scale; RMI, Rivermead Mobility Index; 10 m, 10 m walk test.

PL and Δ10m (ρ=0.99, p=0.002, n=5).

An improvement of 7%, from 45 to 52% of maximum attainable recovery (n=8) at the Fugl-Meyer Lower Extremity test (FM) was found from ~3.5 months (t=0) to half a year (F1) after stroke (n=8) (Figure 2). For FAC a significant increase from 68 to 88% was found and for MI 47 to 53% (not significant) (Table 2). Kwakkel et al*.* (2004) similarly observed an improvement from about 62% to 65% at the FM (n=101), 51 to 59% at the FAC and 58 to 59% at the MI during the first half year after stroke. Both the present results and those of Kwakkel et al*.*  (2004) show that most improvement took place within the first 3 months until half year after stroke, as was also found by others (Wade & Hewer, 1987; Jorgeson et al., 1995).

#### **4.2 Muscle variables**

As would be expected after hemiparetic stroke, PL scored consistently lower than NL on the muscle variables (voluntary extension and flexion torque, triplet torque, voluntary activation and maximal rate of voluntary torque development). Variable results in changes in muscle characteristics were found per subject between t=0 and F2 (Table 3). Also in literature there are different results. Carin-Levy et al*.* (2006) reported, in line with our results, no significant change over time in the strength and muscle mass of both paretic and non-paretic (arm and) leg muscle during the first 6 months after stroke. However, Newham and Hsiao (2001) did observe increased strength throughout the first half year after stroke, while activation failure remained constant. Andrews et al*.* (2003) showed an increase in both PL and NL knee extensor strength from admission (~2 wk post stroke) to discharge (~4 wk post stroke). So, there are no consistent data indicating that muscle variables improve after stroke.

The main limitation of our study is the small sample size at F2 and F3. Although all new patients with stroke in the rehabilitation centre were examined by physicians, a large number were ineligible for our study, because they had severe cognitive and/or communicative problems, medical complications, no hemiparesis of the lower extremity or, conversely, were too heavily paralyzed and had a previous stroke. A considerable number of patients were not willing to participate (or in case of follow-ups to continue), mainly due to their changed life after stroke and/or the intensity of the protocol. Around half of the eligible patients completed the entire protocol (4 measurement days) at t=0. The scores of stroke severity of our patients (FAC median and quartiles 4 (2.25-4)) confirmed that we managed to recruit a very wide a range of patients with stroke at t=0, but this contributed to the difficulty in statistics, besides the great drop-out of patients during the follow-ups. However, studies with smaller samples sizes than ours have detected significant changes in muscle strength over time in NL (Harris et al*.*, 2001) and PL compared to control (Newham & Hsiao, 2001). Thus, it is likely that, if changes in the thigh muscles of our patients had occurred, these must have been small.

#### **4.3 Correlation between muscle variables and functional performance**

The severity of post stroke paresis is related to a person's ability to perform functional tasks; Others found correlations between lower limb isometric knee extension strength and functional performance, like gait distance (Bohannon, 1989) and speed (Bohannon, 1989; Bohannon & Walsh, 1991, 1992; Horstman et al., 2008), sit-to-stand (Bohannon, 2007a; Horstman et al., 2008), transfers (Bohannon, 1988), stair climbing (Bohannon & Walsh, 1991) and balance (Horstman et al., 2008). The new aspect in this study is that we wanted to investigate whether *changes* in functional performance during the first 9 months after stroke related to changes in muscle characteristics.

An improvement of 7%, from 45 to 52% of maximum attainable recovery (n=8) at the Fugl-Meyer Lower Extremity test (FM) was found from ~3.5 months (t=0) to half a year (F1) after stroke (n=8) (Figure 2). For FAC a significant increase from 68 to 88% was found and for MI 47 to 53% (not significant) (Table 2). Kwakkel et al*.* (2004) similarly observed an improvement from about 62% to 65% at the FM (n=101), 51 to 59% at the FAC and 58 to 59% at the MI during the first half year after stroke. Both the present results and those of Kwakkel et al*.*  (2004) show that most improvement took place within the first 3 months until half year after

As would be expected after hemiparetic stroke, PL scored consistently lower than NL on the muscle variables (voluntary extension and flexion torque, triplet torque, voluntary activation and maximal rate of voluntary torque development). Variable results in changes in muscle characteristics were found per subject between t=0 and F2 (Table 3). Also in literature there are different results. Carin-Levy et al*.* (2006) reported, in line with our results, no significant change over time in the strength and muscle mass of both paretic and non-paretic (arm and) leg muscle during the first 6 months after stroke. However, Newham and Hsiao (2001) did observe increased strength throughout the first half year after stroke, while activation failure remained constant. Andrews et al*.* (2003) showed an increase in both PL and NL knee extensor strength from admission (~2 wk post stroke) to discharge (~4 wk post stroke). So, there are no

The main limitation of our study is the small sample size at F2 and F3. Although all new patients with stroke in the rehabilitation centre were examined by physicians, a large number were ineligible for our study, because they had severe cognitive and/or communicative problems, medical complications, no hemiparesis of the lower extremity or, conversely, were too heavily paralyzed and had a previous stroke. A considerable number of patients were not willing to participate (or in case of follow-ups to continue), mainly due to their changed life after stroke and/or the intensity of the protocol. Around half of the eligible patients completed the entire protocol (4 measurement days) at t=0. The scores of stroke severity of our patients (FAC median and quartiles 4 (2.25-4)) confirmed that we managed to recruit a very wide a range of patients with stroke at t=0, but this contributed to the difficulty in statistics, besides the great drop-out of patients during the follow-ups. However, studies with smaller samples sizes than ours have detected significant changes in muscle strength over time in NL (Harris et al*.*, 2001) and PL compared to control (Newham & Hsiao, 2001). Thus, it is likely that, if changes in the thigh muscles of our patients had

stroke, as was also found by others (Wade & Hewer, 1987; Jorgeson et al., 1995).

consistent data indicating that muscle variables improve after stroke.

**4.3 Correlation between muscle variables and functional performance** 

The severity of post stroke paresis is related to a person's ability to perform functional tasks; Others found correlations between lower limb isometric knee extension strength and functional performance, like gait distance (Bohannon, 1989) and speed (Bohannon, 1989; Bohannon & Walsh, 1991, 1992; Horstman et al., 2008), sit-to-stand (Bohannon, 2007a; Horstman et al., 2008), transfers (Bohannon, 1988), stair climbing (Bohannon & Walsh, 1991) and balance (Horstman et al., 2008). The new aspect in this study is that we wanted to investigate whether *changes* in functional performance during the first 9 months after stroke

occurred, these must have been small.

related to changes in muscle characteristics.

**4.2 Muscle variables** 

Fig. 2. Correlations between ΔMRTDstim of PL and ΔFM (ρ=0.99, p=0.01, n=4), between Δtriplet of PL and ΔFM (ρ=0.90, p=0.04, n=5), between Δtriplet of PL and ΔRMI (ρ=0.92, p=0.03, n=5), between ΔMVCf of NL and ΔBBS (ρ=0.92, p=0.03, n=5) and between ΔMVCf of PL and Δ10m (ρ=0.99, p=0.002, n=5).

Functional Recovery and Muscle Properties After Stroke: A Preliminary Longitudinal Study 79

exercise attention controls, long-term training and follow up. Strength measures were reported to improve after resistance training, albeit without clear benefits for functional performance (e.g. gait speed) (Saunders et al., 2004). Therefore, in addition, the strength training may be combined with task-specific functional training (Sullivan et al., 2006; Hubbard et al., 2009), because it has "the potential to drive brain reorganization toward more optimal functional performance" (Shepherd, 2001). When muscles are weak, isometric contractions can be used in the early stages of rehabilitation as a means of improving the muscle's ability to contract. However, once muscle strength reaches a certain threshold, exercises should be biomechanically similar to daily life actions in order to be trained to transfer increased force-generating ability into improved performance (Shepherd, 2001).

The (small) alterations in the muscle variables correlated well with the improvements in scores on tests of functional performance. Although the correlations do not necessarily imply causality, we think (intrinsic) muscle speed and strength are important variables which can potentially be prolific targets to improve during rehabilitation. It is therefore recommended to investigate the effects of strength training of the thigh muscles during at least the first 6 months after stroke. From such an intervention study on functional recovery it can be elucidated whether increasing strength and speed really improves functional

Ada L, Dorsch S & Canning CG. (2006). Strengthening interventions increase strength and

Andrews AW & Bohannon RW. (2003). Short-term recovery of limb muscle strength after acute stroke. *Archives of Physical Medicine and Rehabilitation* Jan;84(1):125-30 Berg K, Wood- Dauphinee S, Williams JI & Gayton D. (1989). Measuring balance in the

Berg KO, Wood-Dauphinee SL & Williams JI. (1995). The Balance Scale: reliability

Berg KO, Wood-Dauphinee SL, Williams JI & Maki B. (1992). Measuring balance in the

Bohannon RW. (1988). Determinants of transfer capacity in patients with hemiparesis.

Bohannon RW. (1989). Is the measurement of muscle strength appropriate in patients with brain lesions? A special communication. *Physical Therapy* Mar;69(3):225-36 Bohannon RW. (2007a). Knee extension strength and body weight determine sit-to-stand independence after stroke. *Physiotherapy: Theory and Practice* Sep-Oct;23(5):291-7 Bohannon RW. (2007b). Muscle strength and muscle training after stroke. *Journal of* 

*Journal of Rehabilitation Medicine* Mar;27(1):27-36

improve activity after stroke: a systematic review. *Australian Journal of Physiotherapy*

elderly: preliminary development of an instrument. *Physiotherapy Canada* 41:304-311

assessment with elderly residents and patients with an acute stroke. *Scandinavian* 

elderly: validation of an instrument. *Canadian Journal of Public Health* Jul-Aug;83

**5. Conclusion** 

performance.

**6. References** 

52(4):241-8

Suppl 2:S7-11

*Physiotherapy Canada* 40:236-239

*Rehabilitation Medicine* Jan;39(1):14-20

Most relations were found within subjects between changes in muscle variables of PL and changes in scores at the BBS, a sitting and standing balance measure and the FM, an impairment measure developed to assess physical recovery after stroke (Sanford et al., 1993). Moreover, changes in flexor strength are positively related with changes in the 10m walking speed, which means the more increase in hamstring strength, the bigger the increase in walking speed. In our cross-sectional study (Horstman et al., 2008) strong significant correlations were found between muscle variables of both PL and NL and various tests of functional performance. However, in the present study if we look within subjects, we hardly see any correlations between changes in muscle variables of NL and changes in scores at the functional performance tests over time. This indicates that longitudinal data are essential to gain the required information regarding which (muscle) variables should be trained to induce improvements in functional performance, because cross-sectional data are not exclusive enough.

A question that remains to be answered is what may have caused the improvements in functional recovery? It is suggested that functional gains experienced by patients with stroke are primarily attributable to spontaneous recovery (changes over time that occur naturally) of functional performance of which eighty percent occurs within six months after the onset of stroke (Lind, 1982). Others state that there is some recovery between 1 and 6 months in almost all acute patients with stroke (Wade & Hewer, 1987) and that at 6 months 60% of people with initial hemiparesis have achieved functional independence in daily activities such as toileting and walking short distances (Mayo et al., 1999; Patel et al., 2000.) To facilitate neuroplasticity and cortical reorganization, it would be interesting to also investigate sensory stimulation in future studies with patients with stroke (Nudo et al., 1996; Johansson, 2000) since sensory impairments of all modalities are common after stroke (Carey, 1995). Moreover, sensory deficits are associated with the degree of weakness and the degree of stroke severity related to mobility, independence in activities of daily living, and recovery (De Haart et al., 2004; Lin, 2005). Addressing sensory deficits that accompany muscle weakness may improve impaired processing of afferent signals which in turn may contribute to improved muscle activation, gait patterns, and responses to perturbation during gait and stance (El-Abd & Ibrahim, 1994).

Secondary changes as a result of stroke could be expected in skeletal muscle, e.g. changes in myofiber type (De Deyne et al*.*, 2004) or number and size of motor units. The latter is already reported in the second week after stroke onset (Jorgensen & Jacobsen, 2001). For instance, a change in muscle fiber composition, characterized by selective type II fiber atrophy and predominance of (slow twitch, oxidative) type I fibers has been shown in paretic muscles (Edstrom, 1970; Scelsi et al., 1984; Dietz et al., 1986; Dattola et al., 1993; Hachisuka et al., 1997), which would lead to concomitant changes in contractile speed of the muscle fibers towards those of slow muscles. We can imagine that such a change in fiber type composition can be combated, for instance by training, during the first year after stroke, so that muscle speed characteristics can be restored. Bohannon concludes in his review (Bohannon, 2007b) that resistance training programs are effective at increasing strength in patients who have experienced a stroke but there is no clear evidence for the effect of strength training on functional activities after stroke (Morris et al*.*, 2004). Main results of Saunders' review (Saunders et al*.*, 2004) include only 4 strength training trials (Inaba et al., 1973; Kim et al., 2001; Ouellette et al., 2004; Winstein et al., 2004) and lack nonexercise attention controls, long-term training and follow up. Strength measures were reported to improve after resistance training, albeit without clear benefits for functional performance (e.g. gait speed) (Saunders et al., 2004). Therefore, in addition, the strength training may be combined with task-specific functional training (Sullivan et al., 2006; Hubbard et al., 2009), because it has "the potential to drive brain reorganization toward more optimal functional performance" (Shepherd, 2001). When muscles are weak, isometric contractions can be used in the early stages of rehabilitation as a means of improving the muscle's ability to contract. However, once muscle strength reaches a certain threshold, exercises should be biomechanically similar to daily life actions in order to be trained to transfer increased force-generating ability into improved performance (Shepherd, 2001).
