Sensorimotor impairments of paretic upper limb correlates with activities of daily living in subjects with chronic stroke

Correspondence to: Dr. Shamay S.M. Ng Department of Rehabilitation Sciences The Hong Kong Polytechnic University Hung Hom, Kowloon, Hong Kong E­mail: Shamay.Ng@inet.polyu.edu.hk AbSTrAcT: The main objective of this study was to investigate the correlations between sensorimotor impairments of paretic upper limb and the hand functions of activities of daily living (ADLs) scores in persons with chronic stroke. This is a cross-sectional study with 19 chronic stroke survivors. Hand function was measured by the Jebsen-Taylor Hand Function Test (JTHFT). Impairments in upper extremity motor function were measured by upper limb items of Fugl-Meyer Assessment (FMA-UE). Forearm muscles strength, handgrip and pinch grip power were assessed using handheld dynamometers. Tactile sensation threshold was measured by monofilaments. Significant differences were found between the affected and unaffected side in the total JTHFT scores, forearm muscle strength, handgrip and pinch grip (p≤0.017), but not the tactile sensation threshold. The total JTHFT scores were found to be correlated with total score of FMA-UE (r s = -0.789), handgrip power (r s = -0.588) and pinch grip power (r s = -0.657) on the affected side, but not the tactile sensation. The total JTHFT scores were correlated with FMA-UE scores, handgrip and pinch grip of the affected side. This is the first study in documenting the correlation between the sensorimotor impairments and JTHFT scores in persons with chronic stroke. Our findings highlights the importance of including upper limb and grip strength training in stroke rehabilitation program in order to improve hand functions in activities of daily living in patients with chronic stroke.

. The possible explanation is that afferent inputs ascending to the somatosensory cortex are required for acquiring motor skills which could be supported by direct anatomical projects had been found from somatosensory cortex to the motor, premotor and pari etal cortices in monkeys (Jones et al. 1978), and these projections could modulate neuronal activity in the pri mary motor cortex (Farkas et al. 1999).Sensory deficits could also affect the motor functions in patients with stroke.Indeed, about 53% of stroke patients were reported to have impaired tac tile sensation (Connell et al. 2008).
The JebsenTaylor Hand Function Test (JTHFT) is a standardized test measur ing hand functions in activities of daily living (ADLs) (Jebsen et al 1969).In the original research by Jebsen et al (1969), JTHFT was used to assess subjects

INTRODuCTION
Stroke is one of the major causes of disability and handicap in adults.It is well documented that stroke induces a wide range of disabilities and func tional limitations and limitations in hand function always occur among stroke survivors.Wade et al (1983) conducted a study to measure the recovery of upper limb function in a group of stroke sub jects.Only 27 out of 92 subjects achieved full recovery two years poststroke.The underlying impairments which caused hand function loss in patients with stroke included muscle weak ness, poor coordination and altered motor pattern (Ng and Shepherd 2000).
In addition to motor impairments, somatosensory deficits are found to be associated with the slower recovery and poorer performance of functional mobility in patients with stroke (Connell with a variety of disorders such as rheuma toid arthritis, hemiparesis and spinal cord injury.They found the test to have moderate interrater reliability (ICC = 0.60) (Jebsen et al 1969).The JTHFT was also capable of differen tiating people with rheumatoid arthri tis, spinal cord injury and hemiparesis from normal subjects (Jebsen 1969).Although the JTHFT is increasingly used in neuro rehabilitation to evaluate the hand functions in patients with stroke, the relationship between sensorimotor impair ments and the JTHFT scores had not been investigated before.Because major aim of stroke rehabilitation is to optimize patients' performances of motor tasks in activities of daily liv ing, clinicians need a reliable and valid measurement that will establish patients' upper limb functions at base line, to monitor patients' progress as a result of treatment.The objective of this study is to delineate possible asso ciation between sensorimotor impair ments of upper limb (i.e.impairments in upper extremity motor function, muscle strength and tactile sensation of affected upper extremity) and hand functions in ADLs (ie, scores of JTHFT) in persons with chronic stroke.The results of the study will provide insight into the design and evaluation of therapeutic programs targeted at improving hand functions in activities of daily living in community dwelling stroke patients.

METHOD
This study was a crosssectional clini cal trial.Nineteen communitydwelling stroke patients were recruited through a local selfhelp group.Inclusion and exclusion criteria are listed in table 1.The characteristics of subjects are sum marised in tables 2 and 3.The study protocol was approved by the ethics committee of the university, and all subjects provided written informed consent after thorough explanation of the study prior to data collection.The study was conducted according to the Declaration of Helsinki for human expe riments.All assessment was explained and administered in Chinese.

OuTCOME MEASuRES
(1) Jebsen-Taylor Hand function Test (JTHfT) The JTHFT consists of seven subtests: writing, cards turning, lifting small objects, simulated feeding, stacking checkers, picking up light cans and pick ing up heavy cans.The time for complet ing each subtest was recorded.The total JTHFT score was calculated by adding the time scores of all subtests.Higher scores indicated slower performance of tasks.Subjects were allowed a maximum of 180 seconds to complete each subtest.If the subject could not complete the task within the time allowed, 180 seconds was taken as the time for completing the task (Eliasson et al 2006) as higher scores indicated slower performance of tasks while lower scores indicated better performance.
(2) fugl-Meyer Assessment of upper Extremity (fMA-uE) FuglMeyer Assessment (FMA) is a wellaccepted validated test for assess ing impairments of motor function in people suffering hemiparesis due to cerebro vascular injuries such as stroke.It was originally developed to assess both upper and lower limb motor func tions and balance (FuglMeyer et al 1975).Only the upper limb items of FMA (FMAUE) were used in the present study in order to evaluate upper extremity motor impairment including reflexes, voluntary control of isolated (3) Maximal isometric muscle strength The Nicholas Hand Held Dynamometer was used to measure upper limb muscle strength.Excellent testretest reliability was reported for healthy subjects (ICC = 0.8200.995)(Magnusson et al 1990).Maximum isometric strength of elbow and wrist flexors and extensors were tested on both affected and unaffected upper limbs.Testing position was listed in table 4. Subjects were instructed to gradually increase force against the tester until maximal contraction was achieved.Maximum contraction was maintained for approximately 5 seconds.Five trials were performed on each muscle group.One minute of rest was given between each trial to minimize the effect of fatigue.The results of the five trials were first averaged.The averaged values were then normalized by body weight for analysis.

(4) Handgrip and pinch grip strength
The Jamar Hydraulic Hand Dynamo meter was used to measure handgrip force and pinch grip force.Excellent interrater reliability was reported in handgrip force (ICC = 0.9960.998,p<0.05) and pinch grip force (ICC = 0.9490.990,p<0.05) in healthy subjects (LindstromHazel et al 2009).Testing positions are listed in table 4.
(5) Tactile sensation SemmesWeinstein Monofilaments was used to evaluate the sensation thresh old to light touch.Excellent interrater reliability (ICC = 0.999) (Novak et al 1993) was reported using monofilament in assessing tactile sensation deficits.Three sites were chosen for the test: tip of index finger, centre of palm, and 1 cm proximal to MCPJ of middle finger on the dorsal side.These points were cho sen because they occupy sites involved in many hand activities.Starting with stiffness of 4.31, the monofilament was applied perpendicularly to subject's skin until the monofilament started to bend then held in this position for approximately 1 second.Subjects were instructed to answer "Yes" once they felt the pressure.If there was no response, the procedure was repeated five times for each monofilament stiffness.The monofilament was applied in a descend ing order until the subject no longer felt the stimulus.The previous filament was applied again to reconfirm the sensory threshold.The lower the scores repre sent the better sensation.

DATA ANALySIS
Results were analysed using SPSS 16.0.0for Windows.MannWhitney U test was used to identify differences in measures between affected and unaffected arms.
The distribution of data was checked using a KolmogorovSmirnov test of normality.As the data was not normally distributed, Spearman's Rho was used to assess the correlations between different outcome measures.Level of significance was set at a = 0.05.For Spearman's Rho statistics, r s = 0.000.25 were classi fied as showing little or no correlation; r s = 0.250.50 as fair; r s = 0.500.75 as moderate to good; and r s >0.75 as good to excellent.

Subjects
Nineteen patients with stroke, an ave rage of 62.0 ± 6.6 years old (16 men; all righthand dominant), poststroke 5.4 ± 0.3 years participated in the study.
Their demographic data and level of functional mobility of our subjects was shown in Table 2.

The Jebsen-Talyor Hand function Test
The total JTHFT scores and scores of all subtests showed significant differ ences between affected and unaffected sides (p<0.001).Total JTHFT scores of the unaffected side (68.61 ±16.99) were lower than those of the affected side (347.07±354.49)(Table 5).

Maximum Isometric Muscle Strength
Comparing affected side with unaffected side, all measurements of elbow and wrist muscle strength showed significant difference (p = 0.017) for normalized wrist extensor, p<0.001 for all others).All measures of maximum isometric strength on the unaffected side were stronger than those on the affected side (Table 5).

Handgrip and Pinch grip Strength
Significant differences were found between affected and unaffected sides in both handgrip and pinch grip strength (p<0.001 for both grips).In both hand grip and pinch grip strength, the strength of the affected sides were significantly weaker than the unaffected sides, with the unaffected sides more than twice as strong as the affected sides for both measures.Ranges for normalized hand grip on the affected and unaffected sides were 0.060.33Nkg 1 and 0.30 0.74Nkg 1 , whereas ranges for norma lized pinch grip were 0.000.13Nkg 1 and 0.050.25Nkg 1 respectively (table 5).

Tactile sensation
On the three points assessed no differ ences between sides were found (p = 0.908 for palm, p = 0.729 for index finger), though the difference between the two sides at MCPJ approaches statis tical significance (p = 0.075) (Table 5).

Correlations with the JTHfT
No significant correlation was found between the total JTHFT scores and age, weight, height, education level, number of episodes of stroke and years since stroke.Significant moderate negative correlations were also found between the total JTHFT scores with norma lized handgrip (r s = 0.588, p = 0.008) and normalized pinch grip (r s = 0.657, p = 0.002) strength on the affected side.
There was no significant correlation found among the total JTHFT scores, tactile sensation and maximum isometric muscle strength of flexor and extensor of elbow and wrist.Strong significant correlations were found between the total JTHFT score on the affected side and total FMA score under Spearman's analysis (r s = 0.789, p<0.001).The subscores of JTHFT were also found to have significant correlations with all but Part I of FMA.Strength of these cor relations ranged from fair to good (r s = 0.482 -0.716).No significant correlation was found between total JTHFT score and muscle strength, grip strength and tactile sensation on the unaffected side (Table 6).

DISCuSSION
Our study is the first study to delineate possible association between sensori motor impairments of upper limb (i.e.impairments in upper extremity motor function, muscle strength and tactile sen sation of affected upper extremity) and hand functions in ADLs (ie, scores of JTHFT) in persons with chronic stroke

Muscle strength
The results obtained from the present study are consistent with previous find ings that there is a significant difference in strength between sides of the stroke subjects (Chae et al 2003).In a review on clinical and physiological studies with stroke patients, the loss of muscle strength following stroke could be due to decreased capacity to activate motor units, reduced number of functioning motor units, reduced motor units firing rate and increased cocontraction of the antagonists during movement (Ng and Shepherd 2000).
Results of this study show that the JTHFT scores have a moderate to good correlation with normalized pinch grip (r = 0.657, p<0.01) and handgrip (r = 0.588, p<0.01), but not with elbow and wrist muscle strength.This could be explained by the fact that JTHFT focuses mainly on evaluating fine motor hand functions instead of gross motor func tions.Effective muscle contraction of hand muscles, as indicated by hand and pinch grip strength, is essential for good hand function.Elbow and wrist muscles however, are less involved in fine hand functions.Consistent with previous find ings (Sunderland et al 1989), our results support the view that grip strength could also be an important indicator of upper extremity function in persons with chronic stroke.Sunderland et al (1989) reported that grip strength has a good prognostic value for persons with acute stroke.It was found that subjects with better grip strength at one month tended to have a better upper limb function at six months, and grip strength was a valid predictor of upper limb function.

Tactile sensation
It was surprising to note that that there was no difference in tactile sensation over the 3 selected sites between affected and unaffected hands, because sensory defi cit is frequently reported by persons with stroke (Connell et al. 2008).The esti mates of prevalence in somatosensory impairment after stroke varied widely from (Connell et al. 2008), and may due to heterogeneity in study populations, the number of somatosensory modalities and body areas assessed.Our results contrasted with the study of Sommerfeld and Arbin, (2004), which found that 39 out of 77 subjects with acute stroke had impaired tactile sensation assessed by light touch with cotton wool.No sig nificant correlation was found between tactile sensation with cotton wool and the total scores of JTHFT for both sides, suggesting that deficits in tactile sensation did not affect hand function.This contradicted the result of a previ ous study by Ebied et al, (2004), which found significant correlation between impaired tactile sensation induced by artificial cutaneous sensation block and writing (p = 0.02).Continuous feedback from afferent input was important for the automatic fine adjustment of hand movement and, the ability to adjust force and direction of movements accurately was essential for performing fine hand functions (Connell et al. 2008).In our study, the three sites tested for tactile sensation may not represent the whole picture of sensory deficiency in indi vi duals with stroke.In addition, differ ent subject types, different assessment method may also affect the result of study.

Correlations between JTHfT and fMA-uE
Significant correlations were found between total JTHFT score and all but the first and fifth items of FMAUE.These items of FMAUE examined upper limb reflexes while the JTHFT examines the functional capacity of hands.One previous study (Woodbury et al 2007) suggested that the reflex items of FMAUE differed from the rest of the test.While the majority of FMAUE items assessed voluntary movements, the reflex items were assessing involuntary responses (Woodbury et al 2007).Since all items in the JTHFT involved subjects performing voluntary movements, it was reasonable that there was no correlation between the JTHFT and reflex items of FMAUE.Similar results with poor cor relation between severity of spasticity and motor functions in stroke patients had been found in some previous studies (Ng and HuiChan 2005).Our negative results may also be due to small sample size and/or to the adoption of compen satory strategy by our stroke subjects, such as increased trunk side flexion to accomplish the upper limb motor tasks.
Strongest correlations were found between total the JTHFT score and sixth and seventh items of FMAUE (r s = 0.716 and r s = 0.661 respectively) which assessed wrist mobility and prolonged isometric muscle contraction, as well as hand functions and grips.The sixth and seventh items of FMAUE are similar in task nature of the JTHFT and also to hand function, as most hand functions such as writing and picking up objects involves the movement of joints of wrist and hand.In addition, strong significant correlation was found between the total JTHFT scores and most items of FMAUE.This could be explained by a number of testing items (ie, the sixth and seventh items) between both measurements

Limitations
Our study did have several limitations.As majority of the subjects recruited in this study were quite active and have a relatively high functional status, results obtained in this study might not be gene ralized to persons with acute stroke or persons with lower mobility level.Further studies might need to recruit subjects with different functional level.The number of subjects that participated in this study was relatively small.This study was a crosssectional study, and it did not prove any causality between the JTHFT score and level of differ ent impairments.We only measured tactile sensation in the present study, whether pinpick sensation or proprio ception generating similar result would need further investigation.In addition, a few subjects with more arm impair ment could not complete some items of JTHFT within the 180 seconds time limit.A score of 180 seconds was given as suggested by Eliasson et al (2006), and this might overestimate the ability of the subjects.All subjects in this study are righthanded, they may never have used the left hand to write.Therefore the measure of writing speed on the left side (JTHFT Item 1) may not really be measuring their functional ability.
Isometric muscle strength measured in this study may not be functional in our assessment protocol, as most ADL tasks involve movement of joints whilst muscle force is exerted.Previous study showed that shoulder flexors strength had a moderate to excellent correla tion with paretic upper limb function (Mercier and Bourbonnais 2004).The present study, however, did not measure muscle strength of shoulder muscles due to limited time, and further studies could include shoulder muscle strength as one of the measurements.

CONCLuSION
The present study showed that there was a significant difference in hand func tions in ADLs, handgrip, pinch grip and maximum isometric strength between affected and unaffected sides, but not tactile sensation.The total JTHFT score was strongly correlated to the total score and all subtests of FMAUE except measurement of reflex (ie, the first and fifth items).Moderate correlation was also identified between the total JTHFT scores and normalized handgrip and pinch grip on the affected side.These findings indicate that motor recovery and muscle strength of the hands are important for hand function in activities of daily living in persons with chronic stroke.It is important for physiotherapist in designing appropriate intervention

Table 2 . Descriptive statistics of demographic data. characteristics (n=19)
rater (r = 0.0980.995)reliability in subjects with chronic stroke.Testretest reliabi lity was also reported to be excel lent (ICC = 0.97).Good correlations were also reported between FMAUE and Arm Motor Ability Test (functional ability r = 0.94, quality of movement r = 0.94) (Chae et al 2003).