NeuroRehabilitation 27 (2010) 351358 DOI 10.

3233/NRE-2010-0619 IOS Press

351

Functional gain in hemorrhagic stroke patients is predicted by functional level and cognitive abilities measured at hospital admission
Ada W.S. Leunga, Stephen K.W. Chengb, Amanda K.Y. Maka,c , Kwok-Keung Leunga,c, Leonard S.W. Lic,d and Tatia M.C. Leea,c,
a b

Laboratory of Neuropsychology, The University of Hong Kong, Hong Kong, China Occupational Therapy Department, MacLehose Medical Rehabilitation Centre, Hong Kong, China c Institute of Clinical Neuropsychology, MacLehose Medical Rehabilitation Centre and The University of Hong Kong, Hong Kong, China d Department of Rehabilitation Medicine, Tung Wah Hospital, Hong Kong, China

Abstract. Background and purpose: Few studies have addressed factors that contribute to functional recovery in people with hemorrhagic stroke. We assessed the value of using pre-training functional level and cognitive abilities measured at admission in order to predict functional gain in a sample of stroke patients. Methods: The Functional Independence Measure (FIM) was administered to 85 hemorrhagic stroke patients. Three multiple regression models were constructed using total gain in FIM scores, gain in scores in the cognitive domain of FIM, and gain in scores in the motor domain of FIM as outcome variables. Predictor variables were age; scores on the Digit Span Forward Test (DSF), Digit Span Backward Test (DSB), and Chinese Auditory Verbal Learning Test (CAVLT); and the FIM scores measured at admission. Results: All regression models were signicant, P s < 0.001, and each explained over 73% of the variance in the FIM gains. Age and admission FIM scores were both signicant predictors in each of the three models. The DSB span score was a signicant predictor of the total FIM and the cognitive FIM gains. The CAVLT recognition score was also a signicant predictor of the cognitive FIM gain. Conclusions: Functional improvement in patients with hemorrhagic stroke after in-patient rehabilitation was predicted by age, pre-training functional level, and cognitive abilities measured at admission. Keywords: Stroke, FIM, functional recovery, learning, memory

1. Introduction Stroke is a major cause of death and disability. With advancements in medical technology in the last decade,

Address for correspondence: Prof. Tatia M.C. Lee, May Professor in Neuropsychology, Laboratory of Neuropsychology, K610, Knowles Building, The University of Hong Kong, Pokfulam Road, Hong Kong, China. Tel.: +86 (852) 2857 8394; Fax: +86(852) 2819 0978; E-mail: tmclee@hkusua.hku.hk.

however, the mortality rate has declined; this has resulted in a dramatic increase in the number of stroke patients requiring functional training. Among the many subtypes of stroke, hemorrhagic and ischemic are the two main classications of the vascular etiology. Clinical outcomes in cases of hemorrhagic stroke have been more severe than outcomes in cases of ischemic stroke; this is due to the difculty of controlling the extent of brain regions invaded by bleeding [1]. Functional recovery, however, shows considerable variation be-

ISSN 1053-8135/10/$27.50 2010 IOS Press and the authors. All rights reserved

352

A.W.S. Leung et al. / Functional gain in hemorrhagic stroke patients

tween patients suffering from hemorrhagic strokes and ischemic strokes. Lipson et al. [16] compared the functional outcome of 110 hemorrhagic and 709 ischemic stroke patients and found that the hemorrhagic stroke patients took a longer time to enter into rehabilitation training. Schepers et al. [26] observed that functional recovery in patients with cerebral infarction happened until 26 weeks post-stroke; whereas patients with intracerebral hemorrhage showed increased functional independence only until 10 weeks post-stroke. In other studies, hemorrhagic stroke patients were found to have a more difcult recovery period, and to suffer more frequent sensory and physical impairments [3,28]. Some studies revealed favorable outcomes for hemorrhage stroke patients (as measured by the Glasgow Outcome Scale) but the incidence of clinically signicant cognitive decits was still higher in those with hemorrhagic than other subtypes of stroke [5]. Although functional recovery varies between hemorrhagic and ischemic stroke patients, the functional outcome in hemorrhagic stroke patients is often associated with poor prognosis. However, there are still only a limited number of studies in the existing literature that address factors which contribute to functional change in hemorrhagic stroke patients. In the current study, we assessed the predictive value of pre-training functional status and cognitive abilities measured at admission for functional gain in hemorrhagic stroke patients after a course of rehabilitative training. The rst objective was to determine how well the pre-training variables, together with patients demographic characteristics, predicted functional gain. The second objective was to identify the cognitive abilities that could signicantly predict functional improvement. We measured three aspects of cognitive abilities attention, learning, and working memory using neuropsychological tests at the admission stage. From a theoretical perspective, studying the extent to which cognitive abilities predict functional improvement allows us to understand the associated neural mechanisms that may contribute to functional recovery.

2. Method We used clinical records in order to retrospectively identify hemorrhagic stroke patients who were admitted to MacLehose Medical Rehabilitation Centre (MMRC) in Hong Kong between September 2002 and March 2007. The project was approved by the Ethics Committee of the Institutional Review Board of The

University of Hong Kong and Hospital Authority Hong Kong West Cluster. At the time they were admitted to MMRC, all of the patients gave their informed consent to allow the use of their clinical data in this study. Hemorrhagic etiologies included in this clinical sample were subarachnoid hemorrhage or intracerebral hemorrhage that did not involve an infarction. Other selection criteria were: 1) primary indication of non-traumatic hemorrhagic stroke without complication, or with one of the following complications: cerebral arterial vasospasm, seizure, and hydrocephalus; 2) mentally stable for rehabilitation training, and attained a score of 15 on the Glasgow Coma Scale at admission to MMRC; 3) documentation of completed Functional Independence Measure (FIM) data; and 4) availability of completed neuropsychological tests; these included the Chinese Auditory Verbal Learning Test (CAVLT), the Digit Span Forward Test (DSF), and the Digit Span Backward Test (DSB). Exclusion criteria were: a primary indication for admission to MMRC for reasons other than hemorrhagic stroke, an incomplete dataset, and death during the hospital stay. Since all patients attending MMRC were transferred from acute-care hospitals, neurological data including diagnosis, site of lesion, and acute medical complications were gathered from medical records documented in the patients acute-care hospitals. Some demographic data documented in MMRC was retrieved for analysis; these data included patients age at the time of admission, length of stay, and incidence of acute medical condition (e.g., seizures, pneumonia, and/or altered level of consciousness), if any, while at MMRC. During their hospital stay, all of the patients participated in active rehabilitation programs ve days per week. The programs included training sessions offered by clinical psychologists, occupational therapists, physical therapists, and speech therapists. Patients functional level was assessed by means of the FIM measured on the day of admission and on the day of discharge; measurements were conducted by occupational therapists who were trained in administering the scale. The FIM has documented content and construct validity for assessing stroke patients [10, 21]. It consists of 18 items; each item requires a rating on a scale of 1 (total assist) to 7 (complete independence), with a total score of 126 indicating total functional independence. Some studies have reported that the FIM measures two statistically and clinically unique domains, namely cognitive and motor [15]. The cognitive domain contains ve items, namely comprehension, expression, social interaction, problem solv-

A.W.S. Leung et al. / Functional gain in hemorrhagic stroke patients

353

ing, and memory; the total score for this domain ranges from 5 to 35. The motor domain contains 13 items, namely eating, grooming, bathing, dressing (upper and lower body), toileting, sphincter control (bowel and bladder), transfer (bed-chair, toilet-chair, and bathtubchair), walking or moving via use of a wheel-chair, and stair climbing; the total score for the motor domain ranges from 13 to 91. In order to calculate the scores for the total FIM gain, the cognitive FIM gain, and the motor FIM gain, we subtracted the admission score from the discharge score for the total FIM, the cognitive FIM, and the motor FIM respectively. Attention was measured by the DSF [2]. It consists of seven pairs of random number sequences (i.e., a total of 14 number sequences) with the number of digits increased from 3 in the rst pair to 8 in the seventh pair. The therapist read each digit aloud at the rate of one digit per second, and the patients were requested to repeat all of the digits in the order in which they were read. The forward span score (DSF-span) consists of the number of sequences in which the digits were correctly repeated. The forward sequence score (DSFsequence) consists of the number of sequences in which the order of the digits was correctly repeated. Working memory was measured by the DSB [6]. The DSB is the same as the DSF, except that the number of digits is increased from 2 in the rst pair to 7 in the seventh pair; patients were also required to repeat the digits in reverse order. The backward span score (DSB-span) consisted of the number of sequences in which the digits were correctly repeated. The backward sequence score (DSF-sequence) consisted of the number of sequences in which the order of the digits was correctly repeated. Learning ability was measured by the CAVLT, which is a validated Chinese version of the Rey Auditory Verbal Learning Test originally developed by Rey [22]. It consists of ve successive presentations of 15 nouns in List A, with each of the ve trials being followed by a free-recall test. After completing the fth trial of List A, an interference list (List B) consisting of 15 nouns is presented; this is followed by a free-recall test of List B. After a 30-minute delay, a free-recall test of List A is then administered, as well as a recognition test of the words in List A (from a mixture of words in Lists A and B). In the current study, two parameters were used to indicate learning ability: 1) the total acquisition score (CAVLT-acquisition), which was the total number of words recalled in the ve free-recall tests of List A; and 2) the recognition score (CAVLT-recognition), which was the number of words recognized in the delayed recall test.

All of the neuropsychological tests were administered by clinical psychologists within one to fteen days of admission to MMRC. 2.1. Statistical analysis In order to conrm that the patients showed functional improvement after the rehabilitation treatment, it was necessary to compare the FIM scores obtained at admission and discharge. To do this, paired-t tests were run for the total FIM, the cognitive FIM, and the motor FIM scores at admission and discharge. Next, the predictive value of the patients functional level and cognitive abilities obtained at admission was calculated with multiple regression by applying the enter method; the FIM gain (discharge FIM scores less admission FIM scores) was used as the outcome variable. Three separate models were computed, which were based on the total FIM gain, the cognitive FIM gain, and the motor FIM gain respectively. For each regression model, the predictors were: age, admission FIM score, DSF span and sequence scores, DSB span and sequence scores, and the CAVLT recognition and acquisition scores. Assumptions of the statistical models were tested by inspecting the correlation matrix of the predictors, and by obtaining tolerance statistics of no less than 0.2 for each predictor.

3. Results 3.1. Patient sample A total of 85 patients (44 female) fullled the inclusion criteria. The mean age of the sample was 53.8 years old (SD = 13.6 years). The average length of stay at MMRC was 42 days. The sample contained 28 cases of subarachnoid hemorrhage in one lesion site; 31 cases of subarachnoid hemorrhage in multiple lesion sites (or with hydrocephalus as a medical complication); 3 cases of intracerebral hemorrhage; and 23 cases of other kinds of hemorrhage other than subdural and epidural hemorrhage. None of the patients had documented acute medical conditions during their stay at MMRC. Further, all of the patients were medically stable and attended the full rehabilitation program. To determine whether any of the demographic characteristics differed among the hemorrhagic subgroups, a Kruskal-Wallis one-way analysis of variance was conducted. Results showed that there were no signicant differences in any of the demographic variables among

354

A.W.S. Leung et al. / Functional gain in hemorrhagic stroke patients Table 1 Demographic characteristics and outcome measures of the study cohort All n = 85 ACoA n = 10 3 7 59.9 (8.5) 40.5 10 189 7.0 (1.9) 6.8 (1.9) 4.9 (1.9) 3.9 (2.1) 11.1 (4.5) 27.4 (12.6) 94.3 (27.5) 115.8 (13.3) 21.5 (19.7) 28.4 (7.2) 31.3 (4.9) 2.9 (3.9) 65.9 (20.8) 84.5 (8.9) 18.6 (16.3) SAH with single lesion AVM MCA n=8 n=6 5 3 44.0 (8.0) 26.5 12 48 7.5 (2.8) 7.1 (2.5) 5.8 (1.8) 5.0 (2.2) 11.1 (4.0) 32.3 (15.4) 102.8 (24.6) 116.1 (11.0) 13.4 (17.9) 28.9 (6.6) 31.3 (3.8) 2.4 (3.8) 73.9 (18.3) 84.9 (7.7) 11.0 (14.3) 4 2 54.8 (11.8) 40.5 22 76 7.8 (1.8) 7.7 (1.9) 6.5 (1.4) 4.3 (1.2) 13.0 (1.8) 35.0 (15.9) 100.7 (14.4) 123.5 (3.2) 22.8 (15.5) 30.2 (4.4) 33.8 (1.8) 3.6 (3.3) 70.5 (13.4) 89.7 (3.3) 19.2 (13.6) ICA n=4 2 2 47.3 (8.8) 63.0 12 157 8.3 (1.5) 8.3 (1.5) 6.0 (2.2) 5.8 (2.4) 13.5 (1.3) 34.5 (10.0) 95.3 (39.1) 115.8 (19.1) 20.5 (20.6) 29.0 (7.3) 33.0 (2.8) 4.6 (4.0) 66.3 (32.8) 82.8 (16.5) 16.5 (16.9) SAH n = 31 18 13 51.1 (14.5) 35.1 10 158 7.1 (2.4) 6.7 (2.3) 4.5 (2.1) 3.7 (2.0) 11.6 (3.4) 28.5 (14.6) 93.3 (25.6) 117.2 (11.2) 23.9 (20.4) 26.7 (6.6) 30.8 (3.8) 4.0 (4.0) 66.5 (20.4) 86.4 (8.8) 19.9 (17.0) ICH n=3 2 1 59.7 (12.5) 76.0 38 146 8.7 (0.6) 8.7 (0.6) 6.4 (1.2) 3.7 (1.2) 15.0 (0) 47.3 (11.0) 66.0 (32.9) 109.0 (10.4) 43.0 (22.5) 27.0 (3.4) 31.3 (0.58) 4.3 (2.8) 39.0 (29.4) 77.7 (9.8) 38.7 (19.6) Other n = 23 10 13 57.4 (15.0) 49.6 13 145 7.1 (2.2) 6.9 (2.0) 5.3 (1.7) 3.6 (1.6) 10.1 (5.0) 26.9 (14.7) 90.7 (25.2) 109.3 (19.2) 18.7 (15.0) 28.7 (5.8) 31.1 (4.0) 2.3 (3.6) 61.9 (22.9) 78.2 (17.8) 16.3 (12.9)

Sex (n) Female Male Age LOS -Average -Min. -Max. DSF-span -sequence DSB-span -sequence CAVLT-recog -acquis Total FIM admission discharge gain Cognitive FIM admission discharge gain Motor FIM admission discharge gain

44 41 53.8 (13.6) 42.0 10 189 7.3 (2.2) 7.0 (2.1) 5.1 (1.9) 3.9 (1.9) 11.4 (4.0) 29.7 (14.5) 93.2 (25.8) 114.9 (14.3) 21.7 (18.6) 28.0 (6.2) 31.3 (3.8) 3.3 (3.8) 65.2 (21.6) 83.6 (12.3) 18.4 (15.6)

0.09 0.24

0.73 0.47 0.12 0.30 0.22 0.32 0.56 0.12 0.48 0.87 0.58 0.47 0.56 0.52 0.47

Other hemorrhage excluding subdural and epidural hemorrhage. SAH with multiple lesion sites and/or one medical complication. ACoA, anterior communicating artery; AVM, arteriovenous malformation; MCA, middle cerebral artery; ICA, internal carotid artery; ICH, intracerebral hemorrhage; LOS, length of stay at MMRC in days; DSF, digit span forward; DSB, digit span backward; CAVLT-recog, Chinese auditory verbal learning test recognition score, CAVLT-acquis, Chinese auditory verbal learning test acquisition score; FIM, functional independence measure. All data reported in means (SDs) unless specied.

the subgroups, 2 = 8.0 to 11.1, df = 6, P 0.09 (Table 1). Table 1 displays the DSF, DSB, CAVLT, and FIM scores. There were signicant differences between the admission and discharge FIM scores, t(84) = 10.7, P < 0.001, for the total FIM score; t(84) = 7.9, P < 0.001, for the cognitive FIM score; and t(84) = 10.8, P < 0.001, for the motor FIM score. This indicates that the patients showed signicant functional improvement after the rehabilitation training. Similarly, a KruskalWallis Test was run on all the neuropsychological test scores; no signicant differences were found among the hemorrhagic subgroups, 2 = 3.6 to 10.0, df = 6, P 0.12. 3.2. Regression models The rst model was computed with variables that predict functional gain (as measured by the total FIM score). As shown in Table 2, the model explained

75.4% of the variance in functional gain (as measured by the total FIM score). The regression model was signicant, F (8, 84) = 29.1, P < 0.001. Signicant predictors were age, P = 0.002, DSB-span, P = 0.048, and total FIM measured at admission, P < 0.001. It should be noted that while the DSB-span was a positive predictor, age and the total FIM score measured at admission were both negative predictors in the regression model. Two additional regression models were computed in order to examine the extent to which the variables used in the rst model would predict the FIM score in the cognitive and in the motor domains. For the cognitive domain, as shown in Table 3, the model explained 73.0% of the variance in functional gain measured by the cognitive FIM score. The regression model was signicant, F (8, 84) = 25.7, P < 0.001. Signicant predictors were age, P = 0.034, CAVLT-recognition, P = 0.005, DSB-span, P = 0.027, and cognitive FIM score measured at admission, P < 0.001.

A.W.S. Leung et al. / Functional gain in hemorrhagic stroke patients Table 2 Variables predicting functional gain measured by the total FIM score 96.31 0.28 0.63 0.51 1.36 1.80 1.15 0.21 0.07 P < 0.001 0.002 < 0.001 0.821 0.583 0.048 0.234 0.511 0.493 95% Condence interval for Lower bound Upper bound 81.54 111.09 0.45 0.11 0.74 0.52 3.94 4.96 6.30 3.57 0.02 3.59 3.06 0.76 0.42 0.84 0.29 0.14

355

Constant Age Total FIM at admission DSF-span DSF-sequence DSB-span DSB-sequence CAVLT-recognition CAVLT-acquisition
P P

< 0.05. < 0.01. Table 3 Variables predicting functional gain measured by the cognitive FIM score 19.41 0.40 0.55 0.20 0.16 0.43 0.13 0.20 0.001 P < 0.001 0.034 < 0.001 0.669 0.754 0.027 0.523 0.005 0.955 95% Condence interval for Lower bound Upper bound 16.52 22.30 0.077 0 0.65 0.44 1.14 0.74 1.19 0.86 0.05 0.80 0.53 0.27 0.06 0.33 0.05 0.05

Constant Age Cognitive FIM at admission DSF-span DSF-sequence DSB-span DSB-sequence CAVLT-recognition CAVLT-acquisition
P P

< 0.05. < 0.01. Table 4 Variables predicting functional gain measured by the motor FIM score 74.54 0.24 0.61 0.50 0.91 1.29 1.19 0.003 0.09 P < 0.001 0.002 < 0.001 0.796 0.676 0.102 0.156 0.992 0.311 95% Condence interval for Lower bound Upper bound 61.86 87.23 0.39 0.09 0.72 0.50 3.36 4.37 5.20 3.39 0.26 2.84 2.84 0.46 0.55 0.55 0.28 0.09

Constant Age Motor FIM at admission DSF-span DSF-sequence DSB-span DSB-sequence CAVLT-recognition CAVLT-acquisition
P P

< 0.05. < 0.01.

For the motor domain, as shown in Table 4, the model accounted for 73.8% of the variance in functional gain measured by the motor FIM score. The regression model was signicant, F (8, 84) = 26.8, P < 0.001. Signicant predictors were age, P = 0.002, and motor FIM score measured at admission, P < 0.001. As in the rst model, the cognitive abilities were positive predictors, whereas age and the FIM scores measured at admission were both negative predictors of functional gain.

4. Discussion The major nding in this study was that patients age, pre-training functional level, and cognitive abilities measured at admission successfully predicted functional gain as measured by the FIM scores. The strength of the prediction was consistent among the three regression models that employed the total FIM, the cognitive FIM, and the motor FIM gain scores. The models also indicated an inverse relationship between pre-

356

A.W.S. Leung et al. / Functional gain in hemorrhagic stroke patients

training FIM scores and functional gain after training; this suggests that hemorrhagic stroke patients who have a poor baseline functional level may indeed be capable of showing progress in rehabilitation. Among the three cognitive functions tested, working memory (measured by the DSB test) and learning ability (measured by the CAVLT) were both shown to predict functional gain, especially in the cognitive domain of the FIM measure. A unique feature of the current research was the incorporation of patients cognitive abilities measured at admission into the prediction of functional gain. Ozdemir et al. [20] used the Minimental State Examination (MMSE) and its subsections (such as orientation, attention, calculation, and recall) to measure stroke patients cognitive abilities in the prediction of functional outcome measured by the FIM. Based on data from 43 patients, they found that only the total score on the MMSE predicted the FIM motor score at discharge. Our study expanded upon this approach by using specic neuropsychological tests to measure different aspect of patients cognitive abilities; therefore, our design provided for a more representative measure of the cognitive status of the patients. It should be noted, however, that the variable length of stay and the variable treatment schedule of the patients meant that testing occurred over the span of a week; some patients were tested within the rst few days after admission, whereas others were tested later in the week. Nevertheless, all measurements were completed before the patients began their rst rehabilitation training session. We showed that indices of learning and working memory could predict functional gain (as measured by the FIM) in hemorrhagic stroke patients. Our ndings are consistent with those reported by Till et al. [30], which also showed a relationship between post-recovery cognitive decline and performance on the AVLT (i.e., the original Rey Auditory Verbal Learning Test). In addition, use of the AVLT as a measure of learning performance has been documented in many clinical studies (e.g., Ross et al. [23]). Working memory is the ability to hold information in mental space for online information manipulation and processing. This ability is important for the production of goal-appropriate behaviors, which are essential for positive functional outcome. Our results indicated that the Digit Span Backward Test (DSB) predicted functional gain as measured by both the total FIM and the cognitive FIM. Previous studies also found that the DSB was a good indicator of the severity of memory impairment in brain damaged patients [13,14,17]. Pre-training functional level has been shown to be a signicant predictor of functional improvement. For

example, Katrak et al. [7] found that cognitive FIM at admission predicted discharge FIM scores and FIM gain for patients with intracerebral hemorrhagic stroke. However, our results show an inverse relationship between the admission FIM scores and the FIM gain. This could indicate that other to-be-determined factors may contribute to hemorrhagic stroke patients functional achievements during the rehabilitation process. For example, psychological issues, such as motivation, depression, and adjustment problems, could be important factors that affect the progress and functional recovery in the rehabilitation process. Indeed, past research has already reported the effect of psychosocial inuence on stroke recovery [24,25,27]. Another plausible explanation of the inverse relationship between the admission FIM scores and the FIM gain observed could be that high FIM scores on admission may suggest the possibility of approaching the ceiling for functional recovery and hence leaving minimal room for further measurable changes during the course of rehabilitation. On the same token, low FIM scores on admission may be associated with more room for improvement and hence allows more room for observable treatment effect reected as FIM gain. Furthermore, previous ndings have suggested that recovery in the rst 2 months poststroke was determined more by spontaneous neurological recovery and less by the impact of therapy [11]. This is particularly applicable to the recovery from hemorrhagic strokes, for the neural functioning could recover partially, if not fully, once when the hematoma has resolved and the edema has subsided [11,26]. Our observation that higher functional gain was predicted by lower pre-training functional level, an observation consistent with that reported by Kelly et al. [9], could relate to the specic pattern of recovery of hemorrhagic stroke patients, the only patient group that was examined in this study. This pattern of recovery describes a quick rebound of functions after the hematoma has resolved, thus laying the path for greater functional gain in these patients who had low FIM scores on admission. Age was also shown to be a signicant inverse predictor of functional improvement; this nding is consistent with previous research [4]. Increasing age has been associated with the need for increased cognitive resources in order to compensate for age-associated cognitive decline [12]. Hence, any additional brain impairments add load onto the already limited cognitive resources for brain functioning. In the current study, one methodological concern is the selection of patients for the study cohort. ODell et al. [19] investigated the functional outcome of 42

A.W.S. Leung et al. / Functional gain in hemorrhagic stroke patients

357

patients with non-traumatic subarachnoid hemorrhage; they found functional gain after a short hospital stay, but did not nd any variables that predicted the functional gain. Although ODell et al. control for the stroke etiology in their sample, other variables such as sample size and the time between injury and rehabilitation admission might have obscured the predictive power. In some studies, stroke cases were analyzed in a larger sample in the unit of hemorrhagic or ischemic stokes regardless of their subtypes [8,16,18]. Some studies, however, did limit their sample to one or two subtypes [7]. In order to retain more participants in our study, we chose to include all patients with hemorrhagic stroke but to control for the presence of medical complications. Previous studies have shown that medical complications have a strong effect on functional outcome [29]. We also conned our prediction of functional gain to in-patient rehabilitation training; we chose to do this in order to avoid loss of follow-up cases in the sample. Future research should include a multicentre longitudinal study that focuses on a subtype of strokes. In addition, future research should make use of a more comprehensive neuropsychological prole in order to predict functional improvement. Overall, the results of the current study demonstrated that functional gain in hemorrhagic stroke patients was predicted by age, pre-training functional level, and cognitive abilities measured during admission to the rehabilitation training. These results may help professionals to draw inferences based on patients admission status, which may then assist them with designing training programmes for the patients with strokes with the aim of promoting functional recovery for these patients.

References

[1] [2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

Acknowledgements and funding This project was supported by a grant from the S. K. Yee Medical Foundation, The CRCG of The University of Hong Kong (200910159024), and a research grant from the National Natural Science Foundation of China (#30828012).

[14]

[15]

Conicts of Interest None.

[16]

D.P. Birkett, in: The Psychiatry of Stroke, (2nd ed.), The Haworth Press, New York, NY, 2008, pp. 517. F.W. Black, Digit repetition in brain-damaged adults: clinical and theoretical implications, Journal of Clinical Psychology 42 (1986), 770782. G.R. De Freitas, G. Devuyst, G. van Melle and J. Bogousslavsky, Motor strokes sparing the leg: different lesions and causes, Archives of Neurology 57 (2000), 513538. L. Denti, M. Agosti and M. Franceschini, Outcome predictors of rehabilitation for rst stroke in the elderly, European Journal of Physical Medicine and Rehabilitation 44 (2008), 311. B.O. Hutter and J.M. Gilsbach, Which neuropsychological decits are hidden behind a good outcome (Glasgow = I) after aneurismal subarachnoid hemorrhage? Neurosurgery 33 (1993), 9991006. S. Kashluba, R.A. Hanks, J.E. Casey and S.R. Millis, Neuropsychologic and functional outcome after complicated mild traumatic brain injury, Archives of Physical Medicine and Rehabilitation 89 (2008), 904911. P.H. Katrak, D. Black and V. Peeva, Do stroke patients with intracerebral hemorrhage have a better functional outcome than patients with cerebral infarction? PM & R: The Journal of Injury, Function, and Rehabilitation 1 (2009), 427433. P.J. Kelly, J. Stein, S. Shafqat, C. Eskey, D. Doherty, Y. Chang, A. Kurina and K.L. Furie, Functional recovery after rehabilitation for cerebella stroke, Stroke 32 (2001), 530534. P.J. Kelly, K.L. Furie, S. Shafqat, N. Rallis, Y. Chang and J. Stein, Functional recovery following rehabilitation after hemorrhagic and ischemic stroke, Archives of Physical Medicine and Rehabilitation 84 (2003), 968972. D. Kidd, G. Stewart, J. Baldry, J. Johnson, D. Rossiter, A. Petruckevitch and A.J. Thompson, The functional independence measure: a comparative validity and reliability study, Disability and Rehabilitation 17 (1995), 1014. G. Kwakkel, B. Kollen and E. Lindeman, Understanding the pattern of functional recovery after stroke: facts and theories, Restorative Neurology and Neuroscience 22 (2004), 281299. T.M. Lee, K.S. Yuen, L.W. Chu and I. Chi, Differential agerelated change of prose memory in older Hong Kong Chinese of higher and lower education, International Journal of Geriatric Psychiatry 19 (2004), 216222. B.E. Leininger, S.E. Gramling, A.D. Farrell, J.S. Kreutzer and E.A. Peck, Neuropsychological decits in symptomatic minor head injury patients after concussion and mild concussion, Journal of Neurology, Neurosurgery and Psychiatry 53 (1990), 293296. M. Leskela, M. Hietanen, H. Kalska, R. Ylikoski, T. Pohjasvaara, R. Mantyia and T. Erkinjuntti, Executive functions and speed of mental processing in elderly patients with frontal or nonfrontal ischemic stroke, European Journal of Neurology 6(1999), 653661. J.M. Linacre, A.W. Heinemann, B.D. Wright, C.V. Granger and B.B. Hamilton, The structure and stability of the Functional Independence Measure, Archives of Physical Medicine and Rehabilitation 75 (1994), 127132. D.M. Lipson, H. Sangha, N.C. Foley, S. Bhogal, G. Pohani and R.W. Teasell, Recovery from stroke: differences between subtypes, International Journal of Rehabilitation Research 28 (2005), 303308.

358 [17]

A.W.S. Leung et al. / Functional gain in hemorrhagic stroke patients L.A. Morrow, N. Robin, M.J. Hodgson and H. Kamis, Assessment of attention and memory efciency in persons with solvent neurotoxicity, Neuropsychologia 30 (1992), 911922. K. Nedeltchev, B. Schwegler, T. Haefeli, C. Brekenfeld, J. Gralla, U. Fischer, M. Arnold, L. Remonda, G. Schroth and H.P. Mattle, Outcome of stroke with mild or rapidly improving symptoms, Stroke 38 (2007), 25312535. M.W. ODell, T.K. Watanabe, S.T. De Roos and C. Kager, Functional outcome after inpatient rehabilitation in persons with subarachnoid hemorrhage, Archives of Physical Medicine and Rehabilitation 83 (2002), 678682. F. Ozdemir, M. Birtane, R. Tabatabaei, G. Ekuklu and S. Kokino, Cognitive evaluation and functional outcome after stroke, American Journal of Physical Medicine and Rehabilitation 80 (2001), 410415. P. Pallicino, W. Snyder and C. Granger, The NIH Stroke Scale and the FIM in stroke rehabilitation, Stroke 23 (1992), 919. A. Rey, Sollicitation de la memoire de xation par des mots et des objets presentes simultanement, Archives de Psychologies 37 (1959), 126139. S.R. Ross, S.R. Millis and M. Rosenthal, Neuropsychological prediction of psychosocial outcome after traumatic brain injury, Applied Neuropsychology 4 (1997), 165170. V.P. Schepers, A.M. Visser-Meily, M. Ketelaar and E. Lindeman, Poststroke fatigue: Course and its relation to personal and stroke-related factors, Archives of Physical Medicine and Rehabilitation 87 (2006), 184188. V.P. Schepers, A.M. Visser-Meily, M. Ketelaar and E. Lindeman, Prediction of social activity 1 year poststroke, Archives of Physical Medicine and Rehabilitation 86 (2005), 14721476. V.P.M. Schepers, M. Ketelaar, A.J.M. Visser-Meily, V. De Groot, J.W.R. Twisk and E. Lindeman, Functional recovery differs between ischaemic and haemorrhagic stroke patients, Journal of Rehabilitation Medicine 40 (2008), 487489. V. Schepers, M. Post, A. Visser-Meily, I. Van De Port, M. Akhmouch and E. Lindeman, Prediction of depressive symptoms up to three years post-stroke, Journal of Rehabilitation Medicine 41 (2009), 930935. C.Y. Su, J.J. Chang, H.M. Chen, C.J. Su, T.H. Chien and H.M. Huang, Perceptual differences between patients with cerebral infarction and intercerebral hemorrhage, Archives of Physical Medicine and Rehabilitation 81 (2000), 706714. P. Tidswell, P.S. Dias, H.J. Sagar, A.R. Mayers and R.D. Battersby, Cognitive outcome after aneurysm rupture: relationship to aneurysm site and perioperative complications, Neurology 45 (1995), 875882. C. Till, B. Colella, J. Verwegen and R.E. Green, Postrecovery cognitive decline in adults with traumatic brain injury, Archives of Physical Medicine and Rehabilitation 89 (2008), S25S34.

[25]

[18]

[26]

[19]

[27]

[20]

[28]

[21] [22]

[29]

[23]

[30]

[24]