Patients with stroke confined to basal ganglia have diminished response to rehabilitation efforts Ichiro Miyai, MD, PhD; Alan D. Blau, PhD; Michael J. Reding, MD; and Bruce T. Volpe, MD Article abstract-Prediction of the functional outcome for patients with stroke has depended on the severity of impair- ment, location of brain injury, age, and general medical condition. This study compared admission and discharge func- tional outcome (Functional Independence Measure, FIM) and deficit severity (Fugl-Meyer, F-M) scores in a retrospective study of patients with similar neurologic impairments: homonymous hemianopia, hemisensory loss, and hemiparesis. CT-verified stroke location was the independent variable: cortical (n = ll),basal ganglia and internal capsule (normal cortex and thalamus, n = 131, or combined (cortical, basal ganglia, and internal capsule, n = 22). By 3 months on average after stroke, all groups demonstrated significantly improved motor function as measured by F-M scores. Patients with cortical lesions had the least CT-imaged damage and the best outcome. Patients with combined lesions and more extensive brain injury had significantly higher FIM scores (p< 0.05) than patients with injury restricted to the basal ganglid internal capsule. Patients with basal ganglidinternal capsule injury were more likely to have hypotonia, flaccid paralysis, and persistently impaired balance and ambulation performance. While all patients had a comparable rehabilitation experience, these results suggest that patients with stroke confined to the basal ganglia and internal capsule benefited less from therapy. Isolated basal ganglia stroke may cause persistent corticothalamic-basal ganglia interactions that are dysfunctional and impede recovery. NEUROLOGY 1997;48:95-101 In several studies rehabilitative intervention has im- matter, but not the basal ganglia, corona radiata, or inter- proved the functional outcome of patients with nal capsule. The basal ganglia lesion (BG) involved the but until there is a precise definition of the caudate, putamen, and corona radiata with or without in- severity of the impairment, lesion location, and ternal capsule damage. The combined lesion (COM) in- quantitative definition of functional outcome, the volved cortex, corona radiata, and the basal ganglia with clinical mechanisms for improvement will remain or without internal capsule damage. The thalamus was unclear. To begin to understand these mechanisms also examined in each scan, and was not damaged in any for functional improvement, this study quantified patient. In order to estimate the volume of injury, each lesion on CT was transposed to the right hemisphere of the neurologic impairment and functional outcome in pa- standardized horizontal brain template^.^ These images tients with lesions confined to the cortex, or to the were then superimposed using NIH Image Version 1.54. to basal ganglia and internal capsule, or to both loca- produce lesion density maps for each group of patients. tions: cortex, and basal ganglidinternal capsule. Reliable and valid scales were used to measure neuro- logic impairment and functional outcome; namely the Methods. We retrospectively reviewed the rehabilitation Fugl-Meyer scale (F-M),'O ambulation endurance," and the outcome of consecutively admitted patients presenting Functional Independence Measure (FIM).12 Each patient with hemiparesis, hemisensory loss, and hemianopia. All was evaluated at the time of admission and discharge. All patients admitted to the Burke Rehabilitation Hospital patients had multidisciplinary rehabilitation in the same from December 1993 to July 1995 were included. Patients stroke unit. Self-care subscore of FIM was the sum of who had had prior stroke, who were not functionally inde- scores for eating, grooming, bathing, dressing upper body, pendent prior to admission, or who had any complications dressing lower body, toileting, bladder management, and requiring transfer to an acute-care hospital were excluded. bowel management (max = 56). Ambulation subscore of These inclusion and exclusion criteria permitted a compar- FIM was the sum of scores for transfers: to bed, to toilet, to ison of the outcome of stroke rehabilitation among patients tub; and walking on level surfaces and on stairs (max = with comparable neurologic deficit^^.^ and with the same 35). Communication subscore was the sum of scores for quantity and quality of rehabilitation experience. Location comprehension and expression (max = 14). Cognition sub- of the brain lesions was classified as cortical, basal score of FIM was the sum of scores for communication, ganglidinternal capsule, or combined based on CT infor- social interaction, problem solving, and memory (max = mation obtained on average 8 2 3 days after the onset. The 35). Total FIM score then was the sum of self-care, ambu- cortical lesion (CX) involved cortical and subcortical white lation, and cognition subscores (max = 126). Total FM From the Department of Neurology, Cornell University Medical College, The Burke Rehabilitation Center, White Plains, Ny. Received May 17, 1996. Accepted in final form June 25, 1996. Address correspondence and reprint requests to Dr. Bruce T. Volpe, Department of Neurology, Cornell University Medical College, The Burke Rehabilitation Center, 785 Mamaroneck Ave, White Plains, NY, 10605. Copyright 0 1997 by the American Academy of Neurology 95 Figure 1. This photomontage depicts lesion density maps (A, D. G) of stroke patients with motor, sensory, and visual deficits, and examples of CT from patients with the largest (B, E, H) and smallest (C, F, I!, lesions within each group. Patients had a cortical lesion (CX, A-C) that in- volved cortical and subcortical white matter, but not the basal ganglia, corona radiata, or internal capsule; a basal ganglia lesion (BG', D-F) that involved the caudate lznd puta- men and corona radiata uth or without internal capsule damage; or a combined lesion (COM, C-I) that involved cortex, corona radiata, and the basal ganglia with or without internal capsule damage. Each le- sion on CT was transposed to the right hemisphere of the standardized horizontal brain templates.' These images were then superimposed us- ing NIH Image Version 1.54 to pro- duce lesion density maps lor each group of patients. (A, CX; D, BG; G, COM). score (max = 202) was the sum of scores for upper extrem- shown in table 1. Patients in the three groups demon- ity (max = 661, lower extremity (max = 34), balance strated comparable age, sex distribution, interval post- (max = 14), joint pain (max = 441, and passive motion stroke, side of stroke, type of stroke (infarction or hemor- (max = 44). Sensation subscore of F-M was omitted be- rhage), Mini-Mental State Examination score and compli- cause approximately half of the patients could not be eval- cations. uated due to aphasia. Table 2 and figure 2 display the dynamic (changes in Statistical analysis relied on an ANOVA in which lesion FIM for all groups. The first analysis attempted to test group, functional outcome scores, and age were factors.13 whether age had an effect on the change in FIM scores. A three-factor mixed ANOVA was performed with group (CX Results. Forty-six of 330 patients had hemiparesis, versus BG versus COM) and age (above and below 65) as hemisensory loss, and hemianopia. All patients had dam- between-subject factors. Time (admission versus dis- age in the cortical or subcortical region that was consistent charge) was the within-subject factor. The dependent vari- with the neurologic deficit, namely, the sensorimotor cor- able was the total FIM score. There was no main effect of tex and the subcortical sensory and motor pathways and age [F(1,40)= 0.014, ns], nor was there any significant age the optic radiations (for the BG group, posterior to the interaction with other variables. Since age had no effect on internal capsule approximately 3.1 cm dorsal to the can- the change in FIM scores over time, it was dropped from thomeatal line, for the CX and COM group, temporal and further analysis. parietal lobes). Further motor evaluation demonstrated The second analysis focused on the change of FIM that flaccidity (all the articular segments of the affected scores for each group over time. There were significant upper and lower extremities) was present in 0 of 11 pa- main effects for group [F(2,40) = 6.107, p < 0.0051, and for tients with CX, 5 of 13 patients (38%)with RG, and 4 of 22 time [F(1,40) = 75.785, p < 0.00011. All group:; improved patients (18%)with COM. CT analysis revealed that 11 significantly from admission to discharge in FIM scores patients had CX lesions, 13 patients had BG lesions, and [CX, F(1,lO) = 23.939, p < 0.001; BG: F(1,12) = 15.431, 22 patients had COM lesions. Figure 1 demonstrates the p < 0.002; COM: F(1,21) = 64.874, p < 0.00111. Impor- lesion density map for each group and examples of the tantly, there was a significant interaction between group largest and smallest injury within each group. The brain and time [F(2,40) = 4.996, p < 0.021. Post hoc analysis injury was always confined to the territory of the middle using the least significant difference test (LSD) revealed cerebral artery, but the thalamus was normal in all scans. that the CX group had a significantly higher admission Demographic features of patients in each group are total FIM score (76 2 5) than either the BG group (54 2 7, 96 NEUROLOGY 48 January 1997 Table 1 Demographic features in stroke patients with motor, sensory, and visual deficits cx BG COM p Value (n = 11) (n = 13) (n = 22) Age (years) 64 2 4 73 2 3 64 2 2 n.s.* Sex (M/F) 318 617 12/10 n.s.t Days after stroke 24 2 3 31 f 5 25 * 5 n.s.;k Side (WL) 912 914 11/11 n.s.1 Type (I/H) 11/0 10/3 2012 n.s.t MMSE 19 * 3 13 2 3 11 t 2 n.s.:b Complications Hypertension 7 (64%) 7 (54%) 15 (68%) n.s.t Ischemic heart disease 4 (36%) 3 (23%) 8 (36%) n.s.t Diabetes mellitus 4 (36%) 3 (23%)) 10 (45%) n.s:t Atrial fibrillation 3 (27%) 4 (31%) 6 (27%) n.s.t Congestive heart failure 2 (18%) 3 (23%) 6 (27%) n.s.t Hypercholesterolemia 3 (27%) 1(8%) 2 (9%) nst COPD 0 (0%) 0 (0%) 2 (9%) n.s.t Depression 2 (18%) 3 (23%) 6 (27%) nst Data are mean 2 SEM.