ALLOCATING REHABILITATION resources, planning

865 ARTICLES Recovery of Ambulation After Traumatic Brain Injury Douglas I. Katz, MD, Daniel K. White, MSPT, Michael P. Alexander, MD, Reva B. Klein, MD ABSTRACT. Katz DI, White DK, Alexander MP, Klein RB. Recovery of ambulation after traumatic brain injury. Arch Phys Med Rehabil 2004;85:865-9. Objectives: To identify variables that are predictive of independent ambulation after traumatic brain injury (TBI) and to define the time course of recovery. Design: Retrospective review of consecutive admissions of patients with severe TBI over a 32-month period. Setting: Brain injury unit in an acute, inpatient rehabilitation hospital. Participants: Of 264 patients screened, 116 met criteria that included the ability to participate in motor and functional evaluation on admission to acute rehabilitation, and the absence of other neurologic disorders or fractures that affect one s ability to ambulate. Intervention: Inpatient rehabilitation on a specialized TBI unit by an interdisciplinary team. Main Outcome Measures: Recovery of independent ambulation and time to recover independent ambulation. Results: Of eligible patients, 73.3% achieved independent ambulation by latest follow-up (up to 5.1mo). Patients who achieved independent ambulation were significantly younger (P.05), had better gait scores on admission (P.05), and tended to be less severely injured based on duration of posttraumatic amnesia (PTA; P.058) than those who did not ambulate independently. There were no differences in recovery based on neuropathologic profile. Mean time to independent ambulation standard deviation was 5.7 4.3 weeks; of those achieving independent ambulation, 82.4% did so by 2 months and 94.1% by 3 months. If not independent by 3 months postinjury, patients had a 13.9% chance of recovery. Multivariate regression analysis generated prediction models for time to independent ambulation, using admission FIM instrument scores and age (38% of variance); initial gait score, loss of consciousness, and age (40% of variance); or initial gait score and PTA (58% of variance), when restricted to just those patients with diffuse axonal injury. Conclusions: Most patients with severe TBI achieved independent ambulation; the vast majority did so within 3 months postinjury. Functional measures, injury severity measures, and age can help guide prognosis and expectations for time to recover. From HealthSouth Braintree Rehabilitation Hospital, Braintree, MA (Katz, White); Department of Neurology, Boston University School of Medicine, Boston, MA (Katz, Klein); Sargent College of Health and Rehabilitation Sciences, Boston University, Boston, MA (White); Department of Neurology, Harvard Medical School, Beth Israel Deaconess Medical Center, Boston, MA (Alexander); Youville Lifecare Rehabilitation Hospital, Cambridge, MA (Alexander); and Memory Disorders Research Center (Alexander) and Harold Goodglass Aphasia Research Center (Klein), Boston University, Boston, MA. Presented in part at the American Physical Therapy Association s Combined Section Meeting, February 15, 2003, Tampa, FL. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Douglas I. Katz, MD, HealthSouth Braintree Rehabilitation Hospital, 250 Pond St, Braintree, MA 02184, e-mail: dkatz@bu.edu. 0003-9993/04/8506-8645$30.00/0 doi:10.1016/j.apmr.2003.11.020 Key Words: Brain injuries; Gait disorders, neurologic; Recovery of function; Rehabilitation. 2004 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation ALLOCATING REHABILITATION resources, planning appropriate treatment, and establishing plausible goals for patients with severe traumatic brain injury (TBI) requires an evidence-based understanding of the prognosis and time course of recovery for key functional limitations such as cognition, arm function, and ambulation. Most research on recovery from TBI has focused on cognition and behavior. We recently reported an analysis of predictors and timing of recovery of arm function after TBI. 1 The best predictors were severity of paresis at admission and overall injury severity; most recoveries occurred within the first 2 months postinjury. The predictors and time course of recovery of ambulation after TBI have not been comparably and directly investigated. Patterns of recovery of gait can only be inferred from global measures of disability and handicap after TBI, such as the FIM instrument and the Glasgow Outcome Scale. 2-5 The absence of data on ambulation after TBI is in striking contrast to the data available on ambulation after stroke. 6-10 The severity of leg paresis and the degree of functional gait disruption at admission are the best predictors of recovery of ambulation after stroke. 6 When recovery does occur, it usually does so within 2 to 3 months after onset. 8-10 Our purposes in this study were to define the time to recovery of ambulation after severe TBI and to identify the variables that predict recovery. METHODS Participants We retrospectively reviewed records of all patients (N 264) admitted to an inpatient rehabilitation TBI unit over a 32- month period from January 1992 to August 1994. Length of inpatient rehabilitation was longer at that time than it is today, which made it possible to review further into recovery with uniform serial evaluations and treatment intensity. We identified a cohort of 116 patients who met inclusion and exclusion criteria. Patients were included if they were responsive and able to participate in motor and functional evaluations at admission. Patients were excluded if they had any other injury factor that affected their ability to ambulate, such as fractures that affected weight bearing, or lower-extremity peripheral nerve injury. Patient characteristics are described in table 1. Measures Indices of overall severity of injury were the Glasgow Coma Scale (GCS) score at acute hospital admission, duration of unconsciousness (loss of consciousness [LOC]), and posttraumatic amnesia (PTA). They were recorded for each case according to a previously published method. 11 Admission FIM score was the measure of overall functional disability. 12 Age and gender were the independent variables. For each patient, we determined on the occurrence of 4 separate neuropathologic

866 RECOVERY OF AMBULATION AFTER TBI, Katz Table 1: Patient Characteristics (N 116) Variable Gender, n (%) Male 78 (67) Female 38 (33) Mean age, y (range) 42.4 (15 87) Mean days from injury to rehabilitation admission (and initial gait evaluation) SD 24 17.1 Mean duration of follow-up SD (d) 57 38.8 Neuropathologic injury subtypes (frequency) DAI 80 Focal cortical contusion 63 Hypoxic-ischemic injury 6 Herniation 15 Abbreviations: DAI, diffuse axonal injury; SD, standard deviation. diagnostic categories: diffuse axonal injury (DAI), focal cortical contusion, hypoxic-ischemic injury, and herniation. Diagnoses were made according to previously published criteria 1 based on acute clinical history and neuroimaging after review of acute hospital records and computed tomography or magnetic resonance imaging scans. Multiple diagnoses are possible in every patient (table 1). Physical therapists evaluated patients gait within 48 hours of admission and documented gait status weekly. The time interval from injury to rehabilitation admission and initial gait evaluation was 24 17.1 days. Gait was rated on an ordinal scale (table 2), which was derived from the standard assessment nomenclature used by physical therapists in a manner similar to other ordinal scales such as the Barthel Index 13 and FIM. 14 The scale we used allows for intermediate ratings. Independent ambulation was defined as a score of 9.5 or 10 (table 2). Gait Rating Table 2: Gait Scale Level of Assistance With Gait 10 Independent 9.5 Independent with assistive device 9 Supervised 8 Supervised/contact guard 7 Contact guard 6 Contact guard/minimum assistance 5 Minimum assistance 4 Minimum/moderate assistance 3 Moderate assistance 2 Moderate/maximal assistance 1 Maximum assist or minimum assist of 2 aides 0 Nonambulatory Table 3: Comparison of Independent Ambulation and Nonindependent Ambulation Patient Groups Independent Ambulators Nonindependent Ambulators P Value LOC (d) 4.0 6.2 4.9 6.5 NS PTA (wk) 3.1 3.1 5.1 5.4.058 Admission gait scale score 5.6 3.0 3.8 3.0.05 Age (y) 40.8 50.4.05 Gender (M/F) 56/29 22/9 NS Inpatient LOS (d) 50 32.9 75 45.8.05 NOTE. Values are mean SD or n. Abbreviations: F, female; LOS, length of stay; M, male; NS, not significant. Statistical Analysis The dependent variables were recovery of independent ambulation and the time required to achieve independent ambulation. Independent variables included injury severity, functional gait scale rating at admission, neuropathologic injury subtypes, gender, and age. Descriptive statistics were used to characterize the proportion of patients achieving independent gait at various postinjury intervals. Patients who recovered independent ambulation were compared with those who did not on the level of initial gait impairment, using the Mann-Whitney U nonparametric test; on injury severity (LOC, PTA) and age, using unpaired t tests; and on gender and frequency of neuropathologic diagnoses, using contingency analysis. For patients who achieved independent gait, mean time to recovery was compared at different levels of initial gait impairment, injury severity (LOC), and age and was also compared between those with or without various pathologic subtypes, using analysis of variance (ANOVA). Stepwise multiple regression determined the best models with which to predict recovery time. Statistical analysis was done with SPSS a for Windows. The criterion for significance was a P value less than.05. RESULTS At the latest follow-up (up to 5.1mo postinjury; mean, 1.9mo), 85 of 116 (73.3%) patients ambulated independently. The group that attained independence in ambulation was significantly younger (P.05) and less impaired (P.05) on initial gait evaluation than the group that did not. There was a trend toward more severe injury based on PTA duration (P.058) in patients who did not regain independent ambulation. The initial gait score was significantly better in younger patients ( 40) ( 2 test 10.3, P.01) and in those less severely injured, based on LOC ( 2 test 9.6, P.05). There was no significant difference in gender between groups (table 3), and there was no difference in neuropathologic profiles between the recovered and nonrecovered groups (table 4). Mean time to recovery of independent ambulation standard deviation (SD) for the entire sample was 5.7 4.3 weeks. Figure 1 illustrates the cumulative percentage of patients achieving independent ambulation at various intervals postinjury among all the study sample patients. When we restricted our analysis to just those who regained independent ambulation within 5.1 months (n 85), we found that 82.4% were independent by 2 months postinjury and 94.1% were independent by 3 months postinjury. A useful perspective for clinicians is knowing the likelihood of recovery at any point after injury. Figure 2 illustrates the moving probability of achieving independent ambulation if a patient is still dependent at different intervals postinjury. For instance, of the 46 patients who were not independently ambulating by 2 months postinjury, just under 33% ultimately achieved independent ambulation. The probability of recovery dropped to less than 14% and 9% at 3 and 4 months postinjury, respectively, for those who were not ambulating independently at those intervals. Prediction of Time to Recovery With univariate analysis, increased injury severity as measured by LOC, was associated with a significantly longer time to achieve independence (P.0001) (fig 3). Increased initial gait severity based on the functional gait scale was also associated with a significantly longer time to recover independence

RECOVERY OF AMBULATION AFTER TBI, Katz 867 Table 4: Comparison of Independent Ambulators and Nonindependent Ambulators by Neuropathologic Injury Subtypes Independent Ambulators % Nonindependent Ambulators % P Values Total 85 100.00 31 100.00 NS DAI 59 69.00 21 68.00 NS Focal cortical contusion 46 54.80 17 54.10 NS Hypoxic-ischemic injury 3 4.00 3 10.00 NS Herniation 11 12.90 4 12.90 NS (P.001) (fig 4). There was no significant effect of age, comparing patients 40 years old or younger (38.4 30.8d) with those over 40 (42.2 29.3d). There was no significant difference in recovery times among different neuropathologic categories. For instance, mean time to recovery for patients with DAI (n 80) was 39.3 30.6 days, compared with 42.3 29.2 days for patients without DAI (n 26) (primarily focal damage). Multivariate Analysis A prediction model was derived using stepwise regression analysis of admission FIM scores, admission FIM cognitive and motor subscores, initial gait impairment, PTA, LOC, GCS score on admission to the acute hospital, and age. Only admission FIM score and age entered the model, accounting for 38% of the variance of time to recover independent ambulation: Recovery time ambulation (days) 72.3.38(age).68(admission FIM). In a stepwise regression analysis, with FIM scores excluded, initial gait severity, injury severity (as measured by LOC), and age all entered as significant predictors of time to recover independent ambulation: Recovery time ambulation (days) 26.5 3.0(LOC [days]) 2.9(gait score) 0.5(age). This model was somewhat better than the one that had admission FIM scores; it accounted for nearly 40% of the variance predicting time to achieve independent ambulation. When a stepwise multiple regression model was derived using just the subgroup of patients with DAI (whether or not combined with other pathology subtypes) (n 80), PTA and initial gait score entered the model. This model predicted up to 58% of the variance of time to independent ambulation for this subgroup. DISCUSSION Impaired ambulation, whether from disorders of balance, power, coordination, cognition, or some combination, is common in patients with moderate to severe TBI. This is particularly true in the subset of patients admitted to inpatient rehabilitation who had not been compromised by orthopedic, peripheral nerve, or other comorbidities. This study is the first description of the time course of recovery of ambulation in patients with TBI. We found that the time to recovery of independent ambulation was highly dependent on injury severity, initial gait impairment, and age. Injury severity was significant, whether measured by traditional clinical measures of TBI severity (LOC, PTA) or by the initial FIM score. These finding are consistent with those of other studies that used similar variables to predict functional motor outcomes in persons with TBI. LOC and age are the variables that best predict recovery of equilibrium reactions in patients with TBI at 3 and 6 months postinjury. 15 Duration of unconsciousness and initial impairment level, but not age, are the best predictors of recovery of arm function. 1 We found no significant effect of neuropathology on recovery. Patients with or without DAI, focal cortical contusion, hypoxic-ischemic injury, or herniation had a similar prognosis and time course of recovery. For recovery of arm function, Fig 1. Cumulative percentage of patients (N 116) achieving independent ambulation at various time intervals postinjury. At 5.1 months, 85 patients were independent ambulators. Fig 2. Probability of achieving independent ambulation for patients who were still dependent in ambulation at various intervals postinjury. *Number of patients not reaching independent ambulation at respective times postinjury.

868 RECOVERY OF AMBULATION AFTER TBI, Katz Fig 3. Time to achieve independent ambulation in relation to duration of unconsciousness. Duration of unconsciousness: <1wk, n 62; 1 3wk, n 7; >3wk, n 4; missing data on 12 (ANOVA, P<.0001). Error bars equal SD. patients with primarily diffuse injury had a more protracted course of recovery than did patients with primarily focal injury. 1 For arm weakness, recovery likely hinges on whether there is damage to a single descending pathway. If damaged by a focal lesion, recovery may be similar to that experienced by stroke patients. If damaged by DAI, recovery will occur as long as synaptic reorganization can succeed. 16 For ambulation, there is no single critical factor for most patients, and recovery will be determined by idiosyncratic mixtures of cognitive, attentional, vestibular, and postural impairments, as well as weakness that can be generated in different manners by different pathologic conditions. The prognosis for independent ambulation may be slightly better for patients with severe TBI than for patients with stroke. Of the patients with TBI, more than 73% recovered independent ambulation within 5 months, compared with between 10% and 64% of patients with stroke who recovered ambulation over a similar period. 6-9,17 These percentages underestimate the proportion of TBI and stroke patients who ultimately recover independent ambulation with longer treatment and follow-up. 18 Within the limited time frame during which most studies are completed, patients with TBI may have a slightly longer window of opportunity to recover than do patients with stroke. In this study, 82% of patients who achieved independent ambulation after TBI could do so at 2 months postinjury and nearly 95% were independently ambulating at 3 months postinjury. The majority of stroke patients who recovered independent ambulation did so within 1 month, and almost all who recovered independent ambulation did so by 2 months. 8 Olsen 9 reported that 95% of those who achieved independent ambulation after stroke did so by 10 weeks. The Copenhagen Stroke Study reported that 80% of patients with stroke achieved their best walking function by 5 weeks and 95% achieved it by 11 weeks, 6 but 50% of the patients did not achieve independent ambulation. The time to recovery for the subgroup that did was presumably even shorter, but it was not separately reported. Patients with TBI appear to have 2 to 4 more weeks during which recovery of ambulation may occur than do patients with stroke. A similar conclusion was reached in contrasting recovery of arm function in patients with TBI to recovery reported in patients with stroke. 1 The fact that patients with stroke and TBI have different outcomes is not surprising, given the differences in the populations and the injuries. Stroke populations are older and the patients are more likely to have limiting medical comorbidities, such as heart disease, lung disease, or arthritis. In a stroke population, impaired ambulation will most commonly be caused by unilateral damage to corticospinal pathways, with variable degrees of postural instability, sensory loss, and spatial neglect. Patients with TBI have a much greater variety of impairments, including paresis, spasticity, ataxia, vestibular impairments, dystonia, and attentional deficits, due to a mix of focal and diffuse pathologic conditions. 19-21 Solitary, unilateral motor pathway damage is unusual in patients with TBI. For instance, large hemorrhages in the capsulostriatal area that potentially affect motor pathways occur in only about 3% of patients with severe TBI. 22 Most such patients have DAI that causes disseminated, bilateral foci of axonal damage. With more severe DAI, there may be greater disruption of motor pathways, especially with concentrated damage to converging networks, deeper in the neuraxis. 23 The potential for recovery and time course of neural reorganization is likely different for diffusely disrupted motor pathways after DAI than for more concentrated damage to motor pathways after stroke. Povlishock et al 16 proposed that diffuse deafferentation after DAI may allow for more complete synaptic reorganization than is possible with large, confluent deafferentation. In DAI, areas of synaptic loss may occur immediately adjacent to intact axons that are capable of reoccupying the denervated sites. 23,24 There are a few caveats in our findings. First, there was no control for separate effects of treatment and natural recovery. Treatment type and treatment intensity are factors that may affect prognosis and rate of recovery. 25,26 Second, the intensity and duration of therapy given the patients may not be achieved in current practice. These patients were on an acute rehabilitation service during their entire follow-up and received therapy 5 to 7 days a week. Independent ambulation was often the discharge goal. In current practice, patients tend to transition sooner to subacute or outpatient settings that provide less intense and less frequent therapy. It is unknown if the same Fig 4. Time to achieve independent ambulation in relation to initial gait score. Initial gait score: 0 3, n 19; 4 8, n 47; 9 10, n 17 (ANOVA, P<.001). Error bars equal SD.

RECOVERY OF AMBULATION AFTER TBI, Katz 869 prognosis over the same time course would result in shorter inpatient stays and less intensive therapy. Third, the main outcome measure in this study was time to independent ambulation. We did not consider other gait-related variables such as speed of walking or cadence, properties that may improve further after a patient achieves independent ambulation. 27 Gait velocity and dynamic balance are important for functional mobility in the home and community. 28 Innovative treatments, such as partial body weight supported treadmill training, can improve the speed and quality of gait long after patients achieve independent ambulation after stroke. 27 Study of these interventions should be extended to TBI. CONCLUSIONS Clinicians can use our findings to guide rehabilitation treatment and anticipate prognosis. Our results can also provide targets for judgments about future studies into recovery of ambulation, such as efficacy of specific treatments or equivalence of different rehabilitation settings. Rehabilitation goals should incorporate the potential for ambulation for patients with TBI, with expectations being modified across time. For instance, based on the findings in our sample, if patients with TBI are not ambulatory by 2 months postinjury, they still have at least a 33% chance of achieving ambulation independence by 5 months. Independent ambulation is a less realistic goal for patients who are not ambulatory by 4 months postinjury. Fewer than 9% of the patients in our study were independently ambulating by latest follow-up if they were not doing so at 4 months. The data suggest that clinicians should be cautious in setting independent ambulation as a therapy goal for patients who are not independent by 3 to 4 months postinjury. Treatment efforts for those patients might focus on other areas of functional mobility. Clinicians should periodically reassess patients for possible ambulatory goals because a small number of patients with TBI will recover independent ambulation after 4 months postinjury. Reassessment of patients with orthopedic or other impairments unrelated to brain injury is particularly important because the impairments may delay recovery. More research is needed to establish parameters to predict later recovery in this subgroup of patients. References 1. Katz DI, Alexander MP, Klein RB. Recovery of arm function in patients with paresis after traumatic brain injury. Arch Phys Med Rehabil 1998;79:488-93. 2. Corrigan JD, Smith-Knapp K, Granger CV. Outcomes in the first 5 years after traumatic brain injury. Arch Phys Med Rehabil 1998;79:298-305. 3. Asikainen I, Kaste M, Sarna S. Predicting late outcome for patients with traumatic brain injury referred to a rehabilitation programme: a study of 508 Finnish patients 5 years or more after injury. Brain Inj 1998;12:95-107. 4. Novack TA, Bush BA, Meythaler JM, Canupp K. Outcome after traumatic brain injury: pathway analysis of contributions from premorbid, injury severity, and recovery variables. Arch Phys Med Rehabil 2001;82:300-5. 5. van der Naalt J, van Zomeren AH, Sluiter WJ, Minderhoud JM. One year outcome in mild to moderate head injury: the predictive value of acute injury characteristics related to complaints and return to work. J Neurol Neurosurg Psychiatry 1999;66:207-13. 6. Jorgensen HS, Nakayama H, Raaschou HO, Olsen TS. Recovery of walking function in stroke patients: the Copenhagen Stroke Study. Arch Phys Med Rehabil 1995;76:27-32. 7. Wandel A, Jorgensen HS, Nakayama H, Raaschou HO, Olsen TS. Prediction of walking function in stroke patients with initial lower extremity paralysis: the Copenhagen Stroke Study. Arch Phys Med Rehabil 2000;81:736-8. 8. Friedman PJ. Gait recovery after hemiplegic stroke. Int Disabil Stud 1990;12:119-22. 9. Olsen TS. Arm and leg paresis as outcome predictors in stroke rehabilitation. Stroke 1990;21:247-51. 10. Patel AT, Duncan PW, Lai SM, Studenski S. The relation between impairments and functional outcomes poststroke. Arch Phys Med Rehabil 2000;81:1357-63. 11. Katz DI, Alexander MP. Predicting outcome and course of recovery in patients with traumatic brain injury admitted to rehabilitation. Arch Neurol 1994;51:661-70. 12. Cook L, Smith DS, Truman G. Using Functional Independence Measure profiles as an index of outcome in the rehabilitation of brain-injured patients. Arch Phys Med Rehabil 1994;75:390-3. 13. Wade DT, Collin C. The Barthel ADL Index: a standard measure of physical disability? Int Disabil Stud 1988;10:64-7. 14. Dodds TA, Martin DP, Stolov WC, Deyo RA. A validation of the functional independence measurement and its performance among rehabilitation inpatients. Arch Phys Med Rehabil 1993;74:531-6. 15. Macpherson V, Sullivan SJ, Lambert J. Prediction of motor status 3 and 6 months post severe traumatic brain injury: a preliminary study. Brain Inj 1992;6:489-98. 16. Povlishock JT, Erb DE, Astruc J. Axonal response to traumatic brain injury: reactive axonal change, deafferentation and neuroplasticity. J Neurotrauma 1992;9(Suppl 1):S189-200. 17. Reding MJ, Potes E. Rehabilitation outcome following initial unilateral hemispheric stroke. Life table analysis approach. Stroke 1988;19:1354-8. 18. Olney SJ, Richards CL. Hemiparetic gait following stroke. Part I: Characteristics. Gait Posture 1996;4:136-48. 19. Katz DI, Black SE. Neurological and neuroradiological evaluation. In: Rosenthal M, Griffith ER, Kreutzer JS, Pentland B, editors. Rehabilitation of the adult and child with traumatic brain injury. Philadelphia: FA Davis; 1999. p 89-116. 20. Haggard P, Miall RC, Wade D. Damage to cerebellocortical pathways after closed head injury: a behavioral and magnetic resonance imaging study. J Neurol Neurosurg Psychiatry 1995; 58:433-8. 21. Basford JR, Chou LS, Kaufman KR, et al. An assessment of gait and balance deficits after traumatic brain injury. Arch Phys Med Rehabil 2003;84:343-9. 22. Katz DI, Alexander MP, Seliger GM, Bellas DN. Traumatic basal ganglia hemorrhage: clinicopathologic features and outcome. Neurology 1989;39:897-904. 23. Blumbergs PC, Jones NR, North JB. Diffuse axonal injury in head trauma. J Neurol Neurosurg Psychiatry 1989;52:838-41. 24. Steward O. Reorganization of neural connections following CNS trauma. Principles and experimental paradigms. J Neurotrauma 1989;6:99-152. 25. Kwakkel G, Kollen BJ, Wagenaar RC. Long term effects of intensity of upper and lower limb training after stroke: a randomised trial. J Neurol Neurosurg Psychiatry 2002;72:473-9. 26. Richards CL, Malouin F, Wood-Dauphinee S, Williams JI, Bouchard JP, Brunet D. Task-specific physical therapy for optimization of gait recovery in acute stroke patients. Arch Phys Med Rehabil 1993;74:612-20. 27. Sullivan KJ, Knowlton BJ, Dobkin BH. Step training with body weight support: effect of treadmill speed and practice paradigms on poststroke locomotor recovery. Arch Phys Med Rehabil 2002; 83:683-91. 28. Perry J, Garrett M, Gronley JK, Mulroy SJ. Classification of walking handicap in the stroke population. Stroke 1995;26:982-9. Supplier a. SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.