Introduction

Spinal cord lesions (SCLs), which include those due to trauma and those resulting from disease (nontrauma), are a major clinical and social problem.1 The incidence of spinal cord injury (SCI) has been reported to be between 20 and 50 cases per million per year.2,3 SCIs usually involve young individuals (half of the cases are under 30 years of age), while diseases usually affect an older population. Walking function, however, is frequently impaired in both groups.4 Recent epidemiological studies5 demonstrated that a larger proportion of new traumatic cases presented as incomplete spinal cord lesions (eg some preservation of sensory and/or motor function below the lesion level) and therefore are more likely to regain walking function. Furthermore, advances in emergency medical care and the acute treatment of spinal cord lesions, for example, with methylprednisolone6,7 suggests efficacy in reducing the severity of neurological impairment, which may also increase the proportion of incomplete lesions with resultant increased probability of recovery of walking function.

Recent reports show that recovery of walking after paralysis of lower extremities muscles, particularly following acute spinal cord injury, is one of the primary goals of the patients. In fact, while individuals with complete SCI may rank bladder and bowel control at their first preference for independence in self-care, independent walking was ranked first by patients, who have the capacity to walk.8 A recent report shows that recovery of walking function in individuals with incomplete SCI was ranked equal with bowel and bladder recover.9

Recovery of locomotion, therefore, has become a primary goal of both pharmacological and rehabilitative treatment. In the last decade, pharmacological treatment10 and rehabilitative approaches11,12 have been initiated to enhance locomotion capacity of SCI patients. Basic science advances13 in regeneration of the central nervous system hold promise of further neurological and functional recovery to be studied in clinical trials.14

In view of these advances, it has become of increased importance to develop new instruments for the precise measure of walking, such as the Walking Index for Spinal Cord Injury (WISCI), for use in clinical trials.15,16

Unfortunately, there is limited information in the literature on the outcome of walking function in large populations following SCI. In recent large clinical trials such as NASCIS III and Model Systems Data Base,17,18 the Functional Independence Measure (FIM) has been utilized to determine ambulation status following acute SCI, but the FIM has limitations in sensitivity.16 Walking outcome has also been reported recently19 utilizing the Benzel Scale,20 which is a modification of the Frankel Grade, but is yet to be validated.

Several scales such as the Barthel Index (BI),21 the Rivermead Mobility Index (RMI),22 and the FIM23 were developed as functional measurements, which include a component of walking for stroke and other disabilities, but have not been validated for SCI. The Modified Barthel Index (MBI) for SCI has been reported by Yarkony et al24 in a large group of subjects from one center, but the MBI has not been utilized in multicenter studies. The Spinal Cord Independent Measure (SCIM) initially reported in 199725 is currently being studied in international studies and is a global scale, with better precision than the FIM and a more detailed description of walking than the RMI, BI and FIM. The WISCI incorporates gradations of physical assistance and devices required for walking and has been shown to have face, criterion and construct validity and is more detailed than the SCIM for walking function.26 We recently reported4 differences in age and spinal cord disease/injury utilizing the WISCI, RMI and BI, which suggested the need for more critical comparison of different walking scales as valid outcome measures.

This study examines the response to change in walking status for subjects with the acute onset of spinal cord disease or trauma during the initial rehabilitation stay by comparing five walking scales and correlating these findings with Lower Extremity Motor Scores (LEMS) and the ASIA impairment Scale (AIS). Our hypothesis is that the WISCI scale will show more incremental change than current scales, and therefore demonstrate criterion validation for use in future clinical trials.

Patients and methods

Subjects

The sample is composed of 284 patients with a recent onset of traumatic and nontraumatic SCLs admitted to the spinal unit of a rehabilitation hospital in Italy between 1997 and 2001. Patients without the cognitive ability to participate in the rehabilitation program were excluded.

Measures

The following general information was collected retrospectively.

  • Onset of lesion-to-admission (LTA) to the hospital refers to onset of trauma or onset of disease at first admission to the acute hospital.

  • Injury variables included etiology, associated injury, medical complications and surgical intervention.

  • Length of stay (LOS) as in-patients has been considered as an outcome measure as well as an independent determinant of other outcomes.

The following admission and discharge data were collected retrospectively from the charts:

  • Neurological evaluation had been performed according to the American Spinal Injury Association (ASIA) Standards,27 including AIS, right and left motor and sensory levels. A subsample of 200/249 subjects was evaluated for total and LEMS. Neurological recovery was defined on the basis of improvement of motor scores (total and LEMS) and AIS.

  • The global FIM24 is an 18-item scale with seven possible functional levels. Only the locomotor (walking/wheelchair) item, which is scored from 1 to 7, was assigned.

  • The BI22 is a global scale with scores that range from 0 to 100, higher scores denoting greater levels of independence. The ambulation item for walking function, which is scored from 0 to 15, was recorded.

  • The RMI23 is a 15-item mobility scale in which some items of the scale evaluate patients' bed mobility and postural transfers, while other items assess patients' walking. Ambulation items, which are scored in three categories (0, 7, 10), were recorded.

  • The SCIM25 assesses three functional domains: self-care, respiration and sphincter management and mobility. Scores for the mobility indoor function range from 0 to 8, and were assigned.

  • The WISCI16 is a new 0–20 level scale, which evaluates walking based on physical assistance, the need of braces and devices. The levels are scored from 0 (patient unable to walk) to 20 (patient walking without braces and/or devices and without physical assistance for at least 10 m (Table 1). In scoring, the level in which the patient is safe as judged by the therapist was assigned.

    Table 1 Walking Index for Spinal Cord Injury (WISCI II)
    Full size table

The BI and RMI had been prospectively recorded on charts. SCIM, FIM and WISCI were retrospectively scored based on the description of walking in the charts by three independent observers (correlations were run with the means of observations by two physical therapists specialized in SCL and one physician neurologist specialized in SCL); such data were available for a subsample of 249/284 patients.

Statistical analysis

The following data were calculated:

  • LEMS, BI, RMI, FIM, SCIM and WISCI changes were calculated based on the difference between rehabilitation discharge and admission.

  • LEMS, BI and RMI efficiency scores were calculated by dividing changes in scores by the respective LOS.

  • Spearman rank correlations were calculated between observers from SCIM, FIM and WISCI demonstrating interrater reliability.

  • For the correlation between ASIA impairment at admission and WISCI at discharge, WISCI scores were divided into three groups according to patients' walking capacity: (1) patients not able to walk; (2) patients walking with physical assistance and (3) patients with independent walking.

  • For concurrent validation, Spearman rank correlations were performed for a subsample (N=76) of those patients with WISCI scores >0 and <20 (0 being unable to stand or participate and 20 being complete independence) and scores on all other scales.

  • Significance was set at P<0.05.

Results

Demographics

The sample consisted of 184 males and 100 females; mean age was 50.4±19.3 years (range 12–86). Mean interval from lesion to admission was 56.9±43.9 days; admission was within 30 days in 96 patients, within 60 days in 90, within 90 days in 47 and over 90 days in 51. Mean length of stay was 98.7±68.13 days. A nontraumatic etiology was present in the majority of the patients (177/284): inflammatory (40), vascular (36), neoplastic (39), degenerative (62); traumatic lesions (107/284): car accident (38), motorcycle accident (15), sport accident (7), acts of violence (6), suicide attempts (6), accidental falls (31). Lesion level: 81 had cervical lesions, 148 thoracic lesions and 55 had lumbar-sacral lesions (including cauda equina damage). At admission, 84 patients had ASIA impairment A, 19 ASIA B, 129 ASIA C and 52 ASIA D.

Concurrent validation

Inter-rater reliability for retrospective data interpretation for SCIM, FIM and WISCI are as follows: for the WISCI and the SCIM, the correlations were 1.00 and for the FIM 0.9 (P<0.001). For the subsample of 76 patients, who were discharged at WISCI level between 1 and 19, the more frequent levels were levels 13 and 16 (see Table 2 for all scale scores).

Table 2 Description of the various locomotion scale scores N=249
Full size table

Spearman rank correlation between the various scales for this subsample is shown in Table 3.

Table 3 Spearmann rank order correlation between the various scales N=76
Full size table

Correlation ASIA impairment/locomotion (WISCI)

In the subsample of 249 patients with all scale scores, the initial ASIA grade was predictive of mobility outcome in WISCI (Table 4a and b).

Table 4 (a) ASIA impairment (AIS) at admission and WISCI levels at discharge and (b) Correlation between ASIA impairment at admission and WISCI levels at discharge
Full size table

Correlation WISCI/LEMS

In the subgroup of 200/249 patients examined with LEMS there was a positive correlation between WISCI (0–20) and LEMS of 0.58 (P<0.001). A subset of 63 of the subsample of 76 (WISCI levels 1–19, eliminating WISCI levels 0 and 20 to control for floor and ceiling effects) had LEMS, which showed a correlation with locomotion outcome at discharge of 0.57 (P<0.001). In addition, there was a positive correlation of LEMS and WISCI levels at discharge for young patients of 0.50 (P<0.01) (N=35<50 years) and for older patients of 0.64 (P<0.01) (N=28 ⩾50 years).

A positive correlation of motor scores at discharge and WISCI at discharge for nontrauma (N=26) of 0.58 (P<0.01) and trauma (N=37) of 0.49 (P<0.01) is shown (one patient was lost to the correlation for incomplete data).

Distribution of LEMS and WISCI levels at discharge

Scattergram showing motor score distribution is displayed in Figure 1.

Figure 1
figure 1

LEMS and WISCI levels distribution

Full size image

Discussion

This study supports our hypothesis and demonstrates that the WISCI scale is more sensitive to incremental change than the RMI, BI, FIM and SCIM and has enhanced concurrent criterion validation.

Table 3 comparing the scales shows that all scale correlations with WISCI fall between 0.67 and 0.97. There is overlap as all scales conceptually measure the domain of mobility; however, each scale differs on criterion of score assignment.

The FIM scale24 (which does not integrate use of devices or braces) was originally developed as a measurement tool assessing burden of care for all types of disability. The locomotion item describes walking and wheelchair locomotion along seven levels of function from total assistance to independence with emphasis on amount of physical assistance needed (for further in-depth discussion see Ditunno et al 2000). The BI22 (which also does not integrate devices or braces) as a total score was developed as a tool for burden of care (feeding, hygene, tranfers, mobility) and recovery scale for bladder and bowel management for stroke patients. The ambulation item describes walking or ambulation along four levels: unable to ambulate, use of wheelchair for 45 m, ambulation with assistance for 45 m and ambulation independent for 45 m. The RMI23 (which is not continuous and does not specify differences in braces or devices) was developed for functional levels of mobility primarily for stroke patients. The walking function items describe three categories (cannot walk, walk with devices and no assistance and walk independent). These three scales (FIM, BI, RMI) show correlations of 0.67–0.70, which is lower than the SCIM at 0.97. This may be due to the primary focus of the scales during the time of original development, which lacked the integration of devices/braces.

The SCIM,25 however, does integrate wheelchair, crutches, canes, leg orthosis, etc, and was developed as a functional level scale specifically for SCI patients. As such, it offers an overall ADL assessment, and the mobility indoors item allows scores from 0 to 8 and shows the highest correlation with the WISCI (0.97), but lacks one of the WISCI levels. In reviewing the pattern of score distributions for the SCIM, we find that although there was a high correlation with the WISCI, there is a large segment of subjects (18/76 about 24%) with WISCI level 13 (ambulates with walker, no braces and no physical assistance) with no comparable SCIM score. In addition, there is a differential hierarchical ranking on WISCI item 18 (ambulates with no devices, with braces and no physical assistance), reflecting an inversion of monotonic function. These issues present concerns regarding the SCIM scale validity for ambulation assessment.

These findings do not invalidate the above scales for global assessment of function, but do emphasize that they are not specific for ambulation. Overall, the WISCI is a more precise and sensitive scale for documenting change in levels of walking along a hierarchical order, integrating devices, braces and physical assistance.

The pattern of the recovery of walking function as measured by the WISCI scale demonstrates some interesting features. First, the majority of subjects fall into three levels, levels 13, 16 and 20 (92/117 or 116), probably because the discharge status from rehabilitation typically is at an independent level. This eliminates levels 1–8, 10, 11, 14 and 17, because these levels are not independent walking function. Our data, which is retrospective, however, shows some of these nonindependent levels (14/117) requiring some assistance at the time of discharge, because this was the maximum walking level achieved after 4–6 months of rehabilitation, but the subjects would need to function in a wheelchair after they were discharged. The 133 subjects discharged at the 0 WISCI level, never progressed beyond the wheelchair.

The strong correlations of the ASIA grades with the WISCI levels (Table 4a and b) further support the construct validity of the scale. The highest correlation is for subjects who are ASIA C and D, thus demonstrating that the WISCI scale primarily is applicable to these subjects and to some ASIA A and B who convert to C. Several subjects remained complete (ASIA A 3/117), but ambulated with a reciprocal gait in long leg braces. Most motor complete subjects (ASIA A and B), who remain motor complete, will only ambulate with braces and will not use reciprocal ambulation unless the legs are locked at the knees and/or they have RGOs. This represents only a small percentage of all SCI subjects, who walk at high-energy requirements.28 These findings are consistent with the restrospective data, reported in the modification of the original WISCI scale from 20 to 21 levels.29

The correlation of LEMS with WISCI levels in this study is in line with the data recently reported by Marino in chronic SCI subjects30 and both support the construct validity of the scale. The concept that walking function should improve as strength in the legs increases after SCI is consistent with validating the hierarchy of the WISCI levels. Waters demonstrated that a threshold of 30 for the LEMS was necessary to achieve community ambulation. Walking function, however, is a very complicated activity, which may be affected by many factors.12 Motor scores alone do not correlate completely with increased walking function, because other factors such as age,4 pain,30,31 balance,32,33 spasticity,34 overweight35 and possibly duration of training may effect walking function. In fact in our patients, motor scores alone show a significant (r=0.57, P<0.001), but not complete correlation with the WISCI.

A subset of 21 patients (mostly young with traumatic lesions) with low LEMS (<30) achieved WISCI levels 13 (walker, no braces, no assistance), 16 (two crutches, no braces, no assistance) and 20 (no devices and no assistance) (see Scattergram, Figure 1), a finding which could have several explanations. Owing to the longer length of stay (4–6 months) of subjects in this study as compared to subjects in the USA, patients may have progressed to higher WISCI levels. A longer length of stay in rehabilitation may result in a higher WISCI level due to training rather than improvement in strength. Perhaps these subjects need to be differentiated by distance, speed and energy requirements as recently reported for different WISCI levels.30,36 These issues will be examined in studies currently in progress.37 Another explanation of differences in response based on different motor scores may be related to the pattern of lower extremity weakness. People with good proximal strength perform better despite low total motor score (hip flexors and knee extensors are the most important muscles for walking).38,39 Finally, it should be noted that Waters38 found that subjects with motor scores below 30 for the lower extremity scores did not achieve community ambulation; he did not examine motor scores for household ambulation. The WISCI levels, however, are based on a 10 m distance (33 ft), which is closer to household ambulation (50 ft) than to community ambulation, which is 150 ft. Therefore, motor scores lower than 30 for the lower extremities may permit household distances, especially if the subjects are young and have had the benefit of additional training time. In these cases, there may be a difference in speed and efficiency for the same WISCI levels and the subjects will need to be examined by methods recently reported in chronic subjects based on velocity and physiologic cost index.30,36 Furthermore, it has been recently reported that, while with training (body weight support), foot-motion of SCI patients tends to recover the shape and the step-by-step reproducibility that characterize normal gait, the specific activation of individual muscles often differs substantially from the normal ones and include the activation of several trunk muscles which are not included in ASIA key muscles evaluation.40

On the other hand, a subset of seven subjects, who performed at lower WISCI levels (4–11) than expected with LEMS of an average of 26, had associated problems of age, weak upper extremities, pain, balance problems and overweight. The effect of age is particularly worthy to note: Burns et al41 and Penrod et al42 reported that subjects over 50 years with ASIA C within a week of injury when compared to younger individuals, did much poorer in recovery of walking function, and speculated that this may have been due to associated medical problems. Scivoletto et al4 analyzed matched cohorts from the same sample of subjects of the present study and found neurologic recovery and ambulation were less in those over 50 years as compared to younger subjects irrespecitve of etiology.