Seven studies (summarized in Table 5) have examined the effect of therapist-assisted (Dobkin et al. 2007; Dobkin et al. 2006; Wernig et al. 1995; Hornby et al.2005a; Harkema et al. 2012; Alcobendas-Maestro et al. 2012; Benito-Penalva et al. 2012) BWSTT in people who had incurred an incomplete SCI <12 months prior (acute/subacute phase) (aggregate N=566). One study also examined overground training and robotic–assisted treadmill training (Hornby et al. 2005a). Treatment time ranged from 90-300 minutes per week and total treatment duration lasted between 3 and 23 weeks.
Although lower levels of study design (non-randomized, non-blinded) suggest that BWSTT in acute/sub-acute SCI yields better outcomes than conventional rehabilitation (Wernig et al. 1995), there exists strong evidence from a single-blind RCT (Dobkin et al. 2006) (n=117) that there are no differences in effects between matched amounts of BWSTT and overground mobility practice in incomplete SCI during inpatient rehabilitation for the locomotor score of the FIM or overground walking speed. These two variables in both groups improved roughly in parallel over the 12 weeks of therapy (Dobkin et al. 2007). In both groups, improvements in walking function were particularly notable in subjects with AIS C (92%) or D (100%). Indeed, as reported by Dobkin et al (2006), the initial AIS classification of subjects is an important indicator of locomotor recovery. Among the subjects who were initially classified as AIS B, those who improved to AIS C within 8 weeks post-injury showed improved walking function while those who remained as AIS B did not (Dobkin et al. 2006). In addition, subjects who entered the trial earlier (< 4 weeks post-injury) had faster walking speeds and endurance post-training. This was particularly the case for subjects who improved in their AIS classification within 4 to 6 weeks post-injury. Preliminary analysis from a smaller RCT (Hornby et al, 2005a) (n = 30) had similar results, showing that there were no differences in motor (lower extremity motor scores) or functional recovery (FIM locomotor subscore or WISCI II score) between those that trained overground, with BWSTT, or with robotic assisted treadmill training. All increases from initial to final evaluations were significant. However, the results from a more recent randomized single-blind parallel-group clinical trial from Alcobendas-Maestro et al. (2012) (N=75) suggested that Lokomat-assisted BWSTT may result in better improvements in the WISCI II, 6MWT, LEMS, and FIM-L scores compared to conventional treatment. However, it is unclear whether the conventional treatment group received an equivalent amount of task-specific locomotor practice, so differences in the intensity and/or specificity of training could account for these contrasting results.
Wernig et al. (1995) showed that 86% (49/57) of incomplete SCI subjects who underwent BWSTT in the acute phase of injury achieved improvements in functional ambulation. They reported that only 50% of the initially non-ambulatory subjects (historical controls) who underwent conventional rehabilitation improved functional ambulation. The results of the remaining 16 historical control subjects who were initially ambulatory were not explicitly reported, although it appears from the article’s bar graphs that they also improved in functional class. Thus, it is possible that the proportion of subjects who improved functional ambulation after Wernig’s conventional rehabilitation may actually have been closer to 70% ([16+12]/40).
A contentious issue in field of gait training has been the appropriateness of the control intervention (Wolpaw 2006). The ‘conventional’ rehabilitation to which BWSTT was compared was not well defined in Wernig’s studies, although it appeared to focus on wheelchair mobility in addition to gait training in parallel bars and using braces (Wernig 2006a). The control group in the large RCT (Dobkin et al. 2006) underwent task-oriented overground gait retraining of equivalent intensity to the BWSTT group and therefore may not have offered enough of a contrast in treatment modality to detect significant differences. Moreover, based on retrospective analysis from the participating centres, it appeared that significantly enhanced locomotor outcomes were achieved with both treatment and control conditions as compared to what was achieved with “standard practice” in the centres prior to the study – although it is likely that pre-study practice of gait therapy varied to some degree across the study centres. On the other hand, the more recent RCT Alcobendas-Maestro et al. (2012) did not clearly state whether the conventional training group received an equivalent amount of overground therapy for walking compared to the Lokomat training group. Despite these difficulties, the important message from this work is that intensive task-oriented gait retraining, whether implemented by BWSTT or overground practice, facilitates the recovery of functional ambulation especially <12 months post-injury. However, there is no strong evidence that one rehabilitation approach is superior to another.
There is level 2 (Alcobendas-Maestro et al. 2012) and level 3 evidence (Wernig et al. 1995) using historical controls that BWSTT is effective in improving ambulatory function. However, two level 2 RCTs (Dobkin et al. 2006; Hornby et al. 2005a) demonstrates that BWSTT has equivalent effects to conventional rehabilitation consisting of an equivalent amount of overground mobility practice for gait outcomes in acute/sub-acute SCI.
- For patients less than 12 months post-SCI, BWSTT may have similar effects on gait outcomes as overground mobility training of similar intensity.