Clinical Benefit of Rehabilitation Training in Spinal Cord Injury

A Systematic Review and Meta-Analysis

Ruimeng Duan, MS; Mingjia Qu, MS; Yashuai Yuan, MS; Miaoman Lin, MS; Tao Liu, MS; Wei Huang, MS; Junxiao Gao, MS; Meng Zhang, MS; Xiaobing Yu, MD

Disclosures

Spine. 2021;46(6):E398-E410. 

In This Article

Results

Study Selection

The initial search in databases yielded a total of 1410, 1644, and 452 English RCT/CT reporting the efficacy of rTMS, FES, ABT, and robot-assisted treadmill training in functional recovery after SCI (Figure 1). After removing duplicated publications and publications not met the inclusion criteria, a total of 31 eligible publications (28 RCTs[5,6,9–14,17,18,24–41] and 3 CTs[2,3,42]) included 1040 participants were included in present study, including 4 articles of rTMS,[13,14,24,25] 6 FES,[26–31] 9 exercise or treadmill training,[3,11,12,17,32–35,42] and 12 locomotor robot.[2,5,6,9,10,18,36–41] Four studies used crossover design.[12,17,32,39]

Figure 1.

The flow diagram of study selection process in this study.

For the grey literatures, we also screened, but no literature meeting the requirements of meta-analysis was included.

Study Characteristics and Participants

A total of 81, 174, 238, and 547 patients were enrolled and were assigned into intervention with rTMS plus gait training versus sham rTMS plus gait training[13,14,24,25] (Table 1), FES hand therapy plus conventional occupational therapy (COT) versus COT alone,[26–31] exercise or treadmill training versus overground walking or no intervention,[3,11,12,17,32–35,41,42] and robotic-assisted treadmill training versus overground walking or no intervention,[2,5,6,9,10,18,36–40] respectively. The interventions ranged from 30 to 60 minutes per session and three to five sessions per week, continued for 4 to 24 weeks. The median age of individuals with SCI ranged from 24 to 59 years and ASIA impairment scale ranged from grade A to D.

Quality Assessment

The averaged PEDro scale score of the 31 included studies was 5.30 (Table 2). Most studies were RCTs (90.32%, 28/30) and had group similarity at baseline (70.97%, 22/31), comparable differences between groups (86.10%, 27/31), primary outcomes from more than 85% of subjects (77.42%, 24/31), and reported variability (96.77%, 30/31). All the studies did not include blind participants and therapists, and most were not intention-to-treat analysis (80.64%, 25/31) and about half were concealed allocation (48.39%, 15/31).

Outcome and Measurement

All studies reported the improvement in functional upper/lower extremity independence, upper extremity function, mobility, and life quality. Patient independence was measured using functional independence measure (FIM) scale[11,12,23,25,27,30,33,34] and SCI independence measure (SCIM) scale.[2,25,30,35,39] The lower or upper extremity functions were measured using ASIA UEMS, UEFI, m-ARAT,[26] and ASIA LEMS,[10,13,25,41] Toronto rehabilitation institute hand function test (TRIHFT),[22,23] Jebsen–Taylor hand function test (JTHFT),[33] and rehabilitation engineering laboratory hand function test (RELHFT).[31] The walking speed and walking distance were measured using 10MWT[5,6,9,13,28,35,36,40,41] and 6WMT,[5,6,9,36,40,41] respectively. Body balance, total motor ability, and functional level were measured using timed up and go test (TUG),[5,6,18,29,35] Craig Handicap Assessment Report Technique (CHART),[28,42] Tinneti scale,[34] Berg balance scale (BBS),[11,41] or walking index for spinal cord injuries version II (WISCI II).[2,9,10,13,24,25,37,38] Spasticity following rTMS was assessed using the MAS.[13,14,24,25] The quality of patient life was assessed using 36-item short-form health survey (SF-36),[34] quality of life index for SCI version III,[12] satisfaction with life scale,[28] and assessment of quality of life-8 (AQoL-8).[30]

Efficacy of rTMS on Lower Extremity Functions in Individuals With SCI

The efficacy rTMS in SCI were confirmed by improvement in 10MWT (MD = 0.09, 95% CI [0.01, 0.16], I 2 = 50%, fixed-effects model; P = 0.03; Figure 2A) and ASIA LEMS scores (MD = 4.41, 95% CI [1.55, 7.27], I 2 = 0%, fixed-effects model; P = 0.003; Figure 2B) compared with controls (sham rTMS or plus gait training). rTMS-induced insignificant spasticity (MAS score) (MD = –0.49, 95% CI [–1.13, 0.14], I 2 = 61%, Random-effects model; P = 0.13; Figure 2C), and WISCI II (MD = –1.52, 95% CI [–3.43, 0.40], I 2 = 25%, Fixed-effects model; P = 0.12; Figure 2D) compared with control.

Figure 2.

The forest plot for the clinical effectiveness of rTMS on the movement of participants with SCI. A, The improvement in 10-meter walking test (10MWT). B, The efficacy in ASIA lower extremities motor score (LEMS). C, The effect on spasticity by MAS score. D, The benefit of rTMS in walking index for spinal cord injuries version II (WISCI II). IV indicates inverse variance; MAS, modified Ashworth scale; rTMS, repetitive transcranial magnetic stimulation; SCI, spinal cord injury.

Efficacy of FES on Extremity Independence

Meta-analysis indicated FES significantly increased upper extremity independence (MD = 2.92, 95% CI [0.37, 5.48], I 2 = 50%, Fixed-effects model; P = 0.03; Figure 3A). There was no obvious difference in overall upper extremity function (including motility, balance, and hand function) (MD = 5.17, 95% CI [–2.23, 12.56], I 2 = 0%, Fixed-effects model; P = 0.17; Figure 3B), lower extremity independence (MD = –0.02, 95% CI [–1.18, 1.15], I 2 = 34%, Fixed-effects model; P = 0.98; Figure 3C), and the life quality of individuals with SCI (MD = 0.07, 95% CI [–0.04, 0.18], I 2 = 0%, Fixed-effects model; P = 0.21; Figure 3D) between the experimental (FES) and control groups.

Figure 3.

The forest plot for the clinical effectiveness of FES on the lower/upper extremity independence and function movement of participants with SCI. A, The effectiveness of FES on the upper extremity (UE) independence. B, The effectiveness of FES on the upper extremity (UE) function. C, The effectiveness of FES on lower extremity (LE) independence. D, The influence of FES on SCI patient quality of life. FES indicates functional electrical stimulation; IV, inverse variance; SCI, spinal cord injury.

Efficacy of ABT on Walking Speed and Distance in Individuals With SCI

The activity-based therapies, including massed practice and treadmill training did not improve 10MWT (MD = –0.01, 95% CI [–0.09, 0.06], I 2 = 0%, Fixed-effects model; P = 0.71; Figure 4A) and 6MWT (MD = 6.46, 95% CI [–9.02, 21.94], I 2 = 64%, Random-effects model; P = 0.41; Figure 4B) in individuals with SCI compared with control (no intervention, overground mobility therapy, self-regulated exercises, or conventional rehabilitation program). Activity-based therapeutic interventions induced equivalent incidence of dropout (OR = 1.36, 95% CI [0.72, 2.55], I 2 = 0%, fixed-effects model; P = 0.34; Figure 5A) and adverse events (OR = 2.73, 95% CI [0.59, 12.54], I 2 = 0%, fixed-effects model; P = 0.20; Figure 5B) with control.

Figure 4.

The forest plot for the clinical effectiveness of activity-based therapy (ABT) on the movement of participants with SCI. The effectiveness of exercise and training on the walking speed (10MWT; A) and walking distance (6MWT; B). IV indicates inverse variance; SCI, spinal cord injury.

Figure 5.

The incidence of activity-based therapy (ABT)-induced adverse events in SCI patients. The incidence of dropout (A) and adverse events (B) during intervention by ABT. IV indicates inverse variance; SCI, spinal cord injury.

Efficacy of Robotic-assisted Training on Movement in Individuals With SCI

Meta-analysis indicated robotic-assisted treadmill training did not increase 10MWT (MD = 0.00, 95% CI [–0.02, 0.03], I2 = 3%, Fixed-effects model; P = 0.74; Figure 6A) and 6MWT (MD = 14.30, 95% CI [–5.80, 34.41], I 2 = 85%, Fixed-effects model; P = 0.16; Figure 6B) compared with control (no intervention, bike, and overground walking). However, robotic-assisted training significantly increased ASIA LEMS score (MD = 5.00, 95% CI [3.44, 6.56], I 2 = 0%, Fixed-effects model; P < 0.00001; Figure 6C) and lower extremity independence (MD = 3.73, 95% CI [2.53, 4.92], I 2 = 28%, Fixed-effects model; P < 0.00001; Figure 6D). No difference was found in the dropout rate between the two groups (MD = 1.05, 95% CI [0.32, 3.40], I 2 = 46%, Fixed-effects model; P = 0.94; Figure 6E).

Figure 6.

The effectiveness of robotic-assisted training on lower extremity (UE) function of participants with SCI. The effectiveness of robotic-assisted training on the walking speed (10MWT; A) and walking distance (6MWT; B). C and D, The effectiveness of robotic-assisted training on the ASIA LE motor scores (LEMS) and LE independence in patients with SCI. E, The incidence of dropout during robotic-assisted training intervention. ASIA indicates American Spinal Injury Association; IV, inverse variance; SCI, spinal cord injury.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....