Exercise and aerobic capacity in individuals with spinal cord injury: A systematic review with meta-analysis and meta-regression

Background: A low level of cardiorespiratory fitness [CRF; defined as peak oxygen uptake (V˙O2peak) or peak power output (PPO)] is a widely reported consequence of spinal cord injury (SCI) and a major risk factor associated with chronic disease. However, CRF can be modified by exercise. This systematic review with meta-analysis and meta-regression aimed to assess whether certain SCI characteristics and/or specific exercise considerations are moderators of changes in CRF. Methods and findings: Databases (MEDLINE, EMBASE, CENTRAL, and Web of Science) were searched from inception to March 2023. A primary meta-analysis was conducted including randomised controlled trials (RCTs; exercise interventions lasting >2 weeks relative to control groups). A secondary meta-analysis pooled independent exercise interventions >2 weeks from longitudinal pre-post and RCT studies to explore whether subgroup differences in injury characteristics and/or exercise intervention parameters explained CRF changes. Further analyses included cohort, cross-sectional, and observational study designs. Outcome measures of interest were absolute (AV˙O2peak) or relative V˙O2peak (RV˙O2peak), and/or PPO. Bias/quality was assessed via The Cochrane Risk of Bias 2 and the National Institute of Health Quality Assessment Tools. Certainty of the evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. Random effects models were used in all meta-analyses and meta-regressions. Of 21,020 identified records, 120 studies comprising 29 RCTs, 67 pre-post studies, 11 cohort, 7 cross-sectional, and 6 observational studies were included. The primary meta-analysis revealed significant improvements in AV˙O2peak [0.16 (0.07, 0.25) L/min], RV˙O2peak [2.9 (1.8, 3.9) mL/kg/min], and PPO [9 (5, 14) W] with exercise, relative to controls (p < 0.001). Ninety-six studies (117 independent exercise interventions comprising 1,331 adults with SCI) were included in the secondary, pooled meta-analysis which demonstrated significant increases in AV˙O2peak [0.22 (0.17, 0.26) L/min], RV˙O2peak [2.8 (2.2, 3.3) mL/kg/min], and PPO [11 (9, 13) W] (p < 0.001) following exercise interventions. There were subgroup differences for RV˙O2peak based on exercise modality (p = 0.002) and intervention length (p = 0.01), but there were no differences for AV˙O2peak. There were subgroup differences (p ≤ 0.018) for PPO based on time since injury, neurological level of injury, exercise modality, and frequency. The meta-regression found that studies with a higher mean age of participants were associated with smaller changes in AV˙O2peak and RV˙O2peak (p < 0.10). GRADE indicated a moderate level of certainty in the estimated effect for RV˙O2peak, but low levels for AV˙O2peak and PPO. This review may be limited by the small number of RCTs, which prevented a subgroup analysis within this specific study design. Conclusions: Our primary meta-analysis confirms that performing exercise >2 weeks results in significant improvements to AV˙O2peak, RV˙O2peak, and PPO in individuals with SCI. The pooled meta-analysis subgroup comparisons identified that exercise interventions lasting up to 12 weeks yield the greatest change in RV˙O2peak. Upper-body aerobic exercise and resistance training also appear the most effective at improving RV˙O2peak and PPO. Furthermore, acutely injured, individuals with paraplegia, exercising for ≥3 sessions/week will likely experience the greatest change in PPO. Ageing seemingly diminishes the adaptive CRF responses to exercise training in individuals with SCI. Registration: PROSPERO: CRD42018104342