Background
Stroke is a common condition that can lead to major disabilities and even death in adults. Stroke has an important impact on various aspects of human functioning, including limiting movement. One frequently observed deficit after stroke is the reduced functioning of the torso of the body. This impairment can, amongst other things, be characterised by reduced mobility, reduced sitting balance, late or reduced reactions to internal and external disturbances, reduced muscle strength and muscle activation patterns of the torso. Movements of the torso and sitting balance are both important for functional independence - that is, the ability to perform daily living tasks such as dressing, eating, and grooming without help. Functioning of the torso can largely forecast the level of recovery and independence after a stroke.
Trunk training aims to regain function of the torso. Trunk training can consist of different elements, such as: strength training of the abdominal and back muscles; exercises that focus on improving the mobility of the torso; or improving lateral or forward balance while sitting, aimed at improving sitting balance.
The torso is the core of the body; it provides a stable basis for control over and movements of the head and extremities. Training of the torso may have a positive effect not only on the functioning of the torso, but also an impact on other outcomes such as activities of daily living, standing balance, walking, and well-being.
Review question
We wanted to find out if training of the torso improves people's activities of daily living, trunk function, standing balance, well-being, and other outcomes, after they have had a stroke.
Search date
We searched nine databases and hand searched the bibliographies of relevant studies published up to 25 October 2021.
Study characteristics
We included 68 studies in which participants were randomly divided into two or more groups, with a total of 2585 participants. The studies compared training of the torso with other therapy or no therapy after a stroke.
Key results
We found that training of the torso may result in improvements of activities of daily living, torso function, standing balance, functional use of the affected arm and hand, movements of the affected lower limb, the ability to walk, and well-being.
Quality of the evidence
The quality of the evidence was very low to low.
There is evidence to suggest that trunk training as part of rehabilitation improves ADL, trunk function, standing balance, walking ability, upper and lower limb function, and quality of life in people after stroke. Core-stability, selective-, and unstable-trunk training were the trunk training approaches mostly applied in the included trials. When considering only trials with a low risk of bias, results were mostly confirmed, with very low to moderate certainty, depending on the outcome.
Previous systematic reviews and randomised controlled trials have investigated the effect of post-stroke trunk training. Findings suggest that trunk training improves trunk function and activity or the execution of a task or action by an individual. But it is unclear what effect trunk training has on daily life activities, quality of life, and other outcomes.
To assess the effectiveness of trunk training after stroke on activities of daily living (ADL), trunk function, arm-hand function or activity, standing balance, leg function, walking ability, and quality of life when comparing with both dose-matched as non-dose-matched control groups.
We searched the Cochrane Stroke Group Trials Register, CENTRAL, MEDLINE, Embase, and five other databases to 25 October 2021. We searched trial registries to identify additional relevant published, unpublished, and ongoing trials. We hand searched the bibliographies of included studies.
We selected randomised controlled trials comparing trunk training versus non-dose-matched or dose-matched control therapy including adults (18 years or older) with either ischaemic or haemorrhagic stroke. Outcome measures of trials included ADL, trunk function, arm-hand function or activity, standing balance, leg function, walking ability, and quality of life.
We used standard methodological procedures expected by Cochrane.
Two main analyses were carried out. The first analysis included trials where the therapy duration of control intervention was non-dose-matched with the therapy duration of the experimental group and the second analysis where there was comparison with a dose-matched control intervention (equal therapy duration in both the control as in the experimental group).
We included 68 trials with a total of 2585 participants.
In the analysis of the non-dose-matched groups (pooling of all trials with different training duration in the experimental as in the control intervention), we could see that trunk training had a positive effect on ADL (standardised mean difference (SMD) 0.96; 95% confidence interval (CI) 0.69 to 1.24; P < 0.001; 5 trials; 283 participants; very low-certainty evidence), trunk function (SMD 1.49, 95% CI 1.26 to 1.71; P < 0.001; 14 trials, 466 participants; very low-certainty evidence), arm-hand function (SMD 0.67, 95% CI 0.19 to 1.15; P = 0.006; 2 trials, 74 participants; low-certainty evidence), arm-hand activity (SMD 0.84, 95% CI 0.009 to 1.59; P = 0.03; 1 trial, 30 participants; very low-certainty evidence), standing balance (SMD 0.57, 95% CI 0.35 to 0.79; P < 0.001; 11 trials, 410 participants; very low-certainty evidence), leg function (SMD 1.10, 95% CI 0.57 to 1.63; P < 0.001; 1 trial, 64 participants; very low-certainty evidence), walking ability (SMD 0.73, 95% CI 0.52 to 0.94; P < 0.001; 11 trials, 383 participants; low-certainty evidence) and quality of life (SMD 0.50, 95% CI 0.11 to 0.89; P = 0.01; 2 trials, 108 participants; low-certainty evidence). Non-dose-matched trunk training led to no difference for the outcome serious adverse events (odds ratio: 7.94, 95% CI 0.16 to 400.89; 6 trials, 201 participants; very low-certainty evidence).
In the analysis of the dose-matched groups (pooling of all trials with equal training duration in the experimental as in the control intervention), we saw that trunk training had a positive effect on trunk function (SMD 1.03, 95% CI 0.91 to 1.16; P < 0.001; 36 trials, 1217 participants; very low-certainty evidence), standing balance (SMD 1.00, 95% CI 0.86 to 1.15; P < 0.001; 22 trials, 917 participants; very low-certainty evidence), leg function (SMD 1.57, 95% CI 1.28 to 1.87; P < 0.001; 4 trials, 254 participants; very low-certainty evidence), walking ability (SMD 0.69, 95% CI 0.51 to 0.87; P < 0.001; 19 trials, 535 participants; low-certainty evidence) and quality of life (SMD 0.70, 95% CI 0.29 to 1.11; P < 0.001; 2 trials, 111 participants; low-certainty evidence), but not for ADL (SMD 0.10; 95% confidence interval (CI) -0.17 to 0.37; P = 0.48; 9 trials; 229 participants; very low-certainty evidence), arm-hand function (SMD 0.76, 95% CI -0.18 to 1.70; P = 0.11; 1 trial, 19 participants; low-certainty evidence), arm-hand activity (SMD 0.17, 95% CI -0.21 to 0.56; P = 0.38; 3 trials, 112 participants; very low-certainty evidence). Trunk training also led to no difference for the outcome serious adverse events (odds ratio (OR): 7.39, 95% CI 0.15 to 372.38; 10 trials, 381 participants; very low-certainty evidence).
Time post stroke led to a significant subgroup difference for standing balance (P < 0.001) in non-dose-matched therapy. In non-dose-matched therapy, different trunk therapy approaches had a significant effect on ADL (< 0.001), trunk function (P < 0.001) and standing balance (< 0.001).
When participants received dose-matched therapy, analysis of subgroup differences showed that the trunk therapy approach had a significant effect on ADL (P = 0.001), trunk function (P < 0.001), arm-hand activity (P < 0.001), standing balance (P = 0.002), and leg function (P = 0.002). Also for dose-matched therapy, subgroup analysis for time post stroke resulted in a significant difference for the outcomes standing balance (P < 0.001), walking ability (P = 0.003) and leg function (P < 0.001), time post stroke significantly modified the effect of intervention.
Core-stability trunk (15 trials), selective-trunk (14 trials) and unstable-trunk (16 trials) training approaches were mostly applied in the included trials.