Exercise for people with a fragility fracture of the pelvis or lower limb: a systematic review of interventions evaluated in clinical trials and reporting quality

Background To aid design of exercise trials for people with pelvic and lower limb fragility fractures a systematic review was conducted to identify what types of exercise interventions and mobility outcomes have been assessed, investigate intervention reporting quality, and evaluate risk of bias in published trials. Methods Systematic searches of electronic databases (CENTRAL, MEDLINE, EMBASE, PEDro) 1996–2019 were conducted to identify randomised controlled trials of exercise for pelvic or lower limb fragility fractures. Two reviewers independently screened titles and abstracts. One reviewer extracted data, a second verified. Two reviewers independently assessed risk of bias. Intervention reporting quality was based on TIDieR, assessed by one reviewer and verified by a second. Narrative synthesis was undertaken. Registration: PROSPERO CRD42017060905. Results Searches identified 37 trials including 3564 participants, median sample size 81 (IQR 48–124), participants aged 81 years (IQR 79–82) and 76% (2536/3356) female. All trials focussed on people with hip fracture except one on ankle fracture. Exercise types focussed on resistance exercise in 14 trials, weight bearing exercise in 5 trials, 13 varied dose of sessions with health professionals, and 2 trials each focussed on treadmill gait training, timing of weight bearing or aerobic exercise. 30/37 (81%) of trials reported adequate sequence generation, 25/37 (68%) sufficient allocation concealment. 10/37 (27%) trials lacked outcome assessor blinding. Of 65 exercise interventions, reporting was clear for 33 (51%) in terms of when started, 61 (94%) for where delivered, 49 (75%) for who delivered, 47 (72%) for group or individual, 29 (45%) for duration, 46 (71%) for session frequency, 8 (12%) for full prescription details to enable the exercises to be reproduced, 32 (49%) clearly reported tailoring or modification, and 23 (35%) reported exercise adherence. Subjectively assessed mobility was assessed in 22/37 (59%) studies and 29/37 (78%) used an objective measure. Conclusions All trials focussed on hip fracture, apart from one ankle fracture trial. Research into pelvic and other lower limb fragility fractures is indicated. A range of exercise types were investigated but to date deficiencies in intervention reporting hamper reproducibility. Adoption of TIDieR and CERT guidelines should improve intervention reporting as use increases. Trials would be improved by consistent blinded outcome assessor use and with consensus on which mobility outcomes should be assessed.