The Effects of Balance Training on Balance Performance and Functional Outcome Measures Following Total Knee Arthroplasty: A Systematic Review and Meta-Analysis

Kenji Doma,Andrea Grant,Jodie Morris

DOI: https://doi.org/10.1007/s40279-018-0964-7

Abstract:Background: Several studies have examined the effects of balance training in elderly individuals following total knee arthroplasty (TKA), although findings appear to be equivocal. Objectives: This systematic review and meta-analysis examined the effects of balance training on walking capacity, balance-specific performance and other functional outcome measures in elderly individuals following TKA. Methods: Data sources: Pubmed, PEDro, Cinahl, SportDiscus, Scopus. Eligibility criteria: Data were aggregated following the population-intervention-comparison-outcome (PICO) principles. Eligibility criteria included: (1) randomised controlled trials; (2) studies with comparative groups; (3) training interventions were incorporated post-TKA; and (4) outcome measures included walking capacity, balance-specific performance measures, subjective measures of physical function and pain and knee range-of-motion. Participants: Elderly individuals (65 + years) who underwent total knee arthroplasty. Interventions: Balance interventions that consisted of balance exercises, which were compared to control interventions that did not involve balance exercises, or to a lesser extent. Participants also undertook usual physiotherapy care in conjunction with either the balance and/or control intervention. The intervention duration ranged from 4 to 32 weeks with outcome measures reported immediately following the intervention. Of these, four studies also reported follow-up measures ranging from 6 to 12 months post-interventions. Study appraisal: PEDro scale. Synthesis methods: Quantitative analysis was conducted by generating forest plots to report on standardised mean differences (SMD; i.e. effect size), test statistics for statistical significance (i.e. Z values) and inter-trial heterogeneity by inspecting I2. A meta-regression was also conducted to determine whether training duration predicted the magnitude of SMD. Results: Balance training exhibited significantly greater improvement in walking capacity (SMD = 0.57; Z = 6.30; P < 0.001; I2 = 35%), balance-specific performance measures (SMD = 1.19; Z = 7.33; P < 0.001; I2 = 0%) and subjective measures of physical function (SMD = 0.46; Z = 4.19; P < 0.001; I2 = 0%) compared to conventional training immediately post-intervention. However, there were no differences in subjective measures of pain (SMD = 0.77; Z = 1.63; P > 0.05; I2 = 95%) and knee range-of-motion (SMD = 0.05; Z = 0.39; P > 0.05; I2 = 1%) between interventions. At the 6- to 12-month follow-up period, improvement in combined measures of walking capacity and balance performance (SMD = 041; Z = 3.55; P < 0.001; I2 = 0%) were significantly greater for balance training compared to conventional training, although no differences were observed for subjective measures of physical function and pain (SMD = 0.26; Z = 2.09; P > 0.05; I2 = 0%). Finally, the training duration significantly predicted subjective measures of pain and physical function (r2 = 0.85; standardised β = 0.92; P < 0.001), although this was not observed for walking capacity and balance-specific performance measures (r2 = 0.02; standardised β = 0.13; P = 0.48). Limitations: A number of outcome measures indicated high inter-trial heterogeneity and only articles published in English were included. Conclusion: Balance training improved walking capacity, balance-specific performance and functional outcome measures for elderly individuals following TKA. These findings may improve clinical decision-making for appropriate post-TKA exercise prescription to minimise falls risks and optimise physical function.