Pain catastrophising and kinesiophobia mediate pain and physical function improvements with Pilates exercise in chronic low back pain::a mediation analysis of a randomised controlled trial

Abstract

QUESTION: How much are the reductions in pain intensity and improvements in physical function from Pilates exercise mediated by changes in pain catastrophising and kinesiophobia? DESIGN: This was a secondary causal mediation analysis of a four-arm randomised controlled trial testing Pilates exercise dosage (once, twice or thrice per week) against a booklet control. PARTICIPANTS: Two hundred and fifty-five people with chronic low back pain. DATA ANALYSIS: All analyses were conducted in R software (version 4.1.2) following a preregistered analysis plan. A directed acyclic graph was constructed to identify potential pre-treatment mediator-outcome confounders. For each mediator model, we estimated the intervention-mediator effect, the mediator-outcome effect, the total natural indirect effect (TNIE), the pure natural direct effect (PNDE), and the total effect (TE). RESULTS: Pain catastrophising mediated the effect of Pilates exercise compared with control on the outcomes pain intensity (TNIE MD -0.21, 95% CI -0.47 to -0.03) and physical function (TNIE MD -0.64, 95% CI -1.20 to -0.18). Kinesiophobia mediated the effect of Pilates exercise compared with control on the outcomes pain intensity (TNIE MD -0.31, 95% CI -0.68 to -0.02) and physical function (TNIE MD -1.06, 95% CI -1.70 to -0.49). The proportion mediated by each mediator was moderate (21 to 55%). CONCLUSION: Reductions in pain catastrophising and kinesiophobia partially mediated the pathway to improved pain intensity and physical function when using Pilates exercise for chronic low back pain. These psychological components may be important treatment targets for clinicians and researchers to consider when prescribing exercise for chronic low back pain.