Twenty-nine RCTs (30 independent comparisons, 2,063 patients) were included in the meta-analysis. None of the studies were cluster RCTs. Seventeen studies reported an adequate method of random sequence generation. Only nine studies reported adequate methods of allocation concealment. Treatment allocation concealment was not adequate in one trial and not reported in 19 trials. Ten trials had blinding of outcome assessors. Thirteen trials used an intention-to-treat analysis. Three trials had baseline imbalance in terms of disease severity (patients in the intervention group had milder disease than those in the control group). Mean follow-up at post-treatment was 10.5 weeks (range four to 52 weeks).
Compared with controls, psychological and/or lifestyle interventions were associated with a significant improvement in depression (SMD -0.28, 95% CI -0.41 to -0.14; Ι²=47.5%; 29 RCTs) and anxiety (SMD -0.24, 95% CI -0.39 to -0.09; Ι²= 56.4%; 26 RCTs).
Based on the results of subgroup analyses, multicomponent exercise training was the only intervention subgroup associated with a significant improvement in depression (SMD -0.47, 95% CI -0.66 to -0.28; Ι²=43.9%; 14 RCTs) and in anxiety (SMD -0.45, 95% CI -0.71 to -0.18; Ι²=63.3%; 11 RCTs). No other intervention subgroups (including cognitive and behavioural therapy) were associated with significant treatment effects. When compared with trials in which severity of depression and anxiety was unknown at baseline, treatment effects did not appear to be larger in trials that included confirmed depressed and/or anxious samples or above threshold samples.
Sensitivity analyses did not significantly alter the results. There was no evidence for publication bias.