Eleven RCTs (n=1,521) were included. Eight studies (n=1,166) compared exercise therapy with control and 3 studies (n=355) compared exercise therapy plus additional booster sessions with control.
Five studies were considered to be of a high methodological quality, four of these evaluated exercise therapy without additional booster sessions. None of the included studies met criteria for blinding of care providers of patients, as this is not possible in studies of exercise therapy.
Pain.
Exercise therapy versus control (8 studies, 4 high quality): 3 high-quality studies and 2 low-quality studies reported small to moderate improvements in pain post-treatment, but these findings were no longer significant at long-term follow-up. One high-quality study and 2 low-quality studies found no significant beneficial effects either post-treatment or at long-term follow-up.
Exercise therapy plus booster sessions versus control (3 studies, 1 high quality): all studies reported significant treatment effects at long-term follow-up.
Self-reported physical functioning.
Exercise therapy versus control (8 studies, 4 high quality): 3 high-quality studies and 2 low-quality studies reported no significant effects either post-treatment or at long-term follow-up. One low-quality study reported no effect post-treatment but a small positive effect at long-term follow-up; another low-quality study reported beneficial effects post-treatment which were no longer significant at long-term follow-up. One high-quality study reported significant moderate beneficial effects both after treatment and at long-term follow-up.
Exercise therapy plus booster sessions versus control (2 low-quality studies): both studies reported a large post-treatment effect that increased at long-term follow-up.
Observed physical function (7 studies).
Exercise therapy versus control (4 studies, 3 high quality): 2 high-quality studies and 1 low-quality study reported significant post-treatment effects which remained in 2 studies but was no longer significant in one of the high-quality studies. The other high-quality study reported no significant treatment effects.
Exercise therapy plus booster sessions versus control (3 studies, 1 high quality): all studies reported small to large significant effects post-treatment and moderate to large effects at long-term follow-up.
Global assessment of effectiveness.
Exercise therapy versus control (1 high-quality study): the study reported a moderate significant post-treatment effect and a small significant treatment effect at long-term follow-up.
Sensitivity analysis based on using a threshold of at least 4 or at least 8 methodological criteria to define a high-quality study did not significantly alter the results.