Nine studies were included (n=550 patients, range 14 to 71). Randomisation was reported as adequate in only two studies. Allocation concealment was reported as: adequate in two studies; inadequate in two and; unclear in five studies. Power calculation, blinding of patients and intention-to-treat analysis were not done in all the studies.
Functional scores: At six weeks, functional scores (Olerud and Molander or Mazur) were significantly higher in the early motion group compared with the cast group (p=0.001; four studies). There were no statistically significant differences in scores between early motion and cast groups at one year (seven studies), 1.5 years (one study) and two years (two studies).
Range of motion: Range of motion was significantly higher in the early motion groups at nine to 12 weeks (dorsiflexion p<0.001, plantarflexion p<0.00001; three studies). No difference in range of motion was found at one year (two studies).
Return to work: Time to return to work was significantly shorter in the early motion group (p<0.00001, three studies).
Wound infections: Rate of wound infection was significantly higher in the early motion group (p=0.002; nine studies). No difference in rate of deep vein thrombosis (DVT) was found between early motion and cast groups (three studies).