The aim of the present study was to test whether there is a significant difference in the clinical outcomes between surgical and non-surgical treatment of mandibular condylar fractures. An electronic search was undertaken in February 2014. Eligibility criteria included clinical human studies, either randomized or not. The search strategy resulted in 36 publications. The estimates of an intervention were expressed as the risk ratio (RR) and mean difference (MD) in millimetres. A statistically significant effect was observed for the outcome of post-treatment malocclusion (RR 0.46, P<0.00001), lateral deviation during maximum inter-incisal opening (RR 0.56, P=0.0001, dichotomous; MD -0.75, P=0.002, continuous), protrusion (MD 0.68, P=0.01), and laterotrusion (MD 0.53, P=0.03) favouring surgical treatment, and for infection (RR 3.43, P=0.03) favouring non-surgical treatment. There was no statistically significant effect on temporomandibular joint pain (RR 0.81, P=0.46) or noise (RR 1.44, P=0.24), or maximum inter-incisal opening (MD 2.24, P=0.14). The test for overall effect showed that the difference between the procedures significantly affected the incidence of post-treatment complications, favouring surgical treatment, when all dichotomous and continuous outcomes were analysed (RR 0.70, P=0.006 and MD 1.17, P=0.0006, respectively).