Purpose. To investigate and compare the probability of recurrence of keratocystic odontogenic tumors (KCOTs) for different variables and treatment protocols. Materials and Methods. An electronic search was undertaken in April 2016 that included clinical series of KCOTs reporting recurrences. Untransformed proportions and meta-analyses were performed to estimate the probability/risk of recurrence, according to several variables. Results. A total of 94 publications were included (6,427 KCOTs, 1,464 recurrences). Probability of recurrence: all lesions, 21.1%; nevoid basal cell carcinoma syndrome, 35.4%; males, 20.3%; females, 19.3%; maxilla, 15.3%; mandible, 21.5%; unilocular, 14.7%; multilocular, 24.4%; marsupialization/decompression, 28.7%; decompression+enucleation±additional therapy, 18.6%; enucleation/curettage, 22.5%; enucleation+peripheral ostectomy, 18.6%; enucleation+Carnoy’s solution, 5.3%; enucleation+cryotherapy, 20.9%; marginal/segmental resection, 2.2%. The recurrence was not statistically significantly affected by lesion location (maxilla vs. mandible, risk ratio [RR] 0.92, P=0.32) or patient’s sex (male vs. female, RR 0.94,P=0.44), but by locularity (unilocular vs. multilocular, RR 0.67, P=0.007). Recurrence risk for surgical managements: marsupialization vs. enucleation (RR 1.65,P=0.0006), marsupialization vs. resection (RR 3.17, P=0.009), enucleation alone vs. enucleation+peripheral ostectomy (RR 1.66, P=0.05), enucleation alone vs. enucleation+Carnoy’s solution (RR 1.94, P=0.03), enucleation alone vs. enucleation+cryotherapy (RR 0.88, P=0.56). Conclusions. KCOTs have a considerable rate of recurrence, which varies significantly according to some clinical, radiographic, and histopathological features, as well as surgical management.