INTRODUCTION: Maxillary non-union following Le Fort I osteotomy is a rare complication in orthognathic surgery. We report our experience and analyze the possible etiological factors. PATIENTS AND METHOD: We retrospectively analyzed the files of 150 patients having undergone Le Fort I osteotomy, between 1996 and 2006, screening for maxillary non-union. We documented patients' features of: sex, age, medical history, indication, orthodontics, osteotomy type(s) and displacement(s), osteosynthesis, quality of dental occlusion, orthodontics, clinical signs supporting a diagnosis of non-union, radiologic examinations, peroperative observations and surgical revision, outcome after surgical revision. RESULTS: Maxillary non-union was observed in four patients (2.6%). They were female patients with a mean age of 34 years (30 to 38 years) without any specific medical history. All underwent bimaxillary osteotomy (including one Le Fort I segmented osteotomy) for a class III (retromaxillary and promaxillary) associated to vertical disharmony (open-bite and/or transverse discrepancy). Maxillary displacements were always associated to advancement (average: 6mm, 4 to 8mm) and a vertical displacement (upward in three cases, downward in one case). In every case the maxilla was osteosynthesized with titanium miniplates, with four holes and 1.5 screws. Postoperative dental occlusion was deficient in every case, requiring surgical revision of orthodontics between four to six weeks after osteotomy. Clinical signs suggesting non-union were in every case maxillary mobility associated in three cases to discomfort. The mean delay between osteotomy and the non-union was 15.5 months (six to 56 months). These signs appeared after infection in three cases (maxillary sinusitis in two cases, dacryocystitis in one case). Conventional radiological (panoramic and lateral cephalometric radiographs) suggested non-union (plate fracture) in two cases and 3D computed tomography was required in every case to make a diagnosis. Surgical revision was decided for all patients (osteotomy site curettage, bone graft and a more stable osteosynthesis using 2.0 screws). Bone healing was achieved in every case after revision surgery. DISCUSSION: The incidence of non-union in our series was superior to that of published data (0.33 to 0.8%). Non-union always appeared in instable occlusion settings, and in three cases because of postoperative infection, but the main risk factor seemed to be osteosynthesis instability. Maxillary mobility is the key sign to screen for. Osteosynthesis material rupture or loosening is present every time but not always visible on standard radiography. 3D CT scan always allows observing non-union. Revision surgery combines curettage, bone graft, and a stronger osteosynthesis.