Mandibular reconstruction with a plate, with or without a vascularised free (bone) flap, is commonly used to treat patients with a segmental mandibular defect. Common complications are loosening of the osteosynthesis screws, malposition, intraoral or extraoral exposure, or infection. To define prognostic factors for premature loss of such plates and improve future planning, we designed a retrospective study of all patients operated on between 2005 and 2011 for reconstruction of a mandibular segmental defect with a reconstruction plate with or without a free vascularised (bone) flap. Prognostic factors collected from medical records were the patient's age, sex, and American Society of Anesthesiologists (ASA) grade; treatment with radiotherapy; whether they had diabetes or smoked; the site of the mandibular defect; whether there was a dental occlusion; the number of screws used on each side, and the use of a free vascularised (bone) flap; and whether the diagnosis was of oral cancer, benign tumour, or trauma. One hundred patients were included, 79 with oral cancer, 19 with benign tumours, and 2 with trauma. In 20 patients the Martin 2.7 reconstruction plate failed. Diabetes and smoking were significant prognostic factors for premature loss of the reconstruction plate with a hazard ratio of 2.95 (95% CI 1.068–8.172), p value=0.04, for diabetes, and 2.42 (95% CI 1.006–5.824), p value=0.05, for smoking. Smokers and diabetic patients have a higher risk of failure after mandibular reconstruction with a 2.7 reconstruction plate.