A retrospective cohort study was conducted to compare fusion rates achieved by H-grafts, posterior vertebral grafts, and no graft for the surgical treatment of thoracolumbar fractures. Ninety-two patients were included in this study. The patients fell into 1 of 3 groups: those who received H-grafts (n = 36), posterior vertebral grafts (n = 30), and no graft (n = 26). Mean follow-up was 38 months (range, 24-51 months). All operations were performed by a single senior surgeon. All patients underwent operative treatment with posterior reduction and instrumentation. Radiographic parameters, estimated blood loss, operative time, and length of hospital stay were compared among patients in the 3 graft groups. Differences were assessed using unpaired t tests. P values < .05 were considered significant. We found no significant difference among groups in age, fracture location, or type of fracture. Patients who received H-grafts or posterior vertebral grafts achieved solid fusion, but spontaneous fusion occurred in only 2 patients who received no bone graft. Most patients with neurological deficits showed significant neurological improvement. Operative time and estimated blood loss were significantly lower in the no-graft group than in the H-graft and posterior vertebral graft groups (P < .05). Mean loss of correction, operative time, and estimated blood loss were lower for patients who received H-grafts than for those who received posterior vertebral grafts (P < .05). The use of an atlas fixation system in combination with a posterior H-graft for the treatment of thoracolumbar fracture is a stable and reliable method that effectively prevents inner fixation failure and reduces bone loss and anisotropy.