SummaryIntroduction. The nonunion rate has been reported between 0.1% and 15%. There are also several predisposing factors for the onset of complications: general factors connected with the patient and specific factors related to the fracture site. The purpose of our study is to review the etiology of nonunion of the clavicle in its atrophic form and investigate the outcomes of the revision treatment in a single step. Materials and methods. Retrospective study on 71 patients suffering from nonunions due to the following treatments: conservative in 13 patients; plate fixation in 12; closed reduction and fixation with K-wire in 24; open reduction and fixation with K-wire. All patients were operated on in beach chair position and classic approach to the clavicle by incising the previous surgical scar. The clinical and radiographic criteria for evaluating the outcomes were: the Short Form (12) Health Survey (SF-12), the Constant Shoulder Score (CSS) and the Disability Disabilities of the Arm, Shoulder and Score (DASH) and radiographic Union Score (RUS) for bone healing. The evaluation endpoint was set at 12 months. Results. Blood and culture tests showed 22 infected nonunions and 49 atrophic or oligoatrophic. In only 10 cases, before surgery, the inflammatory markers were positive. The isolated microorganisms were resistant to common antibiotics. In 70 out of 71 cases, plates and screws on the upper side and fibula allogenic splints at the bottom, associated with cancellous bone grafts taken from the patients' iliac crests, were implanted. In one case, however, it was decided to implant the plate on the front edge of the clavicle and the fibula allogeneic splint on the posterior margin, also associated with a cancellous bone graft taken from the patient's iliac crest. The radiographic bone healing was observed in 107.8 (range 82-160) days for the aseptic nonunions, while in 118.4 (range 82-203) days for the septic ones. The non-healing case was a serious failure that led to asubtotal excision of the clavicle. Conclusions. The importance of classification and study of nonunions are essential to achieve positive outcomes. The guiding principle of our work is that aseptic nonunions heal in the operating room, while infected nonunions can be challenged and defeated on the operating table. Restoring the correct length of the clavicle interconnection between the sternum and the shoulder cingulum is indispensable to avoid functional deficits of the upper limb. The fibula splint and the tricorticale bone graft have both mechanical and strong biological values to quickly heal the nonunion.The return to pre-injury quality of life has to be our main goal.