Atrophic non-unions are usually attributed to impaired blood supply but the events that lead to atrophic non-union remain poorly understood. Recent studies^1,2 have shown that vascularity is not reduced in established non-unions but these studies have not examined vascularity at an early stage. The aims of this study were to: 1) develop and validate a clinically relevant small animal model of atrophic non-union and 2) test the hypothesis that the vessel density of atrophic non-unions reaches that of normal healing bones but at a later time point.

Twenty eight adult female Wistar rats underwent application of a novel circular frame external fixator to the right tibia under general anaesthesia. The fixator construct was standardised, with eight needles that were drilled through the skin into the proximal and distal metaphyses of the tibia. An osteotomy was performed with a 1mm burr under irrigation. The periosteum was removed on 14 of the 28 animals using a scalpel and the intramedullary canal was curetted. Both insults were performed proximally and distally for a distance equivalent to 1 diameter of the tibia. A 1mm gap was introduced at the osteotomy site and the wound was closed. Once the animal had recovered it was allowed unrestricted weight bearing. Anteroposterior X rays were performed every 2 weeks. Animals were killed at 1, 3, 8 and 16 weeks. Callus areas were measured from X rays using an image analysis system. The average callus area was calculated for each rat every 2 weeks as an indicator of callus production. Specimens were fixed, decalcified, embedded in paraffin wax and 6 ìm sections were stained with H& E. Vascularity was assessed immunohistochemically with monoclonal antibody against smooth muscle actin. The total number of blood vessels in the interfragmentary gap was counted.

At 8 and 16 weeks post-osteotomy all animals where stripping and curetting had been performed went on to an atrophic non-union. All animals where this was not performed went on to unite successfully. Histological observations support these radiological findings. Significantly less callus formed in the non-unions than in those that united. There were significantly fewer vessels in the non-unions at week 1 compared to the controls but, by 8 weeks the blood vessel density in the established atrophic non-unions had reached the same level as the vessel density during normal healing.

An atrophic non-union model that closely resembles the clinical situation has been developed and validated in rats. The results support the hypothesis that the number of vessels in atrophic non-unions reaches the same level as in those that unite but at a later time point. It is concluded that diminished vessel density within the first 3 weeks may prevent fractures from uniting.