Surgery
Procaine Penicillin (22 000 IU/kg IM), gentamicin sulfate (6.6 mg/kg IV) and flunixin meglumine (1.1mg/kg IV) were administered preoperatively. The horse was sedated with detomidine hydrochloride (0.1mg/kg IV) and morphine (0.1mg/kg IV). General anaesthesia was induced with diazepam (0.02mg/kg IV) associated with ketamine (2.2 mg/kg) and maintained with isoflurane. The horse was positioned in dorsal recumbency with the head and neck extended. A custom-made wedge was placed at the level of C3-C4 to stabilize the neck in a strict sagittal position with good alignment of the cranial cervical vertebrae. The head was fixed to a separate small table that could be adjusted independently from the surgery table to create additional extension at the level of the atlantoaxial joint(Figure 2) . This enabled to improve the alignment of the two main fragments and reduce the fracture. Radiography was used to determine the amount of head extension needed.
After routine aseptic preparation and appropriate draping, a 20cm-skin incision was made, centred over the axis. The cutaneous colli muscle was bluntly dissected, and the sternohyoid and sternothyroid muscles were separated at the midline to expose the trachea. Dissection was then continued dorsally along the right side of the trachea until the longus colli muscles were reached. The trachea was retracted towards the left and two self-retaining retractors were placed to gain access to the ventral aspect of the axis: A Ricard-Begoin retractor was placed at the cranial end of the wound and a modified Beckman retractor at the caudal end. The atlantoaxial and C2C3 joints were localized using radiography. The aponeurosis of the longus colli muscles was incised from the cranial to the caudal border of C2. Mayo scissors and a periosteal elevator were used to divide the longus colli muscles. An Inge retractor was placed in between the separated longus colli muscles to expose the ventral surface of C2. The fracture line through the cranial mid-body of the axis was palpated and visualized. Palpation and radiography confirmed correct alignment of the axis and fracture reduction. The ventral spinous process of the axis body was slightly flattened using a curved osteoma. A 4.5mm five-hole T-LCP plate was contoured to fit the ventral aspect of the axis. The first standard combi-hole was used as a landmark for plate positioning and placed just caudal to the fracture line. The plate was fixed to the cranial fragment using three locking head screws in the stacked combi-holes. The screw lengths had been determined preoperatively but even so, radiographs were taken strategically during the procedure to make sure that the spinal canal was not entered during drilling or screw placement. The caudal fragment was grasped with bone forceps while a fourth locking head screw was placed through the second standard combi-hole. The third, fourth and fifth standard combi-holes were also filled with locking head screws. Finally, a 4.5mm cortical screw was placed through the first standard combi-hole, instead of a locking head screw, to avoid crossing the fracture line (Figure 3) . The wound was flushed abundantly with sterile saline solution. The longus colli muscles and cutaneous colli muscles were apposed with a simple continuous 2-0 polyglactin 910 suture. The skin was closed with interrupted horizontal mattress 0 polydioxanone sutures and staples. The wound was protected with a stent bandage and an antimicrobial adhesive drape. Recovery was assisted with a single tail rope system and was uneventful. A head and neck bandage was placed over the stent and the horse was put back in the sling after recovery.