Nonunion of the tibia associated with bone loss, previous infection, obliteration of the intramedullary canal or located in the distal metaphysis poses a challenge to the surgeon and significant morbidity to patients. We retrospectively reviewed the records of 24 patients who were treated by central bone grafting and compared them to those of 20 who were treated with a traditional posterolateral graft. Central bone grafting entails a lateral approach, anterior to the fibula and interosseous membrane which is used to create a central space filled with cancellous iliac crest autograft. Upon consolidation, a tibiofibular synostosis is formed that is strong enough for weight-bearing. This procedure has advantages over other methods of treatment for selected nonunions.

Of the 24 patients with central bone grafting, 23 went on to radiographic and clinical union without further intervention. All healed within a mean of 20 weeks (10 to 48). No further bone grafts were required, and few complications were encountered. These results were comparable to those of the 20 patients who underwent posterolateral bone grafting who united at a mean of 31.3 weeks (16 to 60) but one of whom required below-knee amputation for intractable sepsis.

Central bone grafting is a safe and effective treatment for difficult nonunions of the tibia.

We reviewed 78 femoral and tibial nonunions treated between January 1992 and December 2003. Of these, we classified 41 in 40 patients as complex cases because of infection (22), bone loss (6) or failed previous surgery (13). The complex cases were all treated with Ilizarov frames. At a mean time of 14.1 months (4 to 38), 39 had healed successfully.

Using the Association for the Study and Application of the Methods of Ilizarov scoring system we obtained 17 excellent, 14 good, four fair and six poor bone results. The functional results were excellent in 14 patients, good in 14, fair in two and poor in two. A total of six patients were lost to follow-up and two had amputations so were not evaluated for final functional assessment. All but two patients were very satisfied with the results. The average cost of treatment to the treating hospital was approximately £30 000 per patient.

We suggest that early referral to a tertiary centre could reduce the morbidity and prolonged time off work for these patients. The results justify the expense, but the National Health Service needs to make financial provision for the reconstruction of this type of complex nonunion.

The literature on fracture repair has been reviewed. The traditional concepts of delayed and nonunion have been examined in terms of the phased and balanced anabolic and catabolic responses in bone repair. The role of medical manipulation of these inter-related responses in the fracture healing have been considered.

The aim of this randomised, controlled in vivo study in an ovine model was to investigate the effect of cylic pneumatic pressure on fracture healing. We performed a transverse osteotomy of the right radius in 37 sheep. They were randomised to a control group or a treatment group where they received cyclic loading of the osteotomy by the application of a pressure cuff around the muscles of the proximal forelimb. Sheep from both groups were killed at four or six weeks. Radiography, ultrasonography, biomechanical testing and histomorphometry were used to assess the differences between the groups. The area of periosteal callus, peak torsional strength, fracture stiffness, energy absorbed over the first 10° of torsion and histomorphometric analysis all showed that the osteotomies treated with the cyclic pneumatic pressure at four weeks were not significantly different from the control osteotomies at six weeks.

We have managed 21 patients with a fracture of the tibia complicated by bone and soft-tissue loss as a result of an open fracture in 10, or following debridement of an infected nonunion in 11, by resection of all the devitalised tissues, acute limb shortening to close the defect, application of an external fixator and metaphyseal osteotomy for re-lengthening. The mean bone loss was 4.7 cm (3 to 11). The mean age of the patients was 28.8 years (12 to 54) and the mean follow-up was 34.8 months (24 to 75).

All the fractures united with a well-aligned limb. The mean duration of treatment for the ten grade-III A+B open fractures (according to the Gustilo-Anderson classification) was 5.7 months (4.5 to 8) and for the nonunions, 7.6 months (5.5 to 12.5). Complications included one refracture, one transient palsy of the peroneal nerve and one equinus contracture of 10°.