Les pseudarthroses de l'humérus sont particulières car elles siègent sur un membre en décharge où les forces s'exercent essentiellement en rotation.

Trente pseudarthroses de la diaphyse humérale traitées par plaque ont été étudiées avec un recul de 34 mois. Le traitement a consisté en la mise en place d'une plaque vissée avec autogreffe d'os spongieux après repérage du nerf radial.

Tous les patients ont consolidé à un délai moyen de 16 semaines. Les mobilités moyennes du coude étaient de 0/10/130 et de 135° pour l'élévation de l'épaule. Les principales complications étaient une paralysie radiale transitoire spontanément résolutive chez deux patients ainsi qu'une infection.

Plusieurs traitements de la pseudarthrose des fractures de la diaphyse humérale existent : plaque vissée, clou centromédullaire verrouillé ou fixateur externe de type Ilizarov. Ils ont tous leurs avantages et leurs complications propres : paralysie radiale et infection pour la plaque ; taux de consolidation insuffisant et lésion de la coiffe des rotateurs pour le clou ; suppuration sur fiches, paralysies nerveuses et maintien prolongé du matériel pour le fixateur.

Notre préférence va au traitement par plaque qui assure une consolidation rapide dans 95 % des cas au prix de peu de complications dans notre série.

Although aseptic non-union of humeral shaft fractures is generally considered to be an exceptional complication, rates in the literature have varied from 1 to 10%. Factors favoring non-union are often related to technical error or inappropriate therapeutic indication. Several types of treatment (orthopedic, locked centromedullary nailing, ascending pinning, plating, external fixation) can be proposed for humeral shaft fractures. In all cases, a precise technique and proper indication are essential for success. We reviewed the cases of 30 patients who underwent surgery for aseptic non-union of humeral shaft fractures between 1995 and 2000.

Mean patient age was 43 years. Oblique and transverse fractures of the middle third of the shaft predominated. All types of treatment had been used but most of the patients had had ascending pinning. All patients were treated with plate fixatin and a cancellous bone graft after identifying the radial nerve.

Bone healing was achieved in all patients. Mean delay to healing was 16 weeks with good motion of the shoulder (mean elevation 136°) and elbow (mean motion 10-130°). Transient radial paresia recovered spontaneously in two patients. There was one infection. Only two patients complained of a painful arm that was not bothersome for daily activities and did not require long-term analgesia. There were no cases of radial nerve injury. Elbow function improved in 16 patients, was unchanged in 11, and showed limited extension in 3. Shoulder function improved in 15 patients and was unchanged in 15.

Plate fixation is widely described in the literature for the treatment of humeral non-union. The main complications of this treatment are radial palsy and infection, reported in 5% of the series. Several recent reports have therefore advocated locked nailing or external fixation with an Ilizarov device but these techniques are difficult to use and have their own risks of complications. It is difficult to block rotation and the nail may injure the rotator cuff. Pin tract infection, nerve injury, and prolonged external fixation are other disadvantages. We therefore recommend screw plate fixation with a cancellous bone graft. Our good results combined with the very low rate of complications argue in favor of this therapeutic option.