▪ Rationale

Although neither arthrosis nor pain are by any means the inevitable long-term result of untreated scaphoid nonunion,1 pain and arthrosis occur frequently enough for most to recommend an attempt to achieve union. This became widely feasible after Russe published his method of bone grafting in English.2 A union rate of around 90% with this method was reported by several authors2 ^– 5 but it did not allow a correction of the deformity at the nonunion site.4 For this reason iliac crest wedge bone grafts were introduced. This type of graft often requires some type of adjuvant fixation. The most popular method during recent years has been with a double-threaded headless screw.6 ^– 11 The accurate shaping of the bone block and the exact placement of the screw make these operations difficult except in experienced hands. We prefer to treat scaphoid nonunions with a method described by Stark and coworkers,12 using temporary Kirschner-wire fixation and cancellous bone grafting. This is a far less demanding operation, but it still allows correction of the humpback deformity.

▪ Indications

Our main indications for surgery include both symptomatic and nonsymptomatic nonunions. Initial screw fixation of a scaphoid fracture is very rarely done in Norway. We have also performed this as a secondary procedure for persistent nonunions after previous failed surgery. We do not routinely perform preoperative magnetic resonance (MR) studies.

▪ Contraindications

An absolute contraindication to this procedure is collapse of the proximal pole.

▪ Surgical Technique

The technique is performed as described by Stark and coworkers.12 Under tourniquet control a 4 to 6 cm incision is placed over the distal flexor carpi radialis tendon. Just proximal to the distal wrist crease it is angled 45 degrees radially for 2 cm. The flexor carpi radialis tendon is drawn aside and the incision is deepened down to the wrist capsule, which is divided. The wrist joint and the scaphoid are visualized by extending the wrist over a bolster, which brings the nonunion into view ( Fig. 24.1 ). A small osteotome is used to make a 3 by 3 mm cortical window in the larger of the two fragments, for debridement of all fibrous tissue and sclerotic bone ( Fig. 24.2 ). I prefer to use hand tools, such as a curette and chisel, to minimize the danger of thermal injury to the bone. The sclerotic bone can be very hard; hence it is often useful to weaken it with multiple perforations using a sharp awl ( Fig. 24.3 ). All sclerotic bone is debrided, which sometimes entails removing all of the cancellous bone from the proximal fragment so that only the cortical shell remains. If there is a flexion deformity of the bone, this is corrected by forcefully extending the wrist, then passing two 1 mm Kirschner wires from the distal fragment to the proximal to maintain the reduction ( Fig. 24.4 ). The K-wire position can be checked through the cortical window.

If the proximal fragment is very small or if it has been denuded of bone, I occasionally advance the pins into either the wrist joint or the radius. In this case plaster immobilization is maintained until the pins are removed. Cancellous bone chips, harvested from the volar surface of the distal radius through the same incision, are then packed tightly into the defect ( Figs. 24.5 and 24.6 ). I do not follow Stark’s technique of replacing the bone from the cortical window. The wrist capsule is closed with multiple thick, resorbable sutures, and the distal tips of the Kirschner wires are bent and buried under the skin. The skin is closed without a drain.

▪ Postoperative Protocol

Patients are placed in a cast extending from the interphalan-geal joint of the thumb to the upper forearm. Our practice has been to change the cast every 4 weeks to ensure that it does not become too loose because of muscle atrophy. Cast immobilization is discontinued when radiographs show crossing trabeculae across the nonunion site.

At this point the Kirschner wires are removed under local anesthesia with the aid of an image intensifier, and the patient is sent for physiotherapy. The average time is 14 to 16 weeks.

It is often difficult to definitely determine whether the nonunion has healed radiographically. In many cases we continue the cast immobilization for an additional 4 weeks. In other instances we remove the Kirschner wires, apply a new cast, and obtain a computed tomographic (CT) scan. It is also an option to remove the cast and start physiotherapy with the wires still in place. The decision is usually reached after discussion with the patient.