An alternative classification ¹ which is also helpful divides bone deficiency into: (1) deficient humeral or ulnar bone length; (2) compromised bone quality with ballooning of an intact cortex; and (3) periprosthetic fractures.

Presentation, investigations and treatment options

Presentation

The majority of patients requiring this type of revision surgery are elderly, have frequently undergone previous revision procedures and have comorbidities, particularly inflammatory arthritis. They typically present with increasing elbow pain and will often also complain of instability. The instability may result in difficulty controlling the forearm when the upper limb is being used but in severe cases will present as a flail elbow. The patient may at the time of presentation already be using a sling to support the arm. In addition if the elbow has been infected the patient may also complain of current or previous episodes of discharge from the joint.

Clinical examination must involve a full medical assessment in order to identify coexistent comorbidities together with a careful examination of the elbow. The elbow should be inspected for scars and sinuses and should be palpated for areas of tenderness and joint swelling. It is also important to try to identify the course of the ulnar nerve, as this may have been transposed at an earlier operation. Any evidence of sensory or motor ulnar nerve dysfunction should also be noted. The active and passive range of movement must be recorded and note made of any instability.

Clinical Pearl 45.1

It is important to perform a detailed clinical assessment of patients presenting with bone loss secondary to a failed total elbow arthroplasty because most patients are elderly and often have multiple comorbidities.

Investigations

Plain anteroposterior and lateral radiographs of the elbow are essential. These must include all of the prosthesis and must also show the extent of the cement in the proximal humerus and distal ulna. Routine screening blood tests – haemoglobin and full blood count, erythrocyte sedimentation rate, C-reactive protein, urea and electrolytes, and liver function tests – should also be performed. In combination with these investigations if there is any suggestion of previous or current joint infection it is my practice to perform a joint aspiration with the aspirate being sent for microscopy, culture and sensitivity. This investigation is undertaken in the radiology department under fluoroscopy control in order to be certain that a representative sample is obtained from the joint.

Treatments options

The aim of treatment is to intervene at an appropriately early stage so as to minimize bone loss and restore function. The options for achieving this are dependent on the severity of the bone loss and range from revision using a non-customized prosthesis for minimal bone loss, impaction bone grafting, customized prostheses for more extensive defects to autograft and allograft.

Although a one-stage revision is the preferred option for both the patient and the surgeon, this should not be undertaken when there is evidence of previous infection, a positive joint aspiration or if there is clinical suspicion of infection at the time of surgery. In these circumstances a two-stage revision should be performed. The first stage involves removal of the implant, all of the cement and the taking of multiple tissue samples for culture. In addition, once the debridement has been completed antibiotic beads are inserted into the joint cavity. These are made at the time of the first-stage procedure using gentamicin impregnated cement to which additional antibiotics are added. If a joint aspirate has been performed and an organism identified, the appropriate antibiotic to which that organism is sensitive is mixed with the gentamicin cement at the time the beads are made. If no organisms have been identified but there has been a previous infection or infection is suspected at the time of the revision 1 g of gentamicin together with 1 g of vancomycin are added to one mix of Palacos cement in order to produce high concentration wide-spectrum local cover. Postoperatively the arm is supported in a back slab until clinically and on serial blood tests (white cell count, erythrocyte sedimentation rate and C-reactive protein) the infection appears eradicated. Usually this will take at least 3 months but sometimes up to 12 months before I am confident that there is no residual infection. Occasionally if clinical or biochemical assessment suggests persisting infection, I will advise a repeat first-stage revision. When the infection is controlled the second stage procedure is performed with removal of the antibiotic beads and reconstruction of the joint.

Clinical Pearl 45.2

Although a one-stage revision is the preferred option for both the surgeon and the patient, this should not be undertaken if there is any evidence to suggest the presence of infection.

Summary Box 45.1 Treatment options in the presence of bone deficiency

Minimal bone loss

• Revision with a non-customized prosthesis

Advanced bone loss

• Impaction bone grafting

• Customized prosthesis

• Autograft

• Allograft

Surgical techniques and rehabilitation

Although all revisions undertaken in the presence of bone deficiency will present the surgeon with individual challenges, certain principles can be applied which will assist with the surgical procedure.

Patient position: The surgical procedure can be performed with the patient’s arm across the chest but I prefer the lateral position with the arm supported on a bolster and the forearm hanging vertically. This enables exposure of the whole of the humerus and if necessary allows the creation of bone windows to retrieve cement from the proximal humerus. It also provides excellent visualization of the ulnar canal.

Tourniquet: The use of a tourniquet is helpful unless it is anticipated that the skin incision will need to extend too far proximally. It should be well padded to reduce the risk of a radial nerve neuropraxia and should be deflated after a maximum period of 1.5 hours.

Incision and surgical approach: The previous posterior skin incision should be utilized, although it is almost always necessary to extend it particularly if there is a periprosthetic fracture or loss of cortical bone. The ulnar nerve should be identified decompressed and protected. The radial nerve should also be found and protected if it is anticipated that the surgical approach will need to extend to or go proximal to the spiral groove. Several surgical approaches to the elbow have been described ²^–⁴ and it is a matter of preference as to which is used. My own preference⁵ is a modification of that described by Steiger et al.⁶

Implant and cement removal: Having exposed the elbow arthroplasty the implant should be removed together with all of the cement and any membrane lining the intramedullary cavity. Although when there is significant bone loss removal of the prosthesis may be easy, on other occasions either the humeral or ulnar components or both may remain well fixed. In this situation I have found it preferable to make cortical windows using a fine saw rather than attempt to knock out the prosthesis as this invariably results in uncontrolled periprosthetic fractures. Removal of the cement requires the use of fine osteotomes and chisels, and is a slow and laborious process. Having performed a complete debridement of the joint including any membrane lining the joint, the process of reconstruction is determined by the severity of bone loss.

Minimal bone loss

Minimal bone loss is most often encountered in cases of aseptic loosening where the arthroplasty has been in place for several years. The patient will usually present with intermittent symptoms of pain and discomfort around the elbow or pain that is precipitated by certain activities. Radiographs show loosening of either the humeral (Fig. 45.1), ulnar (Fig. 45.2) or both components. Humeral component loosening is thought to result from the posteriorly directed resultant forces that affect the elbow during flexion and extension,⁷ whereas at least some cases of ulnar component loosening result from ulnar component pistoning.⁸ Early revision prior to progressive bone loss is advised (Fig. 45.3).

Figure 45.1 (A,B) Early humeral loosening associated with a Kudo total elbow arthroplasty.

Figure 45.2 Loosening of the ulnar component of a Kudo prosthesis (arrowed).

Figure 45.3 (A,B) Postoperative radiographs following revision for minimal bone loss (preoperative radiograph is shown in Fig. 45.2).

Impaction bone grafting

This method of reconstruction is most useful where there has been extensive loss of medullary bone but with an intact cortical envelope (Fig. 45.4). It can also be used if there are small areas of cortical bone loss provided that these are controlled at the time of impaction grafting. This technique was described by Loebenberg et al⁹ who used a tube system to enable impaction of morcellized auto and allograft bone to restore the medullary bone while maintaining a central canal for insertion of the prosthesis.

Figure 45.4 (A) Pre- and (B) postoperative radiographs following impaction grafting.

The important principle with this method of reconstruction is to achieve firm impaction of the morcellized bone fragments, otherwise there is a risk of implant loosening in the early postoperative period.

Customized prostheses

Custom-made prostheses can be used for more extensive bone loss.¹⁰ With this technique it is important to obtain good-quality, size-marked radiographs together with radiographs of the contralateral elbow so that the manufacturer is able to produce an appropriately sized arthroplasty.

Autograft

Autograft can be used effectively for small bone defects. Larger defects can be treated using fibular strut grafts (Fig. 45.5), although this increases morbidity. I have no personal experience of this technique and prefer to use allograft when dealing with these difficult problems.

Figure 45.5 (A) Pre- and (B) postoperative radiographs following reconstruction using strut grafts.

(Courtesy of Biet Simmons.)

Allograft

Lexer was the first to describe allografts in joint reconstruction.¹¹ He transplanted fresh whole and partial joints removed at amputation into patients with joint deficiencies. Although he claimed a 50% success rate detailed follow-up was incomplete. No further publications on allografts appeared in the literature until the mid to late 1900s, when improvements in bone preservation resulted in renewed interest.^12,¹³ Infection, graft non-union and fractures, however, remained serious complications.

Urbaniak and Black ¹⁴ published their experience of ten partial and total cadaveric elbow reconstructions, nine of which were undertaken for trauma. The authors advised rigid internal fixation. Two years after insertion, however, degenerative changes were present in the joints, with two patients having allograft nonunion and two patients demonstrating elbow instability. One patient had a radial nerve palsy. A further study of 23 allograft elbow replacements followed for 20 years revealed seven allograft non-unions, six unstable elbows, four radial nerve palsies, three infections and one graft resorption.¹⁵ These results suggest that, although early benefit may be achieved by this technique, it is not a viable long-term option for elbow reconstruction. More recently allograft prosthesis joint combinations have been reported,¹⁶ some of which have been reinforced with strut grafts. This technique appears more successful and is my preferred option when bone loss is associated with significant cortical deficiency.

The procedure follows the principles of reconstruction outlined above. I always perform a two-stage revision, the first stage of which involves removal of the arthroplasty (if it has not previously been removed), all of the cement and any intramedullary membrane. In addition, a thorough debridement of the joint is undertaken and an assessment of bone deficiency is made. This involves not only evaluating the bone length deficit, but also making an assessment of the overall size of the bone so that an appropriate allograft replacement can be obtained.

Antibiotic beads are then inserted into the joint space in order to provide a high local concentration of antibiotic and maintain the length of the limb. The wound is closed and the arm placed in an above-elbow backslab with the elbow at 90°. Two doses of third-generation cephalosporin antibiotics are prescribed postoperatively.

The second-stage procedure involves removal of the antibiotic beads and reconstruction of the joint with allograft and a total elbow arthroplasty. It is important at this stage to avoid any further resection of the patient’s bone. Any bone resection that is required to allow insertion of the allograft must be from the allograft otherwise the patient will end up with a more extensive bone deficit. Having approximately corrected the arm length a small step cut should be made in the allograft bone (Fig. 45.6) with a corresponding cut in the patient’s bone. This will help provide rotational stability when the reconstruction has been completed. The allograft is then prepared for insertion of the arthroplasty. Once complete the arthroplasty is cemented into the allograft (Fig. 45.7). This is undertaken outside the patient. The arthroplasty I use for these reconstructions is the Coonrad–Morrey replacement. The allograft prosthesis combination is then cemented into the patient with bone and wires placed around the junctional step cut zone in order to provide additional rotational stability (Fig. 45.8). Strut grafts can also be placed at the allograft host junction to provide additional strength (Fig. 45.9). These are wired in place. The wound is closed and a full-arm backslab applied with the elbow in extension. The arm is elevated in a Bradford sling and two doses of a third-generation cephalosporin are prescribed. The wound is inspected at 48 hours and provided that it is satisfactory the patient is allowed to gently mobilize the elbow.

Figure 45.6 Massive humeral and ulnar allografts prior to insertion of prosthesis.

Figure 45.7 (A,B) Massive humeral and ulnar allograft prosthesis composite prior to insertion into the patient.

Figure 45.8 (A) Pre- and (B) post-operative radiographs following a massive allograft prosthesis reconstruction.

Figure 45.9 (A) Preoperative clinical photograph of a female patient who had undergone multiple previous surgical procedures and presented with breakdown of her wound with exposure of the ulnar component of her total elbow replacement. (B) Preoperative radiographs. (C) Postoperative radiographs showing a massive allograft prosthesis reconstruction with ulnar strut graft reinforcement. (D) Postoperative clinical photograph 4 years following reconstruction.

Mansat et al ¹⁶ reported their experience of using allograft implant reconstructions in 13 patients reviewed after an average of 42 months. The Mayo Elbow Performance Score was excellent in four, good in three, fair in one and poor in five. Five revision procedures were performed, two allograft implants were removed and four patients developed deep infections. My own experience with humeral and ulnar allografts has been encouraging. To date I have managed to restore useful elbow function and have only removed one ulnar allograft for infection.

When osseous deficiency is predominantly on the ulnar side of the joint, Kamineni and Morrey ¹⁷ have found strut allografts useful for reconstruction. They noted in 21 patients followed for an average of 4 years a mean Mayo Elbow Performance Score improvement of 34 points. Four osseous and four soft tissue complications were, however, noted in eight patients.