Complications of Total Elbow Arthroplasty

April D. Armstrong

Leesa M. Galatz

INTRODUCTION

Modern prosthetic arthroplasty has led to dramatic improvements in function and lifestyle of individuals affected with degenerative and inflammatory joint conditions. Hip and knee arthroplasty enjoys a long history of successful results with relatively low rates of complications. Similarly, shoulder arthroplasty in more recent decades has had results approaching those of arthroplasty for the lower extremity. Total elbow arthroplasty (TEA), however, has always been associated with a higher rate of complications. The overall complication rate following TEA has been reported to be as high as 57% (1, 2, 3, 4). Fortunately, recent improvements in patient selection, surgical exposure, component selection and design, rehabilitation protocols, and use of prophylactic infection prevention have helped to decrease these complication rates. Nevertheless, complications remain a significant concern for elbow arthroplasty. Revision surgery is more difficult and the elbow is far less tolerant because of the problems of bone stock preservation and lack of soft-tissue coverage around the elbow.

Major complications of TEA include infection, instability, loosening, periprosthetic failure and component failure, triceps insufficiency, neurologic injury, and wound problems. Critical evaluation of these potential complications has allowed an increased understanding of the risk
factors and issues pertinent to prevention. This chapter will discuss these potential complications, provide guidelines as to how to appropriately recognize and diagnose them, and outline treatment protocols.

INFECTION

Infection is arguably one of the most worrisome complications after prosthetic elbow replacement (Fig. 30-1). Many studies have observed an inordinately high number of infections after TEA (3,5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21). Early studies reported infection rates as high as 11% (18). More recent studies have reported an approximate 2% to 5% rate of infection (7,12,15). This reduced infection rate may be attributed to appropriate use of antibiotic impregnated cement, better handling of soft tissues, and increased awareness of the problem. Regardless, this infection rate is still greater than that for arthroplasty of other joints.

Several factors have been identified that may place certain patient groups at risk for infection. The first is the underlying diagnosis. TEA is commonly performed in patients with inflammatory arthritis, with rheumatoid arthritis being the most common of these conditions. Patients with rheumatoid arthritis may be immunocompromised as a result of the use of oral steroids and other medications. In one series, 50% of the patients with a deep infection were receiving steroid therapy (8). Thus, the use of oral steroids may have some significance in the development of infection after elbow arthroplasty. However, in this series the sample was too small to determine statistical significance. Another inflammatory type of arthritis that may pose a certain risk is psoriatic arthritis. The hip and knee arthroplasty literature contains reports of increased rates of infection in patients with psoriatic arthritis (22,23). Therefore, a preoperative dermatologic consultation may be recommended for patients with psoriatic or other skin lesions.

Patients who have developed posttraumatic arthritis and are considering a TEA may also have a higher risk for developing an infection. This may result from the fact that the soft tissues have been compromised surrounding the elbow, or it may be related to previous surgical procedures on the elbow. In Morrey’s series, 12.2% of the patients that had prior surgery became infected, compared to 7.5% that
had no previous surgery (8). This difference was not statistically significant; however, the numbers themselves are concerning. When the underlying diagnosis and previous surgery together were correlated with deep sepsis, there was a statistically significant relationship.

Figure 30-1 Infected total elbow arthroplasty showing obvious loosening of the humeral and ulnar components.

In general, infection in total joint arthroplasty may be the result of hematogenous seeding from a remote infectious source. Careful medical examination prior to surgical procedures should minimize the risk of a potential source at the time of the index procedure.

Evaluation

Patients presenting with an infected total arthroplasty often give a history of pain and a sudden decrease in progression with regard to their postoperative course. A change in the intensity or character of the pain may also be reported. Night pain or rest pain is considerably worrisome. A history of postoperative wound problems, such as persistent drainage, surrounding erythema, or prolonged postoperative antibiotic administration should raise the level of suspicion for infection. Patients may also report a declining range of motion instead of an improvement, as would be expected. Predisposing patient factors for infection should also be taken into consideration. These complaints should be given special notice because early diagnosis of an infected total elbow may have a significant impact on difficulty of treatment. Specifically, timing may affect the opportunity to retain the prosthesis versus doing a staged implantation or resection arthroplasty.

On physical examination, the wound may be warm and edematous, and in other situations it may look normal. The elbow is generally painful during range-of-motion testing. Obvious signs of infection include drainage from the wound, a sinus tract, and skin necrosis. The history and physical examination should also be directed toward discovering other potential sources for infection such as pneumonia or urinary tract infections.

Early evaluation should include a set of plain radiographs. In some instances of early infection, the radiographs may be completely normal. Radiographic signs of a persistent infection include periosteal or heterotopic bone formation, radiolucent lines surrounding the prosthesis, loosening at the bone-cement interface, and periprosthetic bone resorption. In general, if there is a high clinical suspicion of infection and any of these radiographic findings are present, then the patient should be treated as though he or she has an infection and further radiologic studies are unnecessary.

Additional workup should include a complete blood count, erythrocyte sedimentation rate (ESR), and a C-reactive protein (CRP) level. These tests, however, should always be interpreted in the context of the patient’s clinical situation. The white blood cell count is often normal in patients with an infected total elbow prosthesis. Many of the patients who have undergone TEA have inflammatory arthritis or may have just recently had surgery, thus falsely elevating the ESR. A CRP is more sensitive for infection and is unlikely to be elevated in aseptic cases. However, as previously stated, both the ESR and CRP should be interpreted cautiously in patients with inflammatory conditions or infections at remote sites because both of these markers may be elevated under certain circumstances. The ESR and CRP are also valuable in monitoring a patient’s response to treatment. If the infection is being appropriately and successfully managed, the ESR and CRP levels should decrease.

If the ESR and the CRP are abnormal, then an aspiration of the joint should be considered. Care must be taken, however, not to introduce an infection into an aseptic arthroplasty. Because the elbow has a thin soft-tissue surrounding envelope, if the skin is cellulitic, then the underlying prosthesis is most likely infected as well. A positive gram stain or a cell count of greater than 50,000 cells would be consistent with an infection. Culturing of sinus tracts or drainage is not helpful. These cultures tend to grow mixed organisms, which may or may not be related to the deep prosthetic infection. If a patient has been receiving suppressive antibiotics, it is suggested that the antibiotics be discontinued for 10 to 14 days prior to surgical debridement to ensure that a sensitive organism is not suppressed and to increase the likelihood of identifying the infectious organism. However, if removal of the antibiotics compromises the safety of the patient, then they should be continued.

Management

Nonoperative Management

When an infection at the surgical site of a TEA occurs, aggressive treatment is recommended. There are very few indications for nonoperative treatment. Because the soft-tissue envelope around the elbow is not substantial and the bony and metal structures are so superficial, there is rarely an infection involving the skin or the incisions that does not track down to the prosthesis. If wound problems are detected that suggest infection at a total elbow incision, the prosthesis should be presumed to be infected and treated accordingly.

Operative Management

Operative management of infections after TEA can be divided into three main groups: retention of the prosthesis, staged reimplantation, and resection arthroplasty. It is interesting that the interval from the index procedure to the treatment of the infection has not been shown to be associated with the results of treatment for the infection. Rather, the infecting organism has a significant impact on the results of treatment (8,24). Infections involving Staphylococcus epidermidis failed treatment with irrigation and debridement alone. This is likely related to the propensity of the organism to produce an infective biofilm. S.
epidermidis has been shown to have a high capacity for adherence to orthopedic implants and to produce a significant biofilm within 24 hours. This biofilm can also coat antibiotic impregnated cement and decrease the ability of the antibiotic to affect other organisms. Prostheses infected with Staphylococcus aureus, however, have been treated successfully with irrigation and debridement with retention of the prosthesis (24).

Extensive irrigation and debridement are recommended for the treatment of infection only when there is no radiographic evidence of new radiolucent lines, loss of bone consistent with osteomyelitis, or change in the position of the prosthesis on the radiographs. Components demonstrating gross instability at the time of intraoperative examination must be removed. The objective of extensive irrigation and debridement is to salvage the prosthesis. The rationale for this type of treatment is to preserve the limited bone stock of the distal humerus that could easily be lost during a difficult extraction. Removal of a well-fixed prosthesis with fracture and significant bone loss can lead to significant problems during attempted staged reconstruction, sometimes even prohibiting reimplantation. Therefore, if a component is well fixed and there is no evidence of osteomyelitis, irrigation and debridement are recommended.

The debridement is performed through the standard posterior approach used for the index arthroplasty. The component should be disarticulated, and the bushings should be completely removed. This allows complete exposure of the metallic portion and eliminates the possibility of potential adherence to polyethylene. The wound and any nonviable tissue should be completely debrided. The metallic prosthesis could even be scrubbed with a surgical scrub brush. Antibiotic impregnated cement beads containing 2 to 4 g of tobramycin per package should then be packed into the joint space. A drain is placed in the wound, and the skin is reapproximated using interrupted sutures with monofilament suture.

The patient is brought back to the operating room for repeat irrigation and debridement. At least three debridement procedures are recommended. Cultures should be obtained prior to closure from each irrigation and debridement. If exit cultures are positive after the third debridement, then further debridements are recommended. In one study, the average number of debridements was four, with a range of three to seven. Patients should receive concurrent antibiotic treatment parenterally as determined by the sensitivity of the organism followed by outpatient treatment for a minimum of 4 to 6 weeks.

Exchange Arthroplasty

Either immediate reimplantation or staged reimplantation may be considered when there is evidence of gross loosening. When there is sufficient bone stock for reimplantation, the component and any remaining cement should be removed in a meticulous fashion, taking care not to fracture the humerus and the ulna and to preserve as much bone stock as possible. An antibiotic impregnated cement spacer should be placed between the humerus and the ulna followed by incision closure over a drain. Further irrigation and debridement may be necessary until exit cultures are negative. The patient should be treated with parenteral antibiotics for a minimum of 6 weeks. In some cases it may be advisable to return the patient to the operating room prior to planned implantation to examine the wound and obtain more cultures. If these cultures are negative, the surgeon may then proceed with reimplantation. Antibiotic-impregnated cement should be used to fix the components. If the infection has been successfully eradicated, no more than standard intravenous antibiotics should be necessary after the reimplantation.

Intraoperative frozen section and tissue cultures are mandatory. If the frozen section demonstrates greater than five to ten polymorphonuclear nucleosides per highpowered field, then plans for reimplantation should be aborted. A second antibiotic spacer should be placed, or the patient should then undergo a resection arthroplasty. In either case, another 6 weeks of antibiotics is warranted. Relative contraindications to reimplantation include inadequate soft-tissue coverage, inadequate bone stock, an infecting organism not susceptible to antibiotic treatment, or poor systemic health of the patient.

There are very limited indications for immediate exchange arthroplasty for an infected total elbow. The hip and knee literature is controversial regarding this issue. Some authors have reported good success, whereas others have reported dismal results (25, 26, 27). Immediate exchange arthroplasty should only be considered for patients with a gram-positive organism (other than S. epidermidis), good soft-tissue and bone quality, and good medical health. Antibiotic cement should be used at the time of reimplantation.

Patients with an unstable prosthesis or severe lack of bone precluding staged reimplantation may be managed successfully with resection arthroplasty (8). Patients with a medical condition that precludes multiple irrigation and debridement or reimplantation are also good candidates for resection arthroplasty. Ideally, the ulna should be contained within the condyles because postoperative instability may be reduced if both humeral condyles are preserved (28,29). Postoperatively, the patient is placed in a cast in 90 degrees of flexion to allow for contraction of the soft tissues and maximize chances for a stable elbow. Morrey and colleagues reported satisfactory results in eight out of ten patients who required a resection arthroplasty for infection following a total elbow replacement (8). Furthermore, if instability remains an issue, then reimplantation may be considered at least 1 year following surgery if the patient is free of infection (28).

Long-term antibiotic suppression is also an option for patients in any of these categories. The organism must be susceptible to an oral antibiotic. This is generally only considered in elderly patients who are medically unfit for any
further surgical procedures. The prosthesis should be stable with no radiographic evidence of instability. If progression of the infection leads to loosening of the prosthesis, then surgical removal may be necessary and, in fact, may be easier at that point compared to when the infection was first diagnosed with the component well fixed.

INSTABILITY

Instability following TEA is typically a complication characteristic of unlinked prostheses (Fig. 30-2). Numerous studies have reported instability as a complication following unlinked TEA (5,6,12,18, 19, 20, 21,29, 30, 31, 32, 33, 34, 35). The rate of instability ranges from 0% to 14 % after unlinked TEA (5,6,20,21,29,36). Stability of the elbow is provided by both static and dynamic restraints (29,37, 38, 39). The ulnohumeral articulation, the anterior bundle of the medial collateral ligament, and the lateral ulnar collateral ligament of the elbow are the three most important primary static restraints to the elbow (29,37,38,40,41). Because of the nature of the design, unlinked TEA continues to depend on the static restraints for stability. Dynamic restraints include all of the surrounding elbow musculature. The muscles have recently become increasingly recognized as a substantial factor in instability and have been shown to play an important role in stabilizing the elbow by applying a compressive force across the joint (29,38,39,42, 43, 44, 45).

Etiology

The major factors associated with instability after TEA are patient selection, implant design, surgical technique, and postoperative management. Careful selection of patients is the most important factor in preventing instability following unlinked TEA. Patients with significant bone loss and incompetent soft-tissue structures are at risk for postoperative instability. It is not uncommon for patients with inflammatory arthritis to develop attenuation of soft tissues secondary to repeated effusions and the inflammatory process (35). Patients who have had previous surgery such as radial head resection and synovectomy may also be more at risk. In these patients, important stabilizing structures may have been inadvertently compromised (35). Other “at risk” patients include those with significant malalignment, posttraumatic bone loss or deformity, and flail elbow (29,46).

Implant design has significant implications in determining the ultimate stability of a TEA. There are three classes of TEA: linked, unlinked, and snap-fit linked. The unlinked designs are most commonly associated with instability. These prostheses rely heavily on extrinsic stabilizers, specifically the medial collateral ligament, and on the lateral ligament of the elbow and are less reliant on the intrinsic geometry of the implant for stability (39). Unlinked implants are less constraining or have a more shallow articulating surface and depend on the dynamic compressive forces of muscle activation to coapt the articular surfaces (39).

The normal elbow is dependent on ligamentous, osseous, and dynamic compressive forces for stability. Morrey and An have shown that the osseous articulation provides 31% and 33% of the resistance to valgus stress at the elbow at full extension and 90% of flexion, respectively (47). The bony articulation provides even more support to resist varus stress: 55% with the elbow extended and 75% with the elbow at 90 degrees of flexion (47). Unlinked implants theoretically are designed to replicate normal anatomy and kinematics of the elbow. Therefore, if stability of an unlinked design is primarily reliant on the ligamentous and muscular structures, unusual forces may lead to early failure. Some unlinked designs, such as Souter-Strathclyde, have deeper grooves in the ulnohumeral articulation (48). The deeper groove may improve stability yet risk a potentially higher aseptic loosening rate. The more conforming the articulation, the more important it is that the humeral and ulnar articulating surfaces are aligned. Component malalignment results in an incongruent articulation and implant maltracking. Correct alignment of the ulnar component has been described as particularly difficult (29,37,38). During cementing, immediate reduction of the ulnar component to the humeral component has been recommended to allow for slight corrections in rotation and alignment caused by asymmetric tensions of the intact collateral ligaments (29). Malalignment can lead to abnormal polyethylene wear, eccentric loading of the soft tissues and resultant attenuation, and an increased forced transmission to the bone cement interface with a higher loosening rate. In 2002 Ikavalko and colleagues reported 26 cases of postoperative instability out of 525 cases (5%) of TEA, using a more conforming unlinked prosthesis (Souter-Strathclyde). Dislocation was still the most commonly reported major complication in the study (12).

Nonanatomic design plays a role in creating elbow instability. A nonanatomic total elbow prosthesis should restore the normal anatomic axis of rotation for the elbow and normal valgus carrying angle (14,29,36). This, in turn, should restore the normal length-tension relationship of the surrounding capsular and ligamentous structures, the normal muscle moment arms, and thus stability of the elbow. Earlier designs of unlinked prosthesis did not adhere to these principles, and problems with early loosening and instability developed. Different valgus stem angles for the ulnar component were developed to allow re-creation of the normal valgus carrying angle and to reduce medial laxity, which helped to improve stability (19,20).

Some authors believe that the use of a radial head component will help to improve stability (29,32,35,38,49). Radial head anatomy is difficult to replicate and has been a limiting factor in the development of radial head prostheses to accompany ulnohumeral replacement. Therefore, the radial head must be positioned anatomically or else it will have a tendency to subluxate or dislocate (29). The radial
head acts as a secondary valgus constraint reducing stress on the medial collateral ligament and also provides additional rotational stability. Excision of the radial head has been shown to increase the varus-valgus laxity in TEA (38).

Only gold members can continue reading. Log In or Register to continue