 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1998 The Journal of Bone and Joint Surgery, Inc. Infection after Total Elbow Arthroplasty*  , ROCHESTER, MINNESOTAInvestigation performed at the Department of Orthopedic Surgery, Mayo Clinic and Mayo Foundation, Rochester
The purpose of this study was to review our experience with the treatment of twenty-five infections (in twenty-five patients) after total elbow arthroplasty and to examine indications for salvage of the prosthesis compared with those for resection arthroplasty. The patients were divided into three groups on the basis of treatment. Group I comprised fourteen patients who were managed with multiple, extensive irrigation and débridement procedures with retention of the original components. The primary indication for retention of the prosthesis was evidence that it was well fixed as determined both radiographically and intraoperatively. Group II comprised six patients who had removal of the prosthesis and débridement followed by immediate or staged reimplantation. Group III comprised five patients who were managed with resection arthroplasty. The infection was successfully eradicated in seven of the fourteen elbows that had salvage of the prosthesis with irrigation and débridement. The results were strongly dependent on the causative organism; attempts at débridement failed in the four elbows that were infected with Staphylococcus epidermidis compared with three of the ten that were infected with another organism. Four of the six patients in Group II had successful reimplantation of a prosthesis; in three, the infection had been caused by an organism other than Staphylococcus epidermidis. Only one of the three patients who had a Staphylococcus epidermidis infection had a successful reimplantation. None of the five patients who had a resection arthroplasty had signs of infection at the latest follow-up examination. We concluded that salvage of the prosthesis with extensive irrigation and débridement in the presence of an infection about the elbow can be reasonably successful if the infecting organism is not Staphylococcus epidermidis and if the components are well fixed. When removal of the components is warranted, staged reimplantation can also be highly successful when the infecting organism is not Staphylococcus epidermidis. However, the repeated operations necessary to retain a prosthesis and the high rates of complications seen with this approach—and the relatively good rates of satisfaction obtained with resection arthroplasty—suggest that resection arthroplasty remains the procedure of choice in medically frail patients or in patients for whom function of the elbow is less of a concern.
Refinements in prosthetic arthroplasty have made it a reliable treatment option, with long-lasting results, for restoration of functional motion to severely arthrotic elbows2,4,5,8-10,17,18,20,21. The prevalence of many previously recognized complications, such as premature loosening of the component, instability of the elbow, ulnar neuropathy, and breakage of the component, has been substantially reduced through advances in prosthetic design and operative technique3,10,11. Infection, however, has remained a severe and potentially devastating complication, with reported rates ranging from 1 per cent (three of 2025) to 12 per cent (six of fifty-two8)2,4,7,9-12,17,18,20-22. Options for the treatment of infection have ranged from long-term antibiotic suppression to exchange arthroplasty to resection arthroplasty7,12,22. To date, resection arthroplasty—a procedure that results in relatively good patient satisfaction but poor function—has generally been considered necessary to irradicate the infection. Thus, very little information on other treatment options, including early, extensive irrigation and débridement or exchange arthroplasty, has been reported. The purpose of the present study was to update our experience with the treatment of infection after total elbow arthroplasty. The results of the treatment of twenty-five consecutive infections that developed in twenty-five elbows as a complication of a procedure performed at our institution were reviewed to examine treatment options and to determine the indications for salvage of the prosthesis compared with those for resection arthroplasty.
Seven hundred and fifty-seven consecutive total elbow arthroplasties were performed at our institution from 1981 to 1994. Of these, 591 were primary and 166 were revision procedures. In 1981, the modified Coonrad-Morrey semiconstrained prosthesis was developed and has been used in most of the elbows treated at our institution since. The technique, which has been described previously10, involves a posteromedial approach in which the triceps is reflected in a subperiosteal fashion in continuity with the fascia of the distal part of the forearm. Because of unacceptably high rates of infection, several routine modifications were made in an effort to decrease this complication12. Since 1983, antibiotic-impregnated cement containing tobramycin at a concentration of one gram per package of cement has been used routinely for fixation of the implant in patients who have had a previous operation about the elbow. Preoperative assessment was performed with a high degree of suspicion for previous subclinical infection and included routine measurement of the erythrocyte sedimentation rate, for all patients who had had a previous operation about the elbow, especially as treatment for fracture or non-union. All patients who had had a previous operation about the elbow also had aspiration of the elbow before a primary total elbow arthroplasty was considered. In the present study, aspiration had been performed before the eight index arthroplasties that were done because of post-traumatic osteoarthrosis (Tables I, II, and III). Cultures of all specimens were negative. In addition, attention was focused on decreasing the risk of hematoma by elevating the elbow and maintaining it in extension in an anterior splint for two days postoperatively.
With use of the Institutional Total Joint Registry, we ascertained that twenty-five (3.3 per cent) of the 757 total elbow arthroplasties performed at our institution were complicated by postoperative infection. Twenty-three infections developed after a primary arthroplasty and two, after a revision procedure. Infections that developed after a total elbow replacement performed elsewhere were excluded from this analysis because the pretreatment information that was needed to specifically assess treatment options was inadequate. There were fourteen women and eleven men, and the average age was fifty-eight years (range, forty-two to eighty-nine years). The right elbow was involved in thirteen patients and the left, in twelve. Fourteen patients had involvement of the dominant limb. All patients were seen because of pain, loss of function, erythema, or warmth in the elbow. Patients were considered to have an infection when findings on culture were positive or there was a strong clinical suspicion of infection (based on a high white blood-cell count or erythrocyte sedimentation rate, operative observations, and the appearance of the wound), or both, and supportive pathological findings. All but one patient had a positive finding on culture of material obtained by aspiration or at the time of the operation. The remaining patient had been taking an antibiotic intermittently before a replacement procedure. As previously described22, an infection was considered acute if it had developed less than three months after the operation, as subacute if it had developed between three months and one year postoperatively, and as late if it had developed after one year. The patients were divided into three groups on the basis of the treatment. Group I consisted of fourteen patients who were managed with extensive irrigation and débridement with retention of the original components; Group II consisted of six patients who had removal of the prosthesis and débridement followed by either immediate or staged reimplantation; and Group III consisted of five patients who had resection arthroplasty.
Treatment
Group I The patients were managed with extensive irrigation and débridement with the objective of salvaging the prosthesis. The indications for this type of treatment depended on the surgeon, but a minimum criterion was evidence of a well fixed component. This evidence consisted of no new radiolucent lines, loss of bone consistent with osteomyelitis, or change in the position of the prosthesis on radiographs as well as no gross instability of the components on intraoperative manual stressing at the time of the débridement. Débridement was performed through a standard posterior approach. An essential portion of the initial débridement was the complete disarticulation of the components and removal of bushings (Figs. 1-A and 1-B) in order to remove areas of potential adherence by bacteria, such as the polyethylene, as well as to debride the articulation of the components more thoroughly. This was followed by placement of antibiotic-impregnated cement beads (two to four grams of tobramycin per package). The patients typically were returned to the operating room after four days for repeat irrigation and débridement as well as exchange of the cement beads. In all, at least three irrigation and débridement procedures were performed, and the need for additional procedures was determined by the surgeon and the patient, as indicated by protocol. The second débridement involved removal of specimens for repeat culture, and the third débridement was performed to reassess the soft tissues. Another culture was performed if the previous culture had been positive. Additional irrigation and débridement was performed if cultures were positive after the third débridement or if there were persistent concerns about non-viable soft tissues. The average number of débridements was four (range, three to seven). All patients received concurrent antibiotic treatment parenterally, as determined by the sensitivity of the organism, on an outpatient basis for a minimum of four weeks (average, six weeks). Twelve patients then took ciprofloxacin orally for another four weeks. One patient received chronic ciprofloxacin treatment.
Group II Five patients were managed with a staged exchange arthroplasty, and one, with an immediate reimplantation. The indications for the procedure again depended on the surgeon but it was generally considered when there was evidence of gross loosening in the context of sufficient bone stock for reimplantation. An additional concern was the interval between the onset of symptoms and the operative treatment of the infection. On the basis of previously reported experience with débridement about the knee19, five of the seven patients in Groups I and II who had had symptoms for more than twenty-one days had an exchange arthroplasty. The components and any remaining cement were removed in a meticulous fashion. Antibiotic-impregnated cement (two grams of tobramycin per package) was placed as a spacer between the humerus and ulna, and the incision was closed (Figs. 2-A, 2-B, 2-C and 2-D). The interval between stages for the patients who had a staged procedure ranged from six to forty weeks (average, sixteen weeks), according to the preference of the surgeon. Reimplantation was performed as previously described for primary total elbow replacement with the Coonrad-Morrey semiconstrained implant and antibiotic-impregnated cement. The only modification was that an eight-inch (20.3-centimeter) stem was used in all patients. No additional antibiotics were given intravenously after the reimplantation.
Group III The patients were managed with resection arthroplasty. Again, the indications for the procedure varied. Two patients, one of whom had a stable prosthesis, had refused other treatment options, and two patients had a medical condition that precluded attempts at reimplantation. In the remaining patient, a planned reimplantation was not performed because surprisingly good stability and pain relief were obtained early after the resection. All of the patients were managed with one gram of vancomycin given intravenously every twelve hours for a minimum of four weeks and with ciprofloxacin administered orally for four weeks on the basis of the sensitivities of the organisms. After removal of the components, the elbow was placed in a cast in 90 degrees of flexion; the cast was worn for four weeks to allow contraction of the soft tissues and to provide a measure of stability.
Evaluation of the Patients
Five of the patients died before the latest follow-up evaluation. The average time between the treatment of the infection and the last examination of these patients at our institution was 115 months (range, thirty-four to 189 months). Follow-up data for six patients were obtained from their local orthopaedic surgeon as well as a telephone interview. The average duration of follow-up for the entire series was seventy-one months (range, twenty-four to 189 months).
Group I Seven of the fourteen successful procedures in the entire series were in this group. The treatment failed in all four patients who had an infection with Staphylococcus epidermidis. In contrast, only two of the eight patients who had an infection with Staphylococcus aureus had persistence of that infection. The treatment failed in one of the two patients who had a gram-negative infection. All seven patients in whom the débridement failed had a subsequent resection arthroplasty. There was no additional evidence of persistent infection after the resection. Five of these patients were unsatisfied with the result secondary to poor function. The average elbow score for these five patients was only 35 points. Relative stability was obtained following the resection in the remaining two patients, who had elbow scores of 50 and 60 points and were satisfied with the final result. Six patients, four of whom had a persistent infection, had at least one complication. The most serious complications were irreversible nerve injury (three patients), late fungal infection (one), and breakdown of the wound (one). Of the two patients who had a complication without persistent infection, one had late aseptic loosening of the prosthesis and the other had avulsion of the triceps. Both complications were treated successfully with a revision procedure without recurrence of the infection. At the latest follow-up examination, both patients were satisfied with the result, which was rated as good12. The postoperative elbow scores were good (80 points or more) for the seven patients in whom débridement was successful. However, one patient was unsatisfied with the result at the latest follow-up examination as her expectations had not been met.
Group II The result12 was fair or good for four of the six patients (Table II). The procedure was considered a failure for two of the three patients who had an infection with Staphylococcus epidermidis, and these patients then had resection arthroplasty. Neither patient had any additional evidence of persistent infection. Each had a fair functional result (an elbow score10 of 50 points), and both were satisfied with the outcome at the most recent examination. Complications were reported in only one patient, who had breakdown of the skin over the olecranon and weakness of the triceps, which were treated with a revision procedure. These complications account for the fair result on the basis of the Mayo Clinic performance index for the elbow10 (60 points). The patients who had a successful reimplantation had a dramatic improvement in the elbow score, from an average of 21 points preoperatively to an average of 79 points postoperatively. All six patients were satisfied with the result at the most recent follow-up examination regardless of the outcome with respect to the infection.
Group III None of the five patients had evidence of persistent infection at the time of the latest follow-up. The improvement in the elbow score was equivocal (from an average of 37 points preoperatively to an average of 49 points postoperatively). The average arc of motion was 59 degrees (range, 22 to 100 degrees) preoperatively and 69 degrees (range, 40 to 110 degrees) postoperatively. The elbow was grossly unstable after the resection in four patients and moderately unstable in one. All of the patients needed functional bracing postoperatively. The result12 was good for one patient, fair for two, and poor for two. However, four patients were satisfied with the result (Table III). The only complication was fracture of the distal aspect of the humerus and the proximal aspect of the ulna during removal of the components in one patient. This patient had an infection around a well fixed revision component, but she refused recommendations for débridement. The fractures healed after immobilization, but the result was fair (Figs. 3-A, 3-B, 3-C, 3-D and 3-E).
Little information is available regarding the treatment of infections after total elbow arthroplasty. Early reports focused primarily on identification and characterization of this problem, including epidemiology, risk factors, and recommendations for prophylactic treatment, in attempts to reduce the prevalence11,12,15,16,22. Most experience with treatment has been limited to resection arthroplasty. Of fourteen patients included in a previous study from our institution, ten had resection arthroplasty12 and only three retained the prosthesis through successful débridement (one patient) or reimplantation (two patients). Resection arthroplasty was recommended for most patients. Wolfe et al. treated eleven infections in ten patients with extensive débridement in an attempt to salvage the implant. Eight of the ten patients had a persistent infection, which led to resection arthroplasty in six of them and to arthrodesis in two. The three patients in whom the implant was salvaged were given chronic suppressive antibiotic therapy orally; two had intermittent drainage suggesting persistent infection. Débridement was not considered a viable option. Since the initial reports on infections around total elbow replacements, the clinical presentation of these infections has changed. Increased awareness of this complication has led to a high index of suspicion and, hence, to earlier recognition7. Often, patients are now seen while the infection is still acute, the components are well fixed, and there is no apparent osseous involvement. Removal of the prosthesis in these circumstances may necessitate extensive destruction of the bone, obviating a possible exchange arthroplasty. Wolfe et al. reported that three of eight patients who had a resection sustained a fracture during the procedure. In a study on revision total elbow arthroplasties, eleven of thirty-three patients sustained a fracture or cortical penetration during removal of the cement13. Although reimplantation was not planned, one patient in the present series sustained a fracture of both the humerus and the ulna during resection of a well fixed component (Figs. 3-A, 3-B, 3-C, 3-D and 3-E). It is now recognized that early diagnosis of these infections presents a previously non-existent opportunity for salvage of the prosthesis. Experience with early irrigation and débridement for infections around total knee replacements suggested that both the duration of the infection and the type of infecting organism are strongly associated with the outcome19. In a series of thirty-one knees treated with irrigation and débridement, the average time between the onset of symptoms and the débridement in the knees in which treatment succeeded was twenty-one days compared with thirty-six days in the knees in which treatment failed19. In the context of a relatively short interval from the onset of symptoms to the débridement, our results suggest that a reasonable rate of salvage can be obtained when there is no radiographic or intraoperative evidence of loosening of the prosthesis. The rate of success after irrigation and débridement (Group I) was found to be strongly dependent on the organism involved. All four Staphylococcus epidermidis infections recurred, and the original components had to be removed. Two of the patients who had such an infection were managed on the day that the symptoms developed and the procedure still failed. In contrast, the longest durations of symptoms (thirty and thirty-one days) were associated with Staphylococcus aureus infections, and both were treated successfully with irrigation and débridement. The interval from the index procedure to the treatment of the infection was not shown to be associated with the results of treatment of the infection. Irrigation and débridement was successful in three of the four patients who had a late or subacute infection. However, all of the Staphylococcus epidermidis infections were acute (they were treated within three months after the index procedure) and the treatment still failed, even though the total elbow arthroplasties had been performed as treatment for post-traumatic osteoarthrosis and the patients were not immunocompromised. The poor results associated with the Staphylococcus epidermidis infections may be related to the propensity of the organism to produce more effective biofilms1. Staphylococcus epidermidis has been shown to have a high capacity for adherence to implants and to produce tenacious biofilms within twenty-four hours1,3,14. These biofilms can also adhere to antibiotic-impregnated cement and impair the ability of the antibiotic to kill other organisms3. Several factors may have been important in the successful irrigation and débridement in six of the eight patients who had an infection with Staphylococcus aureus. Unlike in the study by Wolfe et al., multiple débridements were performed, each with an exchange of the antibiotic-impregnated cement. In addition, a complete disarticulation with removal of the pin and bushings was routinely done at the initial débridement. Finally, the average duration of symptoms (nine days) was very short. Despite the relatively high rates of success seen with irrigation and débridement, caution should be employed before salvage of the prosthesis is attempted. There was a high rate of complications (six of fourteen elbows). The repeated débridement of thin, relatively avascular tissues led to either breakdown of the wound or avulsion of the triceps in three patients. More alarmingly, the close proximity of important neurovascular structures led to irreversible nerve injury in three patients. Increased awareness of these complications may reduce their prevalence. Our results suggest that, when removal of the components is necessary, reimplantation can be reasonably successful. Four of the six patients who were managed with reimplantation (Group II) had no signs of persistent infection at the latest follow-up evaluation. Four of the five staged procedures were successful. As in Group I, infection with Staphylococcus epidermidis may have been an important factor in the failure. Reimplantation in two of the three patients who had such an infection failed, as evidenced by persistent infection and drainage from the wound. These patients had a subsequent resection arthroplasty. It should be noted that one of these failures was in a patient who had an immediate reimplantation, which, in retrospect, was an unwise choice. Multiple débridements and longer intervals between stages may be indicated for such patients. In contrast to Groups I and II, all five patients in Group III (resection arthroplasty) had eradication of the infection. There was only one complication, which was functionally unimportant, and four of the five patients were satisfied with the result. Thus, our results suggest that resection remains the standard for treatment of infection and should be the treatment of choice in medically frail patients who cannot tolerate repeat operative procedures or a higher risk of complications and in patients who have loosening of one component or both. Although they were not specific focuses of this study, two other considerations deserve discussion. The overall rate of infection of 3.3 per cent in this series (twenty-five) after the 757 arthroplasties performed at our institution during the present study period, while still high, represents a marked improvement from the 9 per cent rate previously reported from this institution12. Multiple treatment strategies such as the use of antibiotic-impregnated cement, a posteromedial approach, and postoperative control of hematoma, routinely employed since 1983, appear to have been beneficial. In conclusion, extensive irrigation and débridement for the salvage of a total elbow prosthesis associated with infection can be reasonably successful if the infecting organism is not Staphylococcus epidermidis. Additional important indications are a short (less than thirty-day) interval from the onset of symptoms to treatment and the presence of well fixed components at presentation. When removal of the components is warranted, staged reimplantation can also be highly successful in patients who have good bone stock and are medically stable. Resection arthroplasty remains the standard for treatment of infection and is the procedure of choice in medically frail individuals and in patients for whom function of the elbow is less of a concern.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
  CiteULike  Connotea  Del.icio.us  Facebook  Technorati  Twitter What's this?
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |