Suppressive Antibiotic Therapy for Retained Infected Prosthetic Joints

Case Series and Review of the Literature

Abstract

Prosthetic joint infections are optimally treated with a 2-stage revision. However, because of frequent comorbid illness, the prosthesis cannot always be removed. When this is the case, long-term suppressive oral antibiotics are often given to prevent breakthrough infection; however, there is minimal data on this practice. We performed a retrospective case series of patients at our own institution and reviewed published case series on the topic. A systematic review was performed on the available data. The odds ratio for successful outcome in the pooled analysis was 1.3 (95% confidence interval, 0.6-3.1) in favor of suppressive antibiotics, but this difference was not significant. We conclude that, despite the common use of suppressive antibiotic therapy for retained prosthetic joints, there is insufficient data to demonstrate whether this practice reduces failure rates.

Prosthetic joint infections are a challenging problem for orthopedists and infectious disease specialists alike, occurring in 1.5% of total hip arthroplasties and 2.5% of all total knee arthroplasties.¹ When the prosthesis can be replaced with a 2-stage revision, successful outcomes are seen in around 91% of patients.¹ However, when the prosthesis cannot be replaced, outcomes tend to be much worse.¹ Reports of successful outcome when the prosthesis is retained vary widely in the literature, from 23%² to 86%.³ The reason success rates are so much lower when the prosthesis is retained is believed to be due to the presence of a biofilm on the prosthesis.⁴ Antibiotics may penetrate poorly into biofilms,⁵ and microorganisms within a biofilm may have altered growth patterns that prevent antibiotics from effective killing.^6-8 Because of these difficulties, long-term antibiotic suppression is often given to prevent relapse when the prosthesis cannot be removed.^4,9 Despite this being a common practice, good data supporting its efficacy are lacking. We evaluated our own institution's outcomes for antibiotic treatment for retained prostheses and conducted a systematic review of available case series to try and determine if the data support the practice of using suppressive antibiotic therapy in this setting.

METHODS

Selection of Patients for the Case Series

All patients with prosthetic joint infection followed through the Outpatient Parenteral Antimicrobial Treatment Program in the Infectious Disease Group Practice at the University of Colorado Hospital who presented between January of 2003 to January of 2007 were reviewed and included if they met the following criteria: (1) bacterial or fungal pathogen(s) isolated from the joint in which a prosthesis was in place or was being placed and the prosthesis was not removed because of infection. Single-stage exchange of part or all of the hardware was not considered an exclusion criterion. (2) The bacteria or fungi were considered by the Infectious Disease Service to represent infection rather than colonization or contamination. (3) The patients successfully completed a course of intravenous antibiotic therapy appropriate for the isolated pathogen(s). (4) The patients had at least 1 follow-up visit after the completion of intravenous antibiotic therapy in which it could be assessed whether the joint was still functioning or not and whether they were being given suppressive antibiotic therapy.

Selection of Cases for Literature Review

PubMed was used to search for any study evaluating the outcome of prosthetic joint infections in which the prostheses was left in place and treated with incision and drainage and antibiotics. The following search terms were used in combination: "prosthetic joint," "infection," and "suppressive antibiotics." Additionally, when studies that were identified in this manner referenced other relevant studies, those studies were also evaluated. Identified studies were then reviewed and included if they contained sufficient information to identify the number of patients treated with or without suppressive antibiotics and which of these patients had favorable or unfavorable outcomes. Studies were excluded if they did not include both patients who received and who did not receive suppressive antibiotic therapy or did not have at least 1 successful and 1 unsuccessful outcome in each group (because of the inability to calculate an odds ratio with confidence intervals when there is not at least 1 successful and 1 unsuccessful outcome in each group).

Definitions Used

Suppressive antibiotic therapy was defined as the use of oral antibiotics for the prevention of relapse, rather than the treatment of the underlying infection. It is assumed in these patients that the prosthesis remains infected, and oral antibiotics are given indefinitely to keep the symptoms under control but not cure the infection. This is in contrast to the use of oral antibiotics to try and treat and potentially cure infection as is sometimes done with oral rifampin in combination with a second antibiotic for staphylococcal infections.¹⁰ Patients were defined as using suppressive antibiotics if they remained on an oral antibiotic for this purpose up until the last follow-up visit or until there was failure of the prosthesis due to infection. Patients were defined as not taking suppressive oral antibiotics if they were not on an oral antibiotic for this purpose at the last follow-up visit or at the time of prosthesis failure. Thus, patients who started on suppressive oral antibiotics but later stopped them before the last follow-up visit were defined as not receiving suppressive oral antibiotics. Duration of follow-up was defined as the time between the initiation of antibiotics for the infection and the last follow-up visit or contact with the patient. Successful outcome was defined as having a functioning prosthesis without symptomatic infection at the last patient contact in which the status of the prosthesis was addressed. Failure was defined as loss of a functional prosthesis due to infection. If patients were on suppressive therapy, this would be considered a breakthrough infection, and if patients were not on suppressive therapy, this would be considered a relapse.

Statistical Analysis

For the comparison of clinical variables within our own case series, the Fisher exact test was used for the comparison of dichotomous variables and a Wilcoxon test was used for the comparison of continuous variables using SAS, Version 8.2 (SAS Institute, Cary, NC). Odds ratios and 95% confidence intervals were calculated using StatCalc (EpiInfo 2000; Centers for Disease Control and Prevention, Atlanta, Ga). The combined odds ratio and confidence intervals were calculated using the methods of DerSimonian and Laird¹¹ within a Microsoft Excel spreadsheet.

RESULTS

Current Case Series

We identified 18 individuals who met our study criteria. Demographic and clinical data on each patient are provided in Table 1. No statistically significant differences were observed between the group of patients who received suppressive antibiotic therapy and those who did not with regard to age, sex, ethnicity, joint location, age of prosthesis at the time of infection, findings of purulence at surgery, exchange of components at the time of surgery, duration of intravenous antibiotic therapy, use of rifampin for staphylococcal species, or the presence of Staphylococcus aureus. Seven of 9 patients in both groups had a successful outcome at the time of last patient contact.

TABLE 1: Clinical Variables of Patients Included in the University of Colorado Cohort

In the group that was not receiving long-term suppressive antibiotics, 3 patients were originally placed on suppressive antibiotics before electing to stop (patients 2, 5, and 9). All 3 of these patients had more than 1 year of follow-up after stopping suppressive antibiotics without clinical relapse. Two patients in the group not receiving long-term suppressive antibiotics relapsed. Based on identification of the organism and sensitivity patterns, both patients seemed to have the same organism that had caused the original infection (patients 6 and 7).

Two patients in the group that was receiving long-term suppressive antibiotics had breakthrough infections. One patient had a breakthrough infection with a coagulase-negative Staphylococcus whose sensitivity pattern was slightly different from the original isolate (The original isolate was resistant to levofloxacin and trimethoprim/sulfamethoxazole, whereas the breakthrough isolate was sensitive [patient 12].); however, both isolates were sensitive to the doxycycline she was maintained on. The other patient had a breakthrough infection with a coagulase-negative Staphylococcus isolate that was listed as intermediate to levofloxacin, whereas the original isolate was listed as sensitive (patient 15). This patient had been maintained on levofloxacin for suppression.

We then looked to see if there were any differences between patients who relapsed or had breakthrough infection and those that did not. We evaluated for patient age, age of the prosthesis at the time of infection, duration of intravenous therapy, duration of follow-up, presence of purulence at surgery, exchange of hardware, whether S. aureus was present, use of rifampin for Staphylococcus species, ethnicity, joint infected, and sex. We noted a strong trend toward higher failure rates with prostheses that had been in for longer periods (median age of prosthesis, 14.0 vs 114.5 days; P = 0.05). There were no statistical differences in the other variables measured.

Literature Review

Eight case series were identified in the literature that reported outcomes among a series of patients with enough information to identify which of the patients did or did not receive suppressive oral antibiotic therapy for retained prosthetic joints after infection.^2,3,9,12-16 Four studies were excluded because they did not include both patients who received and who did not receive suppressive antibiotic therapy.^2,3,12,15 An additional study was excluded because there were no clinical failures in the group of 8 patients who received suppressive antibiotics for streptococcal prosthetic joint infections, thus preventing the formal calculation of an odds ratio with confidence intervals. In this study, 11 patients did not receive suppressive antibiotics, and 9 of these patients had successful outcomes at the last follow-up visit.¹⁶ This left 3 case series from the literature, plus our own for inclusion in the systematic review (Table 2). All 4 case series had very wide confidence intervals and were statistically nonsignificant. The pooled odds ratio was 1.3 but was again nonsignificant (Fig. 1).

TABLE 2: Case Series Used to Assess the Efficacy of Suppressive Antibiotic Therapy for Retained Prosthetic Joints

FIGURE 1: Systematic review of case series looking at the use of suppressive antibiotic therapy for infection in retained prosthetic joints. The point estimate of the odds ratio is on the right, with the 95% confidence intervals in parentheses. UCH indicates University of Colorado Hospital.

DISCUSSION

To our knowledge, this is the first systematic review of the use of suppressive antibiotics for infected joint prosthesis that cannot be removed. Although all 3 clinician-authors of this paper have generally used suppressive antibiotic therapy in this setting, we acknowledge that this is largely dependent on the personal beliefs of the prescribing physician. In case series reviewed, the number of patients receiving suppressive antibiotics has varied from none to all.^2,3,12,15

Our data are insufficient to answer the question of whether suppressive oral antibiotics should be given to individuals who have a prosthetic joint infection in which the prosthesis is retained. When suppressive antibiotics are given, it is generally on the belief that organisms within a biofilm cannot be fully eradicated and that the oral antibiotic, although not sterilizing the biofilm, prevents the organism in the biofilm from spreading.⁴ However, it is also possible that oral antibiotics will not penetrate adequately into a biofilm,⁵ and if the levels of the antibiotics within the biofilm are below the minimum inhibitory concentration of the organism, then resistance to that antibiotic could develop.¹⁷ It is worth noting that, in our cohort, one of the patients on suppressive antibiotics who failed was on levofloxacin for a coagulase-negative Staphylococcus infection. This organism went from sensitive to intermediate during the course of suppressive therapy. Although it is possible this represented a new infection, all other sensitivities tested were identical, suggesting that it was the same organism that had developed some degree of resistance to the suppressive antibiotic being used.

There are numerous limitations to our data. The first is the very small sample size. Even pooling data from the 3 previous studies that met our criteria, plus our own data, there are only 39 patients who were not on suppressive antibiotics at their last follow-up visit. Additional limitations include the retrospective nature of the studies and the lack of control for other important variables such as whether there was any exchange of hardware at the time of incision and drainage, the duration of the treatment phase of intravenous antibiotics, which intravenous antibiotics were used, the age of the prosthesis, the presence of a sinus tract, which suppressive antibiotic was used, and which organism was present. Additionally, the exclusion of the study by Meehan et al¹⁶ in which 8 of 8 patients on suppressive antibiotics had successful outcomes versus 9 of 11 patients not on suppressive antibiotics clearly skewed the data away from showing benefit with suppressive antibiotic therapy. Finally, the short duration of follow-up on some patients may not allow us to fully appreciate the true rate of relapse because some patients may not present with relapse until 2 or more years after completing therapy.¹⁸

Our study does highlight what others have reported, that up to 50% of patients who have prosthetic joint infection in which the prosthesis cannot be removed can still be effectively managed with incision and drainage followed by antibiotic therapy.^3,9,14-16 Whether long-term suppressive antibiotic therapy should be a part of this treatment will remain unanswered until additional data become available.