Polymicrobial Periprosthetic Joint Infections: Outcome of Treatment and Identification of Risk Factors

Tan, Timothy L. MD; Kheir, Michael M. MD; Tan, Dean D. BS; Parvizi, Javad MD, FRCS

Journal of Bone & Joint Surgery - American Volume:
doi: 10.2106/JBJS.15.01450
Scientific Articles
Abstract

Background: The treatment outcomes of periprosthetic joint infection are frequently dependent on characteristics of the causative organism. The objective of this comparative study was to investigate the prevalence of and risk factors for development of polymicrobial periprosthetic joint infection, and the outcome of surgical treatment of these patients.

Methods: All patients with polymicrobial, monomicrobial, or culture-negative periprosthetic joint infection treated from 2000 to 2014 were identified at a single institution. Ninety-five patients with a polymicrobial periprosthetic joint infection had a minimum follow-up of 12 months. We matched patients with a polymicrobial periprosthetic joint infection with the other cohorts using propensity score matching for several important parameters. Treatment success was defined according to the Delphi criteria; Kaplan-Meier survivorship curves were generated to demonstrate this. A multiple logistic regression analysis was performed to determine risk factors for a polymicrobial periprosthetic joint infection.

Results: Overall, 10.3% (108 of 1,045) of the periprosthetic joint infections treated at our institution were polymicrobial in nature. Patients with a polymicrobial periprosthetic joint infection had a higher failure rate at 50.5% (48 of 95) compared with the monomicrobial periprosthetic joint infection cohort at 31.5% (63 of 200) and the culture-negative periprosthetic joint infection cohort at 30.2% (48 of 159) (p = 0.003). The survivorship of the polymicrobial periprosthetic joint infection group was 52.2% at the 2-year follow-up, 49.3% at the 5-year follow-up, and 46.8% at the 10-year follow-up. Patients with polymicrobial periprosthetic joint infection had a higher rate of amputation (odds ratio [OR], 3.80 [95% confidence interval (CI), 1.34 to 10.80]), arthrodesis (OR, 11.06 [95% CI, 1.27 to 96.00]), and periprosthetic joint infection-related mortality (OR, 7.88 [95% CI, 1.60 to 38.67]) compared with patients with monomicrobial periprosthetic joint infection. Isolation of gram-negative organisms (p < 0.01), enterococci (p < 0.01), Escherichia coli (p < 0.01), and atypical organisms (p < 0.01) was associated with polymicrobial periprosthetic joint infection. Only the presence of a sinus tract (OR, 2.20 [95% CI, 1.39 to 3.47]; p = 0.001) was a significant risk factor for polymicrobial periprosthetic joint infection on multivariate analysis.

Conclusions: This study reveals that polymicrobial periprosthetic joint infection, occurring at a relatively low rate, is associated with poor outcomes when compared with monomicrobial and culture-negative periprosthetic joint infection. Patients with polymicrobial infections were more likely to require a salvage procedure or to have periprosthetic joint infection-related mortality. Polymicrobial periprosthetic joint infection was associated with soft-tissue defects such as a sinus tract and certain types of organisms, which should be considered when administering antibiotics to these patients.

Level of Evidence: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.

Author Information

1The Rothman Institute of Orthopaedics at Thomas Jefferson University Hospital, Philadelphia, Pennsylvania

E-mail address for J. Parvizi: research@rothmaninstitute.com

Article Outline

Periprosthetic joint infection is a serious complication after total joint arthroplasty. The treatment outcomes of periprosthetic joint infection are frequently dependent on the characteristics of the causative organism, including virulence and antibiotic resistance. In some patients with periprosthetic joint infection, multiple organisms may be identified from the synovial fluid or tissue cultures. Specifically, the rate of polymicrobial periprosthetic joint infection is reported to be from 6% to 37%1-4. Although common organisms such as Staphylococcus aureus are frequently isolated in these infections, more virulent organisms such as Enterococcus species, gram-negative bacilli, methicillin-resistant S. aureus (MRSA), and anaerobic bacteria are more commonly associated with polymicrobial infections than monomicrobial infections5. The management of patients with polymicrobial periprosthetic joint infection requires administration of a broad-spectrum antibiotic or often multiple antibiotics to ensure adequate coverage against the infecting organisms4. Hence, treatment of these patients may be associated with higher failure6, mortality7, and treatment cost than those for patients with monomicrobial infections8.

Despite the relative frequency of polymicrobial periprosthetic joint infection, the current literature is sparse with regard to the outcome of treatment of these patients. Furthermore, the risk factors that may predispose a patient to a polymicrobial periprosthetic joint infection are not well documented. The objective of this comparative study was to investigate the prevalence of and risk factors for the development of a polymicrobial periprosthetic joint infection and the outcome of surgical treatment for patients with this infection.

Back to Top | Article Outline

Materials and Methods

We reviewed our prospectively maintained institutional electronic periprosthetic joint infection database, which, at that time, consisted of 1,045 patients with a periprosthetic infection after a knee replacement (n = 578 [55.3%]) or a hip replacement (n = 467 [44.7%]). All patients with a polymicrobial, monomicrobial, or culture-negative periprosthetic joint infection treated between 2000 and 2014 at our institution were identified. Any patient with a periprosthetic joint infection in which ≥2 different organisms were isolated from the synovial fluid or deep-tissue cultures was considered to have polymicrobial periprosthetic joint infection. One hundred and eight patients with polymicrobial periprosthetic joint infection were treated during the study period at our institution. Of these patients, 95 had a minimum follow-up of 12 months.

Periprosthetic joint infection diagnosis was made using the Musculoskeletal Infection Society (MSIS) criteria9. The culture was considered positive if a pathogen was isolated on solid media by the microbiology laboratory; light growth or very light growth was defined as a positive culture finding in this study. According to the standard practice at our institution, every specimen was sent for aerobic, anaerobic, fungal, and acid-fast bacilli culture, with a mean number of 10 samples taken (range, 2 to 24 samples). Both solid media and broth were used. A matrix-assisted laser desorption-ionization time-of-flight (MALDI-TOF) mass spectrometer (Microflex LT, Bruker Daltonics) has been utilized in recent years to confirm the identity of pathogens isolated from culture. Resistant organisms were defined as vancomycin-resistant Enterococcus and MRSA.

The clinical records of these patients were reviewed manually in detail to extract pertinent information that included age, body mass index (BMI), sex, involved joint, Charlson Comorbidity Index10, follow-up findings, nature of the surgical treatment, type of cultured organisms, presence of a sinus tract, and presence of intraoperative purulence.

Treatment success (survivorship) was defined according to the Delphi criteria11, which were based on the following: “(1) infection eradication, characterized by a healed wound without fistula, drainage, pain, or infection recurrence caused by the same organism strain; (2) no subsequent surgical intervention for infection after reimplantation surgery; or (3) no occurrence of PJI [periprosthetic joint infection]-related mortality.”

Back to Top | Article Outline
Statistical Analysis and Matching

All statistical analyses were performed with use of R version 3.1 (R Foundation for Statistical Computing) using the RMS (regression modeling strategies) package for the logistic regression. An alpha level of 0.05 was used to determine significance. Kaplan-Meier survivorship curves were generated for follow-ups of 2, 5, and 10 years. Differences in survivorship were assessed using the log-rank test. A univariate analysis was performed to identify microbiological factors associated with polymicrobial periprosthetic joint infection with an unselected control group. A multiple logistic regression analysis was performed to determine risk factors for the development of a polymicrobial periprosthetic joint infection within the 1,045 periprosthetic joint infections in the complete institutional periprosthetic joint infection database. The Fisher exact test was used to evaluate differences in complication rates, microbiological factors, and treatment success.

To compare the success of treatment for each group, we identified all patients with polymicrobial periprosthetic joint infection and >1 year of follow-up and then matched the cohort with a control group of patients with monomicrobial periprosthetic joint infection at a 1:2 ratio using propensity score matching with respect to the following parameters: joint, sex, age, BMI, Charlson Comorbidity Index, organism resistance (defined by MRSA and vancomycin-resistant Enterococcus), date of the surgical procedure, and years of follow-up. This yielded 200 patients with monomicrobial periprosthetic joint infection. We also performed propensity score weighting to create another control group consisting of culture-negative periprosthetic joint infections; we identified all patients with >1 year of follow-up, and propensity scores were generated from a logistic regression analysis using the culture-negative arm for involved joint, age, sex, BMI, Charlson Comorbidity Index, and date of the surgical procedure. Because BMI and Charlson Comorbidity Index scores had missing values in both arms, multiple imputation was used to generate propensity scores (that is, logistic regression predictions) for the incomplete data sets. The predictive models were pruned using the Akaike Information Criterion so that only variables associated with the outcome remained. Culture-negative infection was defined as a surgical procedure performed to eradicate suspected periprosthetic joint infection by the surgeon in which all aspiration or intraoperative culture results yielded no growth of the infecting organism. This resulted in a cohort of 159 patients with culture-negative periprosthetic joint infections. Although there were no salient differences observed among the three groups using an analysis of variance (Table I), the Charlson Comorbidity Index for the culture-negative group was lower in the pre-multiple imputation analysis.

Back to Top | Article Outline

Results

Overall, 10.3% (108 of 1,045) of periprosthetic joint infections treated at our institution were polymicrobial in nature. Patients with a polymicrobial periprosthetic joint infection had a higher failure rate compared with the matched monomicrobial and culture-negative periprosthetic joint infection cohorts (p = 0.003) (Fig. 1). The overall treatment failure rate was 50.5% (48 of 95) for patients with polymicrobial periprosthetic joint infection compared with 31.5% (63 of 200) for patients with monomicrobial periprosthetic joint infection and 30.2% (48 of 159) for patients with culture-negative periprosthetic joint infection. The survivorship for polymicrobial periprosthetic joint infection, with treatment success as an endpoint, was 52.2% (95% confidence interval [CI], 42.1% to 62.3%) at the 2-year follow-up, 49.3% (95% CI, 39.0% to 59.6%) at the 5-year follow-up, and 46.8% (95% CI, 35.9% to 57.7%) at the 10-year follow-up.

When stratified by treatment, survivorship for polymicrobial periprosthetic joint infection was 55.4% (95% CI, 48.3% to 62.5%) at 2 years, 49.3% (95% CI, 41.8% to 56.8%) at 5 years, and 49.3% (95% CI, 41.8% to 56.8%) at 10 years for two-stage exchange arthroplasty and 43.2% (95% CI, 35.1% to 51.3%) at 2 years, 43.2% (95% CI, 35.1% to 51.3%) at 5 years, and 38.4% (95% CI, 29.9% to 46.9%) at 10 years for irrigation and debridement. Although this result was not significant, there was a trend toward higher treatment success (p = 0.164) for two-stage exchange compared with irrigation and debridement (Fig. 2).

Comparing the monomicrobial, polymicrobial, and culture-negative periprosthetic joint infections treated with two-stage exchange arthroplasty, survivorship with treatment failure as an endpoint was the lowest for polymicrobial periprosthetic joint infection (p = 0.019) (Fig. 3); however, there was no significant difference for the success of irrigation and debridement among the three groups (p = 0.479) (Fig. 4).

Furthermore, patients with a polymicrobial periprosthetic joint infection had a higher rate of amputation (odds ratio [OR], 3.80 [95% CI, 1.34 to 10.80]; p = 0.012), arthrodesis (OR, 11.06 [95% CI, 1.27 to 96.00]; p = 0.029), and periprosthetic joint infection-related mortality (OR, 7.88 [95% CI, 1.60 to 38.67]; p = 0.011) compared with patients with a monomicrobial periprosthetic joint infection. However, there was no significant difference in the rate of retained spacers (OR, 1.23 [95% CI, 0.58 to 2.62]; p = 0.588) or resection arthroplasties (OR, 4.35 [95% CI, 0.78 to 24.19]; p = 0.093) between the patient cohorts (Table II).

In the univariate risk factor analysis, the presence of a sinus tract (OR, 1.93; p < 0.01), isolation of enterococci (OR, 11.36; p < 0.01), isolation of Escherichia coli (OR, 6.55; p < 0.01), infection with atypical organisms (OR, 9.85; p < 0.01), and infection with gram-negative organisms (OR, 6.33; p < 0.01) were significantly associated with polymicrobial periprosthetic joint infection. However, there was no difference between the polymicrobial cohort and the monomicrobial cohort in terms of likelihood of infection with other organisms or the presence of intraoperative purulence (Table III). In the multivariate analysis, the only factor found to be a significant independent risk factor for the development of polymicrobial periprosthetic joint infection was the presence of a sinus tract (OR, 2.20 [95% CI, 1.39 to 3.47]; p = 0.001); no other patient characteristics (including involved joint, sex, age, and BMI) or other comorbidities were identified as significant risk factors.

Back to Top | Article Outline

Discussion

This comparative study revealed that a periprosthetic joint infection caused by multiple organisms, otherwise known as a polymicrobial periprosthetic joint infection, had inferior outcomes compared with those of a monomicrobial periprosthetic joint infection or a culture-negative periprosthetic joint infection at our institution. The cohorts were matched for many important variables that can influence the outcome of surgical treatment of periprosthetic joint infection. Patients with a polymicrobial periprosthetic joint infection were more likely to require salvage operations such as an amputation or arthrodesis. Polymicrobial periprosthetic joint infection also resulted in a higher periprosthetic joint infection-related mortality rate in this cohort.

Although optimal management of patients with polymicrobial periprosthetic joint infection remains unknown, two-stage exchange arthroplasty appears to have a higher treatment success. Our study also found that the presence of a sinus tract was a significant risk factor for polymicrobial periprosthetic joint infection.

The success of surgical treatment of polymicrobial periprosthetic joint infection at our institution was worse than what has been previously reported by Marculescu and Cantey. The latter study also found decreased 2-year cumulative probability of success for patients with a polymicrobial periprosthetic joint infection at 63.8% compared with those with a monomicrobial periprosthetic joint infection at 72.8%5. This difference in the success rate between the study by Marculescu and Cantey and the present study likely relates to the difference in the definition of success. We used the Delphi definition of success for treatment that is more stringent than success being defined as retention of the prosthesis.

When examining the risk factors, we found that the presence of a sinus tract was significantly associated with polymicrobial periprosthetic joint infection. The reason for this finding is that the lack of soft-tissue integrity may allow for entry of organisms into the joint. In this study, using a multivariate analysis, we also found that no other preexistent comorbidities predisposed patients to developing a polymicrobial periprosthetic joint infection. This finding is different compared with that of a prior study, which suggested a link between a preexistent comorbidity, namely rheumatoid arthritis, and polymicrobial periprosthetic joint infection, on the basis of a univariate analysis4.

Interestingly, we found that patients with a polymicrobial periprosthetic joint infection were more likely to be infected with certain organisms, including enterococcal species, E. coli, and gram-negative organisms. Peel et al. found a similar association between polymicrobial periprosthetic joint infection and infection with Enterococcus and gram-negative bacilli4. Thus, the treating surgeon should be aware of this association and should treat the patient with appropriate antibiotics that are likely to cover the aforementioned organisms. For instance, adequate coverage may be attained by the use of a first-generation cephalosporin, cefazolin (which offers excellent coverage of gram-positive bacilli and some gram-negative bacilli, including E. coli), and gentamicin (which offers excellent gram-negative coverage and synergizes with cefazolin to cover Enterococcus)12. The type of organisms isolated from polymicrobial periprosthetic joint infection, including some atypical organisms, may be reflective of the immune status of these patients. Further investigation into the immune function of these individuals may reveal other factors that we did not elucidate in this study. The higher failure rate for polymicrobial periprosthetic joint infection may relate to inadequate antibiotic treatment. Although all patients in our study received appropriate antibiotics to cover the isolated organisms, it is possible that some organisms may have escaped isolation by culture and hence were not covered by the administered antibiotics. Another reason for the higher failure rate of polymicrobial periprosthetic joint infection may relate to the poor status of the soft tissues in these patients, as more patients with a polymicrobial periprosthetic joint infection had a sinus tract.

Although patients with a polymicrobial periprosthetic joint infection had a higher failure rate compared with the matched monomicrobial and culture-negative periprosthetic joint infection cohorts, the rates in these other cohorts were also alarmingly high7,13. Infection, regardless of the type, appears to be a devastating complication for both the patient and the surgeon.

There were several limitations to this study. First, the study was retrospective in nature and thus was subject to its inherent limitations and biases. Although, to our knowledge, this study had one of the largest cohorts in the literature, the number of patients included may have been inadequate for conducting some statistical analyses, and the possibility of a type-II error existed. The patients with a culture-negative periprosthetic joint infection had lower Charlson Comorbidity Index scores, which may, in part, have overestimated the true relative success when compared with patients with a polymicrobial periprosthetic joint infection. Lastly, functional outcome measures could not be routinely obtained. In our attempts to call patients for this measure, we found it extremely difficult to ascertain functional outcomes because of the high morbidity and mortality of this population. However, we did use the Delphi definition for success that takes into account many facets of outcome other than defining success as the ability to retain the prosthesis.

Our study reveals that polymicrobial periprosthetic joint infection, which occurred at a relatively low rate, was associated with a poor outcome when compared with monomicrobial periprosthetic joint infection and culture-negative periprosthetic joint infection. Patients with these infections were more likely to require a salvage operation. In line with prior studies, polymicrobial periprosthetic joint infection was associated with soft-tissue defects such as the presence of a sinus tract. One of the most important findings of this study is that certain types of organisms were likely to be isolated from polymicrobial periprosthetic joint infection, and this fact should be considered when administering antibiotics to these patients.

Investigation performed at The Rothman Institute of Orthopaedics at Thomas Jefferson University Hospital, Philadelphia, Pennsylvania

Disclosure: No external funding source was utilized. On the Disclosure of Potential Conflicts of Interest forms, which are provided with the online version of the article, one or more of the authors checked “yes” to indicate that the author had a relevant financial relationship in the biomedical arena outside the submitted work.

Back to Top | Article Outline

References

1. Pulido L, Ghanem E, Joshi A, Purtill JJ, Parvizi J. Periprosthetic joint infection: the incidence, timing, and predisposing factors. Clin Orthop Relat Res. 2008 ;466(7):1710–5. Epub 2008 Apr 18.
2. Holleyman RJ, Baker P, Charlett A, Gould K, Deehan DJ. Microorganisms responsible for periprosthetic knee infections in England and Wales. Knee Surg Sports Traumatol Arthrosc. 2015 . [Epub ahead of print].
3. Moran E, Masters S, Berendt AR, McLardy-Smith P, Byren I, Atkins BL. Guiding empirical antibiotic therapy in orthopaedics: the microbiology of prosthetic joint infection managed by debridement, irrigation and prosthesis retention. J Infect. 2007 ;55(1):1–7. Epub 2007 Mar 6.
4. Peel TN, Cheng AC, Buising KL, Choong PFM. Microbiological aetiology, epidemiology, and clinical profile of prosthetic joint infections: are current antibiotic prophylaxis guidelines effective? Antimicrob Agents Chemother. 2012 ;56(5):2386–91. Epub 2012 Feb 6.
5. Marculescu CE, Cantey JR. Polymicrobial prosthetic joint infections: risk factors and outcome. Clin Orthop Relat Res. 2008 ;466(6):1397–404. Epub 2008 Apr 18.
6. Jackson WO, Schmalzried TP. Limited role of direct exchange arthroplasty in the treatment of infected total hip replacements. Clin Orthop Relat Res. 2000 ;381:101–5.
7. Zmistowski B, Karam JA, Durinka JB, Casper DS, Parvizi J. Periprosthetic joint infection increases the risk of one-year mortality. J Bone Joint Surg Am. 2013 ;95(24):2177–84.
8. Peel TN, Cheng AC, Lorenzo YP, Kong DCM, Buising KL, Choong PFM. Factors influencing the cost of prosthetic joint infection treatment. J Hosp Infect. 2013 ;85(3):213–9. Epub 2013 Sep 4.
9. Zmistowski B, Della Valle C, Bauer TW, Malizos KN, Alavi A, Bedair H, Booth RE, Choong P, Deirmengian C, Ehrlich GD, Gambir A, Huang R, Kissin Y, Kobayashi H, Kobayashi N, Krenn V, Drago L, Marston SB, Meermans G, Perez J, Ploegmakers JJ, Rosenberg A, Simpendorfer C, Thomas P, Tohtz S, Villafuerte JA, Wahl P, Wagenaar FC, Witzo E. Diagnosis of periprosthetic joint infection. J Arthroplasty. 2014 ;29(2)(Suppl):77–83. Epub 2013 Dec 15.
10. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–83.
11. Diaz-Ledezma C, Higuera CA, Parvizi J. Success after treatment of periprosthetic joint infection: a Delphi-based international multidisciplinary consensus. Clin Orthop Relat Res. 2013 ;471(7):2374–82. Epub 2013 Feb 26.
12. Bourque M, Quintiliani R, Tilton RC. Synergism of cefazolin-gentamicin against enterococci. Antimicrob Agents Chemother. 1976 ;10(1):157–63.
13. Gomez MM, Tan TL, Manrique J, Deirmengian GK, Parvizi J. The fate of spacers in the treatment of periprosthetic joint infection. J Bone Joint Surg Am. 2015 ;97(18):1495–502.
Copyright 2016 by The Journal of Bone and Joint Surgery, Incorporated