Introduction

Total knee and hip arthroplasty have become the standard treatment in end-stage osteoarthritis because these methods reduce pain and improve the functionality of the joint and occasionally the quality of life [1]. Prosthetic joint infection (PJI) is one of the most dreaded complications and occurs in 1–2% of operated patients [2]. PJIs are associated with complex management and an increase in morbidity and mortality [3]. The overall cost to treat bone and joint infections was evaluated to €259 millions to the French health care system in 2008, which presents a major economic burden [3].

Enterococci are Gram-positive facultative anaerobic cocci and are commensals of human and animal digestive microbiota. They have been traditionally implicated in bloodstream infections, endocarditis or urinary tract infections. Enterococci are naturally resistant to many antibiotics, including cephalosporins and aminoglycosides [4]. Enterococcus spp. account for 3–10% of PJIs [5,6,7,8]. Enterococcus faecalis is responsible for the vast majority (90%) of enterococci PJI, but it often occurs as part of a polymicrobial infection [6, 9]. The rate of treatment failure was higher in PJI patients infected with Enterococcus spp. than in patients infected with other species [10, 11]. The polymicrobial nature of these infections, the formation of biofilms and the increased patient age and number of comorbidities could explain this worse outcome [9, 12,13,14,14].

Because of its rarity, the optimal management of PJI due to enterococci remains poorly understood, and guidelines on medico-surgical therapy against these infections are mostly based on the opinions of expert committees [15,16,17]. Given this lack of data, it becomes necessary to carry out additional studies on the treatment of PJI caused by Enterococcus spp. to improve the prognoses of these infected patients.

We therefore conducted a retrospective cohort study and asked the following: (1) Does combination therapy improve outcomes compare to monotherapy in ODRIs due to Enterococcus spp.? (2) Does an antibiotic targeting the biofilm, such as rifampicin, improve the prognosis of patients?

Materials and methods

Study design and participants

We performed a retrospective cohort study in the University Hospital of Caen (France), one of the French reference centres for the management of complex Bone and Joint Infections. All patients with PJI or osteosynthesis-associated infections caused by enterococci from January 2013, to July 2020 were included. Patients who died within 5 days after surgery or who were in palliative care were excluded. Because the diagnosis, management and prognosis are different from other osteosynthesis, we also excluded patients who had osteosynthesis of the hand, foot or vertebra. All patients were followed until July 2020. Cases were first defined using microbiology laboratory data (TD Nexlabs, TECHNIDATA, France). This study was performed using the principles of the Declaration of Helsinki. Approval was granted by the ethics committee of the biomedical research institute of Caen (ID 3164, 03/23/2022).

Definition and outcome

The diagnosis of PJI was defined according to the Philadelphia consensus conference guidelines published in the 2018 “Proceedings of the International Consensus Meeting on Orthopaedic Infections”, which were (i) two positive growths of the same organism using standard culture methods or (ii) sinus tract with evidence of communication to the joint or visualization of the prosthesis or (iii) a combined score ≥ 6 as assessed based on an increase in serum C-reactive protein (CRP) (2 points) or sedimentation rate (1 point), elevated synovial white blood cell count (WBC) (3 points) or polynuclear neutrophils (2 points), histology in favour of an infection (3 points), single peri-sample prosthetic positive (2 points) or positive intraoperative purulence (3 points) [18]. Because of the absence of consensus regarding the definition of osteosynthesis-associated infection, we used similar criteria to those of the PJI definition for the diagnosis of osteosynthesis material infection. Osteoarticular infection was distributed according to the Tsukayama classification: early postoperative infection (< 4 weeks), late postoperative infection (> 4 weeks) or haematogenous origin [19]. Patients with early postoperative infection or haematogenous infection were planned to receive debridement, antibiotics and implant retention (DAIR) procedure via open arthrotomy, with exchange of mobile components if available, followed by antibiotic treatment. Those with late postoperative infection underwent whether 1-stage exchange consisting of infected implant removal and reimplantation of a material followed by antibiotic treatment, or 2-stage exchange procedure consisting of infected implant removal and antibiotic-loaded spacer placement followed by an antibiotic treatment course prior to prosthetic implant replacement. Decision between 1-stage or 2-stage exchange depended on many factors including joint age, pathogens resistance, medical comorbidities, quality of the periprosthetic soft tissue and patient’s preference. Same approach was applied for osteosynthesis associated infection: debridement with implant retention for acute infection, 1-stage exchange or implant removal for chronic infection. Six to 12 weeks of postoperative pathogen-specific antimicrobial therapy following the Philadelphia consensus conference and IDSA guidelines [17, 18] was given. The other factors influencing the choice of treatment were: the chronicity of the infection, pathogens antimicrobial resistance, surgery protocol or patient’s comorbidities. For each patient, the management of PJI or osteosynthesis-associated infection was chosen during a multidisciplinary staff with infectious disease specialists, orthopaedic surgeons and microbiologists. The follow-up period was calculated from the date of infection surgery. During the follow-up, patients underwent routine clinical, biological and radiological evaluations. Clinical success was defined by the absence of clinical failure, and was analyzed at 6 months, 1 year and 2 years. Relapse due to enterococci or other microorganisms, unplanned surgery or death related to PJI or osteosynthesis-associated infection were considered clinical failure. Relapse was defined according to the Philadelphia consensus criteria [18].

Data collection

Information regarding the demographics (age and gender), underlying conditions (chronic obstructive pulmonary disease (COPD), cirrhosis, diabetes mellitus, immunosuppressive treatment, chronic renal failure, tobacco, rheumatoid arthritis), initial surgery (total knee arthroplasty, total hip arthroplasty, osteosynthesis material), delay between the first surgery and the first symptoms of infection, clinical manifestations (fever, sinus tract, local symptoms, joint pain, endocarditis), biological investigations (CRP, WBC), radiological findings, microbiological data (enterococcus species, antimicrobial susceptibility testing, isolation of other microorganisms) and infection management, type of septic surgery (1-stage, 2-stage exchange, DAIR, removal of material), delay between infection and surgery, antibiotic regimen and outcome were collected using the informatics medical records. Electronic medical records (Usv2-Crossway, McKesson, USA) were prospectively filled in by medical and paramedical staff.

Microbiological analysis

The standard microbiological protocol includes the culture of different samples (between 3 and 5 if possible). Briefly, all samples were inoculated on the following media: blood agar, chocolate agar and Schaedler broth with vitamin K3, all from BioMérieux (Marcy l’Etoile, France) under different atmosphere conditions (e.g., anaerobic and aerobic conditions). Biopsies were mechanically grind before to inoculation. Sonication of the removed implants was not used. All cultures performed in agar media were incubated for 5 days and those in Schaedler broths were incubated for 15 days. Species identification was carried out using a Maldi ToF technology (Microflex®, Brucker, Germany). Antimicrobial Susceptibility profiles were determined using the disk diffusion method, or VITEK® XL (bioMérieux, Marcy l’Etoile, France). Clinical susceptibility profiles were evaluated according to the French committee for antibiogram (CA-SFM/EUCAST) [20].

Statistical analysis

Categorical data are expressed as numbers and percentages, and continuous variables are reported as the means and standard deviations (SDs). Raw characteristics were compared between groups using Fisher’s exact test for categorical variables and Mann–Whitney U tests for continuous variables. A p-value less than 0.05 was considered significant and indicated in bold.; all p-values were two tailed. Statistical analyses were performed using R software (The R Foundation for Statistical Computing).

Results

Population and infection characteristics

A total of 36 patients were included, with 24 in the arthroplasty group and 12 in the osteosynthesis material group (Fig. 1). The demographic data and comorbidities of the patients are reported in Table 1. The mean age (SD) was 69.3 years [17], and 52% were male (n = 19). The most frequent underlying conditions were chronic renal failure (16.7%, n = 6), immunosuppressive treatment (13.8%, n = 5), diabetes mellitus (11.1%, n = 4) and rheumatoid arthritis (8.3%, n = 3). The most frequent infections were hip PJI (n = 19, 52.8%), osteosynthesis-associated infections located on the tibia or fibula (n = 11, 30.5%) and knee PJI (n = 4, 11.1%). Twenty-seven patients (75%) had a delayed postoperative infection with a mean delay (SD) of 42.1 months (90.2). Clinical symptoms were pain (71.4%, n = 25), sinus tract (57.1%, n = 20), local symptoms (37.1%, n = 13) and fever (25.7%, n = 9). The mean CRP (SD) level was 85.7 mg/l (93), the mean WBC (SD) was 8.0 G/l (3.0), and 7/33 patients (21.2%) had radiological signs of implant loosening. One patient (2.7%) had a diagnosis of endocarditis according to the Duke criteria [21]. Patients with osteosynthesis were younger (p = 0.002), more often smokers (p = 0.03) and had more local symptoms (p = 0.02), whereas patients with PJI had a longer delay between the last surgery and infection (p = 0.03) and more frequent implant loosening on X-ray (p = 0.03).

Fig. 1
figure 1

Flowchart of the study

Full size image
Table 1 Demographics, comorbidities and infection characteristics of 36 patients with enterococci PJI and osteosynthesis-associated infection
Full size table

Microbiological results

Table 2 shows microbiological data of the cohort. E. faecalis was the most common pathogen (30 isolates, 83.3%), while E. faecium was found in 13.9% (5 isolates). Isolates were susceptible to ampicillin in 90.6% (29/32), vancomycin in 97% (32/33), linezolid in 100% (27/27), levofloxacin in 87.5% (14/16) and rifampicin in 54.8% (17/31) of cases. Infections were polymicrobial for 23 patients (63.9%) with mostly methicillin-sensitive Staphylococcus aureus (MSSA) (56.5%, 13/23) or enterobacteria (52.2%, 12/23). Osteosynthesis-associated infections were more often polymicrobial than PJIs (p = 0.002) and were mostly associated with fermenting Gram-negative bacilli (p = 0.03).

Table 2 Microbiological data of 36 patients with enterococci PJI and osteosynthesis-associated infection
Full size table

Medico-surgical treatment and outcome

Medico-surgical management and outcomes of the 36 patients are described in Table 3. DAIR was performed in 11 cases (30.6%), withdrawal of material in 6 cases (16.7%), 1-stage exchange in 11 cases (30.6%) and 2-stage exchange in 8 cases (22.2%). Antibiotic treatment directed against enterococci included amoxicillin (41.6%, n = 15), fluoroquinolones (30.6%, n = 11), rifampicin (27.8%, n = 10), and/or linezolid (25.0%, n = 9). The duration of antimicrobial therapy was 6 weeks for 14 patients (38.9%) and 12 weeks for 19 patients (52.8%). Combination therapy was prescribed in 60.8% of polymicrobial (14/23) and in 38.5% of monomicrobial infections (5/13), mostly with amoxicillin and rifampicin (13.9%, 5/36) (Table 4). The mean follow-up (SD) was 33.8 months (28). The clinical success rates at 6 months, 1 year and 2 years were 76% (22/29), 67% (18/27) and 59% (13/22), respectively. No relapse due to enterococci was identified. Among 12 patients with osteosynthesis infection, 8 were recovering, 2 were amputated (4 and 18 months after septic surgery) and 2 were in pseudarthrosis and needed new surgery. Two-stage exchanges were more often performed in patients with PJI (p = 0.03), while device removal was more often performed in patients with osteosynthesis-associated infection (p = 0.0005).

Table 3 Medico-surgical treatment and outcome of 36 patients with enterococci PJI and osteosynthesis-associated infection
Full size table
Table 4 Antibiotic treatment and outcome of 36 patients with enterococci PJI and osteosynthesis-associated infection according to the mono or polymicrobial character
Full size table

Univariate analysis of factors associated with clinical failure

An univariate analysis of factors associated with clinical failure is presented in Table 5 and was carried out after excluding 3 patients who died of other causes, 3 who were lost to follow-up and 3 with a postoperative follow-up of less than 1 year. In this analysis, the only variable statistically associated with a reduced risk of clinical failure was a duration of antibiotic therapy of 12 weeks, with a p-value of 0.04. Lower BMI and lower age tended to decrease the risk of clinical failure but the difference was not statistically significant (p = 0.05 and 0.06 respectively). Combination therapy and the use of rifampicin were not associated with a decreased risk of clinical failure.

Table 5 Outcome of 27 patients with enterococci PJI and osteosynthesis-associated infection; univariate analysis
Full size table

Subgroup analysis

We performed an univariate analysis in the group of PJI and the group of osteosynthesis associated infection. We found that lower age (p = 0.02) and 12 weeks of antimicrobial therapy (p = 0.03) were factors significantly associated with clinical success in the PJI group while only lower age was statistically significant (p = 0.009) in the osteosynthesis associated infection group (Tables 6 and 7). The results are, therefore, consistent with the pooled analysis.

Table 6 Outcome of 17 patients with enterococci PJI; univariate analysis
Full size table
Table 7 Outcome of 10 patients with enterococci osteosynthesis-associated infection; univariate analysis
Full size table

Discussion

In our retrospective cohort, patients with ODRI due to enterococci had poor clinical success, and the combination of antimicrobial therapy did not impact the outcome. To the best of our knowledge, this is the first study to include osteosynthesis-associated infections caused by enterococci.

Enterococcus spp. is an opportunistic pathogen that is able to adapt its lifestyle to its environment and modulate the host's immune response to silently persist in multiple niches, such as in bones and joints [22]. Despite frequent coinfections, the clinical manifestations of our patients are reminiscent of infections with slow-growing bacteria such as coagulase-negative staphylococci or corynebacteria [23]. Infections were most often chronic with significant local symptoms contrasting with few systemic signs.

To our knowledge, seven other cohort studies focused on enterococcal PJI with a range of patients included from 31 to 203 [9, 11, 12, 23,24,25,26]. As previously described in bone and joint infections, enterococcal were often polymicrobial, and the most frequent copathogens identified were staphylococci or Enterobacteriales members. Interaction between enterococci and others microorganisms in polymicrobial infections is well known [27]. First, it appears that immune suppression mediated by E. faecalis may promote the virulence of co-infecting organisms during polymicrobial infection. Indeed, it was found that E. faecalis can limit phagocytosis and intracellular killing of other pathogens in a peritonitis model [28] and can suppress cytokine production of the host in an urinary tract infections model [29], allowing to promote colonization of copathogens such as Escherichia coli. Second, inter-species interactions between E. faecalis and S. aureus or P. aeruginosa lead to augmented biofilms biomass in vitro and may also explain the frequent co-occurrence of these microbes in PJI [30, 31]. Third, enterococcal biofilm can serve as gene reservoirs for antibiotic resistance transmission between species by horizontal gene transfer, as demonstrated by the transmission of vanA from E. faecium to S. aureus in a polymicrobial wound biofilm [32]. In clinical practice, polymicrobial PJI are complicated to manage by the need to target pathogens with different susceptibilities at the same time, such as Gram-negative and Gram-positive bacteria.

Enterococci PJIs are considered difficult to treat, as are infections caused by Candida spp. or Staphylococcus aureus. In our study, overall clinical success was only 67% at 1 year. Among the other 7 studies on enterococcal PJI, the overall success rate ranged from 51.7 to 83.7% [24, 25]. This high failure rate should be put into perspective with the definitions of failure. Renz et al. [24] found microbiological success (eradication of Enterococcus spp.) in 83.7% but clinical success (without new surgery for septic or aseptic goals) in only 67.5%. This gap was explained by infections with other microorganisms or with reintervention without documented infection. Among our 9 patients with clinical failures, 4 with new microorganisms were documented, and 5 were culture negative or without microbiological findings. Therefore, no cases of relapse with Enterococcus spp. were found. This high rate of clinical failure contrasting with a low rate of microbiological failure could be linked to co-founding factors such as comorbidities or copathogens.

Two-stage exchange is recommended by the IDSA for difficult-to-treat organisms such as methicillin resistant, Staphylococcus aureus, Candida spp. or enterococci [17]. This strategy involves the removal of prosthetic components and the debridement of infected tissue followed, within 2 weeks to several months after the first stage, by reimplantation of a new prosthesis in a second stage. This approach offers better clinical success but at the cost of multiple surgeries and a possible poorer functional outcome [17]. Among others studies focused on enterococcal PJI, overall success rate was ranged from 45.5 to 83.3% for 1-stage exchange and from 43.8 to 100% for 2-stage exchange [11, 12, 25]. In our study, only 6 patients benefited from 1-stage exchange, with a success rate of 100% (p = 0.07). Due to our small cohort of patients, we cannot draw any conclusions but believe that we must take into account not only microbiological findings in the choice of surgery but also other factors, such as age and stability of the prosthesis, quality of the periprosthetic soft tissues, patient comorbidities and preferences [17]. DAIR was the least effective strategy used in our study, with only 50.0% clinical success at 1 year. These results are consistent with those of other studies that report success rates ranging from 26 to 66% [11, 33]. Only the study by Renz et al. [24] found a success rate of 100% with DAIR, which can be explained by a drastic selection of patients with uncomplicated acute infections. This high failure rate with DAIR can be explained by a therapeutic delay during haematogenous infections and by a lack of knowledge about antibiofilm antibiotics targeting enterococci, which is an essential treatment when the implant remains in place. Further studies are needed to assess the possibility of DAIR and 1-stage exchange in enterococcal PJI.

Although our study was underpowered to detect differences, notably due to the heterogeneity of antimicrobial therapy, a better outcome was not demonstrated for patients with combination therapy and rifampicin in our study. For PJI caused by enterococci, the appropriate antimicrobial therapy remains unclear in the literature. One of the treatments of choice is amoxicillin in monotherapy as recommended by several Infectious Diseases Societies [15,16,17]. Although E. faecalis is susceptible to ampicillin, the low bactericidal activity of antibiotics targeting the cell wall, such as β-lactams, and the high rate of treatment failure have led to the emergence of antibiotic combination therapy [34]. Combination treatments with amoxicillin and gentamicin or ceftriaxone are based on experimental studies showing synergistic activity and are modelled on the management of enterococcal endocarditis [21, 35]. Because rifampicin is a key treatment for staphylococcal PJI notably due to it anti-biofilm activity [17, 36, 37], amoxicillin and rifampicin association is proposed by the French and Spanish Infectious Diseases committees for enterococcal PJI management [15, 16, 38]. In a retrospective cohort study that included 50 episodes of PJI due to enterococci, El Hellou et al. [23] did not find any difference in the outcome between those treated with monotherapy and those receiving combination therapy. Two studies, including the one with the largest case series to date, found better outcomes with rifampicin use [9, 12]. Indeed, Tornero et al. [9] showed in a multicentric cohort study that the use of rifampicin was significantly associated with early infections and increase in the cure rate (60%, p = 0.04). Thompson et al. [12] reported in a study of 49 patients with enterococcal PJI that there was better clinical success in patients receiving rifampicin in combination with another antibiotic (100%, p = 0.04). Other combinations of antibiotics, especially with daptomycin, fosfomycin or linezolid, are emerging treatments for enterococcal bloodstream infections and endocarditis but require further evaluation in PJI [34].

A consistent finding of our study was the association between lower age and clinical success. This result was not found in other studies on enterococcal PJI [9, 11, 12, 23,24,25,26] but higher risk of adverse outcome has been reported after total joint arthroplasty revision in older patients [39] as well as an increased mortality rate related to PJI [40]. Many factors could explain this higher success rate. First, immunosenescence, defined as the decline of the immune function during aging, can result in a compromised immune response against pathogens and greater susceptibility to infections [41]. Second, elderly patients may have geriatric comorbidities (such as incontinence, immobility, lower limb ulcers, arteriopathy, osteoporosis and/or malnutrition) which could contribute to clinical failure due to delayed healing, increased bacterial contamination of the scare or/and increased risk of periprosthetic fracture [42].

Our study has several limitations. First, it is a retrospective single-centre study, confounding factors may exist, and the data cannot be extrapolated to other centres. Second, the number of patients included is small, and unfortunately, almost 20% of patients had incomplete follow-up. In a supplementary analysis considering lost to follow-up as failure, no factor was associated with clinical success (Table S1). However, results of a recent randomized controlled trial showed that antibiotic therapy for 6 weeks was not shown to be noninferior to antibiotic therapy for 12 weeks and resulted in a higher percentage of patients with unfavorable outcomes [43]. Therefore, we suggest treating PJI enterococcal infection with 3 months of antimicrobial therapy. Third, one-year period after septic surgery is too short to judge clinical success, and some patients may have relapsed after the duration of the follow-up. Finally, the inclusion of patients with osteosynthesis-associated infection made our cohort more heterogeneous, which may bias the analysis of the results. However, the inclusion of patients with osteosynthesis-associated infection allowed us to increase the size of our cohort and represent the real life of a complex bone and joint infections centre.

In conclusion, the prognosis of ODRI due to Enterococcus spp. is poor, as previously reported. In our small study, patients with combination therapy and rifampicin use did not demonstrate better clinical success. Further studies are needed to improve knowledge of these infections, which are difficult to treat.