Aim

The primary aim of this in vitro study was to test the efficacy of daptomycin to eradicate staphylococcal biofilms on various orthopedic implant surfaces and materials. The secondary aim was to quantitatively estimate the formation of staphylococcal biofilm on various implant materials with different surface properties.

Aims

To investigate the outcomes of treatment of streptococcal periprosthetic joint infection (PJI) involving total knee and hip arthroplasties.

Aim

Clear differentiation between aseptic failure and prosthetic joint infection remains one of the goals of modern orthopaedic surgery. New diagnostic methods can provide more precise evaluation of the etiology of prosthetic joint failure. With the introduction of sonication an increasing number of culture-negative prosthetic joint infection were detected. The aim of our study was to evaluate culture-negative prosthetic joint infections in patients who were preoperatively evaluated as aseptic failure.

Aim

To evaluate antimicrobial activity of Sb-1 and Pyo-bacteriophage in preventing and eradicating MRSA biofilm in vitro using isothermal micro calorimetry.

Aim

To investigate the antimicrobial activity of a gentamicin-loaded bone graft substitute (GLBGS) in the prevention and eradication of bacterial biofilms associated with prosthetic joint infections (PJI).

Introduction

Despite the lack of data regarding the diagnostic validity of synovial aspiration in Girdlestone hips a Girdlestone-aspiration is often performed before reimplantation to detect a possible persistence of infection during two staged revision total hip arthroplasty (THA). The aim of this study was to assess the diagnostic performance of the synovial aspiration in Girdlestone hips, without a PMMA-Spacer, for the detection of infection persistence prior to THA reimplantation.

Background

The main reasons for hip prosthesis failure are aseptic loosening and periprosthetic joint infection (PJI). The real frequency of PJI is probably largely underestimated because of non-standardized definition criteria, diagnostic procedure, treatment algorithm and other confounders. Therefore, data from joint registries are not reflecting the frequency of PJI and can be misleading; particularly low-grade PJI can be frequently misdiagnosed as aseptic failure. Therefore, prospective clinical studies with standardized protocol, comprehensive diagnostic procedure and sufficient follow-up should be performed. Sonication of explanted prosthesis is highly sensitive for detection of biofilms on prosthetic surface and allows quantitative analysis of biofilm formation. We hypothesize that by using sonication, ceramic components (BIOLOX®delta, BIOLOX®forte) will show higher resistance against biofilm adhesion compared to polyethylene (PE) and metal (CoCrMo).

The aim of the present study was to assess the antibiofilm activity of daptomycin- and vancomycin-loaded poly(methyl methacrylate) (PMMA) and PMMA-Eudragit RL100 (EUD) microparticles against mature biofilms of polysaccharide intercellular adhesin-positive S. epidermidis.

The effect of plain, daptomycin- and vancomycin-loaded PMMA and PMMA-EUD microparticles on S. epidermidis biofilms was assessed by isothermal microcalorimetry (IMC) and fluorescence in situ hybridization (FISH). Biofilms were grown for 48h onto poly-urethane pieces of fixed dimensions. Each sample was washed with PBS in order to remove planktonic bacteria and incubated for 24h with different concentrations of acrylic microparticles (20–1.25 mg/mL). The minimal biofilm inhibitory concentration (MBIC) of the antibiotic-loaded particles was defined as the lowest concentration of particles that was able to prevent heat flow associated to the recovery of the biofilms. After incubation with the microparticles, sessile cocci were hybridized with the pan-bacterial EUB338-FITC and the staphylococci-specific STAPHY-FICT probes and stained with DAPI. Biofilm structure and metabolic state were characterized by fluorescence microscopy.

According to the IMC results, plain PMMA-particles showed no effect on S. epidermidis biofilms, whereas PMMA-EUD-microparticles negatively influenced the recovery of the biofilm probably due to the highly positive charge of these particles. The MBIC of daptomycin-loaded PMMA-microparticles was 20 mg/mL, whereas vancomycin-loaded PMMA microparticles were not able to inhibit biofilm recovery. Adding EUD to the formulation reduced the MBIC of daptomycin-loaded microparticles to 1.25 mg/mL, corresponding to a 16-fold reduction. Regarding the vancomycin-loaded microparticles, EUD caused a further decrease of their antibiofilm activity. The FISH micrographs corroborated the IMC results and provided additional insights on the antibiofilm effect of these carriers. According to FISH, daptomycin-loaded PMMA-EUD microparticles were responsible for the most pronounced reduction in biofilm mass. In addition, FISH showed that both PMMA and PMMA-EUD microparticles were able to attach to the biofilms.

Adding EUD to the formulations proved to be a powerful strategy to improve daptomycin-loaded microparticles antibiofilm activity. In addition, the combination of IMC and FISH was essential in order to fully assess the effect of polymeric microparticles on sessile S. epidermidis. Although the present study enabled gaining further insights on this subject, the nature of these interactions remains unclear. However, this may be a crucial aspect for the enhancement of antibiofilm activity of antibiotic-loaded polymeric microcarriers against mature biofilms.

This work was supported by the Portuguese government (Fundação para a Ciência e a Tecnologia) and FEDER (grant SFRH/BD/69260/2010 and research project EXCL/CTM-NAN/0166/2012) and strategic project PEst-OE/SAU/UI4013/2011.

Objectives

Establishing the diagnosis of implant-associated infections is often difficult, because of variable clinical presentations and lack of uniform diagnostic criteria. Sonication of removed orthopedic devices was shown to have superior sensitivity and specificity for infection. We evaluated the value of microcalorimetry as a quick and reliable tool in the diagnosis of infection in sonication fluid from removed implants.

Objectives

The risk of infection after type III° open fractures is high (10–50%). Preemptive antibiotic therapy may prevent posttraumatic infection and improve the outcome. Recommendations about the type and duration of antibiotic vary among the institutions and it remains unclear whether gram-negative bacilli or anaerobs need to be covered.

In Europe, the most commonly recommended antibiotic is amoxicillin/clavulanic acid.

We retrospectively analyzed microbiology, characteristics and outcome of patients with open type III° fractures treated at our institution.

Objectives: Total ankle replacement (TAR) is increasingly used for treatment of primary or posttraumatic arthritis of the ankle joint, if joint movement is intended to be preserved. Data on characteristics and treatment of ankle prosthetic joint infection (PJI) is limited and no validated therapeutic algorithm exist. Therefore, we analyzed all infections, which occurred in a cohort of implanted ankle prostheses during a 5-year-period.

Methods: Between 06/2004 and 12/2008, all patients with an implanted ankle prosthesis at our institution were retrospectively reviewed. All patients were operated by the same surgical team. Ankle PJI was defined as visible purulence, acute inflammation on histopathology, sinus tract, or microbial growth in periprosthetic tissue or sonication fluid of the removed prosthesis. The surgery on the infected ankle prosthesis and the follow-up were performed by the surgical team, who implanted the prosthesis. A specialized septic team consisting of an orthopaedic surgeon and infectious diseases consultant were included in the treatment.

Results: During the study period, 92 total ankle prostheses were implanted in 90 patients (mean age 61 years, range 28–80 years). 78 patients had posttraumatic arthritis, 11 rheumatoid arthritis and 3 other degenerative disorder. Ankle PJI occurred in 3 of 92 TAR (3.3%), occurring 1, 2 and 24 months after implantation; the causative organisms were Enterobacter cloacae, Streptococcus pyogenes and Staphylococcus epidermidis, respectively. The ankle prosthesis was removed in all infected patients, including debridement of the surrounding tissue was debrided and insertion of an antibiotic loaded spacer. Provisional arthrodesis was performed by external fixation in two patients and by plaster cast in one. A definitive ankle arthrodesis with a retrograde nail was performed 6 to 8 weeks after prosthesis removal. One patient needed a flap coverage. All 3 patients received intravenous antibiotic treatment for 2 weeks, followed by oral antibiotics for 4–6 weeks. At follow-up visit up to 18 months after start of treatment, all patients were without clinical or laboratory signs of infection.

Conclusions: The infection incidence after TAR was 3.3%, which is slightly higher than reported after hip (< 1%) or knee arthroplasty (< 2%). A two-step approach consisting of removal of the infected prosthesis, combined with local and systemic antibiotic treatment, followed by definitive ankle arthrodesis shows good results. Larger patient cohort and longer follow-up evaluation is needed to define the optimal treatment approach for ankle PJI.

Implant-associated infections do not spontaneously cure. The reason for persistence in device-associated infections is the biofilm, a specialized form of bacterial growth on surfaces. The biofilm represents a survival form of bacteria which is highly resistant against most antibiotics, and can persist over months or years as low-grade infection. Bacteria in biofilms enter a metabolically inactive state, embedded in an amorphous substance, called biofilm matrix. Together they form a complex three-dimensional structure with rudimentary communication and circulation systems.

As a rule, only a combined surgical and antimicrobial management can eradicate biofilms and cure implant-associated infection. In selected patients, implant infections can be cured without implant removal with early debridement and long-term antibiotic treatment acting against biofilms. In this presentation, common pitfalls and reasons for treatment failure will be outlined and discussed.

Adequate diagnosis. Every prosthesis with early loosening (< 2 years after implantation) or persistent/increasing pain should rise the suspicion of an implant-associated infection, which may presents without systemic signs of infection (fever), and can manifest without increased leukocyte counts or C-reactive protein). In this case, periimplant infection must be actively searched. Preoperative diagnosis includes aspiration of the synovial fluid of larger joints for Gram stain and culture; cell count and differential; and crystals (if appropriate). Swabs of the sinus tract (fistula) should not be performed, since they may not represent the causing organisms of the bone- or implant-associated infection. Intraoperative diagnosis includes multiple tissue biopsies around the prosthesis (no swabs), histopathology of the periprosthetic tissue and sonication of the removed device or its parts (if the method is available). For low-virulent organisms (such as Staphylococcus epidermidis or Propionibacterium acnes) at least two positive specimens with the same organism or an additional criterion for infection (i.e. acute inflammation in the tissue histopathology, presence of intraoperative pus or fistula) is needed. For slow-growing organisms, long incubation is needed (14 days) in order to culture these slow-growing and fastidious pathogens. Antibiotics should be discontinued at least 2 weeks prior to culture sample in order to minimize false-negative results. Novel diagnostic tests include PCR, which can also detect non-growing microorganisms and can be used for tissue specimens, joint aspirate or sonication fluid.

In conclusion, using the proposed diagnostic and treatment algorithm (Zimmerli et al. NEJM2004; 351: 1645), the long-term success rate of periprosthetic joint infection is 80–90%. To avoid failure, it is important to select the patient for the least invasive, but still appropriate surgical and antibiotic treatment regimen and to avoid mistakes.

Implants are highly susceptible to infection [1]. The infection rate is 0.5–2% after hip or knee arthroplasty [2]. Periprosthetic joint infections are hard to eradicate. Until recently, two-stage exchange or life-long antibiotic suppression therapy has been the rule. However, novel treatment concepts evolved. Whereas no controlled trials on the surgical management of prosthetic-joint associated infection are available, much is known about the optimal antimicrobial therapy. Modern treatment concepts are based on animal studies, in vitro data, observational studies and one single controlled trial comparing two different antimicrobial regimens in patients with orthopaedic device-associated infections [2]. The tissue-cage guinea pig model is especially suitable to analyse the role of antibiotics in device-related infections [3]. In this model, different aspects of implant-associated infections have been studied, namely

host-defense mechanisms around implants [1],

Taken together, these experiments showed that an agent acting on slow-growing and adhering microorganism is needed to eradicate device-associated infection. This requirement is only fulfilled by rifamycins in staphylococcal infection and by fluoroquinolones in infections caused by gram-negative bacilli [5–12]. We could show that even new drugs against methicillin-resistant staphylococci (MRSA), such as linezolid and daptomycin, were not able to eradicate experimental foreign-body infections, when used as single agents. In contrast, in combination with rifampin linezolid cured 60%, and daptomycin 67% of the implant-associated MRSA-infections [9,10]. Based on these in vitro and animal data, observational studies were performed in patients with staphylococcal infections associated to orthopedic devices. We developed an algorithm allowing to choose the optimal therapeutic option for each patient [2]. This algorithm has been validated in cohort studies [13,14]. Finally, in a randomized double-blind controlled trial, the benefit of a rifampin-quinolone combination as compared to quinolone monotherapy has been proven in patients with acute orthopedic implant associated staphylococcal infection undergoing debridement with retention [15].

In conclusion, the favorable role of rifampin has been proven in vitro, in animals and in human studies. Also the newest antistaphylococcal agents must be given in combination with rifampin in order to eliminate infection without removal of the device.

Background: Bacteria form a biofilm on the surface of orthopaedic devices, causing persistent and infection. Little is known about biofilms formation on bone grafts and bone substitutes. We analyzed various representative materials regarding their propensity for biofilm formation caused by Staphylococcus aureus.

Methods: As bone graft beta-tricalciumphosphate (b-TCP, CyclOsTM) and as bone substitute a tantalum metal mesh (trabecular metalTM) and PMMA (Pala-cosTM) were investigated. As test organism S. aureus (strain ATCC 29213) was used. Test materials were incubated with bacterial solution of 105 colony-forming units (cfu)/ml at 37°C for 24 h without shaking. After 24 h, the test materials were removed and washed 3 times in normal saline, followed by sonication in 50 ml Ringer solution at 40 kHz for 5 minutes. The resulting sonication fluid was plated in aliquots of 0.1 ml onto aerobe blood agar with 5% sheep blood and incubated at 37°C with 5% CO2 for 24 h. Then, bacterial counts were enumerated and expressed as cfu/ml. All experiments were performed in triplicate to calculate the mean ± standard deviation. The Wilcoxon test was used for statistical calculations.

Results: The three investigated materials show a differing specific surface with b-TCB> trabecular metal> PMMA per mm2. S. aureus formed biofilm on all test materials as confirmed by quantitative culture after washing and sonication. The bacterial counts in sonication fluid (in cfu/ml) were higher in b-TCP (5.1 x 106 ± 0.6 x 106) and trabecular metal (3.7 x 106 ± 0.6 x 106) than in PMMA (3.9 x 104 ± 1.8 x 104), p< 0.05.

Conclusion: Our results demonstrate that about 100-times more bacteria adhere on b-TCP and trabecular metal than on PMMA, reflecting the larger surface of b-TCP and trabecuar metal compared to the one of PMMA. This in-vitro data indicates that bone grafts are susceptible to infection. Further studies are needed to evaluate efficient approaches to prevent and treat infections associated with bone grafts and substitutes, including modification of the surface or antibacterial coating.

Background: Antibiotic-loaded spacers and cement nails are commonly used in patients undergoing a two-stage implant exchange procedure for treatment of prosthetic joint infection (PJI). During re-implantation 2–6 weeks after implant removal, tissue specimens are collected to document successful eradication of infection. However, these specimens have limited sensitivity, especially in patients receiving antimicrobial treatment. We investigated the value of sonication of removed spacers and cement nails.

Methods: We prospectively included patients in whom a spacer or cement nail was removed from January 2007 through April 2009 during a two-stage exchange procedure. The removed temporary device was sonicated in a container with Ringer’s solution in an ultrasound bath for 5 min at 40 kHz (as described in NEJM2007;357:654). The resulting sonication fluid was cultured aerobically and anaerobically for 10 days. In parallel, > 2 tissue samples were collected for conventional cultures on blood agar plates and enrichment broth. PJI was defined as visible purulence, acute inflammation on histopathology, sinus tract or significant microbial growth in tissue or implant sonication cultures.

Results: In this ongoing study, 28 spacers and 10 cement nails from patients with confirmed PJI were included (median age 75 y; range 49–86 y). All devices were impregnated with antibiotics (gentamicin and/or vancomycin) and were placed in the hip (n=21), knee (n=9) or shoulder joint region (n=7). At the time of explantation, the following pathogens were isolated: coagulase-negative staphylococci (n=19), Staphylococcus aureus (n=7), Streptococcus agalactiae (n=3), Propionibacterium acnes (n=5) and mixed infection (n=4). All patients received systemic antibiotics for a median of 19 days (range 11–42 days) before removal of the spacer/nail. At the time of re-implantation, tissue cultures were negative in all 38 patients, whereas sonication cultures showed growth of Propionibacterium acnes in 2 of 38 patients (5%) with a hip and shoulder spacer, both in significant numbers (150 and 550 colonies/ml sonication fluid, respectively). These organisms were probably present as mixed infection already at the time of explantation, but were missed due to overgrowth due to another organism (S. aureus in one patient and coagulase-negative staphylococci in another). Both patients were not initially treated for the Propionibacterium acnes infection, but the treatment was given after re-implantation.

Conclusion: Sonication of removed spacers is a suitable approach to identify persistent infection in patients with a two-stage exchange. Sonication may replace the current standard approach consisting of multiple tissue specimens in order to document successful eradication of infection.

Background: Negative pressure wound treatment is increasingly used through a Vacuum-Assisted Closure (VAC) device in complex wound situations. For this purpose, sterile polyurethane (PU) and polyvinyl alcohol (PVA) foam dressings are fitted to the wound size and covered with an adhesive drape to create an airtight seal. Little information exists about the type and quantity of microorganisms within the foams. Therefore, we investigated VAC foams after removal from the wound using a validated method (sonication) to detect the bacterial bioburden in the foam consisting as microbial biofilms.

Methods: We prospectively included VAC foams (PU and PVA, KCI, Rümlamg, Switzerland) without antibacterial additions (e.g. silver), which were removed from wounds in patients with chronic ulcers from January 2007 through December 2008. Excluded were patients with acute wound infection, necrotizing fasciitis, underlying osteomyelitis or implant. Removed foams from regular changes of dressing were aseptically placed in a container with 100 ml sterile Ringer’s solution. Within 4 hours after removal, foams were sonicated for 5 min at 40 kHz (as described in NEJM2007;357:654). The resulting sonication fluid was cultured at 37°C on aerobic blood agar plates for 5 days. Microbes were quantified as No. of colony-forming units (CFU)/ml sonication fluid and identified to the species level.

Results: A total of 68 foams (38 PU and 30 PVA) from 55 patients were included in the study (median age 71 years; range 33–88 years, 57% were man). Foams were removed from the following anatomic sites: sacrum (n=29), ischium (n=18), heel (n=13), calves (n=6) and ankle (n=2). The median duration of being in place was 3 days (range, 1–8 days). In all 68 foams, bacteria were found in large quantities (median 105 CFU/ml, range 102–7 CFU/ml sonication fluid. No differences were found between PU and PVA foams. One type of organisms was found in 11 (16%), two in 17 (24%) and 3 or more in 40 (60%) foams. Gram-negative rods (Escherichia coli, Proteus mirabilis, Klebsiella pneumoniae, Acinetobacter baumanii, Pseudomonas aeruginosa) were isolated in 70%, followed by Staphylococcus aureus (20%), koagulase-negative staphylococci, streptococci (8%), and enterococci (2%).

Conclusion: With sonication, a high density of bacteria present in VAC foams was demonstrated after a median of 3 days. Future studies are needed to investigate whether antimicrobial-impregnated foams can reduce the bacterial load in foams and potentially improve wound healing.

Background: Bacteria form biofilms on the surface of orthopaedic devices, causing persistent infections. Monitoring biofilm formation on bone grafts and bone substitutes is challenging due to heterogeneous surface characteristics. We analyzed various bone grafts and bone substitutes regarding their propensity for in-vitro biofilm formation caused by S. aureus and S. epidermidis.

Methods: Beta-tricalciumphosphate (β-TCP, Chro-nOsTM), processed human spongiosa (TutoplastTM) and PMMA (EndobonTM) were investigated. PE was added as a growth control. As test strains S. aureus (ATCC 29213) and S. epidermidis RP62A (ATCC 35984) were used. Test materials were incubated with defined bacterial solution (105 colony-forming units (cfu)/ml) at 37°C for 24 h without shaking. After 24 h, the test materials were removed and washed 3 times in PBS, followed by a standardised sonication protocol (Trampuz et al. 2007, NEJM). The resulting sonication fluid was plated in aliquots of 100μl onto aerobe blood agar with 5% sheep blood and incubated at 37°C with 5% CO2 for 24 h. Bacterial counts were enumerated and expressed as cfu/ml. Sonicated samples were transferred to a microcalorimeter (TA Instrument) and heat flow at 37°C was continuously monitored over a 24h period with a precision of 0.0001°C and a sensitiviy of 200μW. All experiments were performed in triplicates to calculate the mean ± standard deviation. ANOVA analysis was used for statistical calculations.

Results: For S. aureus bacterial counts (log10 cfu/ sample) were significantly higher (p< 0.001) for the porous (β-TCP 7.67 ± 0.17, Tutoplast 7.65 ± 0.15) than for the solid samples (PMMA 6.12 ± 0.18, PE 5.17 ± 0.22). Bacterial density (log10 cfu/surface) was 10^1–10^2 times higher for the S. epidermidis than for the S. aureus. In calorimetry the shape of the heat flow curves was characteristic for the individual strain and was not influenced by the test materials. The time to detection (TTD) was shortest for β-TCP for both strains and TTD was always shorter for S. aureus than S. epidermidis with corresponding material. Cfu/sample calculated from the calorimetric data was concordant with the standard culturing method.

Conclusion: Our results demonstrate biofilm formation with both strains on all tested materials. The calorimetry in all cases was able to detect quantitatively the amount of biofilm. Further studies are needed to see whether calorimetry is a suitable tool also to monitor approaches to prevent and treat infections associated with bone grafts and bone substitutes.

Background: Preoperative aspiration of synovial fluid is not always possible. Intraoperative collection of synovial fluid for determination of cell count and differential may be a useful approach for diagnosing or excluding prosthetic joint infection (PJI). However, the value and accuracy of synovial fluid collected during surgery has not yet been prospectively investigated. We analysed intraoperatively collected synovial fluid cell count for rapid and accurate diagnosis of PJI.

Methods: Unselected patients undergoing total hip or knee revision surgery from 06/06-05/08 were prospectively included. During surgery, synovial fluid was collected for determination of cell count; the cut-off value for PJI was > 1700 leucocytes/mm³ and/or > 65% neutrophils (according to Am J Med2004;117:556). In addition, 6 intraoperative samples were obtained for microbiology and 1 for histopathology. PJI was defined as presence of intraoperative pus, acute inflammation in histopathology, or ≥2 tissue samples growing the same organism.

Results: 59 patients (median age, 69 years; 68% females) with hip (n = 41) and knee (n = 18) prostheses were included, of whom 20 (34%) had PJI. Synovial fluid cell counts were true-positive in 16, true–negative in 38, false–positive in 1 and false–negative 4 patients. Sensitivity of synovial fluid cell count was 80%, specificity 97%, positive and negative predictive values were 94% and 90%, respectively.

Conclusion: The intraoperative synovial fluid cell count analysis is a rapid, inexpensive and reliable test for predicting or excluding PJI, having a ≥90% positive and negative predictive value. This test may be used for making operative decisions during surgery.

Background: The diagnosis of implant-associated infections is difficult due to organisms attached to surfaces as biofilms. We hypothesize that diagnosis can be improved by removing biofilm microorganisms from implant surface by sonication, followed by Gram stain, culture and calorimetric detection in sonication fluid.

Methods: We prospectively included adult patients from May 2005 until December 2006 from whom an orthopedic implant (joint prosthesis or internal fixation device) was removed for any reason. Removed implants were vortexed and sonicated in solid containers 5 min at 40 kHz in 100 to 400 ml Ringer’s solution. The resulting sonicate was plated and incubated on aerobic and anaerobic blood agar and aliquots were in parallel incubated at 37°C for 3 days in an isothermal calorimeter TAM III (TA Instruments, New Castle, DE). Gram stain was performed on sonicates centrifuged at 5000 g for 10 min. Definitive infection diagnosis was of the implant was defined if purulence surrounding the implant, or growth of the same microorganism in ≥2 synovial fluid or intraoperative tissue specimens, or acute inflammation in histopathology, or a sinus tract was present. Sonicate culture was defined positive if > 10 cfu (colony forming units) grew/ml sonicate. Calorimetry was defined positive if heat flow rate increased ≥10 μW above baseline (detection limit ~0.3 μW).

Results: 846 implants (367 joint prostheses and 479 internal fixation devices) were studied, of which 171 (20%) were infected and 675 (80%) were aseptic cases. The sensitivity of intraoperative tissue cultures was 74%, of sonicate culture 89%, of sonicate Gram stain 51%, of sonicate calorimetry 96%. The specificity of all specimens was ≥95%.

Conclusion: Sonicate culture and calorimetry were more sensitive than intraoperative tissue cultures for diagnosing implant infections. With Gram stain of centrifuged sonicate, infection was diagnosed in > 50% cases. Sonicate culture and calorimetry may replace the current approach using multiple intraoperative periprosthetic tissue specimens, whenever the implant or part of it is removed.

We have prospectively studied the outcome of infections associated with implants which were retained and treated using a standardised antimicrobial protocol. Over a period of four years, we studied 24 consecutive patients who had symptoms of infection for less than one year, a stable implant, no sinus tract and a known pathogen which was susceptible to recommended antimicrobial agents. The infections involved hip prostheses (14), knee prostheses (5), an internal fixation device (4), and an ankle prosthesis (1).

Twenty patients had a successful outcome at a median follow-up of 3.7 years (1.8 to 4.7); four had failure of the implant after a median follow-up of 1.2 years (0.3 to 2.5). The probability of survival without failure of treatment was 96% at one year (95% confidence interval (CI) 88 to 100), 92% at two years (95% CI 80 to 100) and 86% at three years (95% CI 72 to 100).

Patients with a short-term infection but with a stable implant, no sinus tract and a known pathogen may be successfully treated by retention of the implant and the use of a standardised regimen of antimicrobial treatment.