Introduction

The use of joint prostheses has become an extremely important advance in modern medicine because it has helped many patients to improve their quality of life. Nevertheless, prosthetic joint infection (PJI) is a rare but severe complication related to these procedures.1, 2 It is well known that bacterial infection after prosthesis implantation causes high morbidity and even mortality among the affected patients. Moreover, these infections frequently require prosthesis removal and a prolonged antibiotic treatment to cure the patients and ensure complete recovery.3, 4

Although a wide range of bacterial species can cause PJI,5 Staphylococcus aureus and Staphylococcus epidermidis are the most commonly isolated pathogens from implants, representing about 2/3 of the cases.6 The increasing number of methicillin-resistant S. aureus (MRSA) and methicillin-resistant S. epidemidis (MRSE) pathogens is a matter of special concern, as this greater prevalence increases the difficulty of antibiotic selection.7 Both MRSA and MRSE strains are also usually resistant to other antibiotics, and although vancomycin remains the elective therapy for these cases, its efficacy has been declining over the last few years.8, 9 For this reason, other antibiotics have recently been considered as alternatives to vancomycin.10, 11, 12, 13, 14, 15 Despite these facts, antimicrobial susceptibility remains extremely variable among these microorganisms, making it necessary to perform individual susceptibility studies for each strain to select the best antibiotic combination for treatment.16

The aim of this study was to evaluate the in vitro susceptibility pattern of S. aureus and S. epidermidis strains isolated from retrieved prosthetic joint implants against antibiotics commonly used in the treatment of PJI.

Materials and Methods

Bacterial strains

Thirty-five clinical Staphylococcus isolates (17 S. aureus and 18 S. epidermidis) as well as two biofilm-producing collection strains, S. aureus 15981 (Valle et al.17) and S. epidermidis ATCC 35984, were included. Clinical strains were isolated from orthopedic devices retrieved from patients with diagnosed PJI (one strain per patient) using a previously described protocol18 between January 2007 and December 2010. Species identification was performed by using the API Staph System (bioMérieux, France). Three S. epidermidis clinical strains were identified as small colony variant strains (SCV). All the strains tested were cultured and kept frozen in skim milk at –80 °C until the experiments were performed.

More details about the tested strains are shown in Table 1.

Table 1 Bacterial strains
Full size table

Antibiotics

Nine antimicrobial agents were selected for susceptibility testing: rifampin, vancomycin, ciprofloxacin, cotrimoxazole, cloxacillin, clindamycin, (Sigma, Munich, Germany), tygecicline (Pfizer, New York, NY, USA), daptomycin (Novartis, Basel, Switzerland) and fosfomycin (ERN, Barcelona, Spain). The antibiotics were prepared according to the instructions published by the Clinical and Laboratory Standards Institute (CLSI)19 and kept frozen at −20 °C until the experiments were performed.

Susceptibility test

In vitro susceptibility assays were performed using the broth microdilution method as described by the CLSI.19 In the case of fosfomycin, agar dilution was used together with broth microdilution. A calcium supplement (CaCl2 50 mg l−1 of final concentration) was added for daptomycin in vitro susceptibility testing. Fosfomycin was tested at the dilution range of 0.06–128 mg l−1; vancomycin, cotrimoxazole, cloxacillin and daptomycin from 0.015 to 32 mg l−1; tygecicline, clindamycin and ciprofloxacin from 0.004 to 8 mg l−1; and rifampin from 0.001 to 4 mg l−1. For MIC determinations, plates were incubated for 24 h, and for the three S. epidermidis SCV strains, the incubation time was extended to 48 h. MIC50, MIC90 and antibiotic non-susceptibility rates were calculated attending to the susceptibility breakpoints published by the European Committee on Antimicrobial Susceptibility Testing (EUCAST) and CLSI19, 20

MBC was calculated by colony counting in trypcase soy agar+5% sheep blood plates after direct plating of the wells-content without visible microbial growth and interpreted according to an internationally accepted definition.21

Results

Table 2 shows the antibiotic susceptibility assay results: MIC50 and MIC90 (mg l−1) as well as the percentage of nonsusceptible bacterial strains.

Table 2 Results of susceptibility studies: MIC50, MIC90 (mg l−1), nonsusceptibility rate and percentage of bactericidal activity
Full size table

Tygecicline and daptomycin showed the highest levels of antimicrobial activity with low MIC90 values (0.5 and 1 mg l−1 for S. aureus and 0.25 and 1 mg l−1 for S. epidermidis, respectively) and no resistant strains were detected.

Fosfomycin also exhibited substantial activity against both bacterial species, with only one S. aureus nonsusceptible strain (5.56% of the total of S. aureus). Rifampin and vancomycin showed high activity, with low rates of nonsusceptibility. Cotrimoxazole and clindamycin showed higher activity for S. aureus than S. epidermidis (5.56 and 11.11%, respectively).

Ciprofloxacin and cloxacillin exhibited a poor antimicrobial effect with a high percentage of nonsusceptible strains in both bacterial species. The high number of MRSA strains in this series is especially remarkable, and was also correlated with S. aureus ciprofloxacin nonsusceptibility.

On the other hand, bactericidal activity rates (also shown in Table 2) revealed the bacteriostatic behavior of rifampin, tygecicline, cotrimoxazole, fosfomycin and clindamycin for most strains.

Vancomycin showed a higher bactericidal effect against S. epidermidis than against S. aureus (94.74% vs 55.56%, respectively), showing strain-dependent behavior. This strain-dependent effect appears also for cloxacillin (50% in both species).

The only two antibiotics that showed bactericidal activity for most strains of both species were ciprofloxacin and daptomycin, both with similar bactericidal activity rates.

Discussion

Treatment of PJI caused by S. aureus and S. epidermidis is still a challenge because of the variable behavior that these microorganisms exhibit against the different antibiotics available and the increasing number of multidrug-resistant strains seen over recent years.22 Efficacy in both diagnosis and treatment of such staphylococcal PJIs could be hampered by the emergence of SCV strains or methicillin-resistant strains (especially MRSA), as these are the two most relevant problems.23, 24 SCV strains can remain undetected if incubation is not prolonged, and this has been recommended for the management of patients with clinical signs and symptoms of infection.25, 26 Moreover, infection by SCV strains should be considered a possibility if a poor response to treatment is detected, as previously reported in PJIs caused by S. aureus.27 The three tested SCV S. epidermidis strains not only required a minimum period of 2 days for growth in conventional media cultures, but also for MIC and MBC determinations, which required a modification of the established microbiological protocols.

In our study, we have detected an elevated number of MRSA (66.67% of S. aureus strains), a fact that confirms the previously reported high prevalence of MRSA.28 On the other hand, all the MRSA strains tested exhibited an identical resistance phenotype, which is the commonest one seen in our hospital in recent years (data not shown) and has been previously characterized in Spanish isolates.29 This finding suggests that the availability of antimicrobials which are active against MRSA has been drastically reduced. Likewise, a high cloxacillin nonsusceptibility rate was detected among the S. epidermidis strains (68.42%).

Vancomycin has traditionally been the first-line agent for PJI caused by multidrug-resistant staphylococci strains. However, owing to the recent increase of the MIC values that has been detected in some strains from both species, its use should be undertaken with caution.30 In this sense, our study has registered one vancomycin nonsusceptible S. aureus strain (5.6%). On the other hand, the bactericidal activity of vancomycin was higher in S. epidermidis compared with that observed in S. aureus (94.74 vs 55.56%), potentially constituting an additional drawback for the use of this antibiotic.

Rifampin, another commonly used antibiotic with outstanding efficacy against these types of clinical isolates, showed high in vitro activity against both staphylococcal species in our series. However, this antibiotic worked as a bacteriostatic agent for most of the tested strains, thus making it necessary to be used in combination with other antimicrobials to avoid resistance development.11, 31 Tygecicline, daptomycin and fosfomycin could be interesting alternatives for antimicrobial therapy. Our results show that tygecicline inhibited all the tested strains, with no bacterial resistance detected. However, MBC data showed that this antibiotic worked as a bacteriostatic agent against most of the tested strains.

Daptomycin has become a viable i.v. alternative to vancomycin because of its high level of antimicrobial and antibiofilm activity, as previously reported.10, 32 According to our data, daptomycin not only inhibited all the tested strains but also showed high bactericidal activity for both bacterial species (77.78% for S. aureus and 78.95% for S. epidermidis). These results are in accordance with those of other studies.33 However, the need for parenteral administration made daptomycin an alternative only for patients requiring i.v. therapy. Other oral alternatives are therefore necessary for treatment of patients outside the hospital setting.16

Finally, fosfomycin seemed to be an efficient agent, with only one S. aureus nonsusceptible strain found (5.56%), although this antibiotic exhibits a mostly bacteriostatic behavior.

Regarding the other tested antibiotics, antimicrobial activity findings generally resulted in differences between both staphylococcal species. Cotrimoxazole, as well as clindamycin, showed a better response against S. aureus than against S. epidermidis. In fact, all the MRSA strains tested in our study were cotrimoxazole-susceptible. This antibiotic has previously exhibited substantial in vitro activity against most MRSA strains.12 Both antimicrobials could be administered orally, thus increasing their attractiveness for the management of some patients.

In conclusion, MRSA strains are an increasingly prevalent cause of PJIs; consequently, the use of more efficient antibiotics is needed. According to our data, rifampin, tigecycline, daptomycin and fosfomycin showed high in vitro activity against most of the tested staphyloccal strains isolated from the PJIs. In vitro results should be considered when determining the therapy for these patients, along with other pathogenic aspects such as biofilm development and prosthesis removal.