Introduction

Finegoldia magna is a gram-positive anaerobic bacterium formerly known as Peptostreptococcus magnus. It is found in the mucosa of the gastrointestinal tract, the female genito-urinary tract and as part of normal human skin [1, 2]. In addition, it can invade the human tissues and cause opportunistic infections such as periprosthetic joint infections (PJI) and soft-tissue abscesses [2,3,4]. In 2018, a patient at the Balgrist University hospital in Zurich, Switzerland, experienced a treatment failure of a F. magna hip PJI due to cefepime resistance that had not been tested in the initial phase of treatment. This case prompted us to investigate the in vitro antimicrobial susceptibility of F. magna, since data remain scarce (Supplementary Table S1). To provide an overview of the local resistance situation and to give recommendations for the treatment of infections caused by this microorganism, susceptibility of clinical F. magna isolates from Zurich, Switzerland, was assessed.

Case presentation

The patient was a 52-year-old woman with diabetes mellitus, chronic kidney insufficiency, and adiposity. She was initially treated for a polymicrobial PJI (Cutibacterium avidum and Citrobacter koseri) after surgical debridement of all infected tissues and explantation of the prosthesis. At time of implantation of the new prosthesis, Pseudomonas aeruginosa and F. magna were found in several intraoperative tissue biopsies. Antibiotic treatment with ciprofloxacin and cefepime was started on hemodialysis and maintained for three months without performing antimicrobial susceptibility testing (AST). Sixteen months later, a relapse with F. magna was diagnosed as the patient presented recurring pain and elevated serum inflammatory parameters. AST revealed susceptibility to amoxicillin, amoxicillin/clavulanic acid, and metronidazole, resistance to clindamycin, and MIC values for cefepime and levofloxacin above the species-unrelated breakpoints (Supplementary Table S2). Because of the relapsing infection, the joint prosthesis was explanted. Intravenous antibiotic treatment with amoxicillin was initiated for two weeks and was followed by metronidazole per os for four weeks prior to implantation of a new hip prosthesis. During the following 24 months, neither a relapse nor a new infection was detected.

Materials and methods

Bacterial isolates

All F. magna strains isolated at the Institute of Medical Microbiology (IMM) of the University of Zurich, Switzerland, between September 2019 and July 2020, were stored in home-made chopped meat carbohydrate broth at room temperature until the beginning of the study. Bacterial identification was performed by MALDI-TOF MS (Bruker Daltonics) using the direct formic acid transfer method [5]. Repeated isolates from same patients were excluded. A total of 57 isolates was analyzed.

AST and analysis of the MIC distributions

AST was performed using E-test strips (ETEST® bioMérieux SA Marcy l’Etoile, France) based on Clinical and Laboratory Standards Institute (CLSI) standard procedure [6]. Briefly, bacterial suspensions corresponding to a McFarland 1.0 were prepared in saline solution from primary sub-cultures and were spread on home-made Brucella agar plates supplemented with 5% laked sheep blood, hemin, and 10 µg/ml vitamin K1. Plates were incubated under anaerobic conditions at 37 °C for 48 h. The following nine antibiotics were tested: benzylpenicillin, amoxicillin/clavulanic acid, cefuroxime, cefepime, levofloxacin, rifampicin, metronidazole, doxycycline and clindamycin. E-tests were chosen because a high correlation between this method and the standard dilution method was shown previously [7, 8]. Internal quality control testing for the antibiotics included in the study performed during 2020–2021 at the IMM is reported in Supplementary Table S4. The median minimum inhibitory concentration (MIC50), the 90th percentile of minimum inhibitory concentration (MIC90), and the MIC range of values were determined for each antibiotic and compared to the following breakpoints. For benzylpenicillin, amoxicillin/clavulanic acid, metronidazole, and clindamycin, the breakpoints from the European Committee on Antimicrobial Susceptibility Testing (EUCAST) for gram-positive anaerobes, Version 10.0, valid from 1st January 2020, were used [9]. For cefuroxime, cefepime, and levofloxacin, EUCAST species-unrelated breakpoints were applied since EUCAST breakpoints for gram-positive anaerobes were not available [9]. For rifampicin and doxycycline, local epidemiological cut-off values (2 mg/l and 1 mg/l, respectively) were determined using “ECOFFinder” version 2.1 provided by CLSI [10].

Detection of rifampicin resistance

Whole-genome sequencing (WGS) was performed on the only clinical isolate with a MIC value for rifampicin above the local ECOFF. DNA extraction, library preparation, and WGS were performed as described recently [11]. Raw sequencing data were trimmed, filtered, and assembled by an in-house bioinformatics pipeline. The complete gene sequence of rpoB was extracted and used to query the Comprehensive Antibiotic Resistance Database (CARD) [12] in order to identify potential point mutations mediating rifampicin resistance.

Results

MIC distributions of nine antibiotics (benzylpenicillin, amoxicillin/clavulanic acid, cefuroxime, cefepime, levofloxacin, rifampicin, metronidazole, doxycycline, and clindamycin) for 57 F. magna isolates are shown in Fig1A–I, while the MIC50, MIC90, and MIC ranges are summarized in Table 1. Based on EUCAST breakpoints for gram-positive anaerobes, F. magna isolates were categorized towards penicillin, amoxicillin/clavulanic acid, clindamycin, and metronidazole as susceptible at standard dosing regimen, susceptible at increased exposure, and resistant with a high likelihood of therapeutic failure even with an increased exposure. F. magna isolates were classified as “wild-type” or “non-wild-type” when the MIC values were below or above the EUCAST non-species-related breakpoints (for cefepime, cefuroxime, and levofloxacin) or in-house determined ECOFF values (for rifampicin and doxycycline).

Fig. 1
figure 1figure 1

Distribution of the MIC values (mg/l) of nine tested antibiotics on 57 F. magna strains. Panel A, benzylpenicillin; panel B, amoxicillin/clavulanic acid; panel C, metronidazole; panel D, clindamycin; panel E, cefuroxime; panel F, cefepime; panel G, levofloxacin; panel H, doxycycline; panel I, rifampicin. Dotted line corresponds to the EUCAST breakpoints for gram-positive anaerobes, dashed line corresponds to the EUCAST species-unrelated breakpoints, and continuous line corresponds to local epidemiological cut-off values

Full size image
Table 1 MIC50, MIC90, and MIC range and resistance pattern of 57 tested F. magna strains are shown considering EUCAST breakpoints for gram-positive anaerobes, EUCAST non-species-related breakpoints (underlined), and self-determined ECOFF values (in italics)
Full size table

All isolates were susceptible to benzylpenicillin, amoxicillin/clavulanic acid, and metronidazole, while 75% were susceptible to clindamycin. F. magna isolates exhibited MICs for rifampicin, doxycycline, and levofloxacin below the susceptibility breakpoint in 98, 72, and 56% of the cases, respectively. For cefuroxime and cefepime, F. magna strains exhibited MICs below the ECOFFs in 75% and only 14% of the cases, respectively. However, based on species-unrelated breakpoints, 93% and 32% of the isolates would be categorized as susceptible at increased dosing regimen.

One isolate exhibited a MIC (3 mg/l) for rifampicin above the distribution of the wild-type population (internally established ECOFF = 2 mg/l, Fig. 1 panel I). Inspection of the rpoB gene (responsible for rifampicin resistance [13]) revealed the presence of a mutation leading to the amino acid substitution Gln486Arg. Moreover, this isolate exhibited MIC values for cefepime, levofloxacin, and cefuroxime above the ECOFFs (Supplementary Table S3).

Discussion

Biofilm infections such as PJIs require surgical debridement to remove infected tissues and prolonged antibiotic treatment to prevent relapse of infection [14, 15]. Antibiotic therapy should be based on AST in order to avoid inadequate treatment. F. magna PJIs are rare and little information is available regarding the best therapeutic options. All 57 F. magna clinical isolates included in this study were susceptible to benzylpenicillin, amoxicillin/clavulanic acid, and metronidazole whereas high rates of isolates showing MICs above the resistance breakpoints for cefepime (68%), levofloxacin (44%), doxycycline (28%), and clindamycin (25%) were observed. Therefore, the latter antibiotics should be taken in consideration only after AST to prevent relapse of infection as documented in our clinical case presentation. Also, MIC values for cefuroxime and rifampicin were above the resistance breakpoints only in 7% and 2% of the cases, respectively.

In accordance with the literature, MIC data of all isolates for penicillin, amoxicillin/clavulanic acid [16,17,18,19], and metronidazole [3, 18,19,20,21,22,23] were within the susceptible range. Penicillin-resistant F. magna strains were found in different countries with rates of 1.7 to 5.7% [18, 20, 21, 24]. Metronidazole-resistant strains have been reported in only a few cases in Germany [18], Kuwait [18], and Spain [17]. As outlined above, we found a high prevalence (68%) of F. magna strains with MIC values for cefepime above the resistance ECOFF. Also, MIC values ranged between 2 and 256 mg/l; the MIC50 and MIC90 were 12 mg/l and 64 mg/l, respectively. In a comparable study, using 30 F. magna strains from diabetic foot infections, Goldstein et al. [25] reported a wide range of cefepime MICs between 1 mg/l and > 32 mg/l with high MIC50 (32 mg/l) and MIC90 (> 32 mg/l).

To our knowledge, we report for the first time AST data for cefuroxime, which is the most common antibiotic used for prophylaxis of implant-associated surgeries [26, 27]. Considering the low rate (7%) of isolates displaying MIC values above the resistance ECOFF and of 18% of isolates that would be classified as susceptible at increased exposure considering species-unrelated breakpoints, cefuroxime deserves better attention and may be discussed in current prophylaxis recommendations.

Comparative analysis with the literature should be performed with caution. While our study included 57 clinical strains isolated in Switzerland, other studies analyzed significantly smaller numbers of isolates. In addition, differences in resistance rates were observed in various countries. For instance, we found a significant lower MIC50 value (0.5 mg/l) for levofloxacin as compared to a Russian study, where a MIC50 of 32 mg/l for 38 F. magna strains isolated from cancer patients was reported [24]. High MIC values for levofloxacin have also been reported in other studies conducted by Veloo et al., Mikamo et al., or Goldstein et al. [28,29,30]. For rifampicin, we found a high MIC50 value (0.38 mg/l) as compared to the only study available in literature from Sweden, where the MIC50 of rifampicin for nine F. magna isolates was as low as 0.064 mg/l [3]. Based on our in-house ECOFF value, we found a moderate portion of strains (28%) falling out the doxycycline wild-type population. This is in stark contrast to the study of Goldstein et al., who reported a high prevalence of non-wild-type strains (85%, n = 100) from the USA, Canada, Germany, Holland, Hungary, and Sweden [19]. The resistance rate of 25% for clindamycin in Switzerland is comparable to that found in 271 F. magna isolates from Slovenia (28%) [18], while significantly lower to what reported in F. magna isolates from Kuwait (50%) [18] and Spain (54%) [17], though smaller numbers were tested.

In conclusion, F. magna clinical isolates are in general susceptible to penicillin, amoxicillin/clavulanic acid, and metronidazole and PJI infections caused by this microorganism can be treated with these antibiotics without prior AST. Treatment with cefuroxime, cefepime, levofloxacin, rifampicin, doxycycline, and clindamycin always requires confirmatory AST due to varying resistance rates.