Introduction

Orthopaedic implant-associated infections (IAI) can result in the need for repeated surgeries, extended hospitalization, prolonged antibiotic usage, and poor clinical outcome [1, 2]. The causative agents of IAI are varied and include both highly virulent bacteria like Staphylococcus aureus with a well-recognized role in IAI, as well as slow-growing, Gram positive anaerobic bacteria (SGAB) like Cutibacterium acnes whose role is more controversial, especially when isolated from presumed aseptic cases [1, 3,4,5,6,7,8,9,10].

Culture of tissue samples obtained from the peri-implant area are traditionally used to detect the causative agents in suspected IAI [11]. However, bacteria causing IAI tend to form biofilms on the surface of implants and can be missed by standard tissue cultures [12, 13]. This may be alleviated by sonication, where biofilms are dislodged into the sonication fluid (SF), which can then be cultured [12,13,14,15]. However, sonication may not be routinely performed during removal of orthopaedic implants in many institutions, especially in non-joint prosthesis hardware, unless an infection is readily suspected by the responsible surgeon. Previous studies have shown that the refrained use of SF cultures might lead to misdiagnosing an implant failure as aseptic, especially in cases with low-grade IAI without overt features of infection [7]. However, the value of utilizing SF culture in all clinically aseptic or apparent mechanical failure cases, is currently unclear [7].

In 2019 we initiated a SGAB project with a focus on sonication and optimized anaerobic growth conditions to explore the prevalence and nature of SGAB in orthopaedic implants. We have previously reported on data obtained in this project with a singular focus on C. acnes [16]. The purpose of this study, which is based on all implants acquired in the previous study, is to evaluate the application of sonication with SF culture in addition to tissue cultures on removed orthopaedic implants, irrespective of the type of implant or reason for removal.

Methods

The study was performed as an exploratory cross-sectional study, based on a research database, with prospective data collection. The implants were collected and processed as described previously [16]. In brief, one hundred orthopaedic implants that were removed between August 2019 and September 2020, irrespective of the type of implant or reason for removal, were collected from the Department of Orthopaedic Surgery, Aarhus University Hospital and transported to the Department of Biomedicine, Aarhus University, Denmark. Implants (n = 33) where tissue samples were not obtained were excluded in the present study leaving a total of 77 implants for evaluation.

Tissue culture

At the discretion of the operating surgeon, tissue samples were obtained for routine culture during implant removal, as described by Kamme and Lindberg [17]. The tissue samples were processed according to standard procedures at the local Department of Clinical Microbiology. The results of tissue sample cultures, as well as relevant implant related data were obtained from the patients’ electronic medical records. Five tissue biopsies were taken in all cases except for four (5%), where less than five biopsies were taken. No bacterial growth was seen in any of these four cases (both in SF and tissue cultures). We defined isolates in tissue cultures as potential contaminants, if a bacterial species was isolated from only one of five tissue biopsies [17].

Sonication fluid culture

All implants were sonicated using a standardized method of 30 s vortexing, sonication for one minute at 100% followed by further 30 s of vortexing [18]. The SF obtained was inoculated on three different agar media (blood agar, fastidious anaerobic agar and reinforced clostridial agar). A total of ten plates per implant were used, with nine plates for anaerobic incubation and one plate for aerobic incubation. This methodology was chosen due to the context of the SGAB project. For aerobic incubation, uncentrifuged SF was used while for anaerobic incubation, both uncentrifuged (three plates) and centrifuged SF (six plates) were inoculated. Incubation was performed up to three days for the aerobic cultivation and 28 days for anaerobic cultivation. All bacterial growth was quantified using colony forming units/ml (CFU/ml) and bacteria were identified with 16 S rRNA gene sequencing [16].

The significance of C. acnes isolates from SF was determined in an in-depth evaluation in the SGAB project where a combined analyses of culture-dependent and -independent methods were used [16]. The culture-dependent methods included single-locus sequence typing (SLST) and whole genome sequencing of all C. acnes isolates obtained from culture [16]. In the culture-independent SLST amplicon-based next-generation sequencing of SF, the various C. acnes SLST types present in the SF were detected directly from the SF and their relative abundancies measured [16]. In instances of discrepancies between the two methods, especially when the C. acnes SLST types isolated from culture were not detected by the culture-independent methods, the C. acnes isolates were considered contaminants [16]. In all other isolates, European Bone Joint Infection Society (EBJIS) criteria for SF culture, defined by bacterial growth ≤ 50 CFU/ml in uncentrifuged SF or ≤ 200 CFU/ml in centrifuged SF was used to identify likely contaminants [19].

The study was registered with Region Midtjylland with reference number 661,624 and carried out in accordance with relevant guidelines and regulations. The need for ethical approval was waived by Central Denmark Region ethical committee. Informed consent was obtained from all patients before their participation in the study. All data reported in the study was prospectively entered into our SGAB project at Aarhus University REDCap system [20], and all patient identifiers were removed before downloading the data for analysis. Data are presented as proportions in percentages.

Results

Out of the 77 implants included in the study, majority (n = 69; 90%) were arthroplasties. Shoulder (n = 34; 44%) was the most common anatomic location, followed by the hip (n = 18; 23%) and knee (n = 17; 22%). Twelve (16%) of the 77 implants included in the study were removed due to infection as noted by the responsible surgeon in the patient chart. The remaining 65 (84%) implants were removed for presumed non-infectious reasons, including mechanical failure and pain.

Bacteria were isolated from SF culture in 35 (46%) of the 77 implants. The reason for implant removal was infection in 11 (31%) of the 35 cases.

Positive tissue cultures were noted in 18 (23%) of the 77 cases. In 10 (56%) of the 18 cases with positive tissue culture the surgeon had registered infection as the reason for removal of the implant.

Data from the 12 (16%) cases in which the implant was removed due to infection are depicted in Table 1. There was concordance between SF and tissue culture results in 11 (92%) cases. In the remaining case (No.2), in which a sinus tract was present, C. acnes was isolated only from SF culture, with a low CFU/mL. Noticeably, in one case, Bacteroides fragilis was isolated from only one of five tissue specimens, while in SF culture it was isolated with > 250 CFU/ml.

Table 1 Details of the twelve implants in the confirmed/suspected IAI group
Full size table

In 41 (63%) of 65 implants removed due to presumed aseptic causes, no bacteria were isolated from SF culture. The corresponding tissue cultures in these implants showed either no bacterial growth in 38 cases or growth that was classified as likely contamination in three cases. In the remaining 24 (37%) implants, 27 bacterial isolates were isolated from SF culture (Table 2 + 3). Eleven (41%) isolates from SF of 11 different implants were classified as indicative of a potential infection according to our study criteria. The remaining 16 (59%) of the 27 isolates were considered likely contaminants (Table 3).

Table 2 Details of the eleven implants in the presumed aseptic group, with isolation of bacteria from sonication fluid culture that potentially represent infection
Full size table
Table 3 Details of the thirteen implants in the presumed aseptic group, with isolation of contaminant bacteria (n=15) on sonication fluid culture
Full size table
Table 4 Reasons for classification of the seventeen bacterial isolates as contaminants
Full size table

Discussion

In this study, we applied sonication with SF culture on 77 removed orthopaedic implants, with concomitant tissue biopsies.

We found that for implants in which the surgeon had noted infection as the cause of removal, SF culture helped improve the diagnostic yield in two patients. In the first patient (Table 1, No. 2), who clinically presented with a sinus tract, the causative agent (C. acnes) was detected only in SF culture and not in tissue culture. In the second patient (Table 1, No. 11) in whom infection was suspected because of raised inflammatory markers, SF culture confirmed the diagnosis and the causative agent. Similarly, in the presumed aseptic group (Table 2), SF culture improved the diagnostic yield in seven patients, six of whom had no growth on tissue culture, and one that had bacterial growth in only one of the five tissue culture specimens.

The use of SF culture in the diagnosis of IAI is widely used internationally and has been studied since 2007 [14]. Although current literature primarily originates from dedicated infection centres or large tertiary referral centres and results are slightly conflicting, with some suggesting that SF culture is superior to tissue culture [12, 14, 21,22,23], while others report better or similar performance of tissue culture [24,25,26,27]. The reported discordance in literature could be due to lack of standardization in the number of tissue samples collected, variations in the sample processing methods, and differences in the growth conditions and CFU/ml cut-offs for positive sonication fluid cultures used in different laboratories.

We also found that there is a substantial risk for the isolation of contaminants in SF culture, especially in the presumed aseptic group, despite taking adequate precautions in the handling of the implants [16] (Tables 4 and 3). Of the 49 bacterial isolates from SF, more than half were considered contaminants according to the criteria used in the study, and all but one of these were from implants removed due to aseptic causes. Cutibacterium species and CoNS accounted for nearly 90% of the contaminants. In comparison, only about a quarter of the bacterial isolates from tissue samples were considered contaminants. A study by Namdari et al. [28] reported a risk of contamination of samples by C. acnes and CoNS that originated from the air in the operating room. In implant removal surgeries, samples for tissue culture are collected relatively soon after skin incision and using strict aseptic methods, while prostheses are exposed to the air in the operating room for a longer duration before retrieval since the extraction can be a tedious and long-standing process. Such differences may explain the higher risk of contamination seen in SF culture of extracted prostheses. In addition, contamination in the laboratory during implant processing and incubation is a potential source of contamination [16]. Use of multiple agar plates, prolonged incubation of SF cultures (28 days versus up to 14 days for most tissue cultures) as well as repeated checking of plates for growth during this period could explain the increased contamination rates of SF cultures in this study.

C. acnes was the most common bacteria isolated from implants in the aseptic group and the second most common organism in the infection group (Tables 1, 2 and 3). All the C. acnes isolates that would represent infection and more than half from the contamination group originated from upper extremity implants (shoulder/elbow). These results correlate with previous studies that have shown that C. acnes has a predilection for the upper extremity [7, 29, 30]. Both et al. also reported that C. acnes was commonly detected from clavicle plates of seemingly infection-free patients, while being absent in fibular implants [31]. We used additional criteria to determine the significance of C. acnes isolates in the study. Interestingly, if the SF culture results were interpreted based on EBJIS cut-off for SF culture, then two Cutibacterium isolates considered to represent infection might have been missed. While the guidelines mention that any positive culture in SF be considered as a ‘potential infection’, it also states that low growth of common contaminants like C. acnes are less likely to represent infection compared to more virulent bacteria like S. aureus [19]. Moreover, C. acnes is often the most common bacteria obtained in culture of implants without prior suspicion of infection [30, 32,33,34,35,36] though the clinical impact of these unexpected positive cultures in the upper extremity is currently unclear [36, 37]. More research into this area is needed. Similarly, three of the Cutibacterium isolates that were considered contaminants in this study would have fallen under the ‘infection confirmed’ category in the EBJIS guidelines [19]. Accordingly, interpretation of positive Cutibacterium species cultures must always be with caution.

The study has several limitations. Due to the exploratory nature of the study, only a limited number of implants were included. The decision to send tissue cultures was dependent on the operating surgeon, further limiting the number of implants included in the study. Another main drawback of the study was that classification of implants into the presumed aseptic group, or the infected group was based on what was written in the patients’ charts by the operating surgeon and not based any internationally recognized criteria like the EBJIS guidelines. This was due to the unavailability of all the required data in the patients’ charts. Finally, we did not perform a clinical follow-up. As such, the clinical relevance of C. acnes in the presumed aseptic group including the isolates considered non-contaminants is still unclear.

Conclusion

Sonication of implants irrespective of reason for removal appears to be of added value in the detection of the causative organisms. However, SF cultures should be interpreted with care especially when Cutibacterium species are isolated, irrespective of the amount of growth.