Does preoperative antibiotic prophylaxis affect sonication-based diagnosis in implant-associated infection?

1 INTRODUCTION

The risk of periprosthetic joint and fracture-related infection (further referred to as implant-associated infection) increases with a growing demand for implants in societies characterized by demographic change.¹ Infection rates range from 1% to 3% in primary joint arthroplasty to 1%–5% in osteosynthesis of fractures.² Perioperative antibiotic prophylaxis (PAP) helps to reduce intraoperative contamination.¹ Nevertheless, most infections occur intraoperatively. The treatment of each implant-associated infection case requires a combination of radical debridement and, if possible, targeted antibiotic therapy.³ However, the causative microorganisms are typically present in a biofilm on the surface of the implants.⁴ This complicates pathogen detection³ because conventional microbiological culture methods from periprosthetic tissue and synovial aspiration are only accessible to planktonic bacteria in free suspension. For this purpose, the application of ultrasound (sonication) was developed. The sonication breaks through the biofilm and allows sampling of vital bacteria for improved microbiological detection in 90%.^{2, 5} For targeted antibiotic therapy, the determination of the pathogen is essential.⁶ In cases of cultural-negative implant-associated infection, however, the standard of using target antibiotic therapy cannot be complied with. Unfortunately, even with the technically modernized methods for implant-associated infections, about 10% to 30% of the pathogen detection procedures remain culturally negative despite other evidence of infection.^{5, 7-9} This rate appears to increase to 40%–60% when patients received preoperative antibiotic treatment.^{5, 10, 11} To improve the pathogen cultivability, various authors consequently recommend an antibiotic break until sampling. It is suggested to discontinue antimicrobial treatment for 2 to 4 weeks before the removal of the infected material or joint aspiration.^{2, 6, 10, 12-14} However, this approach might not be possible in certain situations (e.g., risk of spreading the infection to existing, previously uninfected joints or bone implants).

Further research revealed that also a shorter perioperative single-shot antibiotic prophylaxis affects the cultivability negatively, leading to the approach of renouncing PAP.^{12, 13, 15} On the other hand, two-thirds of the infections are caused intraoperatively by direct colonization of the foreign material with bacteria of the skin flora and the indoor environment.^{2, 16} The use of a single-dose PAP with first- or second-generation cephalosporins administered 30–60 min before incision and air sterilization in the operating room reduced the intraoperative infection rate from 30% to less than 1% because PAP prevents the adhesion of bacteria to the foreign body surface and thus the formation of a biofilm.¹⁷ Previous studies showed that intraoperative tissue cultures in implant-associated infection are not affected by PAP. In a study by Tetrault, a single-dose PAP was administered either before or after tissue sampling and in both study arms indent bacterial cultures could be determined.¹⁸ Pérez-Prieto et al.¹¹  found no significant differences in the microbiological culture positivity rate of 14 sonicates fluids between patients receiving antibiotic prophylaxis and patients who did not and Trampuz et al. proved that even a 2-week break in antibiotic therapy does not result in better culturing from tissue or sonicate fluid compared to untreated samples.²¹

The purpose of this study was to evaluate the influence of PAP on microbiological culture of sonicate fluid cultures (SFCs) for the detection of implant-associated infection using a larger population of patients and diagnostic methods than recommended. We hypothesized that a single dose of preoperative antibiotic prophylaxis does not affect intraoperative culture yield.

2 METHODS

3 RESULTS

We identified 114 patients, of which 24 did not meet the inclusion criteria. The patients were divided into three groups as mentioned before (n = 27 for SINGLE, n = 33 for TREAT, and n = 30 for NO). Sixty-seven of the remaining 90 patients (74.4%) had prosthetic-joint infections and 23 (25.6%) suffered from peri-implant infections of osteosynthesis material components. Demographics, infection, and diagnostic data are shown in Table 1. From 114 identified pathogens, 77 (68%) were detected by SFC and 82 (72%) by tissue culture. The most common microorganisms found in implant-associated infection were staphylococci followed by streptococci and enterococci (Table 2). There were 7 (6%) samples of the SFC and 17 (15%) tissue culture samples with culture-negative results. The detections of 30 (26%) pathogens by SFC and 15 (13%) by tissue culture had to be confirmed by other detection methods (culture of tissue, sonicate fluid, or joint aspiration). Seven pathogens of implant-associated infection were detected by tissue but not sonicate fluid culture, whereas 11 pathogens were detected only by sonicate fluid culture. Six of these 11 pathogens were isolated from patients in the TREAT group and three of them from patients in the SINGLE group (Table 3). No significant differences were found in the frequency of positive cultures for the three patient groups, irrespective of the detection method.

4 DISCUSSION

In the present retrospective, exploratory study we observed that PAP does not affect the microbiological detection rate of intraoperative cultures of sonicate fluid and tissue samples, which supports the hypothesis that a single dose of preoperative antibiotic prophylaxis does not affect intraoperative culture yield. Thus, we confirm a similar research study by Pérez-Prieto and co-workers, who found no significant differences in the microbiological culture positivity rate of 14 sonicate fluids between patients receiving classical antibiotic prophylaxis (first-generation cephalosporin or a glycopeptide between 30 and 60 min before surgery) and patients who did not. In the study, patients were included if they did not receive any kind of antibiotic regimen for 24 h or more during the 14 days before surgery.¹¹ The antibiotic concentration with a single administration of antibiotics seems to be sufficiently high to prevent attachment of planktonic pathogens to the implant surface,¹⁷ yet low enough not to eliminate pathogens in their protective biofilm.¹¹

It has to be pointed out that the retrospective and exploratory data analysis of the current study limits the generalizability of our findings. The design of the present study implies inhomogeneous groups in terms of size and properties. This explains the uneven distribution of the type of explant treated with sonication. Nevertheless, the three study arms were found to be relatively uniform in terms of the number of their associated patients, gender distribution, and other characteristics, which enabled adequate comparability. In this study, databases from various institutes and the hospital information system were used to collect data. Although transmission errors, documentation gaps, and nonrandom sample allocations cannot be ruled out, we cross-validated and double-checked all data sources to reduce the impact of a potential bias. However, according to our results, and contrary to the recommendation of other authors, an omission of PAP before surgery with subsequent pathogen detection in suspected implant-associated infection using sonication appears not indicated. Our results also suggest that culturing of SF and tissue samples is not negatively affected by an antibiotic treatment if administered more than 24 h before surgery, as demonstrated by other studies.¹⁸ Trampuz and co-workers observed that, if antibiotic therapy was continued until the sampling, sonication was more sensitive than other conventional detection methods.⁵ We also found that the detection rates of all proven pathogens (certain and uncertain) were higher in SFC than in tissue cultures. This can be explained by the observation that, in contrast to conventional detection methods, sonication facilitates the detection of biofilm-forming, metabolically inactive microorganisms, which were not influenced by antibiotic treatment.^{3, 5, 23} Furthermore, the fact that late infections with low-virulent pathogens and mature biofilms are predominantly detected by sonication^{1, 23} was confirmed by the current work. In addition, we observed that 11 of 114 infection-relevant pathogens were exclusively detected by sonication and determined to be predominantly chronic late infections and low-virulent pathogens. Most of these patients were part of the TREAT group. According to some authors, 10%–30% of the pathogen detection procedures remain culture-negative in the absence of prediagnostic antibiotic administration despite the use of various microbiological diagnostic methods and the presence of other reliable signs of infection.^{5, 8, 9, 18} If patients additionally received a prediagnostic antibiotic therapy, this rate appeared to rise to 40%–60%.^{5, 10, 11} These observations were not confirmed in the present study. The proportion of culture-negative results remained below 10% for patients with no antibiotic application before surgery for SFC and tissue cultures. An increase of this rate to 40%–60% after prediagnostic therapeutic or prophylactic administration of antibiotics was not observed for any detection method. However, the rate of uncertain pathogen detection by sonication (<50 CFU/ml) increased noticeably. In summary, sonication is not only a pathogen detection method with high sensitivity and low failure rate, but also an excellent tool for the diagnosis of low-virulent late infections, especially after therapeutic antibiotics administration before sampling. Nevertheless, sonication culture cannot be recommended as an exclusive diagnostic tool because the high sensitivity of the process also entails the risk of detecting microbial contamination during implant removal, airborne associated, specimen handling in the laboratory.⁹ In a total of eight cases, a low colony count pathogen infection (<50 CFU/ml) was detected by sonication and had to be considered as contamination. These pathogens were also low virulent microorganisms. In contrast, 30 infectious pathogens, which were detected at <50 CFU/ml by sonication, were also detected by other detection methods (culture of tissue or joint aspiration). Without verification by further infection criteria, these cases would be considered as contaminations by definition. Consequently, in case of suspected implant-associated infection, sonication should not be used as an exclusive diagnostic agent, but rather as a supplement to the histopathological and microbiological examination of peri-implant/periprosthetic samples. Independently of the diagnostic tool, the successful treatment of implant-associated infection depends on an interdisciplinary team approach and an optimized combination of surgical intervention and antimicrobial treatment. According to our results, and contrary to the recommendation of other authors, the omission of PAP and the discontinuation of an ongoing antibiotic therapy before surgery with subsequent pathogen detection in suspected implant-associated infection using sonication is not recommended. This has a significant impact on clinical paths as well as on scientific activities related to biofilm-based orthopedic device infection.

CONFLICT OF INTERESTS

The authors declare that there are no conflict of interests.

AUTHOR CONTRIBUTIONS

Anna Stephan and Maik Stiehler designed the study. Alexander Thürmer organized the microbiological database. Anna Stephan was responsible for data analysis. Anna Stephan, Ingmar Glauche, and Maik Stiehler performed statistics. Stefan Zwingenberger and Jörg Nowotny contributed to data interpretation and drafting the manuscript. Ingmar Glauche contributed to critical review and data interpretation. Maik Stiehler supervised the project, participated in its coordination, and helped to draft the manuscript. All authors read and approved the final manuscript and its submission.