Skip to main content
Download PDF

Characterization of periprosthetic environment microbiome in patients after total joint arthroplasty and its potential correlation with inflammation

BMC Infectious Diseases volume 23, Article number: 423 (2023) Cite this article

Abstract

Aims

Periprosthetic joint infection (PJI) is one of the most serious complications after total joint arthroplasty (TJA) but the characterization of the periprosthetic environment microbiome after TJA remains unknown. Here, we performed a prospective study based on metagenomic next-generation sequencing to explore the periprosthetic microbiota in patients with suspected PJI.

Methods

We recruited 28 patients with culture-positive PJI, 14 patients with culture-negative PJI, and 35 patients without PJI, which was followed by joint aspiration, untargeted metagenomic next-generation sequencing (mNGS), and bioinformatics analysis. Our results showed that the periprosthetic environment microbiome was significantly different between the PJI group and the non-PJI group. Then, we built a “typing system” for the periprosthetic microbiota based on the RandomForest Model. After that, the ‘typing system’ was verified externally.

Results

We found the periprosthetic microbiota can be classified into four types generally: “Staphylococcus type,” “Pseudomonas type,” “Escherichia type,” and “Cutibacterium type.” Importantly, these four types of microbiotas had different clinical signatures, and the patients with the former two microbiota types showed obvious inflammatory responses compared to the latter ones. Based on the 2014 Musculoskeletal Infection Society (MSIS) criteria, clinical PJI was more likely to be confirmed when the former two types were encountered. In addition, the Staphylococcus spp. with compositional changes were correlated with C-reactive protein levels, the erythrocyte sedimentation rate, and the synovial fluid white blood cell count and granulocyte percentage.

Conclusions

Our study shed light on the characterization of the periprosthetic environment microbiome in patients after TJA. Based on the RandomForest model, we established a basic “typing system” for the microbiota in the periprosthetic environment. This work can provide a reference for future studies about the characterization of periprosthetic microbiota in periprosthetic joint infection patients.

Peer Review reports

Article Summary

Article focus

This is the first study that evaluated the periprosthetic environment microbiome in PJI patients after total joint arthroplasty and its potential correlation with inflammation.

Key messages

We built a “typing system” based on the detected microbiota, and different microbiota types are related to different clinical characteristics.

Strengths and limitations

This study identified the periprosthetic microbiota differences between PJI and non-PJI patients. The periprosthetic microbiota can be classified into four types generally: “Staphylococcus type,” “Escherichia type,” “Cutibacterium type,” and “Pseudomonas type.”

The relatively small sample size in this study is a limitation.

Introduction

Periprosthetic joint infection (PJI) is one of the most serious complications after total joint arthroplasty (TJA) and often indicates disastrous outcomes [1]. However, with the increasing number of total joint arthroplasty cases, the incidence of PJI is not going to decrease. In the UK and USA, about 800,000 joint arthroplasties are done annually, with projections to greater than 4 million by 2030 [1]. The annual incidence of PJI is estimated at 1% after hip arthroplasties and ranges between 1 and 2% after knee arthroplasties [1]. Kapadia et al. noted a cost of €95,000 per PJI case, which is five times higher than that for primary arthroplasty [1]. Moreover, PJI patients have less satisfaction with their procedure, only up to 23% are satisfied, and 18% report complete dissatisfaction. Therefore, the projected increase in PJI procedures places a huge burden on patients, clinicians, and the worldwide healthcare system [1, 2].

PJI diagnosis is challenging and identifying the corresponding pathogens is the management core in the diagnosis and treatment of PJI. Traditionally, PJI pathogens have been identified by synovial fluid and periprosthetic tissue culture. However, the false-negative rate of culture is about 20–30% for various reasons such as biofilm formation, inert pathogens, inappropriate culture medium, and antibiotic administration [3]. In recent years, to increase the detection rate of PJI pathogens, some molecular methods with high sensitivity such as multiple PCR, untargeted metagenomic next-generation sequencing (mNGS), and targeted next-generation sequencing (NGS) have been used in PJI diagnosis [4,5,6,7,8,9,10,11,12,13,14,15,16,17,18]. These methods are used to detect pathogens and they often reveal more pathogens than clinical cultures do. In addition, by culturing the sonication fluid of the joint prosthesis retrieved during revisions, the detection rate of polymicrobial PJI also increased [19]. Moreover, studies have found that the discordant rate between preoperative synovial fluid culture and intraoperative synovial fluid culture was about 20% and the culture of specific pathogens such as E. coli and Streptococcusspp. also indicated higher risks of polymicrobial PJI [20, 21]. These studies not only suggested the potential clinical underestimation of polymicrobial PJI but indicated the existence of heterogeneous polymicrobial microbiota in the periprosthetic environment after TJA. However, to our knowledge, few studies have comprehensively studied the heterogeneous microbiota in this environment [22].

In some patients considered to be aseptic loosening clinically, several inert bacteria such P. acneswere shown by mNGS, PCR, prolonged culture time, and the culture of sonication fluid [8, 9, 12]. These phenomena challenge the current notion of the periprosthetic environment as sterile in non-PJI joints after TJA and indicate that microbiota exist in the periprosthetic joint environment despite no significant clinical infection signs. However, most published studies have mainly focused on the suitability of these methods for the diagnosis of pathogens in individual suspected PJI cases [5, 9, 10, 14]. There is no study focusing on the microbiota signature in patients with suspected PJI. In addition, some studies indicated that NGS can be used to detect PJI but the interpretation methods for the results were arbitrary and their diagnostic values were controversial [5, 10]. The major reason can be lack of knowledge about the “relatively normal” periprosthetic microbiota in patients with suspected PJI and ignorance of the microbiota may potentially cause over-treatment because of the high resolution of mNGS to identify pathogens. Hence, a better understanding of the compositional changes in microbiota is necessary so that the role of pathogens in PJI can be further clarified and the bioinformatics pipeline used in interpreting mNGS results can be designed better.

Therefore, we performed a prospective study to shed light on the features of the microbial community in the periprosthetic joint environment after TJA in patients with suspected PJI. Using mNGS to detect microorganisms in the environment surrounding the prosthesis of the revision patient, a classification system based on these major differential microorganisms was established by RandomForest sampling of the major differential microorganisms between PJI and non-PJI. Moreover, our study was also designed to evaluate the specific signatures of the microbiota detected by mNGS in the joint fluid, with PJI confirmed by conventional tests.

Materials and methods

Study design

The approval of the Institutional Review Board and the local Ethics Committee were obtained prior to the commencement of this study and then the study was performed in a tertiary joint center. The patients with suspected joint infection with enough synovial fluid for mNGS and standard diagnostic tests were recruited prospectively for this study.

The definition of patients suspected of having PJI

Based on the institutional protocol, PJI was considered when a patient had one of the following signs or symptoms after TJA.

  1. 1)

    Acute or persistent pain at rest, swelling, redness, or warmth around the joints;

  2. 2)

    Elevated ESR or CRP level; or

  3. 3)

    Implant failure within 5 years after total joint arthroplasty without any reasonable explanation.

All patients provided written informed consent before mNGS tests. In this study, the mNGS lab was blinded to the synovial fluid culture results. Between April 2019 and Match 2020, a total of 77 patients were included in this study. Joint aspiration was performed by two experienced surgeons to obtain synovial fluid for analysis [23]. The obtained joint fluid was sent for mNGS if the patients agreed to participate in the study. The study design is summarized in Fig. 1.

Fig. 1
figure 1

The summary of this study design

Full size image

The PJI diagnosis protocol and the definition of groups

In our institution, ESR and serum CRP was used to screen for PJI. Preoperative joint aspiration was performed in these patients to obtain synovial fluid and then the diagnostic workup was initiated. The methods of preoperative joint aspiration were previously reported by Li et al., and synovial fluid was obtained before the surgeries for joint infection [23]. After joint aspiration, the collected synovial fluid was shipped for mNGS testing, leukocyte esterase (LE) test, synovial fluid analysis (WBC count, polymorphonuclear neutrophils (PMN%)), and anaerobic and aerobic bacterial cultures and fungal culture within 3 h. If PJI can’t be diagnosed based on the 2014 Musculoskeletal Infection Society (MSIS) criteria before joint revisions, intraoperative tissue histology was performed when the joint was revised. Besides, the synovial fluid and 5 periprosthetic tissues obtained during revisions were also sent for cultures during revisions. If PJI was diagnosed based on the 2014 MSIS criteria but no pathogens were identified by cultures (synovial fluid and periprosthetic tissues), the culture-negative PJI was diagnosed. In this study, the diagnosis of PJI was based on the 2014 MSIS criteria.

The patients included in this study were divided into three groups based on the following criteria:

  1. 1)

    Culture-positive PJI group: PJI was diagnosed based on the 2014 MSIS criteria and the pathogens were identified by clinical culture.

  2. 2)

    Culture-negative PJI group: PJI was diagnosed based on the 2014 MSIS criteria but the pathogens were not identified by clinical cultures.

  3. 3)

    Non-PJI group: This group was composed of patients who underwent one-stage aseptic revisions. The one-stage revisions did not fail because of any infection or aseptic reasons such as PJI, loosening, mechanical complications within at least 6-month follow-up after one-stage revision.

Microbiological cultures

The obtained synovial fluid (> 1 ml) was injected into a BacT/ALERT FA fastidious antimicrobial neutralization (FAN) bottle (BioMerieux) for anaerobic bacterial culture and a BacT/ALERT PF Pediatric FAN (BioMerieux) bottle for aerobic bacterial and fungal culture. Each bottle was incubated for 2 weeks, and VITEK-MS (BioMerieux) was used for microorganism identification if pathogens were detected.

If a microorganism was revealed in either the aerobic bottle or anaerobic bottle, the pathogen was recorded as part of the preoperative aspiration results. Antibiotic sensitivity tests were performed by disk diffusion according to laboratory standard protocols. In addition, two to five different intraoperative periprosthetic tissues were also sent for microbiological cultures during revisions.

Metagenomic next-generation sequencing and analysis

Volumes of at least 0.5 mL samples were collected from the subjects. To improve the efficiency of pathogen detection, the samples were first enriched in small solutions (~ 200 μL) and centrifuged at 15,000 × g for 10 min at 4 ℃. Then 200 μL of enrichment solution was used for nucleic acid extraction and purification with a nucleic acid extraction kit combined with magnetic beads (Sagene, Guangzhou, China); the magnetic bead method had been optimized, as compared with the precipitation method and adsorption column method (Sagene, Guangzhou, China). The metagenomic library was respectively constructed according to the protocol of the library construction kit, Nextera XT (Illumina, USA). The extracted DNA was first divided into ~ 300 bp fragments followed by the addition of different index sequences. During sequencing, no specific DNA fragments were amplified. The library size and quantification were analyzed using an Aglient 2100 bioanalyzer system, and the accurate quantification was detected by qPCR (Bio-Rad CFX96, USA). After the libraries were mixed in equal amounts, high-throughput sequencing was performed on the Illumina Nextseq 550 DX sequencing platform (sequencing strategy: PE150), an FDA-approved sequencer. During the process of library construction and sequencing, no target probes were spiked into the DNA library to capture specific DNA fragments in the microorganisms. In addition, a negative control group was used to eliminate possible DNA contamination during library construction and sequencing.

Statistical analysis

After the sequencing was finished, raw data were filtered by FastQC software, including removing the reads containing the sequencing adapters, the reads containing more than 10% N, and the low-quality reads containing less than 50% of low-quality bases (Q-value ≤ 10). The remaining reads were used for next-step analysis. Human-related reads were removed by aligning with the human genome reference sequence (version: GRCh38), using BWA (http://bio-bwa.sourceforge.net/) software, and then a proprietary microbial pathogen database (Kraken2/Bracken) was used for analysis to obtain the identification and quantitative results of pathogenic microorganisms. The microorganisms reference databases were downloaded from NCBI (ftp://ncbi.nlm.nih.gov/genom es/). Then, possible DNA contamination in the reagent was filtered by SourceTracker [24]. Relative abundance data were profiled in comparison analysis, and clinical relevance analysis was mainly based on R. Alpha-diversity was calculated based on Shannon indexes at the species level. Bray–Curtis distance was used to perform principal coordinate analysis (PCA) based on the species levels. The Circos graph (www.circos.ca) was plotted based on the top 10 genera with the highest relative abundance in the microbiota. We used the RandomForest bag (R, version: 4.1.1) to build a RandomForest model and extracted the top five major features (multiple displacement amplification, MDA method) and the “typing system” was built based on these major microorganisms. Then the abundance of these major microorganisms was translated into relative abundance based on these five microorganisms for further analysis. We used Corrplot for interactively analyzing microbiome data and clinical data. The correlation was evaluated by the Wilcoxon rank-sum test. Clustering was dependent on the Ward distance.

Results

The study design and the characteristics of patients included in this study

There were 42 and 35 patients in the PJI group and non-PJI group, respectively. The median age in the PJI group and non-PJI group was 65 years and 67 years, respectively. In this study, only hip and knee joints were included. There were 30 and 26 knee joints in the PJI group and non-PJI group, respectively. The details of the patients included in this study are summarized in Table 1.

Table 1 Characteristics of patients included in this study
Full size table

The distribution and diversity of the periprosthetic microbiota

The characterization of periprosthetic microbiota was different between the PJI group and the non-PJI group. As shown in Fig. 2, Shannon’s index and Simpson diversity index in the PJI group were lower than that in the non-PJI group. Similarly, as shown in the PCA analysis (Fig. 2c), the beta-diversity in these two groups was significantly different, and the samples can be classified into two groups based on the Bray-Curtis distance. Some pathogens in the PJI group were more likely to show higher abundance than that in the non-PJI group, such as S. aureus. However, some pathogens in the PJI group had a lower abundance than that in the non-PJI group, such as P. acnes. The distribution of microorganisms in PJI and non-PJI groups is summarized in Fig. 2. Generally, there were some overlaps in microbiomes between the PJI group and non-PJI groups. The PJI cohort was then divided into culture-positive PJI and culture-negative PJI based on the clinical culture results, and Shannon’s index in the culture-negative PJI group was higher than that in the culture-positive PJI group.

Fig. 2
figure 2

PJI patients and non-PJI patients harbor a distinct microbiome. a Boxplots of microbial Alpha-diversity (Shannon's index) for the synovial fluid from PJI patients and non-PJI patients. b Boxplots of microbial Alpha-diversity (Shannon’s index) for the synovial fluid from culture-positive PJI patients and culture-negative PJI patients. c PCoA plot of the microbiota using Bray–Curtis distance metric of Beta-diversity in the synovial fluid of PJI patients and non-PJI patients. d Genus-level distribution of the microbiota in the PJI patients and Non-PJI patients. e PCoA plot of the microbiota using Bray–Curtis distance metric of Beta-diversity in the synovial fluid of culture-positive PJI patients and culture-negative PJI patients. F The Venn plot shows the overlaps among the non-PJI group, culture-positive PJI group, and culture-negative PJI groups at the specie level. g CPCoA plot of the microbiota using Bray-Curtis distance metric of Beta-diversity in the synovial fluid of culture-negative PJI patients, culture-positive PJI patients, and non-PJI patients

Full size image

The construction of a “typing system” of periprosthetic microbiota based on RandomForest model

The clustering of microbiota distribution suggested that there can be different “microbiota types” in the periprosthetic environment after TJA. We extracted the major microorganisms from the microbiota based on the RandomForest model (Supplementary Fig. 1). According to the major microorganisms of the microbiota extracted from the RandomForest model and the abundance of these microorganisms detected in the periprosthetic microbiota, the microbiota can be classified into four types corresponding to different clinical signatures:

  1. 1.

    Cutibacterium type: The details of the microbiota distribution in this type are shown in Fig. 3. In this type, the specific highly abundant microorganisms were Cutibacterium spp., and the clinical diagnosis of this type was often non-PJI, indicating that the clinical signs of infection in patients with this type (ESR, CRP) were not obvious.

  2. 2.

    Staphylococcus type: In this type, the highly abundant microorganisms were Staphylococcus spp., and corresponding patients were more likely to be diagnosed as having PJI. Clinically, in this type (a total of 11 cases), 10 cases were diagnosed as PJI based on the 2014 MSIS criteria.

  3. 3.

    Escherichia type: In this type, the most abundant microorganisms were Escherichia spp., and the corresponding patients were likely to be diagnosed as having PJI.

  4. 4.

    Pseudomonas type: In this type, the most abundant microorganisms were Pseudomonas spp., and the corresponding patients were likely to be diagnosed as having PJI.

Fig. 3
figure 3

The genus-level distribution of the microbiota in the PJI patients and non-PJI patients. A Heatmap showing the relative abundance of the microbiota (top 10) clustered by the Ward method based on candidate microbiota. B Bar plot denoting the relative abundance of five bacteria selected by the RandomForest model at the genus level for each sample. The x-axis indicates the number of patients with PJI or non-PJI. The y-axis represents the detection ratio of target microbiomes based on the average relative abundance. C The Circos plot shows the distribution of the relative abundance of the top 10 microorganisms within each sample at the species level

Full size image

To further show the distribution of these microorganisms within each sample, a Circos plot was built at the level of genus (Fig. 3) according to the top 10 most abundant microorganisms. The proportions of different periprosthetic microbiome types in PJI and non-PJI groups are shown in Table 2.

Table 2 The proportion of different periprosthetic microbiome types in PJI and non-PJI groups
Full size table

The clinical data on different types of microbiota in the periprosthetic environment

To further classify the association between the microbiota and clinical signatures, we integrated the clinical omics and the abundance data about these microorganisms. The abundance of Staphylococcus spp. is positively correlated with the levels of ESR, serum CRP, and the presence of clinical PJI. However, the abundance of Cutibacterium spp. is negatively correlated with the levels of ESR, CRP, the presence of clinical PJI, the abundance of Staphylococcus spp., and Escherichia spp. The details about the relationships are summarized in the cluster rectangles (Fig. 4 and Table 3). In addition, the relationship between different types of periprosthetic microbiota and clinical diagnostic results is shown in the Sankey plot (Fig. 4). As shown in this figure, the “Staphylococcus type” was more likely to appear in the infection group compared to other types while the “Cutibacterium type” was more likely to appear in the non-infection group.

Fig. 4
figure 4

The relationship between different microbiota types and clinical data. A Heatmap denoting the relative abundance of five bacteria selected by the RandomForest model at the genus level for every five subsets (subset_red, subset _green, subset _blue, subset _purple, subset _pink) clustered by Ward’s method based on candidate microbiota in the synovial fluid of patients. B The heatmap shows the relationship between the clinical characteristics and microbiota characteristics. *P < 0.05; *P < 0.01; ***P < 0.001. C Synovial fluid with a predominance of different microorganisms has different clinical diagnoses. Sankey’s diagram shows the relationship between the types based on microbiota and clinical diagnosis. Shannon: Shannon–Wiener diversity Index. Simpson: Simpson’s diversity index

Full size image
Table 3 Clinicopathological characteristics of different microbiota types
Full size table

External validation of the typing profile and its correlation with inflammation

To further clarify the dedicated and specific correlation between the ‘typing profile’ and inflammation. External verification was performed in the other cohort (including 168 patients). In this cohort, the metagenomic data generated from 168 patients, were obtained from NCBI (PRJNA436717).

The microorganism abundance tables of these samples were obtained according to the bioinformatic pipeline mentioned in the ‘methods part’ and then, the relative abundance of the “major microorganisms” which were selected in the previous RandomForest model was extracted. We tested the diagnostic accuracy of the pre-built RandomForest model for PJI diagnosis compared to the MSIS criteria in this validation cohort and the AUC of the RandomForest model for PJI diagnosis was 0.81. Besides, the alpha-diversity in the non-PJI group was significantly higher than that in the PJI group (Fig. 5a), which was consistent with the results we obtained in the exploration cohort.

Fig. 5
figure 5

Metagenomic analysis of the correlation between periprosthetic microbiota structure and inflammation. aThe alpha-diversity in the non-PJI group was significantly higher than that in the PJI group. (Chao1: Chao1 richness estimator; ACE: ACE estimator; Shannon: Shannon Wiener index; Simpson: Simpson diversity index). *P < 0.05 (Mann–Whitney U-test). The levels of synovial fluid WBC count (b), synovial fluid PMN% (c), ESR (d), serum C-reactive protein (e) in patients with those four types of periprosthetic microbiota. Box plots reflect median and IQRs with whiskers bounding non-outlier values. *P < 0.05 (Mann–Whitney U-test)

Full size image

After that, the patients in the validation cohort were classified into 4 types based on the relative abundance of the selected ‘major microorganism’: “Staphylococcus type,” (71 cases) “Escherichia type,”(24 cases) “Cutibacterium type,”(17 cases) and “Pseudomonas type.”(54 cases). The incidence of PJI in the ‘Staphylococcus type’ was 78.9% (56/71), significantly higher than that in the ‘Cutibacterium type’ (78.9% vs 0%, P < 0.05). Then, we evaluated the levels of host inflammatory responses to these different types of periprosthetic microbiota. We found that the levels of ESR, serum CRP and synovial fluid WBC count, synovial PMN% in the patients with ‘Staphylococcus type’ were significantly higher than that in the patients with ‘Pseudomonas type’, followed by ‘Escherichia type’ and ‘Cutibacterium type’ (Fig. 5bcde). The intensity of the inflammatory response shows a stepwise distribution in patients with different types of periprosthetic microbiota. The details were shown in Fig. 5 and Table 4. Moreover, the percentage of cases with “Staphylococcus type” in the acute PJI group was higher than that in the chronic PJI group (Cohort1: 71.4% vs 18.5%; P = 0.043; Cohort2: 62.4% vs 24.7%; P < 0.0001). However, the percentages of cases with “Cutibacterium type” in the acute PJI group was lower than that in the chronic PJI group (Cohort1: 11.1% vs 0; P = 0.036; Cohort2: 20.5% vs 0; P < 0.0001). These data indicated that chronic and acute PJI were of different microbiota types, suggesting different microbiota composition between acute and chronic PJI (Fig. 6).

Table 4 Clinicopathological characteristics of different microbiota types in the validation cohort
Full size table
Fig. 6
figure 6

The pie graphs showing the composition of microbiota types in chronic and acute PJI in two cohorts. The patients (77 patients) in cohort1 were recruited by the authors. The cases (168 cases) in cohort 2 were obtained from the public database (PRJNA436717)

Full size image

Discussion

This study systematically examined the characterization of the periprosthetic environment microbiome in 77 patients with suspected PJI after TJA. Our study suggested that the characterization of the periprosthetic environment microbiome was significantly different between the PJI group and the non-PJI group. Results show that the diversity of microbiota in the PJI group was significantly lower than that in the non-PJI group. In general, bacterial biodiversity can be a relatively good marker in some “populated” environments [25]. From this point of view, the increase of bacterial biodiversity can potentially indicate a relatively “healthier” periprosthetic environment. Moreover, based on the features extraction by the RandomForest model, the periprosthetic microbiota in patients after TJA can be classified into four different types: Cutibacterium type, Staphylococcus type, Escherichia type, and Pseudomonas type. In the patients with the Cutibacterium type, the clinical signs and symptoms were relatively “weak” and these patients were more likely to be classified clinically as non-PJI patients based on the 2014 MSIS criteria compared to the other types. Conversely, in the “Staphylococcus type”, PJI was more likely to be diagnosed after undergoing septic revisions. Staphylococcus spp., including S. aureus and coagulation-negative staphylococcus, are the leading causes of PJI in clinical practice. E. coliis a gram-negative pathogen and the main gram-negative bacteria causing PJI [1]. As the literature reports, Cutibacteriumspp. is an emergent pathogen causing a low-grade infection (CRP and other clinical and laboratory parameters are often negative) [12, 13]. In addition, in some presumed aseptic loosening patients, Cutibacterium spp. are also isolated by prosthesis sonication fluid culture [22, 26]. Pseudomonasspp. are a group of conditionally pathogenic bacteria, commonly found in hospital infections [27, 28].

To our knowledge, this is the first study that evaluated the characterization of the periprosthetic environment microbiome in patients with suspected PJI after TJA. Consistent with clinical findings, some microorganisms of high abundance in the microbiota were more likely to be isolated in clinical cultures [10, 29]. Considering the relatively low abundance of these microorganisms in the non-PJI patients, these microorganisms may reach an ecological “peak” in the development of PJI after TJA followed by obvious clinical PJI symptoms and signs.

To further clarify the microbiota composition in suspected PJI patients after TJA, we extracted the major organisms contributing most to the microbiota composition to depict the microbial community and established a “typing system” based on the RandomForest model [30]. Generally, the microbiota composition in suspected PJI patients can be grouped into four different types: Cutibacterium type, Staphylococcus type, Escherichia type, and Pseudomonas type. In the patients of the Cutibacterium type and Pseudomonas type, PJI is not likely to be diagnosed as compared with those of the Staphylococcus type and Escherichia type. In addition, the levels of some clinical inflammatory indicators (serum ESR, CRP, and synovial fluid WBC) were increased with the increasing abundance of Staphylococcus spp. and decreased with the decreasing abundance of Cutibacterium spp. It also indicated that the inflammatory response was more intense in the Staphylococcus type than that in the Cutibacterium type. Clinically, some studies detected Cutibacteriumspp. in the sonication fluid of prostheses retrieved from assumed aseptic loosening cases [22, 26, 31], suggesting that Cutibacterium spp. can be “relatively normal” microorganisms in patients with suspected PJI, and it may not usually cause clinically obvious PJI after TJA because the virulence of Cutibacterium spp. is lower than that of Staphylococcus spp. However, if Staphylococcus spp. become highly abundant in the periprosthetic microbiota, the clinical inflammatory response can be severe because Staphylococcus spp. is highly virulent and can cause obvious local and systemic inflammatory responses. Staphylococcusspp. is also the most common pathogen that causes PJI [1].

In clinical practice, positive culture and the presence of a sinus communicating with the prosthesis were thought to be the “gold standard” for PJI diagnosis [32,33,34,35]. However, culture-negative PJI still accounts for about 20–30% of all PJI in patients [3]. In these patients, molecular diagnostic methods such PCR and mNGS are used to identify the offending pathogens and the use of these techniques is increasing [6, 14, 18]. At present, the interpretation of mNGS sequencing results for PJI diagnosis remains controversial, and diagnostic accuracy is heterogeneous [6, 14, 36, 37]. The major reason can be the lack of knowledge about the microbiota in patients with suspected PJI [22]. Our study explored this problem and showed that PJI is not likely to be diagnosed clinically if another highly abundant microorganism is detected by mNGS, such as Cutibacteriumspp. This is consistent with previous studies showing that these pathogens are likely to be isolated in the sonication fluid of the prosthesis retrieved during aseptic revisions [13, 19]. In addition, several studies focusing on the utility of RCR in PJI diagnosis also reported similar results [18]. In this study, the microbiota with these special characteristics were classified into Cutibacterium type, and corresponding patients had weak inflammatory responses (low levels of synovial fluid WBC count, PMN%, ESR, and serum CRP). Therefore, we hypothesize that these pathogens can be the “relatively native microbiota” that colonize the periprosthetic environment after TJA, and these may not usually cause obvious inflammatory responses or trigger significant septic symptoms. More importantly, these pathogens should be considered when the accurate interpretation pipeline of mNGS results is designed because the misleading interpretation of these pathogens can cause overuse of antibiotics, especially in culture-negative PJI.

In the validation cohort, the results strengthen the conclusion that the microbiota diversity in the non-PJI patients was higher than that in PJI patients. Besides, the corresponding hosts showed obvious different inflammatory responses to different types of periprosthetic microbiota. The host inflammatory response to ‘Staphylococcus type’ was significantly more severe than ‘Pseudomonas type’, followed by ‘Escherichia type’ and ‘Cutibacterium type’. From the view of microbiota, the hosts show different inflammatory responses to the microbiota when the periprosthetic environment was dominated by different microorganisms. Clinically, nucleic acid sequence of pathogens could still be detected by mNGS even when the culture results were negative and it suggests that metagenomic can be used as a complementary approach to conventional cultures for PJI diagnosis, just as other pieces of literature reported [5, 10, 38]. Taken together, the composition of microorganisms in the periprosthetic environment is not only related to PJI but also reflects the levels of host inflammation. In other words, the disturbances of the microbiome might affect the development of PJI potentially and their composition may also be promising in monitoring the treatments outcomes of PJI patients or providing a reference for clinical diagnosis and prognosis.

With the increasing use of clinical mNGS for PJI diagnosis, the exploration of periprosthetic microbiota can contribute to the establishment of the bioinformatic pipeline for mNGS results [13, 36, 39, 40]. For example, if the detected periprosthetic microbiome is classified as Cutibacteriumtype and the clinical tests cannot provide strong evidence for PJI, the aseptic loosening is more likely to be considered based on the PJI diagnosis criteria [19, 41]. In addition, prior studies have revealed that some of the cases classified as aseptic may be infectious in origin but either was not investigated or escaped diagnosis of PJI using the available modalities [38, 42]. Taken together, these phenomena suggest the need for further research into the role of Cutibacterium in PJI and presumed aseptic loosening. In this study, 17 patients (classified as Cutibacterium type) with a clinical diagnosis of aseptic loosening did not develop PJI within postoperative 6-month follow-up. Correspondingly, if the periprosthetic microbiome was classified as the Staphylococcus type, PJI can be highly suspected based on the PJI diagnosis criteria, and septic revisions can be an option. In this study, PJI was confirmed by the 2014 MSIS criteria in 10 out of 11 cases of the Staphylococcus type. Therefore, the bioinformatic pipeline based on the “typing system” has promising value in helping clinical decision-making.

There are still some limitations in this study. First, the samples were collected in the same hospital, and selection bias was inevitable. Therefore, the classification system proposed in this study still needs to be validated extensively. Second, we studied only the composition of periprosthetic microbiota, and the function of the microbiota was not evaluated in this study because of the relatively low sequencing volume. Our follow-up study will try to interpret microbiota function with the use of metabolomics [36].

Overall, our study sheds light on the composition of the periprosthetic environment microbiome in patients after TJA, and the microbiota could be classified into four types: Staphylococcus type, Escherichia type, Cutibacterium type, and Pseudomonas type. Different types correspond to different levels of inflammatory responses and clinical features. This result can be a reference for the design of appropriate mNGS reporting criteria in clinical practice and aid in future studies on the periprosthetic microbiome in patients with PJI.

Availability of data and materials

All data and materials were in full compliance with the journal’s policy. And the data were obtained in Department of Orthopedic Surgery, The First Medical Center, Chinese PLA General Hospital. The datasets used and during the current study are available from the corresponding author on reasonable request. The datasets generated and/or analysed during the current study are available in the NGDC repository, [ACCESSION NUMBER: PRJCA016981].

Abbreviations

TJA:

Total joint arthroplasty

PJI:

Periprosthetic joint infection

References

  1. Kapadia BH, Berg RA, Daley JA, Fritz J, Bhave A, Mont MA. Periprosthetic joint infection. Lancet. 2016;387(10016):386–94.

    Article  PubMed  Google Scholar 

  2. Labek G, Thaler M, Janda W, Agreiter M, Stöckl B. Revision rates after total joint replacement: cumulative results from worldwide joint register datasets. J Bone Joint Surg Br. 2011;93(3):293–7.

    Article  CAS  PubMed  Google Scholar 

  3. Kalbian I, Park JW, Goswami K, Lee YK, Parvizi J, Koo KH. Culture-negative periprosthetic joint infection: prevalence, aetiology, evaluation, recommendations, and treatment. Internat Orthop. 2020;44(7):1255–61.

    Article  Google Scholar 

  4. Bémer P, Plouzeau C, Tande D, Léger J, Giraudeau B, Valentin AS, et al. Evaluation of 16S rRNA gene PCR sensitivity and specificity for diagnosis of prosthetic joint infection: a prospective multicenter cross-sectional study. J Clin Microbiol. 2014;52(10):3583–9.

    Article  PubMed  PubMed Central  Google Scholar 

  5. Fang X, Cai Y, Shi T, Huang Z, Zhang C, Li W, et al. Detecting the presence of bacteria in low-volume preoperative aspirated synovial fluid by metagenomic next-generation sequencing. Internat J Infect Dis. 2020;99:108–16.

    Article  CAS  Google Scholar 

  6. Flurin L, Wolf MJ, Greenwood-Quaintance KE, Sanchez-Sotelo J, Patel R. Targeted next generation sequencing for elbow periprosthetic joint infection diagnosis. Diagnos Microbiol Infect Dis. 2021;101(2): 115448.

    Article  CAS  Google Scholar 

  7. Gomez E, Cazanave C, Cunningham SA, Greenwood-Quaintance KE, Steckelberg JM, Uhl JR, et al. Prosthetic joint infection diagnosis using broad-range PCR of biofilms dislodged from knee and hip arthroplasty surfaces using sonication. J Clin Microbiol. 2012;50(11):3501–8.

    Article  PubMed  PubMed Central  Google Scholar 

  8. Huang Z, Li W, Lee GC, Fang X, Xing L, Yang B, et al. Metagenomic next-generation sequencing of synovial fluid demonstrates high accuracy in prosthetic joint infection diagnostics: mNGS for diagnosing PJI. Bone Joint Res. 2020;9(7):440–9.

    Article  PubMed  PubMed Central  Google Scholar 

  9. Indelli PF, Ghirardelli S, Violante B, Amanatullah DF. Next generation sequencing for pathogen detection in periprosthetic joint infections. EFORT Open Reviews. 2021;6(4):236–44.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Ivy MI, Thoendel MJ, Jeraldo PR, Greenwood-Quaintance KE, Hanssen AD, Abdel MP, et al. Direct detection and identification of prosthetic joint infection pathogens in synovial fluid by metagenomic shotgun sequencing. J Clin Microbiol. 2018;56(9):e00402–18.

  11. Ryu SY, Greenwood-Quaintance KE, Hanssen AD, Mandrekar JN, Patel R. Low sensitivity of periprosthetic tissue PCR for prosthetic knee infection diagnosis. Diagnos Microbiol Infect Dis. 2014;79(4):448–53.

    Article  Google Scholar 

  12. Salar-Vidal L, Achermann Y, Aguilera-Correa JJ, Poehlein A, Esteban J, Brüggemann H, et al. Genomic analysis of Cutibacterium acnes strains isolated from prosthetic joint infections. Microorganisms. 9:1500.

  13. Salar-Vidal L, Martin-Garcia M, Auñón A, Esteban J. Cutibacterium spp. isolated from orthopaedic implant-associated infection: a not-so-slowly growing organism. Enfermedades infecciosas y microbiologia clinica (English ed). 2021;39(6):287–90.

    PubMed  Google Scholar 

  14. Thoendel MJ, Jeraldo PR, Greenwood-Quaintance KE, Yao JZ, Chia N, Hanssen AD, et al. Identification of Prosthetic Joint Infection Pathogens Using a Shotgun Metagenomics Approach. Clin Infect Dis. 2018;67(9):1333–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Van Belkum A, Gros MF, Ferry T, Lustig S, Laurent F, Durand G, et al. Novel strategies to diagnose prosthetic or native bone and joint infections. Expert Rev Anti-Infect Ther. 2022;20(3):391–405.

  16. Widerström M, Stegger M, Johansson A, Gurram BK, Larsen AR, Wallinder L, et al. Heterogeneity of Staphylococcus epidermidis in prosthetic joint infections: time to reevaluate microbiological criteria? Eur J Clin Microbiol Infect Dis. 2022;41(1):87–97.

    Article  PubMed  Google Scholar 

  17. Yang B, Fang X, Cai Y, Yu Z, Li W, Zhang C, et al. Detecting the presence of bacterial RNA by polymerase chain reaction in low volumes of preoperatively aspirated synovial fluid from prosthetic joint infections. Bone Joint Res. 2020;9(5):219–24.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Zhu H, Jin H, Zhang C, Yuan T. Intestinal methicillin-resistant Staphylococcus aureus causes prosthetic infection via “Trojan Horse” mechanism: evidence from a rat model. Bone Joint Res. 2020;9(4):152–61.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Trampuz A, Piper KE, Jacobson MJ, Hanssen AD, Unni KK, Osmon DR, et al. Sonication of removed hip and knee prostheses for diagnosis of infection. New Engl J Med. 2007;357(7):654–63.

    Article  CAS  PubMed  Google Scholar 

  20. Echeverria AP, Cohn IS, Danko DC, Shanaj S, Blair L, Hollemon D, et al. Sequencing of circulating microbial cell-free DNA can identify pathogens in periprosthetic joint infections. J Bone Joint Surg AM. 2021;103(18):1705–12.

    Article  PubMed  Google Scholar 

  21. Li H, Xu C, Hao L, Chai W, Jun F, Chen J. The concordance between preoperative aspiration and intraoperative synovial fluid culture results: intraoperative synovial fluid re-cultures are necessary whether the preoperative aspiration culture is positive or not. BMC Infect Dis. 2021;21(1):1018.

    Article  PubMed  PubMed Central  Google Scholar 

  22. Carr C, Wilcox H, Burton JP, Menon S, Al KF, O’Gorman D, et al. Deciphering the low abundance microbiota of presumed aseptic hip and knee implants. PLoS ONE. 2021;16(9): e0257471.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. Li R, Li X, Ni M, Zheng QY, Zhang GQ, Chen JY. Anatomic landmark-guided hip aspiration in the diagnosis of periprosthetic joint infection. Orthoped. 2021;44(1):e85–90.

    Article  Google Scholar 

  24. Knights D, Kuczynski J, Charlson ES, Zaneveld J, Mozer MC, Collman RG, et al. Bayesian community-wide culture-independent microbial source tracking. Nat Methods. 2011;8(9):761–3.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Sędzikowska A, Szablewski L. Human gut microbiota in health and selected cancers. Intern J Molecul Sci. 2021;22(24):13440.

    Article  Google Scholar 

  26. Zhang Q, Xi Y, Li D, Yuan Z, Dong J. The yield of sonication fluid culture for presumed aseptic loosening of orthopedic devices: a meta-analysis. Ann Palliative Med. 2021;10(2):1792–808.

    Article  Google Scholar 

  27. Elborn JS, Blasi F, Haworth CS, Ballmann M, Tiddens H, Murris-Espin M, et al. Bronchiectasis and inhaled tobramycin: a literature review. Respir Med. 2022;192: 106728.

    Article  PubMed  Google Scholar 

  28. Fagotti L, Tatka J, Salles MJC, Queiroz MC. Risk factors and treatment options for failure of a two-stage exchange. Curr Rev Musculoskel Med. 2018;11(3):420–7.

    Article  Google Scholar 

  29. Kuo FC, Lu YD, Wu CT, You HL, Lee GB, Lee MS. Comparison of molecular diagnosis with serum markers and synovial fluid analysis in patients with prosthetic joint infection. Bone Joint J. 2018;100-b(10):1345–51.

    Article  PubMed  Google Scholar 

  30. Liao H, Zhang Y, Guo W, Wang X, Wang H, Ye H, et al. Characterization of the blood and cerebrospinal fluid microbiome in children with bacterial meningitis and its potential correlation with inflammation. mSystems. 2021;6(3):e0004921.

    Article  PubMed  Google Scholar 

  31. Figa R, Muñetón D, Gómez L, Matamala A, Lung M, Cuchi E, et al. Periprosthetic joint infection by Propionibacterium acnes: Clinical differences between monomicrobial versus polymicrobial infection. Anaerobe. 2017;44:143–9.

    Article  PubMed  Google Scholar 

  32. Amanatullah D, Dennis D, Oltra EG, Marcelino Gomes LS, Goodman SB, Hamlin B, et al. Hip and knee section, diagnosis, definitions: Proceedings of International Consensus on Orthopedic Infections. J Arthrop. 2019;34(2s):S329-s37.

    Article  Google Scholar 

  33. McNally M, Sousa R, Wouthuyzen-Bakker M, Chen AF, Soriano A, Vogely HC, et al. The EBJIS definition of periprosthetic joint infection. Bone Joint J. 2021;103-b(1):18–25.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Parvizi J, Gehrke T. Definition of periprosthetic joint infection. J Arthrop. 2014;29(7):1331.

    Article  Google Scholar 

  35. Rothenberg AC, Wilson AE, Hayes JP, O’Malley MJ, Klatt BA. Sonication of arthroplasty implants improves accuracy of periprosthetic joint infection cultures. Clin Orthop Related Res. 2017;475(7):1827–36.

    Article  Google Scholar 

  36. Akhbari P, Jaggard MK, Boulangé CL, Vaghela U, Graça G, Bhattacharya R, et al. Differences between infected and noninfected synovial fluid. Bone Joint Res. 2021;10(1):85–95.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Kildow BJ, Ryan SP, Danilkowicz R, Lazarides AL, Penrose C, Bolognesi MP, et al. Next-generation sequencing not superior to culture in periprosthetic joint infection diagnosis. Bone Joint J. 2021;103-b(1):26–31.

    Article  PubMed  Google Scholar 

  38. Tarabichi M, Shohat N, Goswami K, Parvizi J. Can next generation sequencing play a role in detecting pathogens in synovial fluid? Bone Joint J. 2018;100-b(2):127–33.

    Article  CAS  PubMed  Google Scholar 

  39. Alt V, Rupp M, Langer M, Baumann F, Trampuz A. Can the oncology classification system be used for prosthetic joint infection?: The PJI-TNM system. Bone Joint Res. 2020;9(2):79–81.

    Article  PubMed  PubMed Central  Google Scholar 

  40. Yin H, Xu D, Wang D. Diagnostic value of next-generation sequencing to detect periprosthetic joint infection. BMC Musculoskel Disord. 2021;22(1):252.

    Article  CAS  Google Scholar 

  41. Hipfl C, Mooij W, Perka C, Hardt S, Wassilew GI. Unexpected low-grade infections in revision hip arthroplasty for aseptic loosening: a single-institution experience of 274 hips. Bone Joint J. 2021;103-b(6):1070–7.

    Article  PubMed  Google Scholar 

  42. Goswami K, Parvizi J. Culture-negative periprosthetic joint infection: is there a diagnostic role for next-generation sequencing? Expert Rev Mole Diagnos. 2020;20(3):269–72.

    Article  CAS  Google Scholar 

Download references

Acknowledgements

The authors would like to thank Tao Deng for his participation in the data collection process of this study.

Funding

National Key Research and Development Program of China (No.2020YFC2004900).

Youth Project of National Natural Science Foundation of China (No.82102585).

Author information

Authors and Affiliations

  1. Medical School of Chinese PLA, Beijing, People’s Republic of China

    Hao Li

  2. Department of Orthopedic Surgery, The First Medical Center, Chinese PLA General Hospital, 28 Fuxing Road, Beijing, People’s Republic of China

    Hao Li, Jun Fu, Chi Xu, Libo Hao, Jiying Chen & Wei Chai

  3. Department of Orthopedics, 305 Hospital of PLA, Beijing, People’s Republic of China

    Niu Erlong

  4. Senior Department of Orthopedics, Fourth Medical Center, Chinese PLA General Hospital, Beijing, People’s Republic of China

    Rui LI & Wei Chai

Authors
  1. Hao Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Jun Fu
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Niu Erlong
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Rui LI
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Chi Xu
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Libo Hao
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Jiying Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Wei Chai
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

Study design: Hao Li. Data collection/validation: Hao Li and Rui Li. Data analysis: Hao Li. Result interpretation: Hao Li. Reporting and editing: Jun Fu and Chi Xu. Final approval of the version to be submitted: Jiying Chen and Wei Chai. Project guarantor: Jiying Chen and Wei Chai. The author(s) read and approved the final manuscript.

Authors’ information

All authors come from the Department of Orthopedics, Chinese People’s Liberation Army General Hospital (301 Hospital), 28 Fuxing Rd., 100,853, Beijing, PR China. ErLong Niu come from Department of orthopedics,305 Hospital of PLA, Beijing, People's Republic of China.

Corresponding authors

Correspondence to Jiying Chen or Wei Chai.

Ethics declarations

Ethics approval and consent to participate

This study was performed in accordance with the ethical standards in the 1964 Declaration of Helsinki and ethical approval was obtained by the Ethics Committee of Chinese PLA General Hospital. Institutional review board approval was obtained prior to the commencement of this study and informed consents were obtained before revisions.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1: Supplementary figure 1.

The top 30 variables with highest mean decrease accuracyvalues in random forest model.

Additional file 2: Appendix 2.

The survival curve of the association between microbiota types and clinical outcomes.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Li, H., Fu, J., Erlong, N. et al. Characterization of periprosthetic environment microbiome in patients after total joint arthroplasty and its potential correlation with inflammation. BMC Infect Dis 23, 423 (2023). https://doi.org/10.1186/s12879-023-08390-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12879-023-08390-x

Keywords

BMC Infectious Diseases

ISSN: 1471-2334

Contact us