Inclusion criteria: 1. the diagnosis of PJI was in accordance with the diagnostic criteria of the American Musculoskeletal Infection Society in 2013 (Patel et al., 2016); 2. the infected joint fluid was sent for routine bacterial culture and mNGS testing at the same time; 3. the follow-up time was more than 1 year.

Exclusion criteria: 1. mental, language, psychological and other obstacles can not cooperate with the treatment; 2. combined malignant tumors, autoimmune diseases, etc.; 3. patients who died during the follow-up process.

The study included 15 patients in total, with 11 being male and 4 being female; ages ranged from 28 to 87 years, with an average of 63 years. Type of replacement prosthesis: all were biologically fixed prostheses(Titanium prosthesis, ceramic or polyethylene lining, ceramic femoral head). Infection occurred 6–42 months after replacement surgery, with an average of 22.7 months; the main clinical symptoms were elevated local skin temperature, formation of sinus tracts, and pus flow. The infection lasted between 15 and 115 days, averaging 37.6 days. The preoperative Erythrocyte sedimentation rate (ESR) was 20~88 mm/1 h, with an average of 47.9 mm/1 h. The C-reactive protein (CRP) was 4.7~318.0 mm/L, with an average of 130.1 mg/L. The white blood cell count (WBC) was (3.6~24.6) ×109/L, with an average of 10.1×109/L; The absolute neutrophil count (ANC) was (1.9~67.2) ×109/L, with an average of 12.0×109/L. The general information is shown in Tables 1 and 2.

The two assays were categorized with reference to bacterial operational taxonomic units (OTUs), and there were three categories (nine types) of OTUs: (1) Thick-walled phylum: OTU1 (Staphylococcus aureus, Staphylococcus epidermidis, Staphylococcus simulans, Staphylococcus simulans, and Streptococcus hemolyticus), OTU3 (Enterococcus faecalis), OTU4 (Streptococcus mitis, Streptococcus oralis, pyogenic streptococcus, Streptococcus dysgalactiae); ② Aspergillus phylum: OTU5 (Pseudomonas aeruginosa), OTU6 (Baumannii), OTU7 (Enterobacter chuandaensis); ③ Other groups: OTU2 (Mycobacterium Houstonese, Mycobacterium tuberculosis), OTU8 (Desulfomonilaceae), OTU9 (Aspergillus).

SPSS 25.0 statistical software was used for analysis. Measurement information was tested by Shapiro-Wilk test, all conformed to normal distribution, and the data were expressed as mean ± standard deviation; count information was expressed as rate, and was tested by four-cell tabular chi-square test or Fisher’s exact probability method; and the test level was α = 0.05.

Bacterial culture method tested positive in 10 cases (66.7%) with a total of 12 pathogenic bacteria detected. Among the positive results, two pathogens were detected in one patient, three pathogens in one patient, four pathogens in one patient, and the rest were infected with a single pathogen. Single Gram-positive cocci were found in 7 cases (70.0%); Gram-positive cocci mixed with Gram-negative cocci in 1 case (10.0%); Gram-positive cocci mixed with fungi in 1 case (10.0%); and Gram-negative cocci mixed with Mycobacterium tuberculosis in 1 case (10.0%).

The mNGS test was positive in 15 cases (100%), with a total of 19 pathogenic bacteria detected. Of these, 8 specimens were detected with 2 or more different bacteria, and 7 cases were infected with a single pathogen. Single Gram-positive cocci in 8 cases (53.3%); single Gram-negative cocci in 1 case (6.7%); Gram-positive cocci mixed with Gram-negative bacilli in 2 cases (13.3%); Gram-positive cocci mixed with fungi in 1 case (6.7%); Gram-positive cocci mixed with Mycobacterium tuberculosis in 1 case (6.7%); Gram-positive cocci mixed with fungi and Mycobacterium tuberculosis was detected in 1 case simultaneously (6.7%); Gram-negative cocci mixed with Mycobacterium tuberculosis in 1 case (6.7%). See Table 3 and Figure 1.

There was a statistically significant difference in the positive detection rates between the two methods (P < 0.05). In our group, 10 cases (66.7%) tested positive using both the bacterial culture method and mNGS test, with 5 cases showing identical pathogenic bacterial types, resulting in a 50.0% compliance rate. The time taken from sample delivery to obtaining results was (3.07 ± 0.96) days for the bacterial culture method and (1.67 ± 0.49) days for the mNGS test, with the difference being statistically significant (t=5.03, P = 0.000).

All patients were adjusted antibiotic treatment in full dosage and duration according to the test results. MNGS detected Enterobacter chuandaensis and Streptococcus hemolyticus on the basis of bacterial culture in 1 patient, and the antibiotic treatment regimen was adjusted to penicillin in combination with meropenem; in 1 patient with a negative bacterial culture, mNGS detected Staphylococcus epidermidis and Mycobacterium tuberculosis, and the antibiotic treatment regimen was changed to vancomycin in combination with rifampicin; and in 1 patient mNGS detected Desulfomonilaceae on the basis of bacterial culture, and antibiotic treatment was added to fluconazole; 2 patients had negative bacterial culture, mNGS detected Staphylococcus aureus and Staphylococcus epidermidis respectively, and the antibiotic treatment was changed from broad-spectrum antibiotics to rifampicin combined with levofloxacin; 2 patients’ mNGS detected pyogenic streptococcus and Streptococcus dysgalactiae on the basis of bacterial culture, and the antibiotic regimen was added to the original antibiotic regimen was changed to rifampicin combined with levofloxacin; 1 patient mNGS detected Aspergillus on the basis of bacterial culture, and antibiotic treatment was added with voriconazole; 1 patient with negative bacterial culture and Pseudomonas aeruginosa detected on the basis of mNGS, and was given ceftazidime combined with amikacin; (deleted) In 2 patients with negative bacterial cultures, viral infection was identified by mNGS analysis, and antiviral therapy was combined with valacyclovir on the basis of bacterial-sensitive antibiotics. In the remaining patients, the detection of pathogenic bacteria by the two methods differed slightly, but the antibiotic regimens were adjusted according to the test results(That is, the results of the two methods are different, but the choice of treatment regimen is the same based on the results of the test).

According to the results of bacterial culture and mNGS test culture, we took corresponding sensitive antibiotics or combined with surgery for the patients. Three days after treatment, patients’ ESR was 15–56 mm/1 h, mean 34.8 mm/1 h; CRP was 3.3-68.2 mg/L, mean 25.2 mg/L; WBC was (4.5-12.0) × 109/L, mean 7.4 × 109/L; ANC was (2.6-10.4) × 109/L, mean 5.5 × 109/L. Patients were followed up, and their follow-up period was 13 to 82 months, with a mean of 40.7 months. In 1 patient, the infection returned three months after undergoing one-stage revision surgery, and was given a stage II revision surgery of the hip joint, and was followed up for 5 years afterward, and no signs of infection were seen. There were no signs of infection in the remaining 14 cases, resulting in an infection control rate of 93.3%.

As a key preoperative minimally invasive diagnostic method, synovial fluid culture has the advantage of rapidly obtaining evidence of pathogens to guide initial treatment, but its sensitivity is susceptible to antibiotic use. Prosthetic ultrasound fluid culture is the intraoperative “gold standard” for the diagnosis of PJI, which significantly improves pathogen detection by disrupting the biofilm, especially in low virulence infections and after antibiotic exposure, but must be dependent on prosthesis removal surgery. Intraoperative tissue culture is highly specific, but sensitivity is limited by sampling variability, and all three have the shared limitations of long culture cycles and insensitivity to caustic bacteria.

Microbial culture methods are recognized as the diagnostic standard for PJI, but have the disadvantage of detecting only a number of limited pathogens at a time, and suffer from problems such as inappropriate culture methods, the possibility of false positives in a single culture, and missing the optimal therapeutic window because of the long culture time, which in turn affects the outcome of the treatment (Linjie et al., 2021; Yilong et al., 2021). MNGS is a powerful tool for obtaining pathogenicity evidence, and it is a nontargeted, broad-spectrum pathogenicity screening. MNGS has excellent detection sensitivity in mixed infections or infections with negative culture of pathogenic bacteria (Wang et al., 2020; Liling et al., 2023). However, due to the complexity of PJI, the diversity of pathogens, and the complex interrelationship of microbial and host factors, the use of mNGS in the diagnosis and management of PJI is still in the exploratory stage. A study of 341 patients with PJI found that the sensitivity of NGS sequencing was 94% for detecting joint fluids of the affected limbs using two methods, while the sensitivity of routine bacterial culture was only 70% (Hantouly et al., 2023). In this study, we found that the positive rate of mNGS detection was significantly higher than that of bacterial culture method in patients with postoperative PJI after THA, indicating that mNGS can improve the detection rate of pathogenic bacteria in postoperative PJI after THA, and can not only detect pathogenic bacteria consistent with the bacterial culture method, but also detect more potential pathogenic bacteria, which can effectively avoid the occurrence of false-negative events. Once an infection occurs, its progression is usually rapid, and traditional testing methods often take 3–7 days to produce results, which is very unfavorable to early diagnosis and antibiotic application (Chang et al., 2023). In this study, the mNGS test required significantly less time than bacterial culture, indicating that the mNGS test is valuable for early clarification of pathogenic bacteria and guidance of antibiotic treatment in patients with PJI after THA.

In this study, there were 10 patients with inconsistent results of bacterial culture method and mNGS detection, of which 10 cases were detected by mNGS for more species. One patient had a history of antibiotic use within 1 month, which resulted in false-negative culture results, which might be related to the fact that the pathogenic bacteria were killed by antibiotic treatment but their DNA was still active in the short term, and the pathogenic bacteria could still be detected by mNGS. In this study, mNGS detection revealed Staphylococcus aureus、Staphylococcus epidermidis、Streptococcus dysgalactiae、Pseudomonas aeruginosa、Enterobacter chuandaensis、pyogenic streptococcus、Streptococcus hemolyticus、Mycobacterium tuberculosis、Desulfomonilaceae, whereas the results of the bacterial cultures were negative, which indicates that the detection rate of these types of bacteria by mNGS is higher than that of the culture method (Kurtz et al., 2018; Maimaiti et al., 2021; Maimaiti et al., 2023). Three cases mNGS detected Staphylococcus aureus and Staphylococcus epidermidis respectively, but the bacterial culture method failed to detect them, which may be related to the formation of biofilm. In this study, viral fragments were detected in 2 patients for whom we administered antiviral therapy. It is suggested that in specific cases where patients present with clinical signs highly suggestive of a viral etiology and where conventional tests are negative, detection of the corresponding viral DNA by mNGS may provide the critical evidence needed to justify targeted antiviral therapy trials.

CRP and ESR are important inflammatory indicators of postoperative infection, usually patients’ CRP level can be normalized 2 weeks after surgery, if patients’ CRP level is still higher than normal 1 month after surgery, PJI is considered. Similarly, if patients’ ESR maintains a high level for 3 months consecutively after surgery, infection is considered (Fengnian et al., 2021). For superficial infections that have not yet involved the level of the joint cavity, we can adopt a conventional antibiotic regimen, i.e., vancomycin combined with cephalosporin (Zhang et al., 2020), and then adjust the use of bacterial susceptible antibiotics according to the results of tissue culture. Deep-seated infections, often manifested by biofilm-protected bacterial adherence to the prosthesis (Wildeman et al., 2020), are more drug-resistant, and conservative treatment with antibiotics alone is no longer able to control the infection, and surgical intervention is often required. If the infection does not exceed three weeks postoperatively, debridement with prosthesis preservation can be performed and antibiotics can be used in conjunction with bacterial culture results; if the infection lasts longer than three weeks postoperatively, surgical removal of the prosthesis is required. Hip revision surgery is often complex and difficult, with a high chance of postoperative complications, so thorough debridement of the joint cavity and adequate use of sensitive antibiotics are particularly important. Clinical revision of PJI of the hip mainly includes one-stage and two-stage revision. One-stage revision involves one-time removal of the infected prosthesis, placement of a new prosthesis after thorough debridement, and cemented fixation with long-acting antibiotics, combined with treatment with sensitive antibiotics for more than 6 weeks postoperatively to achieve infection control. It has been shown that one-stage revision surgery for PJI can also achieve infection cure rates comparable to those of two-stage revision surgery, as well as reduce surgical trauma, which is more conducive to functional recovery of the affected limb and improved patient satisfaction (Saul et al., 2023; Xie et al., 2024). Classical two-stage revision surgery provides satisfactory results (Kildow et al., 2022) and is usually performed in two stages, completing the removal of the infected prosthesis and the cemented absences with antibiotics, as well as the placement of a new prosthesis. In this study, eight patients underwent a one-stage revision, one patient was found to have recurrence of infection at 3-month follow-up, and a two-stage revision was performed, with eventual control of the infection; the remaining patients showed no signs of infection at later follow-up. For the choice of antibiotics, we mainly used β-lactams (amoxicillin, vancomycin, etc.), macrolides, and artificial antibiotics (quinolones, sulfonamides, etc.), relying on the high accuracy of the genetic test, and the infection was finally controlled in 14 of the 15 patients in this study, with a high cure rate of 93.3%.

In conclusion, mNGS can rapidly and accurately find the causative organisms in patients with periprosthetic infections after total hip arthroplasty, and combined with surgery and antibiotic use, it can effectively improve the cure rate of PJI after THA. However, this study still has its limitations: (1) the number of cases included in this study is still small, and the follow-up time is short, so the results may be biased; (2) this study is a retrospective study, and the experimental design is relatively simple, and clinical studies with a larger sample size and longer follow-up time are needed to further verify the clinical efficacy; (3) Limited by the different surgical management of patients, this study was unable to collect and compare sonication and intraoperative tissue cultures in a uniform manner; (4) The broader implementation of mNGS as a routine diagnostic tool for PJI, however, faces challenges related to cost, the need for specialized laboratory infrastructure and bioinformatics support, and a longer turnaround time compared to conventional culture, which may currently restrict its use to tertiary referral centers or complex cases.