The patient, a 75-year-old female with a past medical history of hypertension, underwent total left hip arthroplasty at another hospital in 2010. Three years later, she developed recurrent ulcers and oozing from the surgical incisions. She received treatment with antibiotics (specific pharmacological strategies unknown) and underwent surgical procedures, including drainage of abscesses, one-stage revision, and vacuum sealing drainage. Despite these interventions, her symptoms persisted, and bacterial cultures were negative during treatment. The patient was first seen at Jiangsu Provincial Hospital of Chinese Medicine in August 2019 (as shown in Fig. a, b). We performed a two-stage revision on the patient. In the first stage, we debrided the joint and removed all components. We then replaced the left hip with vancomycin bone cement. After the first surgery, the patient received vancomycin intravenously for two weeks and took oral levofloxacin and rifampicin for three months for infection control. Three months later, the patient’s incision had healed well, and two repeat blood and CRP tests showed no abnormalities. We therefore proceeded with a second-stage revision. During the operation, we observed slight inflammatory synovial tissue hyperplasia in the joint cavity. Although the incision had a good postoperative recovery, the patient continued taking oral levofloxacin and rifampicin for three months to prevent infection. Bacterial cultures taken during treatment were negative. In May 2020, the patient returned to our hospital due to yellowish cloudy secretions exuding from the incision of the hip, accompanied by superficial proliferation of granulation tissue (as shown in Fig. c). Given the history of recurrent incisional infections, we decided to perform a one-stage revision, which involved the removal of all components and cement and replacement with a bone cement prosthesis. Following surgery, we sent tissue from the sinus tract to BGI Genomics (Shenzhen, China) for metagenomic next-generation sequencing (mNGS) testing. The test revealed an infection with M. houstonense. Due to the inability to obtain drug susceptibility test results from cultures of the joint fluid, we prescribed a combination of clarithromycin and cefoxitin orally for three months based on the relevant literature to treat the infection []. Subsequently, the patient experienced intermittent infections, prompting us to perform several retention-prosthesis debridement procedures while performing bacterial cultures with infected tissues. The bacterial cultures revealed a variety of highly drug-resistant bacteria, including Staphylococcus capitis, Staphylococcus haemolyticus, and Staphylococcus epidermidis. Based on the susceptibility testing results, we had modified the anti-infective regimen to include intravenous drip meropenem for one week, followed by oral therapy consisting of rifampicin, clarithromycin, and cefoxitin. In November 2020, the patient’s incision exhibited clinical signs of infection again, with yellowish cloudy secretions. After consultation, we discontinued the antibiotics for two weeks and performed a prosthetic debridement and revision surgery. During the intraoperative exploration, the femoral stem did not show any loosening, so we retained the femoral stalk prosthesis and only removed the acetabulum and ball-head prosthesis (as shown in Fig. d) and placed a cemented acetabular cup after thorough debridement. We placed the removed joint prosthesis and wash fluid in a sterile environment and processed the joint prosthesis specimen by vortexing for 30 s, ultrasound at 40 Hz for 5 min, and repeating vortexing for 30 s to obtain the wash fluid. A portion of the wash fluid was sent for mNGS testing, and the mNGS results indicated infection by Mycobacterium houstonense (The test reports were shown in ). Then, we injected 10 ml of wash fluid into blood culture bottles (one bottle for aerobes and one bottle for anaerobes). Additionally, we took a portion of the wash fluid and centrifuged it at 3500 rpm for 5 min. The resulting precipitates were inoculated on Columbia blood agar medium. After 72 h of culture, small dry colonies were grown; after continuation for 24 h, the cultures showed dry and wrinkled colonies. Automated mass spectrometry (MODI-TOF MS) identified them as Mycobacterium fortuitum. In addition, the aerobic bottle became positive after 102.7 h, and pathogenic bacteria were subcultured and identified as M. fortuitum, consistent with the mNGS test result (mNGS only identified M. houstonense, which belongs to the group of M. fortuitum). A broth microdilution method was used to determine the minimal inhibitory concentration (MIC) according to CLSI M24 A2 [], and the drug susceptibility testing results showed that M. houstonense was sensitive to amikacin and moxifloxacin (as shown in Table ). Based on the sensitivities and clinical experience, we replaced the antibiotics with oral moxifloxacin and clindamycin for three months. Thereafter, no abnormalities in routine blood or CRP results were observed. The patient was followed for 24 months without signs of any infection (as shown in Fig. e, f).