This retrospective, cross-sectional study evaluated 2776 sterile orthopedic samples obtained from 1393 cases at our institution between January 2013 and October 2019. Samples were collected using sterile technique from patients with high clinical suspicion of orthopedic infections, including PJI, infection around the implant, surgical site infection, septic arthritis, osteomyelitis, pyogenic spondylitis.
A total of 1234 cases (2391 samples) without a prosthetic joint were excluded from the evaluation. Of the 159 cases (385 samples) with a prosthetic joint, 109 (299 samples) were diagnosed as PJI according to the Musculoskeletal Infection Society criteria [2] and were included in the study. Of these 109 PJI cases, 82 (204 samples) were culture-positive PJI and 27 (95 samples) were culture-negative PJI. Culture-negative PJI was defined as a negative culture result for all samples (both preoperative and intraoperative). There were 27 males and 55 females in the culture-positive group and 13 males and 14 females in the culture-negative group. The mean age was 70 years (range 18-87) in the culture-positive group and 72 years (range 18-86) in the culture-negative group. The mean follow-up period for the culture-positive group was 39 months (range, 6–87) and that for the culture-negative group was 26 months (range, 6–78). The following parameters were examined when collecting a culture sample: serum white blood cell (WBC) count, C-reactive protein (CRP) level, erythrocyte sedimentation rate (ESR), D-dimer, American Society of Anesthesiologists (ASA) physical status classification, comorbidity, type of implant, type of sample, type of infection, use of antibacterial drugs within 2 weeks before collecting culture samples (not surgery), type of antibiotic, treatment methods, and treatment success rates. Type of infection was classified as early postoperative, acute hematogenous and chronic infection according to Tsukayama et al [13]. Treatment methods were divided into one-stage revision arthroplasty, two-stage revision arthroplasty, debridement antibiotics and implant retention (DAIR), resection arthroplasty, and observation. Although a standardized protocol was not used to determine treatment methods, multiple different surgeons used similar principles in determining treatment of choice. One-stage revision arthroplasty was performed when it was not culture-negative results, not infection with antibiotic-resistance organism, and there were no soft tissue problems [4, 5]. Two-stage revision arthroplasty treated with debridement of the infected joint, removal of the prosthetic components and placement of antibiotic-loaded hydroxyapatite block or cement spacer in the first stage. Antibiotics were administered for 6-12 weeks. Inflammatory markers such as ESR, CRP level, and WBC count and joint aspiration fluid were obtained. If there was no evidence of infection, revision arthroplasty was performed. Two-stage revision arthroplasty were mainly performed for chronic infections. Early postoperative and acute hematogenous infections were mainly treated with DAIR. In cases with the patient’s general condition was poor, a resection arthroplasty or observation was chosen.
A standardized protocol for antibiotic treatment was not used, however, we used similar principles. For culture-positive PJI patients, intravenous antibiotics were administrated for 10–14 days, and then oral antibiotics were administrated with evaluation of antibiotic sensitivity. The administration period was approximately 6–12 weeks in total. For culture-negative PJI patients, antibiotics with broad spectrum activity were selected, and the administration period was approximately 6–12 weeks in total.
Treatment success was assessed using the Delphi consensus criteria: (1) infection eradication, characterized by a healed wound without fistula, drainage, or pain, and no reinfection by the same organism strain; (2) no subsequent surgical intervention for infection after reimplantation surgery; and (3) no occurrence of PJI-related mortality [14].
Culture methods
One or several samples were obtained from tissue, aspiration fluid, synovial fluid, and the drain. In most patients suspected of PJI, we obtained puncture from periprosthetic joint at first, and 3–5 tissue biopsy samples from different places were obtained in patients who underwent surgery. There are cases which puncture could not be obtained and surgery was not performed, therefore, there are cases of only puncture sample and only biopsy sample. In this study, cases which all samples obtained in one patient were culture-negative were defined as culture-negative PJI. All samples were sent to the microbiology laboratory. Standard culture was performed at first. If samples were not cultured using standard culture methods, an enrichment culture method was performed using broth culture medium for 5 days according to a previous study [15]. Briefly, standard culture was performed using 5% sheep’s blood agar plates containing peptone and sodium chloride. The enrichment culture comprised semi-solid Gifu anaerobic medium containing peptone, soy peptone, protease peptone, beef extract, yeast extract, liver extract, glucose, starch, L-tryptophan, L-cysteine hydrochloride, thioglycolic acid sodium salt, L-arginine, vitaminK1, hemin, potassium dihydrogen phosphate, and sodium chloride. The standard culture method was conducted for 24 h at 35°C/5% CO2. Enrichment culture was performed for 5 days at 35°C in ambient air. When the enrichment culture was negative, it was identified as culture-negative. Bacteria were identified by analysis of biological properties using MicroScan WalkAway and a combo panel (Beckman Coulter, Brea, CA).
Statistical analysis
The PJI cases were divided into culture-positive and culture-negative groups. The two groups were compared using Student’s t-tests (continuous variables) or Chi-square test (categorical variables). Residual analysis was performed after Chi-square tests to observe the significance difference between groups. Risk factors for culture-negative PJI were evaluated using univariate and multivariate logistic regression analysis. Univariate logistic regression analysis was performed to identify the independent influence of each variable in Table 1. Baseline variables with a p-value < 0.10 in univariate analysis were included in multivariate analysis. A p-value < 0.05 was considered significant.