Clinical characteristics and risk factors for culture-negative periprosthetic joint infections

Background Culture-negative periprosthetic joint infections (PJIs) can complicate diagnosis and management of PJI. This study aimed to identify risk factors for culture-negative PJI and differences in clinical characteristics between culture-positive and culture-negative PJI group. Methods This retrospective, cross-sectional study evaluated PJI cases obtained between January 2013 and October 2019 at our institution. These PJI cases were divided into culture-positive and culture-negative groups and then compared. The demographics, laboratory findings, and details of patient’s clinical characteristics were investigated. Univariate and multivariate logistic regression analysis were performed to investigate risk factors for culture-negative PJI. Results A total of 109 PJI cases were included in the analysis: 82 (75%) culture-positive and 27 (25%) culture-negative. The mean serum white blood cell (WBC) count, C-reactive protein level, and erythrocyte sedimentation rate in the culture-negative group were significantly lower than those in the culture-positive group (p < 0.05). There were no significant differences between the two groups regarding history of prior antibacterial administration or treatment success rates. Multivariate analysis identified a low serum WBC count as a risk factor for culture-negative PJI (odds ratio = 0.78; 95% confidence interval [CI] = 0.63–0.97; p = 0.027). Conclusions A low serum WBC count is a risk factor for culture-negative PJI, but prior antimicrobial therapy is not. The results suggest that PJI cases with lower levels of systemic inflammation are likely to be culture-negative; therefore, the possibility of a culture-negative result should be considered in suspected cases of PJI with low inflammatory markers, regardless of prior antibiotic exposure.


Background
Periprosthetic joint infection (PJI), one of the most serious complications of joint arthroplasty, occurs in 0.8-1.9% of knee arthroplasties and in 0.3-1.7% of hip arthroplasties [1]. Diagnosis of PJI depends on various factors such as bacterial culture, clinical findings, serum laboratory data, and synovial fluid examination [2]. Early and accurate diagnosis of infection is essential to enable appropriate treatment, guided by internationally-recognized diagnostic criteria [3]. Identification of the infecting organism is particularly crucial for appropriate treatment; this includes antibiotic therapy based on the results of drug susceptibility tests and choosing surgical treatment [4,5]. However, many cases of culture-negative PJI are encountered in a clinical setting; indeed, they account for 7-42.1% of PJI cases [6][7][8][9]. In the absence of culture results and evaluation of antibiotic sensitivity, selecting an appropriate antibiotic therapy or treatment strategy occurs without direct evidence of the responsible pathogens; therefore, culturenegative PJI is a significant clinical issue.
Studies report that culture-negative PJI is caused by several factors, including prior administration of antibiotics before obtaining culture samples, a biofilm around the implant, and culture conditions (e.g., culture medium or incubation time) [10]. In particular, the prior use of antimicrobials is likely to contribute to culturenegative PJI [7,8,11,12]; however, the other risk factors for culture-negative PJI are unclear. In addition, few studies report the clinical characteristics of culturenegative PJI cases compared with culture-positive cases.
The clinical questions addressed by this study are (1) what are the risk factors for culture-negative PJI? (2) Is there any difference of in the clinical characteristics of culture-positive and culture-negative PJI cases?

Methods
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, and pyogenic spondylitis. Figure 1 shows the number of cases in each culturepositive and culture-negative PJI groups. 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 International Consensus Meeting criteria (Tables 1, 2, and 3) [2,[13][14][15] 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 meeting the diagnostic criteria for PJI and being culture negative result for all samples (both preoperative and intraoperative). Table 4 shows the demographic characteristics of each group. 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 55 months (range,  and that for the culture-negative group was 41 months (range, 18-92). The following parameters were examined when collecting a culture sample: serum white blood cell (WBC) count, C-reactive protein level (CRP), 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, 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. [16]. 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. 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 PJIrelated mortality [17].

Culture methods
One or several samples were obtained from tissue, aspiration fluid, synovial fluid, and the drain. 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 [18]. 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% CO 2 . 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). The possibility of tuberculous, fungal, or viral infection was also investigated.

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 chisquare 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. A post hoc power analysis was performed to evaluate that number of 82 and 27 samples was appropriate or not. As the results of calculation from effect size of WBC count (main evaluation parameter) and α err prob 0.05, the power was decided as 0.77 for Student's t tests, and we thought this power must be appropriate sample size. Statistical analysis was performed using BellCurve for Excel version 2.21 (Social Survey Research Information, Tokyo, Japan).

Discussion
This study demonstrated that 25% of PJI cases investigated were culture-negative. There were some differences in clinical characteristics between the culture-positive and culture-negative PJI groups. Levels of systemic inflammatory markers such as serum WBC count, CRP level, and ESR in the culture-negative group were significantly lower than those in the culture-positive group. The treatment success rate of culture-negative PJI was no different from that of culture-positive PJI. A low serum WBC count was identified as a risk factor for culture-negative PJI, but prior antimicrobial therapy was not.
In the current study, the observation that several serum inflammatory markers (all of which are important as first-line tests for diagnosis of PJI) [19] were significantly lower in the culture-negative group than those in the culture-positive group has important clinical implications. In agreement with our results, Choi et al. reported that serum ESR in culture-negative patients was lower than in culture-positive patients [7]. In addition, a previous study by Kheir et al. demonstrated that culturenegative PJI cases had lower CRP values than PJI cases caused by Gram-negative organisms, antibiotic-resistant organisms, Staphylococcus aureus, and Streptococcus species; they also had lower WBC counts than PJI cases caused by Staphylococcus aureus and Streptococcus species [20]. The results of our study suggest that culturenegative PJI is caused by less virulent organisms and that the organism count is lower, resulting in less severe systemic inflammation. In this regard, the ideal cut-off value for these markers may need reconsideration for more accurate screening of PJI [21].
We found that the single risk factor for culturenegative PJI was a low serum WBC count. Although several risk factors for culture-negative PJI have been described [7,8,11,12], serum WBC count was not highlighted. This result suggests that culture-negative PJI should be considered in suspected cases with low levels of inflammatory markers. In such cases, additional diagnostic approaches such as polymerase chain reaction or alpha-defensin tests may be required [22,23].  Prior use of antibiotics had no effect on culturepositive or culture-negative PJI. In addition, the multifactorial logistic regression model did not identify prior antimicrobial therapy as a risk factor for culturenegative PJI. Several studies demonstrate that perioperative administration of prophylactic antibiotics has no effect on culture yield [24][25][26][27]. Tetreault et al. reported that prophylactic antibiotics administered before skin incision had no effect on the results of cultures obtained intraoperatively [24]. A prospective study by Bedencic et al. revealed no differences in diagnostic yield between cultures taken before and after administration of antibiotics to patients with suspected PJI [25]. Thus, prior use of antibiotics may not affect culture results. By contrast, several studies demonstrate that prior use of antimicrobial therapy is a risk factor for culture-negative PJI [7,11,12]. Malekzadeh et al. reported that antimicrobial therapy before diagnosis of PJI is associated with increased odds (odds ratio, 4.7) of being culture-negative [11]. In addition, Ibrahim et al. showed that pre-operative use of antibiotics was a risk factor for culturenegative PJI (odds ratio, 4.1) [12]. A reasonable explanation for a causal association between antibiotics administration and culture-negative results is that antibiotic pressure can induce a viable but nonculturable state in a biofilm [28]. Thus, prior antibiotics use in suspected cases of PJI is a controversial issue; therefore, further studies are required.
In this study, the treatment success rate of culturenegative PJI was similar to that of culture-positive PJI. In agreement with our results, previous studies reported similar outcomes for culture-positive and culturenegative PJI patients [12,29]. In addition, Choi et al. reported that the success rate of infection control was higher in the culture-negative group [7]. Thus, culturenegative PJI may not necessarily be a negative prognostic factor for PJI. Certainly, culture-negative PJI was associated with less virulent organisms and a lower organism count; therefore, the clinical outcome of culture-negative PJI might not be poor despite the lack of culture results and antibiotic sensitivity. On the contrary, inadequate treatment may lead to unsuccessful results for culturenegative PJI. In fact, the rate of treatment success was greater for patients with 2-stage exchange than for those  with irrigation and debridement [30]. In addition, selection of antibiotics is difficult in the absence of information about the causative organism. This study has several limitations. First, the rate of prior antibiotics use was high: 61% and 56% in the culture-positive and culture-negative groups, respectively. A previous study reported a rate of prior use of antimicrobial therapy of 64% in the culture-negative group and 24% in the culture-positive group [11]. In our study, many PJI cases were referred from other hospitals, and many had already been treated with antibiotics. This limitation might have led to selection bias. Second, our sample sizes were relatively small because PJI is a rare condition in a single center. Further multicenter studies need to be performed with more cases. Third, while 5 days culture period was applied in this study, an international consensus meeting recommended that routine cultures should be maintained for 5-14 days [31]. It is, therefore, possible that longer culture duration may have yielded different results. However, Kheir et al. reported that most organisms were cultured within 5 days [32]. The current study was conducted on the assumption that most organisms would be cultured within 5 days and that these durations would limit the risk of isolating contaminant organisms. Fourth, molecular and modern techniques such as sonication of implants, polymerase chain reaction, alpha-defensin tests, and next-generation sequencing were not used in our study; therefore, the proportion of culture-negative PJI might be otherwise. Fifth, the follow-up period of 18 months might be short to assess treatment success of PJI. Treatment success was assessed using the Delphi consensus criteria in our study. The criteria agreed on the definition of midterm follow-up (5 years) and long-term follow-up (10 years or more); however, no agreement was achieved regarding short-term follow-up [17]. Thus, the follow-up period of short-term assessment is a controversial issue. In cases with follow-up of 18 months or more, the infection has not recurred even though the antimicrobial therapy has been terminated. Therefore, we considered that the treatment success in short-term period can be assessed with the follow up of 18 months. Sixth, four Candida albicans infections were found in this study. Fungal PJIs differ in treatment and clinical outcomes compared to bacterial PJIs [33,34]; therefore, this might bias the observation. Finally, there is a significant difference in follow-up period between the culture-positive group and culture-negative group. In addition, treatment methods were not standardized; this may have affected the treatment success rates. In fact, four culture-negative PJI cases were treated with one-stage revision arthroplasty, although it is considered a contraindication for culturenegative PJI [4]. These cases were diagnosed as aseptic loosening before surgery, but pathological results were positive and they were diagnosed as PJI after surgery. These were treated as PJI after surgery, the infection was eradicated.

Conclusions
We identified a single risk factor for culture-negative PJI and key differences between culture-positive PJI and culture-negative PJI patients. A low serum WBC count was associated with culture-negative PJI, but prior antimicrobial therapy was not. A clinical characteristic associated with culture-negative PJI was low levels of inflammatory markers. There was no difference in treatment success rates between the groups. It is possible that culture-negative PJI should be considered in suspected cases of PJI with low inflammatory markers, regardless of prior antibiotics administration.