Role of asymptomatic bacteriuria on early periprosthetic joint infection after hip hemiarthroplasty. BARIFER randomized clinical trial

To evaluate preoperative asymptomatic bacteriuria (ASB) treatment to reduce early-periprosthetic joint infections (early-PJIs) after hip hemiarthroplasty (HHA) for fracture. Open-label, multicenter RCT comparing fosfomycin-trometamol versus no intervention with a parallel follow-up cohort without ASB. Primary outcome: early-PJI after HHA. Five hundred ninety-four patients enrolled (mean age 84.3); 152(25%) with ASB (77 treated with fosfomycin-trometamol/75 controls) and 442(75%) without. Despite the study closed without the intended sample size, ASB was not predictive of early-PJI (OR: 1.06 [95%CI: 0.33–3.38]), and its treatment did not modify early-PJI incidence (OR: 1.03 [95%CI: 0.15–7.10]). Neither preoperative ASB nor its treatment appears to be risk factors of early-PJI after HHA. ClinicalTrials.gov Identifier: Eudra CT 2016-001108-47

Bacterial colonization of the genitourinary tract as an infection cause of hip prostheses due to a hematogenous seeding or skin contamination by continuity has been suggested. This asymptomatic colonization is called asymptomatic bacteriuria (ASB), and its prevalence reaches 30-50% in older women in long-term care facilities [6]. Published studies demonstrated that preoperative ASB treatment in elective total hip and knee arthroplasties has no impact in early-PJI rates [7][8][9][10][11]. However, its impact on HHAs is controversial. A singlecenter study concluded that treating ASB in geriatric patients with a femur fracture decreases the risk of PJIs [12].
We evaluate preoperative ASB treatment's impact on the cumulative incidence of early-PJI in patients undergoing HHA for a hip fracture. We hypothesized that preoperative ASB treatment in these populations could decrease the incidence of early-PJI caused by Gram-negative bacilli (GNB).

Patients and methods
BARIFER was a phase IV, multicenter, randomized, openlabel, and parallel-group clinical trial conducted at 11 sites in Spain designed to evaluate the impact of treating ASB on the incidence of early-PJI in HHA.
All patients provided informed consent. Protocol approval was obtained from an independent ethics committee at each site. The trial (EudraCT 2016-001108-47) was performed under the principles of the Declaration of Helsinki. Adherence to the Consolidated Standards of Reporting Trials [13] (CONSORT) is supported by the completed checklist provided as Supplementary material.
Patients >18 years requiring HHA for fracture were recruited. Exclusion criteria include any concomitant infection requiring antibiotics and hip fractures treated with screws or THA.
Urine analysis was performed before HHA surgery. ASB referred to a urine culture growing ≥10 5 colony-forming units/ mL of a bacterial species in a patient lacking symptoms of a urinary tract infection (UTI). Standard procedures identified all microorganisms isolated. Antimicrobial susceptibility was performed by microdilution (Vitek bioMérieux, France). The MIC values of fosfomycin were interpreted according to EUCAST criteria 2012 (version 2.0) guidelines (www.eucast.org).
Participants with ASB were randomly assigned in a 1:1 ratio, centralized, and stratified by center, to receive 3 g of fosfomycin-trometamol (oral route) vs. no treatment, between 24 and 6h before surgery. A parallel follow-up cohort of HHA candidates without ASB was established.
Preoperative antibiotic prophylaxis was decided according to each center protocol (Supplementary Table 1). All patients were followed for three months after HHA or until early-PJIs or death was diagnosed, whichever occurred first.
PJIs occurring within 3 months after HHA were considered early-PJIs [14]. Patients were diagnosed with a PJI following diagnostic criteria established by the Infectious Diseases Society of America [15]. In the case of early-PJI, a new visit was completed in which the microorganism causing the infection was recorded.
The primary outcome was cumulative incidence of early-PJI after preoperative ASB treatment. Secondary analyses included global incidence of ASB and early-PJI, risk factors for early-PJI, and fosfomycin treatment safety.

Statistical analysis
Categorical variables were presented as numbers and percentages, and quantitative variables as a median and interquartile range or a mean and standard deviation, as appropriate. Comparative analyses were performed using X 2 or Fisher's test for categorical variables and Student's t test or Mann-Whitney U test for continuous variables. The level of significance was set to p < 0.05. Predictors of early-PJI were determined by univariate analysis. The Kaplan-Meier method was used to describe cumulative probability early-PJI stratified by study group.
The EAST program calculated the sample size. We assumed a prevalence of ASB up to 20% in men and 50% in women, an incidence of 9% of early-PJI, and an expected 50% reduction with fosfomycin treatment with a test power of 90% and alpha error of 0.05. We needed 1394 patients (697 in each treatment group). An interim analysis was planned to stop the study if it would not be possible to test the hypothesis. Analyses were performed with the STATA 15.1 software (StataCorp, TX, USA) in the intention-to-treat (ITT) population.

Results
A total of 594 patients were included from September 2016 to November 2018. Overall, 420 (71.0%) were women, and the mean age was 84.3 years. ASB was diagnosed in 152 (25%) patients, 77 treated with fosfomycin and 75 untreated controls. Figure 1 shows the flow chart of patients' distribution.
Patients with ASB versus the non-ASB group mainly were women, with a higher Charlson comorbidity index score and more commonly with urinary incontinence (  considered at risk for Fosfomycin resistance as they were under chronic antibiotic prophylaxis with Fosfomycin-trometamol for recurrent cystitis. Therefore, they were not randomized were susceptible to fosfomycin. Table 2 compares baseline characteristics of treated and untreated patients with ASB. HHA implants were 65.46% cemented with antibiotics (64% with single-antibiotic and 36% with dual-antibiotic Vancogenx®).   Overall, 558(93.9%) patients (140 with ASB and 418 without) completed three months of follow-up (Table 3). Early-PJI rate was 2.5% (15 of 594 patients). Of these 15 patients, 4 (2.7%) showed previous ASB, but only two received fosfomycin (Table 3). Our trial showed that treating preoperative ASB does not modify the incidence of early-PJI (OR: 1.03 [95%CI: 0.15-7.10], p= 0.9787). Of note, all early-PJI occurred within 60 days after HHA (Fig. 2). Table 4 shows the etiology of the 15 early-PJIs. We observed a lack of correspondence between ASB and early-PJI causing microorganisms. Univariate analysis of risk factors for early-PJI is presented in Table 5. Preoperative ASB was not a predictor of early-PJI (OR: 1.06 [95%CI: 0.33-3.38], p= 0.9228).
AEs related to fosfomycin occurred in 4 patients, all of them of mild intensity. Three patients suffered from nausea, and one reported dizziness (Supplementary Table 3).

Discussion
Identifying potentially modifiable preoperative risk factors of PJIs is of great interest. Experts traditionally recommended treating ASB before THA [16][17][18][19], although the latest published studies contradict this recommendation [7,8,10,11]. There are only two previous randomized controlled trials addressing this in THA and HHA [7,8]. Our findings suggest that preoperative ASB treatment does not impact on the reduction of early-PJI after HHA. BARIFER is the first randomized trial that only enrolled this subgroup of patients. ASB prevalence in our cohort was 25% higher than in THA candidates [7,8,16] and consistent with data reported for HHA [20]. Female sex, adjusted Charlson index, and urinary incontinence are significantly more prevalent in the ASB group as previously reported [7]. In our trial, almost 90% of the identified GNB causing ASB were susceptible to fosfomycin as previously published [21,22]. The efficacy of a single dose of fosfomycintrometamol for uncomplicated lower UTI maybe be comparable to standard regimens with fluoroquinolones or trimethoprim/sulfamethoxazole [23] and easier to administer. On this basis, it was chosen as preoperative treatment. Fosfomycin has a good tolerability with a low incidence of adverse events (AEs), mainly mild and transient gastrointestinal symptoms [23]. This coincides with our study as only four patients experienced associated nausea or dizziness.
Only four patients with ASB showed an early-PJI which represents an incidence of 2.7%. Although this is lower than expected [4,7,8], it is consistent with the latest data collected in the VINCat registry (surveillance database of nosocomial infections in Catalonia) [5]. When investigating risk factors for early-PJI, our study focuses on preoperative ASB. Among our series, ASB is not a risk factor for early-PJI unlike other published data stating that, although the risk of PJI is not influenced by ASB treatment, there seems to be an increased risk of PJI in this population [7]. It should also be noted that in no case, the microorganism causing ASB was the same as the one causing early-PJI and this has also been described by other authors [7,24]. Our experience shows that ASB treatment does not modify the incidence of early-PJI. Although we observed a delay from HHA surgery to onset of infection of about 10 days higher in patients treated with fosfomycin, the exceptionally low number of events prevents us from reaching any conclusion. Consequently, since we could not demonstrate a potential benefit in treating preoperative ASB, we do not recommend systematic urinalysis screening and treatment.
Besides, the percentage of antibiotic-loaded cement used is also significant. Published studies show that it reduces the rate of PJIs in HHA with no associated increase in complications [25][26][27]. This approach could justify a global reduction of early-PJ rates compared to our previous incidence between 2011 and 2013 [4].
Finally, global mortality in our study is high (9%) and can be explained by the population's age and comorbidity, particularly among those with ASB, as evidenced by the high Charlson comorbidity index values [1,28].
The main limitation of our study is the small sample size. The difficulty of obtaining the informed consent signed and all study requirements at least 6 h before surgery made our inclusion rate slow. We did an interim analysis that showed that it would not be possible to test the hypothesis so we decided to end the study. It is also possible that we overestimate ASB and early-PJI after HHA incidences since our calculations were based on our previous experience [4] and data published regarding ASB prevalence in the elderly [29]. ASB and early-PJI after HHA incidences were lower than expected so the study might be underpowered to confirm the hypothesis. The study's main strengths are its randomized design and recruiting geriatric patients (often underrepresented in clinical trials) all of them undergoing HHA.
In conclusion, our results suggest that ASB appears not to be an independent risk factor for early-PJ, and its treatment did Fig. 2 Distribution of the time to early-PJI according to study group. Early-PJI, early periprosthetic joint infection not reduce the incidence of early-PJI after HHA. Therefore, we cannot recommend routine screening and treatment of preoperative ASB in HHA surgery.
Part of this study was presented at the XXIII Congreso Nacional de la Sociedad Española de Enfermedades Infecciosas y Microbiología Clínica, which took place in Madrid, on May 23-25, 2019.