111. Duration of Antibiotic Therapy after Debridement and Implant Retention in Patients with Periprosthetic Joint Infections

Abstract Background Debridement, antibiotics, and implant retention (DAIR) is appropriate for select acute postoperative and hematogenous periprosthetic joint infections (PJIs). However, the optimal duration of antimicrobial therapy in patients treated with DAIR has not been defined. Therefore, we aimed to identify the ideal duration of parenteral and oral antibiotics after DAIR. Methods We performed a retrospective study of patients >18 years of age with hip or knee PJI managed with DAIR between January 1, 2008, and December 31, 2018, at Mayo Clinic. PJI was defined using criteria adapted from the International Consensus Meeting on PJI. The outcome was defined as either PJI recurrence or unplanned reoperation due to infection. Joint-stratified Cox proportional hazards regression models with time-dependent covariates were used to assess nonlinear effects of antibiotic duration. Hazard ratios were computed based on prespecified time points for comparison, whereas p-values represented the overall effect across the entire range of durations. Results There were 247 unique episodes of PJI in 237 patients during the study period. Parenteral antibiotics were given in 99.2% of cases (n=245). This was followed by chronic oral antibiotic suppression in 92.2% (n=226) with a median duration of 2.2 years (1.0-4.1). DAIR failed in 65 cases over a median follow-up of 4.4 years, with a 5-year cumulative incidence of 28.1%. After adjustment for risk factors, there was no significant association between duration of parenteral antibiotics and treatment failure (p=0.203), with no difference between four versus six weeks (HR 1.11; 95% CI 0.71-1.75) (Figure 1). However, both use and longer duration of oral antibiotic therapy was associated with a lower risk of failure (p=0.006). To account for the possibility that this association was driven by results during early follow-up, conditional analyses at one- and two-year follow-up were performed. Both showed a significantly lower risk for a longer duration of antibiotics (Figure 2). Figure 1. Time-Dependent Analysis of Parenteral Antibiotic Duration Figure 2. Time-Dependent Analysis of Oral Antibiotic Suppression Duration Conclusion After DAIR, efficacy from four weeks of parenteral antibiotics was no different from six weeks when followed by chronic oral antibiotic suppression. Our results could not establish an optimal duration but suggested that continuing suppression portends a lower risk of failure of DAIR. Disclosures Elie Berbari, MD, Uptodate.com (Other Financial or Material Support, Honorary unrelated to this work) Matthew P. Abdel, MD, Stryker and AAOS Board of Directors (Board Member, Other Financial or Material Support, Royalties) Aaron J. Tande, MD, UpToDate.com (Other Financial or Material Support, Honoraria for medical writing)


Session: O-23. New Developments in Antibiotic Efficacy
Background. Debridement, antibiotics, and implant retention (DAIR) is appropriate for select acute postoperative and hematogenous periprosthetic joint infections (PJIs). However, the optimal duration of antimicrobial therapy in patients treated with DAIR has not been defined. Therefore, we aimed to identify the ideal duration of parenteral and oral antibiotics after DAIR.
Methods. We performed a retrospective study of patients >18 years of age with hip or knee PJI managed with DAIR between January 1, 2008, and December 31, 2018, at Mayo Clinic. PJI was defined using criteria adapted from the International Consensus Meeting on PJI. The outcome was defined as either PJI recurrence or unplanned reoperation due to infection. Joint-stratified Cox proportional hazards regression models with time-dependent covariates were used to assess nonlinear effects of antibiotic duration. Hazard ratios were computed based on prespecified time points for comparison, whereas p-values represented the overall effect across the entire range of durations.
Results. There were 247 unique episodes of PJI in 237 patients during the study period. Parenteral antibiotics were given in 99.2% of cases (n=245). This was followed by chronic oral antibiotic suppression in 92.2% (n=226) with a median duration of 2.2 years (1.0-4.1).
DAIR failed in 65 cases over a median follow-up of 4.4 years, with a 5-year cumulative incidence of 28.1%. After adjustment for risk factors, there was no significant association between duration of parenteral antibiotics and treatment failure (p=0.203), with no difference between four versus six weeks (HR 1.11; 95% CI 0.71-1.75) ( Figure  1). However, both use and longer duration of oral antibiotic therapy was associated with a lower risk of failure (p=0.006). To account for the possibility that this association was driven by results during early follow-up, conditional analyses at one-and two-year follow-up were performed. Both showed a significantly lower risk for a longer duration of antibiotics ( Figure 2). Background. Identifying infectious etiology is essential for appropriate patient management, including antibiotic use. A host-protein signature for differentiating bacterial from viral infection has exhibited robust performance (AUC of 0.9, 95% CI 0.86-0.95) in prior studies. Performance data was lacking for a broad pediatric population recruited in emergency departments (EDs) and urgent care centers (UCCs).
Methods. Non-immunocompromised children were recruited prospectively from 5 EDs and 3 UCCs in the U.S. and 1 ED in Israel between May 2019 and August 2020. Eligibility required physician's clinical suspicion of acute infection and reported fever. Reference standard etiology was adjudicated by experts based on clinical, laboratory, radiological, microbiological and follow-up data. For the primary analysis, experts blinded to one another, to the host-signature results and also to procalcitonin and CRP, classified cases as bacterial or viral. For the secondary analysis, experts blinded to one another and the host signature results, were permitted to classify cases as bacterial, viral or indeterminate; indeterminates were removed from the secondary analysis. Host signature (comprising TRAIL, IP-10 and CRP; MeMed BV®) was measured using a rapid platform (MeMed Key®) generating a bacterial likelihood score (0-100) in 15 minutes.
Results. The study cohort comprised 162 children (median age, 5.5 yrs; interquartile range, 8.5), of whom 69 (43%) presented within 2 days of symptom onset and 37 (23%) were hospitalized for a median of 3 days. Respiratory tract infection was the predominant syndrome (11% lower and 44% upper). Host signature attained AUC 0.87 (0.74-1) and 0.92 (0.79-1) in the primary and secondary analysis, respectively. With higher the signature score, there was a significantly higher likelihood of bacterial infection (p< 0.001; Table 1). The 3 bacterial infections assigned score < 35 (false negative) would have been identifiable by physical examination (Table 2).
Increasing host signature score is associated with increasing likelihood of bacterial infection across both the primary and secondary cohort The performance of the host signature score in differentiating between bacterial and viral infection was evaluated by allocating children to one of five score bins and within each bin according to their adjudication label and determining if there is a meaningful increase in the relative likelihood of bacterial infection across the bins based on the Cochrane-Armitage test of trend. PPV, positive predictive value. NPV, negative predictive value. *Includes patients adjudicated as non-infectious Three children assigned a bacterial adjudication label and a score of 35 or less (false negatives) have bacterial infections identifiable in physical exam