What is the evidence base of used aggregated antibiotic resistance percentages to change empirical antibiotic treatment? A scoping review

Systematic review What is the evidence base of used aggregated antibiotic resistance percentages to change empirical antibiotic treatment? A scoping review Ali Auzin 1 , * , Menoeska Spits 2 , Evelina Tacconelli 3 , Jos e Rodríguez-Ba~ no 4 , Marlies Hulscher 5 , Eddy Adang 1 , Andreas Voss 6 , Heiman Wertheim 1 1) Radboud University Medical Centre, Nijmegen, the Netherlands 2) Utrecht University, Utrecht, the Netherlands 3) University of Verona, Italy 4) Infectious Diseases and Microbiology, Hospital Universitario Virgen Macarena and Medicine Department, University of Seville/Biomedicine Institute of Seville, Spain 5) Scientific Centre for Quality of Healthcare (IQ Healthcare), Radboud Institute for Health Sciences, Radboud University Medical Centre, Nijmegen, the Netherlands 6) Canisius Wilhelmina Ziekenhuis (CWZ), Nijmegen, the Netherlands article info Article history: Received 3 August 2021 Received in revised form 8 November 2021 Accepted 4 December 2021 Available online 11 December 2021 Editor: A. Kalil Keywords: Antibiotic resistance Antibiotic use Antimicrobial stewardship Guidelines Thresholds abstract Objectives: Antibiotic resistance requires continuous monitoring by experts to decide whether empirical antibiotic therapies (EATs) should be replaced by alternative antibiotics. The exact moment and criteria for this change are unclear and generally based on consensus between experts. This scoping review aims to identify from the literature the resistance thresholds used for a change in EAT and the criteria on which they are based. Methods: Scoping review for which a comprehensive structured literature search was conducted. Rayyan, software for systematic reviews, was used for the screening of abstracts and titles. Data sources were Pubmed and a hand-search of reference lists and grey literature. Papers were eligible if they concerned any type of bacterial infectious disease and mentioned or defined antibiotic resistance thresholds for decision-making purposes for EAT. The inclusion and analysis of articles was done by two researchers; any conflicts were resolved through discussion or by consulting a third reviewer. Results: We identified 3146 unique papers. Following title/abstract screening, 125 papers were comprehensively read, and 16 papers were included. The included papers gave thresholds for urinary tract infections, respiratory tract infections, meningitis, skin and soft tissue infections, gonorrhoea, and bone and joint infections. Six criteria were found that were commonly used to base the thresholds on. These were: disease severity, efficacy of treatment, adverse drug events, risk of Clostridioides difficile infection, costs, and increased resistance. The number of criteria used to define each threshold varied from one to six between papers. Conclusions: The thresholds used for EATs are few, commonly based on expert opinion estimates, and can therefore have broad ranges. Used criteria underlying reported thresholds are heterogenous and require standardization. Considering the rising trend in resistance, there is a clear need for rigid tools to determine thresholds in order to support guideline development with the best and timely evidence. Ali Auzin, Clin Microbiol Infect 2022;28:928 ©2021 The Authors. Published by Elsevier Ltd on behalf of European Society of Clinical Microbiology and Infectious Diseases. This is an open access article under the CC BY license (http://creativecommons.org/ licenses/by/4.0/). *Corresponding author. Ali Auzin, Geert Grooteplein Zuid 10, P.O. Box 9101, 6500 HB Nijmegen, the Netherlands. E-mail address: [email protected] (A. Auzin). Contents lists available at ScienceDirect Clinical Microbiology and Infection journal homepage: www.clinicalmicrobiologyandinfection.com https://doi.org/10.1016/j.cmi.2021.12.003 1198-743X/©2021 The Authors. Published by Elsevier Ltd on behalf of European Society of Clinical Microbiology and Infectious Diseases. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Clinical Microbiology and Infection 28 (2022) 928e935 Introduction One major challenge in the field of medical microbiology and infectious diseases is antibiotic resistance (ABR). Most infections requiring antibiotic therapy are treated empirically with antibiotics covering the likely causative pathogens of the infection before laboratory results (if any) are known. Empirical antibiotic treatment (EAT) should be effective against the majority of microorganisms causing infections. If the causative pathogen is resistant to the administered EAT, the infection may progress and lead to a lengthened hospital stay and an increase in morbidity and mortality. If microbiological testing is done, the EAT can be switched and directed towards the detected pathogen (directed therapy) in individual patient care. At an aggregated (e.g. regional or national) level, resistance data are continuously monitored to verify whether local or national selected EATs for specific infectious disease syndromes are still appropriate. Recommendations for EATs are published in local and (inter)national guidelines, using available evidence and expert consensus. Recommendations are usually based on several criteria, including epidemiology and the susceptibilities of the causative organisms, efficacy of antibiotic options, disease severity, adverse drug events, and costs [1e3]. Infectious diseases are dynamic, and resistance levels change over time, possibly requiring regular changes in EAT recommendations. The optimal timing of such epidemiologically based adjustments is unknown and depends on the judgement of guideline committees who should monitor whether their recommendations are still valid. Clear guidance and tools are needed for these committees to ensure that empirical choices are optimal and timely. We need a definition of an antibiotic-resistance threshold in surveillance data above which EAT recommendations should be adjusted for specific infections. Although thresholds are used, consensus is lacking concerning their objective determination. As a result, EAT adjustment might be initiated too late, or experts might arrive at different recommendations. This study presents a systematic scoping review of the existing literature on the use and estimation of resistance thresholds used for EAT recommendations, if any. It describes the thresholds that are applied for various infectious diseases, as well as the evidence on which they are based and the criteria that are used to determine them. The insights generated by this review could contribute to the improvement and standardization of selecting the most optimal EAT. Methods This scoping review follows in part the guidelines of the Preferred Reporting Items for Systematic reviews and Meta- Analysis (PRISMA) statement for the search strategy [4], Prospero ID: CRD42020147888. Information sources and search strategy The MEDLINE database was screened for articles published between 1st January 1976 and 3rd May 2021 using a search strategy combining terms relating to “antibiotic or antibacterial agents”, “bacterial infections”, and “thresholds, level, or rates”. The full list of search terms is presented in the Supplementary Material. References were processed using Endnote x9 (Clarivate Analytics, Philadelphia, Pennsylvania, USA). The references of full-text articles were searched manually for additional potential studies. For articles not available online, they were obtained by contacting the authors. Eligibility criteria The review considered published articles that (a) concerned any type of bacterial infectious disease syndrome (IDS) and (b) mentioned or defined resistance thresholds/levels for decisionmaking purposes for empirical antibiotic treatment. If an abstract suggested that the article contained data on thresholds, the paper was included for full-text screening. Studies were included after full-text screening if they (a) were written in English, (b) were published in full, and (c) defined the threshold through expert consensus or other means (e.g. modelling studies). The eligibility criteria were intentionally broad in order to maximize the sensitivity of the search, given the expected paucity of available literature on the subject. A manual search was conducted of grey literature (e.g. guidelines mentioning ABR thresholds), and relevant sources were included. A manual search was also conducted of antibiotic policy documents regarding IDS (e.g. urinary tract infections (UTIs), bacterial respiratory tract infections (RTIs), bacterial meningitis, bacterial skin and soft tissue infections (SSTIs)). Infections for which the primary intervention was not antibiotic treatment and for which antibiotics were supportive (e.g. abscess, cholecystitis) were not considered. As a prerequisite, policy documents were required to have been compiled by the European Society for Clinical Microbiology and Infectious Diseases (ESCMID), the Infectious Diseases Society of America (IDSA), or other prominent international guideline-makers. National guidelines were not considered. Study selection Study selection was performed using Rayyan (http://rayyan.qcri. org/), and all duplicates and non-English studies were excluded. The titles and abstracts of all records were screened independently by two reviewers (AMA and MS) to asses full-text eligibility. Disagreements were resolved by involving a third reviewer (HFLW). Assessment of study quality Study quality was assessed independently by two reviewers (AMA and MS) using various tools developed by the Critical Appraisals Skills Program (CASP). Disagreements were discussed in meetings and, if not resolved, a third reviewer was involved (HFLW). Data extraction Data were extracted independently by two reviewers (AMA and MS), and all results were cross-checked. The data extracted by the two reviewers contained descriptive data such as study characteristics. The data concerning the threshold were categorized as follows: the threshold, the evidence base, IDS/pathogens, antibiotics, and criteria used. Results The inclusion and exclusion processes are presented in Fig. 1.Of 3146 unique papers identified in the initial database search, 112 met the criteria for full-text review, and six defined thresholds and were included for the review. In addition, 13 guidelines on various infections were screened: UTIs (uncomplicated (uUTI) and complicated (cUTI) (n¼2) [3,5], RTIs (n¼4) [1,6e8], bacterial meningitis (n¼3) [9e11], SSTIs (n¼2) [12,13], prosthetic joint infections (n¼1) [14], and sepsis (n¼1) [15]. Four guidelines mentioned thresholds and were included [1,3,5,6]. References of A. Auzin et al. / Clinical Microbiology and Infection 28 (2022) 928e935 929 eligible papers identified an additional six papers that defined thresholds and were included. The 16 papers that were included yielded a total of 19 thresholds, four of which concerned the same threshold and were excluded. We thus identified 16 unique thresholds for the following syndromes: UTI (n¼8) [2,3,5,16e20], RTI (n¼3) [1,6,21], meningitis (n¼1) [22], SSTI (n¼1) [23], bone and joint infections (n¼2) [24,25], and gonorrhoea (n¼1) [26](Table 1). Urinary tract infections Thresholds in guidelines and review For UTIs thresholds were determined for two antibiotics: cotrimoxazole and ciprofloxacin. The IDSA guideline recommends cotrimoxazole if resistance among Escherichia coli strains causing uUTI is <20% [3]. This threshold was first defined in 1999 [2] and based on expert opinion, taking clinical studies into account [27e30]. The 2010 updated IDSA guideline adds a >10% ciprofloxacin threshold for complicated UTI, although no scientific evidence for this threshold is provided. The 2018 European Agency for Urology (EAU) guideline recommends the same thresholds of 20% (uUTI) for cotrimoxazole and 10% (cUTI) for fluoroquinolones (FQs) for empirical treatment [15]. The same thresholds are mentioned in a 2020 review [20]. Cost-effectiveness studies to determine thresholds Four cost-effectiveness studies pertained to UTIs. Although the thresholds vary, all studies define thresholds for cotrimoxazole or FQs, and all use similar methodologies (Supplementary Material). In 2001, Le et al. found that FQ is more cost-effective when cotrimoxazole resistance in E. coli exceeds 22% [16], based primarily on the cure rates for cotrimoxazole and FQ. Perfetto et al. performed a similar study in 2004, considering several case scenarios with varying resistance rates for ciprofloxacin and cotrimoxazole in E. coli isolates. If the cotrimoxazole resistance rate exceeded 4.3%, ciprofloxacin was more cost-effective at the resistance rate of 1% for ciprofloxacin in E. coli. The resistance rate of ciprofloxacin had a significant impact on the variation of the threshold, as the threshold for switching to ciprofloxacin became 13.3% when a resistance rate of 10% was applied for ciprofloxacin [18]. In a 2007 Fig. 1. Flow chart showing the inclusion and exclusion processes. A. Auzin et al. / Clinical Microbiology and Infection 28 (2022) 928e935930 Table 1 Characteristics of articles and thresholds defined Author Year Infectious disease syndrome Document type Antibiotics Bacteria Threshold Comments Warren et al. [2] 1999 UTI IDSA Guideline Cotrimoxazole/ trimethoprim E. coli Cotrimoxazole/trimethoprim threshold of 10% or 20% indicates switch to FQ Recommendation based on expert opinion Le TP et al. [16] 2001 UTI Cost minimization and sensitivity analysis Cotrimoxazole and fluoroquinolones. E. coli Cotrimoxazole threshold of 22% indicates switch to FQ Modelling study. Fluid threshold, changing according to FQ resistance and cure rates Bonkat G et al. [5] 2018 UTI EAU guideline Cotrimoxazole, fluoroquinolones Uropathogens Uncomplicated UTI: cotrimoxazole 20% Pyelonephritis: FQ resistance 10% Recommendation based on expert opinion. Perfetto E et al. [18] 2004 UTI Cost minimization Cotrimoxazole E. coli Depending on FQ resistance, 1% FQ resistance equals threshold of 4.3%; 10% FQ resistance equals threshold of 13.3% Modeling study. Recommendations according to IDSA guidelines are costeffective Gupta K et al. [3] 2011 UTI IDSA Guideline Cotrimoxazole/ trimethoprim E. coli Cotrimoxazole/trimethoprim threshold of 20% indicates switch to FQ Recommendation based on expert opinion McKinnell JA et al. [17] 2011 UTI Cost minimization and sensitivity analysis Nitrofurantoin, cotrimoxazole, ciprofloxacin E. coli Switch to nitrofurantoin: 17% (cotrimoxazole), 12% (ciprofloxacin) Switch from cotrimoxazole to ciprofloxacin: 10% Emphasizes 2011 IDSA guideline decision to use nitrofurantoin as first-line drug in many communities Sadler S et al. [19] 2017 UTI Cost-effectiveness Trimethoprim, fosfomycin, nitrofurantoin E. coli Trimethoprim 30% fosfomycin becomes costeffective At a trimethoprim threshold of 35%, both fosfomycin and nitrofurantoin become costeffective Modelling study Bader et al. [20] 2020 UTI Review Cotrimoxazole, fluoroquinolones Uropathogens Cotrimoxazole threshold of 20% in uUTI, FQ resistance 10% in cUTI Recommendation based on expert opinion Mandell LA et al. [6] 2007 RTI/CAP IDSA and ATS Consensus guideline Macrolides S. pneumoniae Macrolide-resistant Streptococcus pneumoniae threshold of 25% Recommendation based on expert opinion Kalil et al. [1] 2016 RTI/HAP/VAP IDSA and ATS Consensus guideline Antibiotics with activity against MRSA Anti-pseudomonal antibiotics MRSA P. aeruginosa HAP: MRSA prevalence >20% or unknown VAP: MRSA prevalence >10e20% or unknown, anti- MRSA antibiotic should be considered Prevalence of Gram-negative isolates resistant to the agent being considered for monotherapy >10% or unknown, two anti-pseudomonal antibiotics from different classes should be considered Recommendation based on expert opinion Babela RT et al. [21] 2017 RTI/Sinusitis Cost minimization and sensitivity analysis Macrolides S. pneumoniae Macrolide resistant Streptococcus pneumoniae threshold of 13.8% Modelling study Kaplan SL [23] 2005 CA-MRSA Review Antibiotics with activity against MRSA MRSA MRSA prevalence of 10e15% Recommendation based on expert opinion Bradley JS et al. [22] 1997 Meningitis Review Cephalosporins S. pneumoniae Complete penicillin resistance combined with >5% cephalosporin resistance or unknown resistance prevalence, then vancomycin or rifampicin should be added to empirical treatment Recommendation based on expert opinion WHO [26] 2012 Gonorrhoeae WHO global action plan to control the spread and impact of antimicrobial resistance in Neisseria gonorrhoeae Ceftriaxone N. gonorrhoeae If proportion of tested samples is 5% or more, steps should be taken to review and modify the national guidelines for STI treatment and management Recommendation based on expert opinion Nada et al. [24] 2014 Osteomyelitis in children Review Anti-staphylococcal penicillin or cefazolin Clindamycin CA-MRSA 10% CA-MRSA, switch to vancomycin or clindamycin (only when clindamycin resistance 25%) Recommendation based on expert opinion Saavedra-Lozano et al. [25] 2017 Bone and joint infections in children European Society for Paediatric Infectious Diseases guideline Anti-staphylococcal penicillin or cefazolin CA-MRSA 10e15% CA-MRSA switch to empirical therapy that includes this pathogen Recommendation based on expert opinion UTI, urinary tract infection; EAU, European Agency for Urology; IDSA, Infectious Diseases Society of America; RTI, respiratory tract infection; CAP, community-acquired pneumonia; HAP, hospital-acquired pneumonia; VAP, ventilator-associated pneumonia; FQ, fluoroquinolone; CA-MRSA, community-acquired methicillin-resistant Staphylococcus aureus; STI, sexually transmitted infection. A. Auzin et al. / Clinical Microbiology and Infection 28 (2022) 928e935 931 cost-effectiveness study comparing nitrofurantoin and FQ, McKinnell et al. demonstrated that the EAT with nitrofurantoin is more cost-effective when the prevalence of FQ-resistant E. coli exceeds 12% [17]. Both studies were performed in the USA. In a 2017 cost-effectiveness study for cystitis comparing trimethoprim, fosfomycin, pivmecillinam, and nitrofurantoin in a UK setting, Sadler et al. found that fosfomycin became more costeffective when the prevalence of trimethoprim-resistant E. coli exceeded 30%, with nitrofurantoin becoming cost-effective when the prevalence of trimethoprim-resistant E. coli exceeds 35% [19]. Respiratory tract infections Thresholds in guidelines The joint consensus guidelines of the American Thoracic Community and IDSA recommend thresholds for community-acquired pneumonia (CAP, 2007), and for hospital-acquired pneumonia (HAP) and ventilator-associated pneumonia (VAP, 2016). The resistance threshold of 25% for macrolide-resistant Streptococcus pneumoniae was defined, likely based on expert opinion, for CAP in outpatient settings, specifically for high-level macrolide resistance (MIC 16). The use of macrolides for the treatment of CAP is discouraged above this threshold [6]. In hospital settings, thresholds have been defined for methicillin-resistant Staphylococcus aureus (MRSA) and Pseudomonas aeruginosa for HAP and VAP. The threshold for MRSA prevalence in HAP is set at 20%, above which the EAT should be changed to an agent with activity against MRSA [1]. The published VAP threshold for MRSA is broad (10e20%) as compared to that for HAP, above which it is advised to change to an antibiotic with activity against MRSA [1]. Similarly, a threshold is recommended for the EAT in VAP regarding Gram-negative microorganisms, using P. aeruginosa as a proxy. When the prevalence of Gram-negative bacteria with resistance to an agent rises more than 10% in ICU settings, the EAT should be changed to a treatment containing two antibiotics from different classes with activity against P. aeruginosa [1]. For all HAP and VAP recommendations, the EAT should be changed when the prevalence is unknown or when the patient has risk factors for antibiotic resistance. However, more concrete tools are needed to do this. The reasoning behind the lower threshold for Gram-negatives is that they are more frequently implicated in VAP than are Gram-positives. All three of these thresholds are based on low-grade evidence (i.e. expert opinion) involving a compromise between appropriate antibiotic coverage and avoidance of superfluous treatment, potentially leading to undesirable effects [1]. Cost-effectiveness study to determine thresholds One cost-effectiveness study was found regarding ABR thresholds in RTI (Supplementary Material). In this 2017 study, Babela et al. compared oral cephalosporins and macrolides for the treatment of S. pneumoniae in rhinosinusitis, describing the ABR threshold of 13.8% prevalence of macrolide-resistant S. pneumoniae for rhinosinusitis. Above this threshold, treatment with cephalosporins was more cost-effective [21]. The threshold of 13.8% is subject to change in a manner similar to that reported in the previous cost-effectiveness studies. Similar to the previous costeffectiveness studies, the following criteria were considered: cost, antibiotic coverage, and treatment efficacy. Meningitis In a 1997 review on meningitis, Bradley et al. discussed the treatment options that were current at that time, including a suggestion for a resistance threshold for meningitis [22]. The threshold for meningitis cases in communities with complete penicillin resistance in S. pneumoniae combined with intermediate or complete resistance to ceftriaxone or cefotaxime was set at >5%, although the exact criteria are unclear. The authors state that either vancomycin or rifampicin should be added to the EAT when resistance exceeds the threshold. The threshold appears to be based on in vitro studies and expert opinion [22]. The fact that this threshold is lower than those for UTI and HAP is likely due to the severity of the disease and risk of mortality in case of EAT mismatch. Gonorrhoea In its global action plan concerning the spread and impact of antimicrobial resistance in Neisseria gonorrhoeae (2012), the WHO states that a 5% resistance to cephalosporins warrants a change in treatment guidelines and more intensive surveillance. They further state that a sudden and unexpected increase in key population groups (e.g. sex workers, men who have sex with men) where the disease is more prevalent is also a trigger for change, even if resistance remains below the 5% threshold [26]. Community-acquired MRSA (CA-MRSA) infections For SSTIs, a threshold of 10e15% prevalence of CA-MRSA was determined by expert opinion [23]. Above this threshold, clindamycin or cotrimoxazole is recommended for suspected CA-MRSA SSTI. Local clindamycin and cotrimoxazole resistance levels should allow for this. The same threshold is applied to the prevalence of clindamycin resistance in CA-MRSA, recommending to not use clindamycin above this threshold. The 2017 guideline of the European Society for Paediatric Diseases advise a threshold of 10e15% for CA-MRSA for bone and joint infections in children. Nada et al. advise a threshold of 10% for CAMRSA in children with osteomyelitis. They also advise using clindamycin only when the resistance to this antibiotic is 25%. These thresholds are based on expert opinion, with weak supporting evidence. Criteria for thresholds Seven criteria for thresholds were considered by various papers: efficacy, antibiotic coverage, disease severity, increase in ABR, adverse drug events, risk of Clostridioides difficile infections, and cost (see Table 2). One paper contained three different thresholds, which we assessed separately [1]. The cotrimoxazole threshold of 20% for UTIs was based on expert opinion (1999 IDSA guidelines) and used in both the updated 2011 guidelines and the subsequently published EAU guidelines. We treated them as identical, because they use the same criteria. The majority of the papers (8/12) mentioned three or fewer criteria. No papers used all the criteria identified. The most prevalent criterion was antibiotic coverage (11/13 thresholds), followed by cost (8/13), efficacy (6/13), disease severity (6/13), increase in ABR (5/13), adverse drug events (5/13), and C. difficile infections. The number and type of criteria used to define thresholds were quite heterogenous, even when defined for the same infection and antibiotic combination. Discussion This scoping review systematically reports on the literature on antibiotic resistance thresholds that should trigger a change in EAT recommendations. In general, the evidence base is weak. Most thresholds (n¼11) were found in expert opinion/consensus-based A. Auzin et al. / Clinical Microbiology and Infection 28 (2022) 928e935932 papers, consisting largely of various antibiotic policy documents, with high variation in the number of criteria considered. No clinical trials assessing thresholds were found. Just five cost-effectiveness modelling studies defined thresholds and were the most scientific approaches we were able to find. However, in contrast to the expert-opinion-based papers, the modelling studies considered only cost, efficacy, and antibiotic coverage, in line with their methodology. Antibiotic coverage is, nevertheless, self-evident and actually a prerequisite. Efficacy of any alternative empirical choice would be more relevant. If the coverage is sufficient but efficacy inferior for other reasons, this needs to be weighed against the reduction in efficacy due to resistance. The heterogeneity in the criteria used in the different papers posed a problem when comparing thresholds. The criteria that experts considered for their thresholds were: disease severity, antibiotic coverage, efficacy of treatment (cure and failure rates), adverse drug events (e.g. allergies), risk of C. difficile infection, costs, and increased resistance. The use of these criteria varied between papers, even concerning the same IDS/threshold combination, thus complicating the comparison of these thresholds for a given IDS. The question is raised whether thresholds for the same IDS would have differed if the different authors had considered the same criteria. Uniformity in the use of criteria is thus essential when determining thresholds for common syndromes and commonly used antibiotics to ensure that thresholds are defined uniformly. Disease severity is an important criterion for determining thresholds: the greater the severity the lower the threshold. In severe disease the consequences of treatment failure are greater. This is confirmed by the observations that thresholds are higher for non-severe diseases like cystitis (20% for cotrimoxazole), decreasing for more severe illnesses like pyelonephritis/cUTI (10% for ciprofloxacin) and meningitis (5% for ceftriaxone). The low threshold of 5% for N. gonorrhoeae could be explained by the high communicability of this specific pathogen in certain populations (or subpopulations), instead of the disease severity; a considerable public health problem is another key criterion. It should be noted that multi-resistant pathogens like N. gonorrhoeae generally have few alternative options. More than half of the thresholds identified were for UTI as UTIs are common and sufficient surveillance data are available. Of all thresholds identified in this review, the threshold for cotrimoxazole has the most solid evidence base in clinical studies [3,5,31], in which the clinical failure rate in the group with resistance to cotrimoxazole was relatively high (40e50%). These studies demonstrated that when cotrimoxazole resistance is 10e15%, cure rates of cotrimoxazole EAT are comparable to EAT with ciprofloxacin or nitrofurantoin [22,27e29]. The role of MICs in the definition of thresholds needs more consideration. The 25% threshold for high-level macrolide resistance in S. pneumoniae (MIC 16 m g/mL) was defined through expert opinion with no explicit mention of criteria (see Table 2)[6]. The effect of low-level resistance (MIC of 1e8 m g/mL) was not considered. In a 2019 theoretical modelling study of low-level resistance [32], Daneman et al. argue that some studies have called the difference in clinical failure rate between high-level and low-level macrolide resistance into question [33]. In addition, when the resistance rate of high-level and low-level macrolide resistance are combined for S. pneumoniae, they could exceed the 25% Table 2 Overview of criteria used to define thresholds Source Threshold Criteria used a Efficacy b Disease severity Increase in ABR c Adverse drug event C. difficile Costs IDSA guidelines 1999 [2], 2011 [3], EAU guideline 2018 [5], Bader et al. [31] Cotrimoxazole (UTI) >20% Yes Yes Yes Yes Not mentioned Not mentioned IDSA guidelines 2011 [3], EAU guideline 2018 [5], Bader et al. [20] Ciprofloxacin (UTI) >10% Yes Yes Yes Not mentioned Not mentioned Not mentioned Le et al. [16] Cotrimoxazole (UTI) >22% Yes Not mentioned Not mentioned Not mentioned Not mentioned Yes Perfetto et al. [18] Cotrimoxazole (UTI) >4.3% Yes Not mentioned Not mentioned Not mentioned Not mentioned Yes McKinnell et al. [17] Cotrimoxazole (UTI) >10% Yes Not mentioned Not mentioned Not mentioned Not mentioned Yes Sadler et al. [19] Trimethoprim (UTI) >30% Yes Not mentioned Not mentioned Not mentioned Not mentioned Yes IDSA and ATS consensus guideline CAP 2007 [6] Macrolides (CAP) >25% Not mentioned Not mentioned Not mentioned Not mentioned Not mentioned Not mentioned IDSA and ATS consensus guideline HAP/VAP 2016 [1] MRSA (HAP) >20% Not mentioned Yes Yes Yes Yes Yes IDSA and ATS consensus guideline HAP/VAP 2016 [1] MRSA (VAP) >10e20% Not mentioned Yes Yes Yes Yes Yes IDSA and ATS consensus guideline VAP 2016 [1] Gram-negatives (VAP) >10% Not mentioned Yes Yes Yes Yes Yes Babela et al. [21] Macrolides (rhinosinusitis) >13.8% Yes Not mentioned Not mentioned Not mentioned Not mentioned Yes Kaplan et al. [23] MRSA (SSTI) >10e15% Not mentioned Not mentioned Not mentioned Not mentioned Not mentioned Not mentioned Bradley et al. [22] Cephalosporins (meningitis) >5% Not mentioned Yes Not mentioned Not mentioned Not mentioned Not mentioned WHO [26] Ceftriaxone (gonorrhoea) 5% Yes Not mentioned Not mentioned Not mentioned Not mentioned Not mentioned Nada et al. [24] Anti-staphylococcal penicillin/ cefazolin (CA-MRSA) 10% Yes Not mentioned Not mentioned Yes Not mentioned Not mentioned Saavedra-Lozano et al. [25] Anti-staphylococcal penicillin/ cefazolin (CA-MRSA) 10e15% Yes Yes Not mentioned Not mentioned Not mentioned Not mentioned ABR, antibiotic resistance; UTI, urinary tract infection; CAP, community-acquired pneumonia; HAP, hospital-acquired pneumonia; VAP, ventilator-associated pneumonia; CAMRSA, community-acquired methicillin-resistant Staphylococcus aureus; SSTI, skin and soft tissue infection. a Only criteria explicitly mentioned in the text. b Efficacy: the maximum response that can be achieved with a drug or treatment. c Increase in ABR: the potential impact on antibiotic resistance associated with the use of the selected antibiotic treatment. A. Auzin et al. / Clinical Microbiology and Infection 28 (2022) 928e935 933 threshold defined in the guideline. Therefore, the ‘hidden’clinical failures due to low-level macrolide resistance should also be considered when determining resistance thresholds. Moreover, clinical breakpoints and the definition of susceptibility/resistance could potentially be different when using different sources (European Committee on Antimicrobial Susceptibility Testing (EUCAST) or National Committee for Clinical Laboratory Standards (NCCLS)). Preferably, thresholds should be well-defined and as accurate as possible. The 10e20% for MRSA in VAP is rather broad, and it is unclear in which scenarios the threshold of 10% versus 20% should be applied [5]. The threshold for MRSA and Gram-negatives in VAP patients is important, given the severity of the infection and the potentially substantial impact of a poor EAT choice on mortality. Broad thresholds are a challenge to implement; more precise estimates for various syndromes and settings are needed to really improve patient outcomes. For individual patient care you need ‘wiggle room’to decide on antibiotic change. However, on a population level and for guideline development to decide on empirical antibiotic choice, this is should be as precise as possible. Considering the above, we propose considering a composite resistance threshold that covers the most common pathogens and their susceptibilities for specific infections with known efficacies of various antibiotic treatments. The context surrounding a threshold becomes complete when a clear recommendation is given for one or more alternative antibiotics, based on a clear set of criteria. Any threshold and EAT recommendation should include de-escalation indications for when the antibiogram becomes available. A small set of recommendations for antibiotic policymakers to consider when defining thresholds, based on our review, is presented in Fig. 2. In conclusion, EAT thresholds are sparse while there is a multitude of infectious disease syndromes. Most thresholds are based on weak evidence, if any at all, and are usually based on expert opinion drawing on heterogeneous set of criteria. Considering the rising trend in antibiotic resistance, there is a clear need to provide tools for determining thresholds to support guideline development. We propose developing a systematic multidisciplinary approach to determine various thresholds for various classes of antibiotics for a range of prioritized IDSs. This approach should be supported by high-quality clinical datasets on ABR and surveillance at the local, national, and international levels. The need for well-defined thresholds suggests the importance of epidemiological and clinical surveillance of ABR [34]. Author contributions AA: methodology, investigation, drafting of the manuscript. MS: investigation, approval of the manuscript. ET, JR-B and AV: funding acquisition, supervision, editing and approval of manuscript. MH and EA: methodology, supervision, editing and approval of manuscript. HW: conceptualization, methodology, investigation, supervision, drafting, editing and approval of the manuscript. Transparency declaration The authors declare no conflict of interest. This research was funded by the EPI-Net COMBACTE-MAGNET project. We thank the Innovative Medicines Initiative Joint Undertaking for supporting the EPI-Net COMBACTE-MAGNET project (grant agreement number 115737), resources of which include a financial contribution from the European Union Seventh Framework Program (FP7/ 2007e2013) and European Federation of Pharmaceutical Industries and Associations in-kind contribution. The funder had no role in the study design, data collection, analysis, interpretation of data or writing the manuscript. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.cmi.2021.12.003. Fig. 2. Recommendations to consider when defining thresholds. A. Auzin et al. / Clinical Microbiology and Infection 28 (2022) 928e935934 References [1] Kalil AC, Metersky ML, Klompas M, Muscedere J, Sweeney DA, Palmer LB, et al. Management of adults with hospital-acquired and ventilator-associated pneumonia: 2016 clinical practice guidelines by the infectious diseases society of America and the American thoracic society. Clin Infect Dis 2016;63: e61e111. [2] Warren JW, Abrutyn E, Hebel JR, Johnson JR, Schaeffer AJ, Stamm WE. Guidelines for antimicrobial treatment of uncomplicated acute bacterial cystitis and acute pyelonephritis in women. Clin Infect Dis 1999;29:745e58. [3] Gupta K, Hooton TM, Naber KG, Wullt B, Colgan R, Miller LG, et al. International clinical practice guidelines for the treatment of acute uncomplicated cystitis and pyelonephritis in women: a 2010 update by the Infectious Diseases Society of America and the European Society for Microbiology and Infectious Diseases. Clin Infect Dis 2011;52:e103e20. [4] Moher D, Liberati A, Tetzlaff J, Altman DG, The PG. Preferred reporting Items for systematic reviews and meta-analyses: the PRISMA statement. PLOS Med 2009;6:e1000097. [5] Bonkat G, (Chair) RB, Bruy ere F, Cai T, Geerlings SE, K€ oves B, et al. Guidelines associates: T. Mezei AP, B. Pradere, R. Veeratterapillay. EAU Guidelines on Urological infections; 2019. [6] Mandell LA, Wunderink RG, Anzueto A, Bartlett JG, Campbell GD, Dean NC, et al. Infectious Diseases Society of America/American Thoracic Society consensus guidelines on the management of community-acquired pneumonia in adults. Clin Infect Dis 2007;44:S27e72. [7] Woodhead M, Blasi F, Ewig S, Garau J, Huchon G, Ieven M, et al. Guidelines for the management of adult lower respiratory tract infectionsdfull version. Clin Microbiol Infect 2011;17:E1e59. [8] Torres A, Niederman MS, Chastre J, Ewig S, Fernandez-Vandellos P, Hanberger H, et al. International ERS/ESICM/ESCMID/ALAT guidelines for the management of hospital-acquired pneumonia and ventilator-associated pneumonia: guidelines for the management of hospital-acquired pneumonia (HAP)/ventilator-associated pneumonia (VAP) of the European Respiratory Society (ERS), European Society of Intensive Care Medicine (ESICM), European Society of Clinical Microbiology and Infectious Diseases (ESCMID) and Asociaci on Latinoamericana del T orax (ALAT). Eur Respir J 2017;50:1700582. [9] Tunkel AR, Hasbun R, Bhimraj A, Byers K, Kaplan SL, Scheld WM, et al. 2017 infectious diseases society of America's clinical practice guidelines for healthcare-associated ventriculitis and meningitis. Clin Infect Dis 2017;64: e34e65. [10] Tunkel AR, Hartman BJ, Kaplan SL, Kaufman BA, Roos KL, Scheld WM, et al. Practice guidelines for the management of bacterial meningitis. Clin Infect Dis 2004;39:1267e84. [11] van de Beek D, Cabellos C, Dzupova O, Esposito S, Klein M, Kloek AT, et al. ESCMID guideline: diagnosis and treatment of acute bacterial meningitis. Clin Microbiol Infect 2016;22:S37e62. [12] Stevens DL, Bisno AL, Chambers HF, Dellinger EP, Goldstein EJC, Gorbach SL, et al. Practice guidelines for the diagnosis and management of skin and soft tissue infections: 2014 update by the Infectious Diseases Society of America. Clin Infect Dis 2014;59:e10e52. [13] Sartelli M, Guirao X, Hardcastle TC, Kluger Y, Boermeester MA, Ras¸ a K, et al. 2018 WSES/SIS-E consensus conference: recommendations for the management of skin and soft-tissue infections. World J Emerg Surg 2018;13:58. [14] Osmon DR, Berbari EF, Berendt AR, Lew D, Zimmerli W, Steckelberg JM, et al. Diagnosis and management of prosthetic joint infection: clinical practice guidelines by the Infectious Diseases Society of America. Clin Infect Dis 2013;56:e1e25. [15] Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M, Ferrer R, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock. Intensive Care Med 2016;43:304e77. 2017. [16] Le TP, Miller LG. Empirical therapy for uncomplicated urinary tract infections in an era of increasing antimicrobial resistance: a decision and cost analysis. Clin Infect Dis 2001;33:615e21. [17] McKinnell JA, Stollenwerk NS, Jung CW, Miller LG. Nitrofurantoin compares favorably to recommended agents as empirical treatment of uncomplicated urinary tract infections in a decision and cost analysis. Mayo Clin Proc 2011;86:480e8. [18] Perfetto EM, Keating K, Merchant S, Nichols BR. Acute uncomplicated UTI and E. coli resistance: implications for first-line empirical antibiotic therapy. J Manag Care Pharm 2004;10:17e25. [19] Sadler S, Holmes M, Ren S, Holden S, Jha S, Thokala P. Cost-effectiveness of antibiotic treatment of uncomplicated urinary tract infection in women: a comparison of four antibiotics. BJGP Open 2017;1. bjgpopen17X101097. [20] Bader MS, Loeb M, Leto D, Brooks AA. Treatment of urinary tract infections in the era of antimicrobial resistance and new antimicrobial agents. Postgrad Med 2020;132:234e50. [21] Babela R, Jarcuska P, Uraz V, Kr cm ery V, Jadud B, Stevlik J, et al. Decision and cost analysis of empirical antibiotic therapy of acute sinusitis in the era of increasing antimicrobial resistance: do we have an additional tool for antibiotic policy decisions? Neuro Endocrinol Lett 2017;38:9e26. [22] Bradley JS, Scheld WM. The challenge of penicillin-resistant Streptococcus pneumoniae meningitis: current antibiotic therapy in the 1990s. Clin Infect Dis 1997;24:S213e21. [23] Kaplan SL. Treatment of community-associated methicillin-resistant Staphylococcus aureus infections. Pediatr Infect Dis J 2005;24:457e8. [24] Harik NS, Smeltzer MS. Management of acute hematogenous osteomyelitis in children. Expert Rev Anti Infect Ther 2010;8:175e81. [25] Saavedra-Lozano J, Falup-Pecurariu O, Faust SN, Girschick H, Hartwig N, Kaplan S, et al. Bone and joint infections. Pediatr Infect Dis J 2017;36: 788e99. [26] WHO. Global action plan to control the spread and impact of antimicrobial resistance in Neisseria gonorrhoeae. 2012. [27] Arredondo-García JL, Figueroa-Dami an R, Rosas A, J auregui A, Corral M, Costa A, et al. Comparison of short-term treatment regimen of ciprofloxacin versus long-term treatment regimens of trimethoprim/sulfamethoxazole or norfloxacin for uncomplicated lower urinary tract infections: a randomized, multicentre, open-label, prospective study. J Antimicrob Chemother 2004;54: 840e3. [28] Gupta K, Hooton TM, Roberts PL, Stamm WE. Short-course nitrofurantoin for the treatment of acute uncomplicated cystitis in women. Arch Intern Med 2007;167:2207e12. [29] Iravani A, Klimberg I, Briefer C, Munera C, Kowalsky SF, Echols RM. A trial comparing low-dose, short-course ciprofloxacin and standard 7 day therapy with co-trimoxazole or nitrofurantoin in the treatment of uncomplicated urinary tract infection. J Antimicrob Chemother 1999;43:67e75. [30] Kavatha D, Giamarellou H, Alexiou Z, Vlachogiannis N, Pentea S, Gozadinos T, et al. Cefpodoxime-proxetil versus trimethoprim-sulfamethoxazole for shortterm therapy of uncomplicated acute cystitis in women. Antimicrob Agents Chemother 2003;47:897e900. [31] Gupta K, Stamm WE. Outcomes associated with trimethoprim/sulphamethoxazole (TMP/SMX) therapy in TMP/SMX resistant community-acquired UTI. Int J Antimicrob Agents 2002;19:554e6. [32] Daneman N, Low DE, McGeer A, Green KA, Fisman DN. At the threshold: defining clinically meaningful resistance thresholds for antibiotic choice in community-acquired pneumonia. Clin Infect Dis 2008;46:1131e8. [33] Daneman N, McGeer A, Green K, Low DE. Macrolide resistance in bacteremic pneumococcal disease: implications for patient management. Clin Infect Dis 2006;43:432e8. [34] https://www.ecdc.europa.eu/en/about-us/partnerships-and-networks/ disease-and-laboratory-networks/ears-net. A. Auzin et al. / Clinical Microbiology and Infection 28 (2022) 928e935 935