A retrospective observational cohort study conducted at Hospital de la Santa Creu i Sant Pau in Barcelona, Spain, has demonstrated that weighted-incidence syndromic combination antibiograms provide superior empirical antimicrobial coverage estimates for adult emergency department patients with bloodstream infections compared to traditional cumulative susceptibility reports. Analyzing 1,716 consecutive bacteremia episodes attended between January 2022 and December 2024, the investigation revealed that conventional cumulative antibiograms omitted crucial pathogen diversity by excluding species that failed to meet standardized minimum isolate reporting thresholds. In contrast, the weighted-incidence syndromic approach incorporated every recovered pathogen, weighting in vitro susceptibility directly against local syndromic prevalence to generate precise regimen-level coverage metrics. By capturing both frequent and uncommon pathogens across monotherapies and dual-agent empirical regimens, the research indicates that syndromic antibiograms overcome structural blind spots in hospital surveillance reporting, offering clinicians and stewardship teams an evidence-based framework for tailoring early antimicrobial prescribing to specific clinical presentations upon emergency presentation.
The researchers established their investigation upon the recognized clinical urgency surrounding bloodstream infections, which remain a primary driver of morbidity and mortality across global healthcare settings. Because microbiological culture identification and phenotypic susceptibility confirmations require multiple days to finalize, initial antimicrobial regimens must be administered empirically at emergency presentation. Previous clinical investigations cited by the authors emphasize that inadequate initial therapy directly correlates with prolonged hospitalizations and elevated patient mortality, making accurate local epidemiological intelligence indispensable for emergency decision-making. Historically, clinical teams have relied on hospital-wide cumulative antimicrobial susceptibility test reports constructed in accordance with recommendations from the Clinical and Laboratory Standards Institute and the Spanish Antibiogram Committee. However, conventional cumulative reports summarize susceptibility purely at the individual species level without accounting for relative pathogen frequencies within distinct clinical syndromes. Furthermore, standard cumulative reporting fails to provide clinical stratification based on patient exposure history and cannot evaluate whether multi-agent regimens achieve simultaneous coverage against all organisms in polymicrobial infections.
To evaluate how syndromic modeling addresses these operational deficiencies, the investigative team examined the clinical characteristics and acquisition routes of 1,716 consecutive adult bloodstream infection episodes representing patients with a median age of 78 years, with an interquartile range from 67 to 86 years. Within this emergency cohort, 947 patients were male, accounting for 55.2 percent, and 419 individuals, representing 24.4 percent, resided in nursing homes or long-term care facilities. Clinical severity was underscored by 271 patients, or 15.8 percent, requiring vasopressor hemodynamic support, while 219 individuals, representing 12.8 percent, died within thirty days of admission. Applying epidemiological criteria proposed by Friedman and colleagues, the authors classified 1,191 episodes, or 69.4 percent, as healthcare-associated infections. Anatomical source evaluation identified urinary tract infections as the predominant origin, accounting for 721 episodes or 42 percent, followed by intra-abdominal infections in 347 episodes representing 20.2 percent, and bacteremia of unknown origin in 295 episodes representing 17.2 percent.
The microbiological investigation recovered 1,913 clinically relevant microorganisms from patient blood cultures utilizing the automated BactAlert VIRTUO detection system, with organism identification achieved through matrix-assisted laser desorption ionization-time of flight mass spectrometry. Gram-negative bacilli predominated across the cohort, comprising 1,344 isolates or 70.3 percent, while Gram-positive bacteria represented 557 isolates or 29.1 percent, and yeasts accounted for 12 isolates representing 0.6 percent. Among individual species, Escherichia coli was isolated most frequently, totaling 804 isolates or 42 percent, followed by Klebsiella pneumoniae with 193 isolates representing 10.1 percent, Staphylococcus aureus with 133 isolates representing 7 percent, Streptococcus pneumoniae with 98 isolates representing 5.1 percent, and Pseudomonas aeruginosa with 83 isolates representing 4.3 percent. Extended-spectrum beta-lactamase production occurred in 294 of 1,181 Enterobacterales isolates, representing 24.9 percent, while 38 isolates produced AmpC beta-lactamases and 7 carried carbapenemases. Additionally, methicillin resistance was identified in 37 of 133 Staphylococcus aureus isolates, representing 27.8 percent, while seven Pseudomonas aeruginosa isolates exhibited extensive drug resistance.
When constructing the traditional cumulative antibiogram following national recommendations, the hospital applied the standard minimum threshold requiring thirty bloodstream isolates per species, which immediately restricted standalone analysis to seven common organisms accounting for 75.4 percent of all isolated pathogens. Even after grouping low-frequency Enterobacterales species, coagulase-negative staphylococci, and viridans group streptococci, the standard cumulative report captured only 87.3 percent of recovered pathogens, completely omitting more than 75 distinct bacterial species that each caused fewer than thirty bacteremia episodes. In marked contrast, the weighted-incidence syndromic combination antibiogram synthesized the entire microbiological spectrum by multiplying each pathogen’s empirical probability by its observed in vitro susceptibility to specific regimens and summing those products across all causative organisms. The model evaluated eight widely utilized monotherapies alongside six combination regimens incorporating metronidazole, aminoglycosides, levofloxacin, and vancomycin, assuming active coverage whenever at least one regimen component was active against monomicrobial isolates or against all pathogens in polymicrobial episodes.
Overall empirical coverage estimates derived from the syndromic model revealed significant performance differences across standard monotherapies and dual-agent combinations. Among single-agent regimens, the highest estimated coverage across all emergency bloodstream infections was achieved by imipenem at 94.8 percent, meropenem at 91.3 percent, piperacillin-tazobactam at 89.6 percent, and ertapenem at 85.8 percent. For cephalosporin regimens, ceftriaxone monotherapy yielded an empirical coverage rate of only 65.1 percent, reflecting the substantial local burden of extended-spectrum beta-lactamase producers and resistant Gram-negative bacilli. However, the addition of an aminoglycoside markedly augmented ceftriaxone coverage, elevating overall activity to 88.7 percent with amikacin and 81.8 percent with gentamicin. Adding levofloxacin to ceftriaxone achieved an intermediate coverage level of 76.7 percent, whereas the anaerobic coverage provided by metronidazole offered minimal overall improvement, reaching 67.5 percent. Conversely, combining carbapenems with additional agents yielded narrow coverage increments, as meropenem combined with amikacin reached 92.8 percent, while meropenem paired with vancomycin achieved 97.6 percent across the unstratified emergency cohort.
Stratification by epidemiological exposure revealed that empirical coverage was consistently and substantially lower in healthcare-associated bloodstream infections than in community-acquired episodes. The most pronounced coverage discrepancies emerged for ceftriaxone, which dropped from 81.4 percent in community-acquired bacteremia to 58.2 percent in healthcare-associated presentations, ciprofloxacin, declining from 79.8 percent to 59.1 percent, and amoxicillin-clavulanic acid, falling from 78.4 percent to 59.4 percent. Even piperacillin-tazobactam experienced a measurable decrease from 95.8 percent down to 87 percent in healthcare-associated infections. This divergence became exceptionally prominent in urinary-source bacteremia, where ceftriaxone coverage plunged from 82.5 percent in community infections to 50.9 percent in healthcare-associated cases. Within urinary bacteremia, monotherapy coverage peaked with imipenem at 97.3 percent, meropenem at 95.2 percent, piperacillin-tazobactam at 90.6 percent, and ertapenem at 90.3 percent. Conversely, respiratory-source bacteremia demonstrated high susceptibility across broad classes, with meropenem achieving 97.5 percent, imipenem reaching 96.7 percent, piperacillin-tazobactam achieving 95 percent, and fluoroquinolones reaching 95 to 95.8 percent, while amoxicillin-clavulanic acid and ceftriaxone each provided 87.5 percent.
Further subgroup evaluations highlighted specific clinical niches where tailored combinations proved essential. In catheter-related and cardiovascular-source bacteremia, meropenem monotherapy exhibited markedly reduced activity, achieving only 66.7 percent and 72.4 percent coverage, respectively. However, supplementing meropenem with vancomycin restored empirical coverage to 95.6 percent in catheter-associated episodes and 100 percent in cardiovascular infections, reflecting the concentration of resistant Gram-positive pathogens in endovascular foci. Similarly, residence in a nursing home or long-term care facility emerged as an impactful risk factor, driving ceftriaxone coverage down to 48.9 percent compared with 70.4 percent among non-residents. Polymicrobial bacteremia, which occurred in 144 episodes involving 330 organisms, also exhibited suppressed empirical coverage across multiple monotherapies under the strict criterion requiring every co-infecting pathogen to be covered. Ertapenem coverage fell from 87.2 percent in monomicrobial cases to 79.1 percent in polymicrobial infections, and meropenem dropped from 92.4 percent to 85.8 percent, although meropenem plus vancomycin maintained high efficacy across both monomicrobial and polymicrobial presentations at 97.8 percent and 97 percent, respectively.
In an exploratory analysis examining clinical outcomes across stratified syndromic coverage tiers, the researchers observed that receiving an empirical regimen with higher predicted coverage was strongly associated with achieving adequate in vitro therapy. Adequate therapy was documented in 72.4 percent of patients receiving regimens with low coverage below 70 percent, 87.9 percent for moderate coverage between 70 and 90 percent, and 94.4 percent for high coverage exceeding 90 percent, demonstrating statistical significance at p less than 0.001. Paradoxically, crude clinical outcomes displayed an inverse pattern: vasopressor dependency increased progressively from 8.8 percent in the low-coverage tier to 21.6 percent in the moderate tier and 26.2 percent in the high-coverage tier. Similarly, unadjusted 7-day and 30-day mortality rates were highest among patients who received high-coverage regimens. The authors emphasized that this observed association represents severe confounding by indication rather than causal harm, because treating emergency physicians preferentially prescribed potent broad-spectrum combination regimens to patients presenting with hemodynamic collapse, septic shock, and advanced baseline clinical severity.
The investigators contextualized their findings within international antimicrobial stewardship practice, citing foundational syndromic antibiogram work by Hebert and subsequent intensive care studies by Randhawa and Liberati. The research team noted that while syndromic models provide nuanced probabilities, they must function alongside multidisciplinary stewardship reviews rather than as automated prescribing algorithms that default to the highest numerical percentage. Statistical modeling was conducted using the AMR package in R software, utilizing non-parametric bootstrap resampling across 1,000 iterations to generate 95 percent highest density intervals and manage analytical uncertainty. The study acknowledged several inherent limitations, including its single-center retrospective design, rule-based phenotype imputations for untested pathogen-drug combinations, and the absence of pharmacokinetic drug exposure, organ toxicity, and source control data. Ultimately, the authors concluded that integrating weighted-incidence syndromic combination antibiograms into hospital guideline updates enables emergency physicians to reconcile syndrome-specific epidemiology and multidrug resistance risks with responsible stewardship, preserving carbapenems and glycopeptides while ensuring robust empirical therapy for critically ill bacteremia patients.
Subject of Research: Evaluation of weighted-incidence syndromic combination antibiograms for empirical therapy of bloodstream infections in the emergency department
Article Title: Beyond the Cumulative Antibiogram: A Weighted-Incidence Syndromic Combination Antibiogram to Guide Empirical Therapy of Bloodstream Infections in the Emergency Department
Article References: Batista, C. S. P., Escolà -Vergé, L., Navarro, F., & Rivera, A. (2026). Beyond the Cumulative Antibiogram: A Weighted-Incidence Syndromic Combination Antibiogram to Guide Empirical Therapy of Bloodstream Infections in the Emergency Department. Infectious Diseases and Therapy, 15(12), 3321-3336. https://doi.org/10.1007/s40121-026-01437-x
Image Credits: AI Generated
DOI: 10.1007/s40121-026-01437-x
Keywords: bloodstream infections, emergency department, weighted-incidence syndromic combination antibiogram, cumulative antimicrobial susceptibility test, antimicrobial stewardship, empirical antibiotic therapy, Enterobacterales, Pseudomonas aeruginosa, extended-spectrum beta-lactamases, meropenem, imipenem, piperacillin-tazobactam, ceftriaxone, amikacin, vancomycin, healthcare-associated infections, polymicrobial bacteremia, urinary tract infections
News Source: Ophelia Keating. (November 20, 2026). Syndromic Antibiogram Models Improve Empirical Antibiotic Coverage Estimates in Emergency Bacteremia. Scienmag.



