2.6 Antibacterial Therapy for Suspected Sepsis: Is Earlier Always Better?

Sepsis management has always emphasized “time is life,” with early recognition and rapid antibiotic initiation as cornerstones. However, accumulating high-quality evidence reveals this principle does not universally apply. Emerging data suggest that for suspected sepsis without infectious shock, blindly pursuing the 1-hour bundle may lead to overtreatment, increased resistance, and delayed diagnosis of non-infectious emergencies. This article reviews the background of sepsis therapy, interprets a 2025 decision analysis model on antibiotic initiation timing in the ED, and proposes a stratified decision strategy for critical care settings.

1. Background and Challenges in Sepsis Therapy

Sepsis is defined as life-threatening organ dysfunction caused by a dysregulated host response to infection. Over the past two decades, international guidelines like the Surviving Sepsis Campaign have emphasized early antibiotics, particularly the 1-hour bundle requiring blood cultures and broad-spectrum antibiotics within 1 hour of triage [1]. This strategy yields significant survival benefits in septic shock but faces scrutiny in broader suspected sepsis populations.

Recent retrospective cohorts and meta-analyses explore the time-outcome relationship. Early intervention correlates with improved survival in septic shock, but for non-shock suspected sepsis, delaying antibiotics 1–3 hours for diagnostic evaluation does not increase mortality [8, 10].

The greatest challenge lies in the lack of specific early diagnostic criteria. Fever, leukocytosis, and tachycardia are non-specific; many critical conditions (acute heart failure, pulmonary embolism, drug reactions, autoimmune crises) are easily misdiagnosed as sepsis initially [2, 3]. Data show 30%–50% of ED patients initially diagnosed with sepsis and started on antibiotics are ultimately non-infectious [4, 5]. Indiscriminate 1-hour bundles in these patients offer no benefit, may mask underlying conditions, and increase adverse drug reactions [6].

2. Core Study Interpretation: Rapid vs. Deferred Antibiotic Initiation Decision Model

A 2025 study by Hill et al. in Intensive Care Medicine utilized a decision analysis model to weigh two outcomes in suspected sepsis: benefits of early empiric therapy (earlier pathogen coverage, reduced shock/mortality risk) vs. harms of premature/unnecessary use (adverse events, C. difficile risk, resistance selection pressure) [11]. This model provides critical guidance when diagnosis is uncertain and evidence is incomplete.

In a mortality-only restricted model, rapid initiation favored outcomes. However, when incorporating net utility (avoiding unnecessary antibiotics, adverse events), results shifted significantly. Sensitivity analysis indicated rapid initiation is only stably optimal when true bacterial infection probability exceeds 90%. Below this threshold, a brief diagnostic window to gather evidence and improve accuracy makes deferred initiation preferable in overall net benefit. Furthermore, if key diagnostic workup takes ≤1.33 hours, deferred initiation remains optimal. The study concludes antibiotic timing depends on: (1) probability of infection, and (2) time required to obtain key evidence. In hemodynamically stable suspected patients, rapid assessment followed by targeted therapy aligns better with net clinical benefit.

The model compared two strategies: ① Rapid initiation: Broad-spectrum antibiotics immediately (0–3 h) without waiting for further results. ② Deferred initiation: Antibiotics delayed 0–6 h to allow CT, biomarkers, etc., initiating only when evidence is robust.

The decision tree operates in two steps: first, determining true bacterial infection probability; second, evaluating outcome differences between strategies in infected vs. non-infected populations. Two models were built: a restricted model (mortality only) and a primary model (mortality + antibiotic management factors like avoiding unnecessary use and adverse events).

3. Stratified Decision-Making for Antibiotic Initiation in Critical Care

Critically ill patients have higher true infection probabilities but also frequently present with non-infectious inflammatory/organ dysfunction mimics (pancreatitis, PE, hemorrhage, drug reactions, cardiogenic shock). A stratified approach is recommended: complete risk stratification within 1 hour of identification, and gather key evidence within 1–3 hours to improve diagnostic certainty and reduce unnecessary use.

(1) Step 1: Identify High-Risk Patients Requiring Immediate Initiation
Target patients requiring vasopressors, showing clear hypoperfusion (oliguria, mottling, rising lactate), having a clear infection focus with rapid deterioration, or being immunocompromised with high suspicion. These patients have high true infection probability and high delay penalties. Obtain cultures first, then immediately start empiric therapy [12].

(2) Step 2: Complete Diagnostics for Hemodynamically Stable Suspected Sepsis
Priority should be given to tests that will change management within 1–3 hours. After standardized specimen collection, perform imaging to locate infection foci while actively ruling out non-infectious emergencies (PE, acute heart failure decompensation, hemorrhage, hypovolemia, metabolic crises) [2, 3]. Conduct short-cycle reassessment (30–60 min for vitals, lactate trend, oxygen demand). If infection evidence strengthens → initiate antibiotics promptly. If evidence is weak with plausible non-infectious explanations (post-op absorption fever, clear drug fever) → hold antibiotics, implement active monitoring (close vitals, repeat lactate/inflammatory markers, rapid reassessment). Escalate if evidence strengthens or condition deteriorates.

Crucially, for suspected sepsis, overall net benefit depends less on saving minutes for the first dose and more on completing an antibiotic “time-out” at 48–72 hours. If evidence supports infection, promptly identify the site/pathogen and de-escalate. If evidence refutes infection or the course is non-infectious, discontinue antibiotics decisively. Kam et al. demonstrated that early (48–72 h) de-escalation or discontinuation based on microbiology/clinical assessment in suspected sepsis is safe and correlates with lower in-hospital mortality and complications [13]. New evidence supports this: patients initially suspected of sepsis but later ruled out should have antibiotics stopped promptly, rather than continued for “insurance” [14].

Conclusion: “Sooner is better” does not apply universally to suspected sepsis. For high-risk patients (e.g., septic shock), rapid empiric therapy remains standard. For hemodynamically stable patients with uncertain diagnoses, a brief diagnostic window followed by targeted therapy is superior. The 2026 management focus should emphasize risk stratification, dynamic reassessment, and integrating diagnostic stewardship with antimicrobial management to minimize unnecessary broad-spectrum use safely.
(Huang Man, Zhang Piao, The Second Affiliated Hospital, Zhejiang University School of Medicine)

References

  1. Evans L, Rhodes A, Alhazzani W, et al. Surviving Sepsis Campaign 2021. Crit Care Med. 2021;49(11):e1063-e1143.

  2. Singer M, Deutschman CS, Seymour CW, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA. 2016;315(8):801-810.

  3. Van Vught LA, Klein Klouwenberg PM, Cremer OL, et al. Likelihood of infection in presumed sepsis at ICU admission. Crit Care. 2015;19(1):319.

  4. Prescott HC, Iwashyna TJ. Improving sepsis treatment by embracing diagnostic uncertainty. Ann Am Thorac Soc. 2019;16(4):426-429.

  5. Littell JM, Guirgis F, Driver B, et al. ED patients meeting sepsis criteria not diagnosed at discharge. Acad Emerg Med. 2021;28(7):745-752.

  6. Klompas M, Calandra T, Singer M. Antibiotics for sepsis: finding the equilibrium. JAMA. 2018;320(14):1433-1434.

  7. Rhee C, Kadri SS, Dekker JP, et al. Prevalence of resistant pathogens in culture-proven sepsis. JAMA Netw Open. 2020;3(4):e202899.

  8. Freund Y, Canceilla de Abreu M, Lebal S, et al. Effect of the 1-h bundle on ED suspected sepsis mortality. Intensive Care Med. 2024;50(7):1086-1095.

  9. Philippe AL, Lebal S, Canceilla de Abreu M, et al. Time to antibiotic & mortality in ED suspected sepsis. Eur J Emerg Med. 2025;32(2):109-115.

  10. Sterling SA, Miller WR, Pryor J, et al. Impact of antibiotic timing on severe sepsis/shock outcomes. Crit Care Med. 2015;43(9):1907-1915.

  11. Hill TM, Kerivan LT, Vilain KA, et al. Decision analysis: rapid vs deferred antibiotics in ED suspected sepsis. Intensive Care Med. 2025;51(5):841-848.

  12. Kumar A, Roberts D, Wood KE, et al. Duration of hypotension & survival in septic shock. Crit Care Med. 2006;34(6):1589-1596.

  13. Kam KQ, Chen T, Kadri SS, et al. Antibiotic de-escalation in suspected sepsis. Clin Infect Dis. 2025;80(1):108-117.

  14. Klompas M, Lapaol V, Taylor SP. Antibiotics for suspected sepsis: who & when? Intensive Care Med. 2025;51(6):1142-1144.