Ferrer et al, 2014: Antibiotic timing in Severe Sepsis

Journal Club 012

Author: Harini Bala
Reviewer: Chris Nickson

Ferrer R, Martin-Loeches I, Phillips G, Osborn TM, Townsend S, Dellinger RP, Artigas A, Schorr C, Levy MM. Empiric antibiotic treatment reduces mortality in severe sepsis and septic shock from the first hour: results from a guideline-based performance improvement program. Crit Care Med. 2014 Aug;42(8):1749-55. doi: 10.1097/CCM.0000000000000330. PubMed PMID: 24717459.


  • What is the relationship between timing of antibiotic administration and mortality in patients with severe sepsis or septic shock?



  • Retrospective analysis of a large dataset collected prospectively by the Surviving Sepsis Campaign database from 165 multi-national ICUs in Europe, USA and South America.


  • n = 17,990 patients were included in the analysis out of 28,150 patients with severe sepsis and septic shock between January 2005 to February 2010
    • 457 patients received no antibiotics
    • 832 received antibiotics but were missing timing of antibiotics and
    • 8,871 patients received antibiotics prior to suspected sepsis.

Inclusion criteria (all of the following)

  • Those admitted to ICU having a suspected site of infection
  • 2 or more SIRS criteria
  • 1 or more organ dysfunction
  • Once severe sepsis or septic shock was identified patients were eligible for antibiotics

Exclusion criteria

  • If patients received no antibiotics in the first 6 hours
  • If no timing of first antibiotics documented in the notes
  • Receiving antibiotics prior to the diagnosis of sepsis


  • Time to first antibiotic administration within 6 hours of sepsis identification and the effect on mortality
  • They also looked at different patient characteristics and the antibiotic timing
  • Length of stay, location (ED, Ward, ICU) organ dysfunctions (number and type), infection site, baseline organ dysfunctions and indices of shock


  • Results of the analysis of this large population of patients with severe sepsis and septic shock demonstrate that a delay in first antibiotic administration was associated with increased in hospital mortality
  • In addition there was a linear increase in the mortality for each hour delay in antibiotic administration
  • The adjusted hospital mortality odds ratios (OR) steadily increase from 1.00 to 1.52 as time to antibiotic administration increases from 0 to 6 hours where 0-1 hour is the referent group
  • The probability of mortality increases from 24.6% to 33.1%
  • The results underscore the importance of early identification and treatment of septic patients in the hospital setting
  • adjusted hospital mortality is shown in the table below (NNH is calculated 1/(EER-CER) e.g. for >6h = 1/(0.331 – 0.246):
Time OR p-value Mortality 95% CI NNH
<1h 1.00 24.6% 23.2–26.0 N/A
1–2h 1.07 0.165 25.9% 24.5–27.2 77
2–3h 1.14 0.021 27.0% 25.3–28.7 42
3–4h 1.19 0.009 27.9% 25.6–30.1 30
4–5h 1.24 0.006 28.8% 25.9–31.7 24
5–6h 1.47 <0.001 32.3% 28.5–36.2 13
>6h 1.52 <0.001 33.1% 30.9–35.3 12



  • Kumar et al 2006 that showed a linear increase in mortality with an increase in time to administration of antibiotics after the onset of hypotension in patients with sepsis. this study suggests that the same relationship is present for all patients with severe sepsis, who weren’t shocked and still shows an increase in mortality.
  • Strengths: used a large multi-center and multi-national data set, used GEE population averaged logistic regression to control for expected covariates, included both the raw and the adjusted data.
  • A definitive RCT answering this question will never be performed as delaying antibiotic administration is unethical


  • potential for residual confounding as this is a retrospective study, e.g. reliant on accurate data entry, early antibiotic use could be a surrogate for better overall care, some patients groups (e.g. immuncompromised) may get delayed antibiotics due to difficult in recognising severe sepsis in these patients – the accuracy and completeness of the SSC data set is uncertain
  • did not assess appropriateness of the antibiotic prescribed, and inappropriate or inadequate antibiotic choices may confound the results
  • Reasons why antibiotics were delayed eg: delay in prescribing or system factors, were not assessed
  • did not assess whether appropriate source control was performed
  • Unlike the ED patients, for patients admitted to the ICU from the medical and surgical wards and for patients in ICU the time of diagnosis of sepsis was determined by chart review, which is subject to biases in data entry and other factors (e.g. clinician seniority, timing of observation measurement) that may affect the time taken to recognise severe sepsis
  • time of severe sepsis recognition, rather than onset, is used as ‘zero time’. This is largely of academic interest as there is no feasible means of doing this and it is unlikely to affect the importance of antibiotic timing.
  • uncertainty as to whether the patients included in this dataset resemble the Australasian population


  • Severe sepsis is a time dependent condition and this study builds on previous literature which supports that early identification and treatment of sepsis with antibiotics decreases mortality.
  • We should aim to give antibiotics as early as possible following the recognition of severe sepsis.

References and Links

  • LITFL CCC — Antibiotic timing
  • Levy MM, Dellinger RP, Townsend SR, Linde-Zwirble WT, Marshall JC, Bion J, Schorr C, Artigas A, Ramsay G, Beale R, Parker MM, Gerlach H, Reinhart K, Silva E, Harvey M, Regan S, Angus DC. The Surviving Sepsis Campaign: results of an international guideline-based performance improvement program targeting severe sepsis. Intensive Care Med. 2010 Feb;36(2):222-31. doi: 10.1007/s00134-009-1738-3. Epub 2010 Jan 13. Review. PubMed PMID: 20069275; PubMed Central PMCID: PMC2826633.
  • van Zanten AR. The golden hour of antibiotic administration in severe sepsis: avoid a false start striving for gold*. Crit Care Med. 2014 Aug;42(8):1931-2. doi: 10.1097/CCM.0000000000000363. PubMed PMID: 25029127.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.