Time to antibiotics in sepsis
© BioMed Central Ltd. 2010
Published: 1 March 2010
Mortality from sepsis increases with delay in administration of antibiotics. The Surviving Sepsis Guidelines suggest that antibiotics are administered within 1 hour of diagnosis of severe sepsis and septic shock. This retrospective study investigated the time to appropriate antibiotics in patients with sepsis and identifies reasons for delay where possible.
All adult patients in our institution who had positive blood cultures in 2008 were identified. Included were patients with sepsis, severe sepsis, or septic shock. Excluded were the isolation of low-pathogenicity organisms from blood culture (for example, coagulase-negative Staphylococcus). Case notes were reviewed to identify the time between diagnosis of sepsis and administration of antibiotics. Time zero was taken as the time recorded in the notes by the doctor that the patient was seen. Times of prescription and administration of antibiotics were taken from the prescription chart.
There were 1,477 sets of positive blood cultures in 1,098 patients. Following exclusions there were 503 episodes of positive blood cultures in 494 patients. Three hundred and seventy-five fulfilled the criteria and were analysed: 52 did not fulfil entry criteria, in 16 cases sepsis was not suspected, and 60 sets of notes were incomplete or unavailable. Average age was 69 years. Mean time between diagnosis and antibiotic administration was 8.1 hours, median 4 hours. A delay of 3 hours or more between diagnosis and administration of antibiotics occurred in 218 cases (58%). The 28-day mortality for all patients overall was 21%; 19% in the no-delay group compared with 24% in the delayed group. The 1-year mortality was 35% for both groups. Causes of delay in administration included prescription error, awaiting senior review, and patient movement between departments.
These retrospective data demonstrate that there is a delay in administering antibiotics patients with sepsis. These data are likely to be an underestimation as an accurate time zero is difficult to ascertain.
- Kumar A, Roberts D, Wood KE, et al.: Crit Care Med. 2006, 34: 1589-1596. 10.1097/01.CCM.0000217961.75225.E9View ArticleGoogle Scholar
- Dellinger RP, Levy MM, Carlet JM, et al.: Crit Care Med. 2008, 36: 296-327. 10.1097/01.CCM.0000298158.12101.41View ArticleGoogle Scholar
- Bone RC, Sibbald WJ, Sprung CL, et al.: Chest. 1992, 101: 1481-1483. 10.1378/chest.101.6.1481View ArticleGoogle Scholar