DOI: 10.1097/CCM.0000000000001219

,  

,  

PMID: 26376254

Issn Print: 0090-3493

Publication Date: 2015/10/01


Print

Appropriate Antibiotic Treatment in Severe Sepsis and Septic Shock: Timing Is Everything*

Marya D. Zilberberg; Andrew F. Shorr

    loading  Checking for direct PDF access through Ovid

Excerpt

Timely administration of empiric antibiotics that cover the potential pathogen(s) in patients with severe infections remains the mainstay of successful treatment. Over a decade of research indicates that when such treatment is not timely, the patient’s outcomes, including hospital mortality and length of stay (LOS), worsen significantly (1–9). In the setting of sepsis specifically, the Surviving Sepsis Campaign guideline recommends starting broad-spectrum coverage within the first hour of recognition of severe sepsis or septic shock (10). Despite this aggressive evidence-driven threshold of 1 hour, the sepsis bundle liberalizes this timeframe to 3 hours. A review of actual prescribing practices in this area suggests that fully one third of patients requiring aggressive initial therapy are left without appropriate coverage beyond even this 3-hour limit (11). This is a missed opportunity, given the well-characterized relationship indicating that for every hour’s delay in administration of appropriate antimicrobials, there is a substantial penalty in patient mortality (12, 13). The relationship between antibiotic treatment delay and hospital resource use, however, has been less reliably characterized. In the current issue of Critical Care Medicine, Dr. Zhang et al (14) attempt to do just that.
In this single-center retrospective cohort study, the investigators hypothesized that time to appropriate antimicrobial treatment in the setting of culture-positive severe sepsis and/or septic shock would prolong both ICU and hospital LOS. Despite the shortcomings of this retrospective single- center project, such as limited generalizability, the authors used appropriate enrollment and analytic techniques to address bias, misclassification, and confounding, all issues that can plague observational, and particularly retrospective, studies. In the end, the authors reported a small increase in LOS of about 0.1 days per each 1-hour delay in the administration of appropriate therapy. In other words, waiting 6 hours to institute appropriate antibiotic treatment following identification of severe sepsis or septic shock increased LOS on average by more than half of a day. Notably, this estimate remained durable across several sensitivity analyses.
It is true that this does not appear to be a particularly large attributable increase in the LOS for each individual patient. However, given that 30% of the patients in the study by Zhang et al (14) received appropriate therapy 18 hours or longer after the recognition of their sepsis, in aggregate this small individual hourly delay can quickly add up to a gargantuan total. In this cohort alone, one would estimate that the additional days in the hospital associated with a delay in therapy totaled over 600 days. Applying this estimate in turn to the national burden of sepsis, estimated at 1 million annually in the United States, with one third of cases subject to an over 3-hour delay in therapy, would yield well over 99,000 extra days in the ICU and the hospital each year. When looked at in aggregate, the seemingly small individual impact of a delay in treatment is astounding.
What are the barriers to eliminating or at least contracting this delay in treatment? Because the pathway to inappropriate selection of empiric coverage is at least partly through antimicrobial resistance of organisms, a clinician who is unaware of current local patterns of resistance lacks a crucial tool for stratifying his/her patient’s risk of harboring a resistant organism (15). Complicating the situation further is the fact that clinicians encounter both rare examples of resistance among common organisms, such as carbapenem-resistant Enterobacteriaceae, and relatively uncommon organisms with very high rates of resistant, such as Acinetobacter spp., where carbapenem resistance is seen in over 60% of all isolates. Because both cases represent a rare event, a clinician is less likely to cover these organisms empirically.

Related Topics

    loading  Loading Related Articles