Sepsis Guidelines and Antibiotics: Empiric Recommendations Going a Bit Too Far?

Paul G. Auwaerter, MD


May 16, 2017

Hello. This is Paul Auwaerter with Medscape Infectious Diseases. The new Surviving Sepsis Campaign guidelines,[1] released earlier this year, can take a considerable amount of time to look over and digest. The impact of this kind of document should not be underestimated, as often the recommendations are put into bundles for clinical care or are adopted by regulatory agencies for monitoring how well institutions are taking care of their patients.

As an infectious diseases specialist, I focused on a couple of things during my review, and I thought I might discuss these briefly. The first is that in the 2012 iteration, many of us grew accustomed to the idea of having a cascade of terms to facilitate early recognition of sepsis: SIRS [systemic inflammatory response syndrome], sepsis, severe sepsis, and septic shock.

Now, this has been changed a bit so that there is a SOFA [sequential organ failure assessment] calculator. This is a rather complex equation in which you enter data to see whether the patient is heading into trouble with high mortality. The qSOFA is a quick score with a more limited set of criteria. This will take a little getting used to. Many people have kept to the SIRS criteria and other terms, mainly because those have been set in motion and change will take a while.

An element of the 2012 document that's continued in the new guidelines is the goal of administering antibiotics within 1 hour of the recognition of severe sepsis or septic shock. This is based on an observational study[2] from more than 10 years ago and others, and is biologically plausible, although a meta-analysis[3] showed no difference in mortality between people receiving antibiotics within 3 hours versus 1 hour. With sepsis being so highly heterogeneous, and with many delivery-of-care variables, it's very hard to know for sure whether this 1-hour goal is meaningful for all patients or only for some patients.

Starting antibiotic therapy as early as possible is a reasonable goal. Whether a randomized trial will ever address this question is probably unlikely, but the true worth of this practice will continue to be a subject of some discussion.

The new guidelines have an expanded section on antibiotic therapy. The authors of this document tried to cover a tremendous amount of ground which may require some refinement. My sense is that they tried to cover a bit too much and are a bit too prescriptive on the large number of potential situations that people might face. I worry that it might be confusing or lead to inadvertent or prolonged therapy.

Let me give you some examples of weak recommendations that are based on low levels of evidence. This includes using procalcitonin levels to either stop antibiotic therapy when there is no infection or to shorten the course of therapy. Recommendations are given to try to limit antibiotic duration for sepsis to 7 -10 days, but a longer duration of therapy might be needed if you are dealing with Staphylococcus aureus or endocarditis, or a shorter duration if the patient has pyelonephritis or an intra-abdominal infection that resolves quickly.

These situations, with so many variables, are often hard to put into guidelines. The efforts for de-escalation of therapy, hewing to antibiotic stewardship and source control, are laudable, and very few people would argue with these. However, there is a rather complex concept in this document about combination therapy for the sickest patients with scores indicating a 25% mortality. If you have someone with only 15% mortality, you use only monotherapy. I am concerned that there may be an easy default to combination therapy. It hasn't been well proven —certainly not for gram-negative infections —that there is a need for double coverage, and the advice is to only give it for a "few" days. I'm not sure how long that is.

Given wide heterogeneity, it may be difficult to put all of these factors into any kind of formal guidance. I thought the guidelines might have gone too far and tried to address too many potential variables. If you have someone with a severe pneumonia—versus an intra-abdominal infection or meningitis—these infections require so many different considerations that it is very hard to put them under the category of sepsis, even if you are trying to address empiric therapy.

The guidelines will hopefully lead to earlier recognition of sepsis, avoidance of unintended mortality, and so on. These are good goals. However, the antibiotic recommendations are a bit too complex, mainly with respect to initial overtreatment and a lack of consideration for other conditions. We must try to avoid the kind of lockstep issues that sometimes come into play with guidelines. I'm interested in your thoughts as well. Thank you for listening.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.