The National Institute of Clinical Excellence in the UK has been a real boon to the NHS, despite attempts to limits its influence by the pharmaceutical companies, the explicit evaluation of risks and benefits of various interventions allows appropriate use of resources.
This year they produced a guideline (and a shorter version they call a guidance) regarding the use of antibiotics for early onset infection in the newborn. The guideline is 320 pages long, and is a complete evaluation of the published literature. Even the guidance is 40 pages long, of which the first 26 are the meat of the guidance. Although I think in general it is very well done (such as the review of clinical signs which should lead to a sepsis evaluation), there are 2 things that I disagree with.
The first is the choice of antibiotics for preliminary treatment of a suspected sepsis in a full term newborn infant. NICE promote the use of benzylpenicillin and gentamicin. In their review of case series from the UK they note that up to 6% of cases of early onset sepsis are due to listeria. Listeria are not covered by benzylpenicillin, but require ampicillin. The reasoning behind the choice of benzylpenicillin is that the spectrum is narrower, therefore less risk of promoting resistance. While this is true, that is done at the expense of having 6% of true sepsis untreated. Is the use of 36 hours of antibiotics in a subgroup of newborns with suspected sepsis an important source of evolutionary pressure for the selection of resistant organisms? Ampicillin will also often treat community acquired E Coli also, another potential benefit. There is inadequate justification for the choice to my mind, and I will continue to promote ampicillin and gentamicin.
The second recommendation that I don’t agree with is the recommendation to take a CRP at baseline and to repeat 18 to 24 hours later. The problem with that guidance is that there are no clear recommendation about what to do with the results! The recommendation is that if the results are ‘reassuring’ you can stop antibiotics after 36 hours if cultures are negative. But what is a ‘reassuring’ result and trend of CRP? Is there any evidence that an infant with negative cultures and a non-‘reassuring’ trend in CRP benefits from prolonged antibiotic treatment? The observational studies on which the reliance on CRP are based have defined culture negative sepsis according to whether the antibiotics were continued, which often depended on the CRP! So we have a sort of circular argument, non-reassuring CRP leads to prolonged treatment, therefore the infant has culture negative sepsis, therefore we have to measure CRP in order to determine whether or not to stop antibiotics.
In the meantime we have no data that culture negative sepsis benefits from antibiotics, nor that CRPs are cost-effective or affect any measurable outcome.
The problem with this argument is that I don’t know where to go from here. Culture negative sepsis exists. Infants with negative cultures but who have all of the clinical signs of systemic infection are a real phenomenon; but do they benefit from antibiotic treatment? I dunno. Is a CRP a better decision making tool than a clinical exam? I dunno. Oh dear there are too many things that I don’t know, but at least I get to travel a lot to great places to tell people that!
Off to Melbourne tomorrow, but I will try to post a few times while I am gone.