Intended for healthcare professionals


What is the optimal duration of antibiotic therapy?

BMJ 2006; 332 doi: (Published 08 June 2006) Cite this as: BMJ 2006;332:1358

Antibiotic therapy: suggestion for necessary trials

Dear Sirs,

In BMJ 2006; 332:1358 (issue June 10, 2006) your author John Paul writes
about the optimal duration of antibiotic therapy. He is right that for
certain situations (surgical prophylaxis, treatment of gonorrhoea,
treatment of uncomplicated urinary tract infections in women) there is an
empirical basis for duration of antibiotic therapy.

In some other points, however, it would be open and a lot of
variables have to be juggled, including the patient’s immune status, the
infecting agent, choice of antibiotic, nature of the septic focus and
others. It is evident that for such a multitude of variables simple rules
like e.g. this diagnosis needs five days, another one 10 days, are

Unfortunately, there are only few well-founded articles on the optimal
duration of antibiotic therapy, but all authors agree that a prolonged
exposure to antibiotics has to be avoided by means of side effects,
worsening resistance situation und costs. In this situation - as an
internist and nephrologist dealing with serious septic infections and
pneumonias every day – I always kept to the following rule: I continuously
controlled all possible parameters of inflammation, i.e. fever,
leukocytosis, left shift and specially the CRP. My practical experience
showed that after normalization of blood count and CRP the treatment could
be stopped. However, decrease of fever and normalization of blood count
alone are not sufficient indicators, a discontinuation can result in
relapses; so the serial determination of CRPs is to be advised as a most
important sign – a very cheap measure.

As there is no study which could stand up to scientific criticism I cannot
contribute more than this personal experience to this subject. An
adequate trial should be designed as follows: A large number of patients
with different serious infections will be randomised; in the first group
the antibiotic therapy will be discontinued after normalization of CRPs,
in the second group according to personally experience and view of the
doctor. The criterion would be the number of relapses in both groups and
the consumption of antibiotics. However, such a trial would be biased in
the way that in the second group the treatment would last longer and
perhaps despite more antibiotic consumption safer.

Finally, who should pay such a trial? There will not be any interest by
the pharmaceutical industry to support studies, which show, how you can
spare antibiotics.

Prof. Dr. med. D. Hoeffler


Competing interests:
None declared

Competing interests: No competing interests

30 June 2006
Dietrich Hoeffler
Vicepresident Drug commission of the German Medical Association
64331 Weiterstadt GERMANY