The current issue of the American Journal of Medicine (the Green
Journal) has an article titled Top Myths of Diagnosis andManagement of Infectious Diseases in Hospital Medicine. This is
one of the better articles pertaining to antibiotic stewardship that
I have seen. Ten myths are listed. They are not complete myths
(exceptions apply to just about all of these principles); rather,
they are misconceptions.
Myth one:
antibiotics do no harm. High-level data refute this myth. For
example, a recent metaanalysis showed
that the use of procalcitonin guidance to shorten the duration of
antibiotic therapy was associated with lower mortality.
Myth two: antibiotic
durations of 7, 14, or 21 days are typically necessary. Although
these are common recommendations, evidence is lacking. In many
situations (and there are exceptions) shorter duration therapy is as
good as longer duration. Examples include 3 to 5 days for community
acquired pneumonia; eight days for nosocomial pneumonia; 5 to 7 days
for pyelonephritis; four days for intraabdominal infection; five days
for acute exacerbations of COPD and 5 to 6 days for cellulitis. There
are notable exceptions. Certain deep-seated and difficult to
eradicate infections are not candidates for either shorter duration
or procalcitonin guidance for discontinuation. These include
tuberculosis, meningitis, prosthetic joint infections, staphylococcal
bacteremia, endocarditis and invasive fungal infections. The same
caution applies to some immunocompromised patients.
Myth three: if one
drug is good two (or more!) must be better. This requires nuance.There are some indications for combination therapy. They are
exceptions rather than the rule. The main indication for combination
therapy is initial empiric treatment for life-threatening infection
such as sepsis. The rationale is to cover all likely pathogens.
De-escalation is appropriate if and when culture and sensitivity
results indicate that a single agent would be appropriate. This
principle is also applied in meningitis where in patients 50 years
of age or older we include listeria coverage such that in non pen
allergic patients ampicillin is added to the combination of
ceftriaxone and vancomycin. Also in meningitis the combination of
ceftriaxone and vancomycin accommodates the possibility of relative
resistance of strep pneumo which might cause treatment failure with
cephalosporin monotherapy. In community acquired pneumonia
requiring hospitalizations the guidelines call for combination
cephalosporin and macrolid therapy. For patients admitted to ICU it
is recommended that MRSA coverage be added (this is in the IDSA
MRSA guideline, not the pneumonia guidelines).
A frequently asked
question is what to do about serious gram negative infections.
Traditionally “double coverage“ with two gram-negative agents has
been used. For the most part this is not supported by evidence. One
exception is in the initial (empiric) antipseudomonal coverage for
HAP/VAP., the guidelines for which indicate double coverage
initially which should be de-escalated later if microbiologic data
allow. In contrast, the CAP guidelines for patients with
pseudomonas risk recommend monotherapy---not double coverage--from
the start.
Myth four: oral
antibiotics are not as good as IV antibiotics for hospitalized
patients. This is, in general, a myth but there are exceptions. IV
therapy is not inherently better than oral if there’s adequate
bioavailability with the oral agent and if susceptibilities allow
switch to an oral agent. Cautions apply in bacteremia. Some of
these (eg staphylococcal bacteremia ) require intravenous therapy for
the entire course. In other bacteremic infections step down to oral
agents may be appropriate. Examples include certain streptococcal
bacteremias and gram-negative bacteremic urinary tract infections.
In such cases a switch to oral therapy can be considered as early as
day four.
Myth five: bacteria
in the urine signifies a UTI and should be treated. If it is
asymptomatic treatment is only warranted in pregnancy and patients
about to undergo a urologic procedure.
Myth six: history of
penicillin allergy means the patient can never receive a beta lactam
antibiotic. Former thinking was that there’s a 10 to 15% cross
sensitivity rate between penicillin and cephalosporins. More recent
findings indicate that the cross sensitivity rate to penicillin
allergy is more like 3% for cephalosporins and 1% for carbapenems.
The article provides some general principles for decision making in
patients with purported penicillin allergy. For reactions that are
mild and non specific a cephalosporin can be given. If the reaction
was anaphylactoid either an alternative antibiotic to a cephalosporin
or penicillin desensitization is recommended. Severe non anaphylactic
reactions such as Stevens-Johnson syndrome, toxic epidermal
necrolysis and DRESS syndrome are in a different category. In those
cases the use of any beta lactam is contraindicated as is penicillin
desensitization.
Myth seven:
antibiotics for surgical prophylaxis should be continued for 24
hours or more .
Myth eight:
antibiotics must be continued for as long as drains are in place.
Although clinical judgment is required here there is no robust
evidence to support such a practice.
Myth nine:
nitrofurantoin can be used for UTIs only if the creatinine clearance
is greater than 60. This is in accordance with product labeling but
the data indicate 30 may be more reasonable cut off.
Myth ten:
fluoroquinolones are first line agents for many infections. We now
have mounting evidence of adverse effects (CNS toxicity,, tendon
rupture, dysglycemias, irreversible neuropathy, QT prolongation and
aortic dissection) such that fluoroquinolones have been relegated to
a lower position in the sequence. They should be used only when
safer and equally effective alternatives are not available.
In a related
antibiotic stewardship topic this article from Cinical Infectious Disease looked at the
utility of MRSA PCR screening. Negative PCR can allow for
discontinuation or avoidance of MRSA therapy such as vancomycin in
many situations. Those studied in the article were bloodstream
infections, intra-abdominal infections, respiratory infections, wound
infections and urinary tract infections. Negative predictive value
was 93% or better in all those situations.
