Are Antibiotics Useful for Small Skin Abscesses? (NEJM Journal Watch)

NEJM Journal Watch:
Are Antibiotics Useful for Small Skin Abscesses? Now There’s an Answer

Comments from Dr. Paolo:

Some initial thoughts on this study:

  1. Even taken at face value this is one study in a vast sea of
    literature that has found no distinction between antibiotic therapy versus
    simple incision and drainage.  If we are to assume the 7% difference in
    rate of cure as defined in the article (more to come) then difference in
    cure rate is a solitary data point that will be summated within the larger
    body of literature currently demonstrating no significant first pass cure
    rate (rates of subsequent skin and soft tissue infections aside).  This may
    move the summated needle and change the end point of the combined
    literature but it is not, in and of itself against a backdrop of negative
    studies, compelling on its own.
  2. Primary outcome was defined by a clinical test of cure assessment at
    7-14 days via a non-validated internally developed outcome measure.  This
    instrument has never been studied, externally validated, or assessed to
    determine the secondary generalizability and inter-rater reliability of the
    assessment tool.  Though the study mentions a “good inter-rater
    reliability” there is no external means to assess this instrument and it is
    the (very) subjective crux upon which the whole study rests.
  3. 34 of the total cohort was lost to follow-up and not included in an
    intention to treat analysis.  69 individuals were never assessed at the
    test of cure visit (or were excluded prior to the test of cure visit) and
    were included in the mITT-2 population.
  4. The secondary outcomes are nonsense and there are 12 of them and they
    should be dismissed out of hand and ignored.  There is no reason to include
    multiple uncontrolled, nonsensical secondary outcomes that cannot possibly
    be controlled for and will, defintionally, have significant differences
    discovered simply by the volume of the parsing of the initial dataset.
  5. The point estimate of this study offers a NNT of 14 to (and this is
    important) demonstrate clinical cure as assessed by a non-validated
    instrument at 7-10 days post incision and drainage (with no standardization
    of the I&D techniques).  This means that at best 13 people will receive no
    benefit whatsoever from the exposure to high dose antibiotics and will
    demonstrate no changes in a (non-validated) clinical assessment
    instrument.  Further if the lower bound of the confidence interval were
    true this would yield an NNT of 50 in order to achieve a benefit on the
    (non-validated) clinical assessment instrument.   There were no harms found
    in this study (NNH) to counterbalance the NNT for us to determine the
    likelihood of harm from the prescription of antimicrobials.  I am not sure
    why this occurred but we are aware of high rates of allergic reactions,
    c-diff, antimicrobial resistance, and diarrhea to counterbalance to alleged
    benefit of antimicrobial prescriptions.
  6. There were no distinctions between invasive disease between each group
    or other severe outcomes that would bias towards the treatment with
    antimicrobials.  If you believe this study and the point estimated
    difference then this study does not (as it alleges using a non-validated
    instrument) argue for treatment of all comers to the ED with initial
    antimicrobials rather than I&D alone (which cured the vast majority of
    patients)—in fact it argues that the initial treatment SHOULD be I&D alone
    followed by a return visit (or PCP follow-up) in 7-10 days for
    reassessment.  At the return visit antimicrobial therapy can be added to
    the small and select group who are deemed treatment failures rather than
    the wanton and inappropriate prescription of antibiotics to all comers
    (based upon this study and its non-validated clinical assessment