RT Journal
A1 Haider AW, Luna M, Patel S, Gaziano J
T1 ANtibiotic use and risk of myocardial infarction
JF JAMA
JO JAMA
YR 1999
FD December 1
VO 282
IS 21
SP 1997
OP 1999
DO 10.1001/jama.282.21.1997
UL http://dx.doi.org/10.1001/jama.282.21.1997
AB 
First, confounding remains a real explanation for the associations.
It would be helpful if the authors presented cardiovascular risk factors by
antibiotic user status. In addition, it would be useful to see a comparison
of crude, unadjusted risk ratios with adjusted estimates. This would permit
the reader to assess the impact of confounding and gauge the possibility of
residual confounding. Second, exclusion of the traditional risk factors of
hypertension, diabetes mellitus, and hyperlipidemia may introduce selection
bias. Patients who more often seek medical attention and thus have the opportunity
to be given antibiotics also may be more likely to have their hypertension,
diabetes, and hyperlipidemia diagnosed. These conditions might not be diagnosed
as frequently among those who less often seek medical attention as reflected
by their lower rates of antibiotic use. Eliminating all individuals with these
risk factors from the analyses could have the effect of selectively reducing
the rate of these known cardiovascular risk factors from the group receiving
antibiotics and thus reducing the rate of cardiovascular events. It would
be useful to see the rates of these factors by antibiotic status as well as
analyses that do not exclude those with these risk factors for comparison
with the analyses presented in the article. Third, after allowing for multiple
testing, the apparent associations could be merely due to chance; this is
particularly the case for quinolones, for which the association is only marginally
significant. Finally, recent prospective studies have failed to show a significant
association of serologic evidence of chronic infections with Chlamydia pneumoniae, Helicobacter pylori,
and cytomegalovirus infection and risk of coronary artery disease.2- 4