Determination of the incidence of tuberculosis in low-income countries.

Correspondence Aspirin Use in Infective Endocarditis To the Editor—Chan et al. [1], are to be congratulated for their prospective in-terventional study of aspirin use in patients with infective endocarditis (IE). Their study did not fin that aspirin reduced the frequency of major embolic events and showed a trend towards an increase in the number of major bleeding episodes in the intervention group. As acknowledged in their recent article [2], the initial study may have underestimated the benefi of aspirin, because the intervention was started relatively late after symptomatic onset of IE. Thus, they carried out a post-hoc analysis of observational data comparing patients taking long-term aspirin (in various doses) prior to the onset of IE, who continued using aspirin after the diagnosis, versus " placebo-control " patients with IE who did not receive aspirin before or after the diagnosis of IE. Data from this most recent study appear to support the authors' prior contentions that aspirin use in IE does not reduce em-bolic complications, although it potentially increases the risk of major hemorrhage. Importantly, these data are in contradistinction to a recent, well-conducted , retrospective study from the Mayo Clinic that showed significantl reduced rates of emboli in patients with IE who had been taking aspirin prior to the diagnosis [3]. Moreover, many of the limitations extant in the initial article [1] are perpetuated in this follow-up investigation and serve to substantially confound the authors' conclusions. Firstly, the animal studies cited showing the benefi of aspirin in improving IE outcomes are specifi for IE caused by Staphylococcus aureus [4–6]. This paradigm was recently underscored by a prospective investigation [7] showing significan reduction in the frequency of S. aureus (but not streptococcal or enterococcal) bacteremia in patients with indwelling hemodialysis catheters who had been taking long-term aspirin. Thus, a major shortcoming in the clinical data reported in both papers by Chan and colleagues is the relatively small number of S. aureus IE cases. This small sample size prevents a valid statistical analysis of potential salutary impacts of aspirin on IE complications. Secondly, initial studies by Kupferwas-ser et al. [4] were careful to point out that the antimicrobial and anti-embolic impacts of aspirin in animals with established S. aureus IE depended on the dose. Thus, there appears to be a prominent " Goldi-locks effect, " in which too little or too much aspirin may cause paradoxically diminished impacts on outcome metrics in …