In Reply: Dr Levien and colleagues and Dr Keipert and colleagues suggest that our meta-analysis should have included additional studies.1,2 The first article is not indexed in PubMed, EMBASE, or the Cochrane Library; moreover, it is not a controlled study, the major criterion for inclusion in our meta-analysis. Regarding the second study, as shown in the PubMed record, it was not fully coded with MeSH terms until July 2, 2008, and therefore could not have shown up on the March 25 term-based PubMed search. The study was present in EMBASE on July 2, 2008, but evidently was not present on March 25, 2008; we cannot determine when it was entered. The study was still not present in the Cochrane Library as of July 2, 2008. Regardless, the study had no fatalities or MIs in either study arm and therefore would have had no impact on our reported relative risks.
Natanson C, Lurie P, Wolfe SM. Hemoglobin-Based Blood Substitutes and Risk of Myocardial Infarction and Death—Reply. JAMA. 2008;300(11):1295-1299. doi:10.1001/jama.300.11.1298