Copyright 2011 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2011
We thank Nakao and colleagues for their letter. A selection bias in meta-analyses is mainly due to errors in the search strategy to screen the trials to be considered. These errors are related to an arbitrary selection of the trials or to the choice of factious search and eligibility criteria to select them. As a result, some studies are excluded from the analysis although they should be considered.
To avoid a selection bias, in our article1 we predefined and clearly reported the process for study selection, in keeping with the PRISMA check list. Our search keywords were generic and impartial and could not systematically exclude trials with specific characteristics. The meta-analysis by Verdecchia et al,2 which evaluated our same outcome (ie, heart failure [HF]), was only used as a control.
Sciarretta S, Tocci G, Palano F, Volpe M. Network Meta-analysis of Heart Failure Prevention by Antihypertensive Drugs—Reply. Arch Intern Med. 2011;171(5):472-473. doi:10.1001/archinternmed.2011.58