The specter of publication bias: Adjustment for publication bias in the evidence on cardiac death associated with passive smoking in nonsmoking women

Abstract

Several meta-analyses and reviews have been published on the risk of cardiac death in association with passive smoking. A recent meta-analysis by Kaur et al. [ [1] Kaur S. Cohen A. Dolor R. Coffman C.J. Bastian L.A. The impact of environmental tobacco smoke on women's risk of dying from heart disease: a meta-analysis. J Womens Health (Larchmt). 2004; 13: 888-897 Crossref PubMed Scopus (22) Google Scholar ] in conjunction with the American Heart Association guideline development for women, however, did not assess publication bias as well as most other meta-analyses. Meta-analyses are subject to bias because smaller or non significant studies are less likely to be published, and most meta-analyses do not consider the effect of publication bias on their results [ [2] Sutton A.J. Duval S.J. Tweedie R.L. Abrams K.R. Jones D.R. Empirical assessment of effect of publication bias on meta-analyses. BMJ. 2000; 320: 1574-1577 Crossref PubMed Scopus (944) Google Scholar ]. A funnel plot is a simple scatter plot of the effect estimates from individual studies against some measure of each study's size or precision, and the plot should approximately resemble a symmetrical (inverted) funnel in the absence of bias [ [3] Sterne JA, Egger M, Moher D, editors. Chapter 10: Addressing reporting biases. In: Higgins JP, Green S, editors. Cochrane Handbook for Systematic Reviews of Intervention. Version 5.0.1 (updated September 2008).The Cochrane Collaboration, 2008. Available from www.cochrane-handbook.org. Google Scholar ]. Although funnel plot asymmetry has long been equated with publication bias [ [3] Sterne JA, Egger M, Moher D, editors. Chapter 10: Addressing reporting biases. In: Higgins JP, Green S, editors. Cochrane Handbook for Systematic Reviews of Intervention. Version 5.0.1 (updated September 2008).The Cochrane Collaboration, 2008. Available from www.cochrane-handbook.org. Google Scholar ], the funnel plot should be seen as a generic means of displaying small-study effects: a tendency for the effects estimated in smaller studies to differ from those estimated in larger studies [ [4] Sterne J.A. Gavaghan D. Egger M. Publication and related bias in meta-analysis: power of statistical tests and prevalence in the literature. J Clin Epidemiol. 2000; 53: 1119-1129 Abstract Full Text Full Text PDF PubMed Scopus (1449) Google Scholar ]. Several statistical tests for funnel plot asymmetry (small-study effects) have been proposed, and Duval and Tweedie [ [5] Duval S. Tweedie R. Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics. 2000; 56: 455-463 Crossref PubMed Scopus (7800) Google Scholar ] developed a nonparametric method (‘trim and fill’ algorithm) for estimating the number of missing studies that might exist in a meta-analysis and the effect that these studies might have had on its outcome. We revisited the evidence [ [1] Kaur S. Cohen A. Dolor R. Coffman C.J. Bastian L.A. The impact of environmental tobacco smoke on women's risk of dying from heart disease: a meta-analysis. J Womens Health (Larchmt). 2004; 13: 888-897 Crossref PubMed Scopus (22) Google Scholar ] on cardiac death associated with passive smoking and adjusted for publication bias using the ‘trim and fill’ algorithm [ [5] Duval S. Tweedie R. Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics. 2000; 56: 455-463 Crossref PubMed Scopus (7800) Google Scholar ].