Response to letter: On the validity of meta-analysis

Article Outline

• References
• Copyright

In their communication, Pambrun et al. [1] comment on the well-known problems of literature-based meta-analysis in general and, especially, on the problem of including duplicate information from different publications in a recent literature-based meta-analysis on the diagnostic accuracy of transient elastography for the staging of liver fibrosis [2].

As authors of this literature-based meta-analysis, we agree with the well-known problem. Nevertheless, we did check thoroughly if the included full-article publications refer to duplicate use of patient data, did not include obvious duplications, and contacted authors of the included abstract publications. In addition, our meta-analysis was based on a large number of publications; we used robust statistical procedures. Even complete exclusion of the critical articles mentioned by Pambrun et al. would result in only minor changes of the results of our meta-analysis.

Also patient-based meta-analyses have their own disadvantages. Publication or selection bias is the most prominent one and statistical procedures to test for publication bias have limited power. Furthermore, the selected study populations may be more homogeneous and the heterogeneity of a general population may be underestimated.

There is a broad discussion on the differences between literature-based and patient-based meta-analysis. We feel that both approaches have their own areas of application. They could and should complement each other to obtain a full picture of the underlying clinical questions. Therefore, we look forward to a full publication of the patient-based meta-analysis of Bouteloup et al. [3] to complete our knowledge on the diagnostic accuracy of transient elastography for the staging of liver fibrosis.

Back to Article Outline

References 

1. Pambrun E, Bouteloup V, Thiébault R, Asselineau J, De Lédinghen V, Perez P. On the validity of meta-analyses: exhaustivity must be warranted, exclusion of duplicate patients too. J Clin Epidemiol. [in this issue]
   • View In Article
2. Friedrich-Rust M, Ong MF, Martens S, Sarrazin C, Bojunga J, Zeuzem S, et al. Performance of transient elastography for the staging of liver fibrosis: a meta-analysis. Gastroenterology. 2008;134:960–974
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (1175 KB)
   • CrossRef
3. Bouteloup V, Thiébault R, De Lédingen V, Perelz P. Méta-analse des performances diagnostiques de l'élastométrie impulsiennelle (Fibroscan) dans l'évaluation de la fibrose hépatique chez les patients attaints d'une hépatite chronique C. Rev Epidemiol Sante Publique. 2008;56:S70
   • View In Article