Using individual patient data in meta-analyses of glucose-lowering studies

Excerpt

To the Editor: The Collaborators on Trials of Lowering Glucose (CONTROL) Group [1] have performed an excellent and important meta-analysis of the impact of intensive glucose-lowering on cardiovascular outcomes. The fact that the trialists of the four major glucose-lowering studies [2‐5] were involved in the collaboration has the advantage of providing detailed understanding of the similarities and differences between studies, both in terms of protocol and for classifying and adjudicating pre-existing cardiovascular disease (CVD) and individual endpoints. We would nevertheless like to suggest some ways in which the analyses might have been improved, together with some comments on the presentation of their data. …