|Home | About | Journals | Submit | Contact Us | Français|
Tight control of HbA1c levels has been enshrined in the QOF (quality and outcomes framework) of the new general practitioner contract as being an evidence based proposal. Multiple drugs are licensed on the understanding that they reduce HbA1c levels and that this is a good thing. So should it surprise us that a meta-analysis of trials of rosiglitazone shows a raised risk of myocardial infarction and an increase in cardiovascular deaths?1
The reduction in diabetes related end points, mortality and stroke from using metformin is not explicable on the basis of glycaemic control.2
The data of UKPDS 33, which compared tight glycaemic control with sulphonylureas or insulin with conventional treatment, showed little benefit from tight control.3 The outcomes that did show some clinical benefit were cataract extractions, retinal photocoagulation, and non-fatal myocardial infarction and all cause mortality—that is, if you can call absolute risk reductions of between 1 and 3 per 1000 patient years as being clinically relevant.
The wonder with drug licensing is that we continue to accept surrogate end points in trials to license new treatments for conditions for which we already have treatments. Show me better data or accept that the control of blood glucose means metformin--anything else is merely for symptom control.
Competing interests: IRT has coauthored with Adrian Edwards, Glyn Elwin, and Rhys Williams a paper on explaining risk information over the internet to patients with diabetes, which was funded by the BMJ Group.