The study was a double-blind, randomised controlled trial involving over 4,000 treatment naive patients who were followed up for a median of 4 years. Patients were randomly assigned to one of the three study drugs (rosiglitazone, metformin or glyburide). A dose increase was required if fasting plasma glucose was above 7.8mmol/L at any visit. Initiation and maximum doses were defined as follows:
- Rosiglitazone, initially 4mg daily increased to a maximum of 8mg daily in divided doses
- Metformin, initially 500mg daily increased to a maximum of 2,000mg daily in divided doses
- Glyburide, initially 2.5mg daily increased to a maximum of 15mg daily in divided doses
The primary outcome of the study was monotherapy failure defined by fasting plasma glucose levels or more than 10mmol/L.
A higher than expected drop out rate (~40%) and a lower than expected event rate contributed to the decision to extend the follow-up period and increase the number of patients enrolled in the study. These changes reduced the statistical power of the study from 90% to 83%. The impact of this reduction in power upon the findings will be the topic of debate among statisticians.
Over the course of the study, Kaplan-Meier estimates of monotherapy treatment failure were 15% in the rosiglitazone group, 21% in the metformin group and 34% in the glyburide group. An analysis of the difference between the rosiglitazone and metformin groups could not exclude the effect of bias introduced by the high withdrawal rates.
Rosiglitazone was associated increases in presentation with oedema and use of loop diuretics than either of the other two study drugs but had fewer gastrointestinal side effects than metformin and fewer episodes of hypoglycaemia than glyburide. All of these differences were statistically significant.
There were no differences between groups in all cause death. Heart failure data were also collected showing no statistical difference between rosiglitazone and metformin but this diagnosis was investigator reported and the study was not designed to evaluate cardiovascular disease outcomes.
The authors conclude that additional studies are required to investigate whether the differences found in this study translate into longer-term benefits on disease progression and cardiovascular outcomes. They also state that, "the potential risks and benefits, the profile of adverse events, and the costs of these three drugs should all be considered" when selecting an appropriate treatment.
Action: The cardiovascular benefits of metformin are well proven. This study highlights the need for further investigation into the potential long-term benefits of treatment with rosiglitazone. In the meantime, metformin remains the first-line choice among hypoglycaemic agents.
|« MeReC Bulletin - Contraception||Lancet reports on stopped cholesterol trial »|