Five RCTs (n=32,629) were included in the review. The authors reported that trial quality in terms of randomisation, allocation, treatment of dropouts and intention-to-treat analysis was satisfactory; three of four trials were open label and one was double blind. Trial duration varied from 2.9 to 11.1 years.
The weighted mean difference in glycated haemoglobin (HbA1c) between intensive and conventional hypoglycaemic therapy was -0.9%. Intensive therapy was associated with a significant reduction in the incidence of cardiovascular events (OR 0.89, 95% CI 0.83 to 0.96; five RCTs) and myocardial infarction (OR 0.85, 95% CI 0.78 to 0.93; five RCTs). There was no evidence of a difference in the incidence of stroke, chronic heart failure, all cause mortality or cardiovascular mortality.
Intensive therapy was associated with a significant increase in BMI (WMD 1.2kg/m2, 95% CI 0.3 to 2.2; four RCTs) and a significant increased risk of severe hypoglycaemia (OR 3.01, 95% CI 1.47 to 4.60; five RCTs). Exclusion of one, two, or both of two trials that measured hypoglycaemia in different ways did not significantly alter these results,
In meta-regression, a higher BMI and a longer duration of diabetes at enrolment were associated with a negative effect of intensive therapy on cardiovascular mortality (BMI p=0.001; duration of diabetes p=0.02). A significant correlation was found between the risk of severe hypoglycaemia and cardiovascular death in the intensive therapy group (figures reported as graphs).
There was no evidence of significant publication bias.
Heterogeneity assessments were not reported.