Meta-analysis: effects of intensive glucose control on cardiovascular outcomes in type 2 diabetes

Intensive blood glucose control in patients with type 2 diabetes decreases the risk of some cardiovascular disease (CVD), but does not decrease the risk of death in the medium term and increases the risk for severe hypoglycaemia, according to this systematic review and meta-analysis: recent studies with stringent control suggest an increased risk for cardiovascular death.

Type 2 diabetes is a major risk factor for CVD. Intensive control of blood glucose reduces microvascular complications in people with type 2 diabetes, but its effect on clinical CVD is uncertain. Earlier trials suggested benefits, however later studies found no effect or adverse effects. There were fewer than expected events in later studies, so the authors of this paper carried out a meta-analysis of the available trials to clarify the effects of intensive control on CVD as a whole and on a range of individual clinical outcomes.

They searched the literature using Medline, and experts contacts and reference lists only. Eligible studies were large (>500 participants), randomised controlled trials comparing intensive blood glucose control with conventional treatment in patients with type 2 diabetes, that had clinical CVD as a primary endpoint. From these, they extracted data on all-cause mortality, and CVD mortality, along with coronary heart disease (CHD), congestive heart failure (CHF), and stroke events. Additionally, they recorded single endpoints for fatal and non-fatal myocardial infarction (MI) and strokes, and peripheral artery disease. The main safety outcome was severe hypoglycaemic events. Individual patient data were not obtained.

The initial literature search identified 341 reports, of which 304 were excluded on the basis of title and abstract. From the remaining 37 reports, 5 papers were eligible for analysis (23 duplicate reports, 5 not type 2 diabetes, 2 n<500, and 2 did not include groups of interest). The five eligible studies included 27,802 participants (range 753 to 11,140) and had durations of between 3.4 and 10.7 years. The two earlier studies (UKPD 33 and 34) recruited newly diagnosed patients whereas participants in more recent studies (ADVANCE, ACCORD and VADT) had established diabetes (average duration from 7.9 to 11.5 years).

Analysis of the summary results found that intensive control was associated with a 10% reduction in risk of CVD compared to conventional control (relative risk [RR], 0.90; 95% CI, 0.83 to 0.98). The risk difference (RD, per 1000 patients over 5 years of treatment) for CVD was -15 (95% CI, –24 to –5), and intensively treated patients also had a reduced risk of CHD (RR, 0.89; 95% CI, 0.81 to 0.96; RD, –11; 95% CI, –17 to –5).

Intensive control was associated with a 16% overall reduction in risk of non-fatal MI, however the absolute reduction (i.e. RD) was 9 events per 1000 patients over 5 years of treatment. There were no significant effects on fatal MI, or on fatal or non-fatal stroke, or peripheral artery disease.

Intensive control was not associated with a reduced risk of cardiovascular death (RR, 0.97; 95% CI, 0.76 to 1.24; RD, –3; 95% CI, –14 to 7), nor was it associated with reduced all-cause mortality (RR, 0.98; 95% CI, 0.84 to 1.15; RD, –4; 95% CI, –17 to 10).

Overall, intensive treatment was associated with a doubled risk of severe hypoglycaemia (RR, 2.03; 95% CI, 1.46 to 2.81; RD, 39; 95% CI, 7 to 71), absolute increase of 39 events per 1000 patients over 5 years). Most of this increase came from the three later studies, in which the absolute increase was 54 events per 1000 patients over 5 years.

The authors conclude that in patients with type 2 diabetes, intensive blood glucose control reduces the risk for some CVD, such as non-fatal MI, but does not (over the period studied) reduce cardiovascular or all-cause death. It does, however, double the risk of severe hypoglycaemia. They note that the earlier trials suggested a reduction in cardiovascular death, however the later studies, which had more stringent blood glucose control, suggested an increase. There were significant differences between the trials in the treatments used in both active and control groups, and in the patients recruited. Limitations include use of summary rather than individual patient data, and the short duration of the more recent studies.

Ann Intern Med, published early online 21 July 2009; 151(6) (link to abstract)