Most, but not all, epidemiological studies have shown an association between glycemic control and the development of cardiovascular disease. However, the association of glycemic control with cardiovascular disease is considerably weaker than the association of glycemic control with the microvascular complications of diabetes, such as retinopathy and nephropathy. It must be recognized that epidemiological studies can only demonstrate associations and that confounding variables could account for the association between poor glycemic control and cardiovascular disease. For example, patients with poor glycemic control may not undertake other preventive measures that could reduce cardiovascular disease such as exercise, healthy diet, etc. Therefore, randomized studies are essential in determining the role of glycemic control on cardiovascular disease.

Unfortunately, there are very few randomized studies of the effect of glycemic control on cardiovascular disease. Early randomized studies, such as the UGDP and VA cooperative study, did not demonstrate a reduction in cardiovascular events in aggressively treated patients and in fact suggested that improvements in glycemic control (VA cooperative study) or the use of certain drugs to treat diabetes (oral sulfonylureas in UGDP) may actually increase the risk of cardiovascular disease. More recent studies, the DCCT in patients with Type 1 diabetes and the Kumamoto study in patients with Type 2 diabetes, while demonstrating a decrease in cardiovascular events in the subjects randomized to improved glycemic control did not have enough cardiovascular disease events to demonstrate a statistically significant reduction (DCCT studied a population at low risk for cardiovascular disease and the Kumamoto study had a very small number of subjects). In contrast, both the DCCT and the Kumamoto study clearly demonstrated that improvements in glycemic control resulted in a reduction in microvascular disease. Very recently a long term follow-up of the DCCT has demonstrated that improvement in glycemic control reduces cardiovascular disease. The initial DCCT compared intensive vs. conventional therapy for a mean of 6.5 years. At the end of the study a very large proportion of subjects agreed to participate in a follow-up observational study (Epidemiology of Diabetes Interventions and Complications- EDIC). During this follow-up period glycemic control was relatively similar between the intensive therapy and conventional therapy group (glycosylated hemoglobin 7.9% vs. 7.8%) but during the trial there was a large difference in glycosylated hemoglobin levels (7.4% vs. 9.1%). After a mean 17 years of observation the risk of any cardiovascular event was reduced by 42% and the risk of nonfatal myocardial infarction, stroke, or death from cardiovascular disease was reduced by 57% in the intensive control group. This study demonstrates that intensive glycemic control (for 6.5 of the 17 years of observation) is sufficient to have long term beneficial effects on the risk of developing cardiovascular disease. This beneficial effect was not entirely due to the prevention of microvascular complications as the differences between the intensive and conventional treatment groups persisted after adjusting for microalbuminuria and albuminuria.

The UKPDS studied a large number of patients with Type 2 diabetes at high risk for cardiovascular disease. In this study improved glycemic control, with either insulin or sulfonylureas, reduced cardiovascular disease by 16%, which just missed being statistically significant (p=0.052). In the UKPDS the improvement in glycemic control was modest (HbA1c reduced by approximately 0.9%) and the 16% reduction in cardiovascular disease was in the range predicted based on epidemiological studies. Similarly, the DiGami study, which used insulin infusion during the peri MI period to improve glycemic control followed by long-term glycemic control, demonstrated that survival post MI was significantly improved by good glycemic control. While this study focused on a very selected population and time period (patients undergoing a MI) the results are consistent with the hypothesis that improvements in glycemic control will reduce cardiovascular disease. However, a recently reported DiGami 2 study did not confirm the benefits of tight glucose control beginning in the peri MI period on outcomes. It most be noted though that the differences in glucose control achieved in DiGami 2 were much smaller than planned and the number of patients recruited was less than anticipated. Together these deficiencies could account for the failure to demonstrate significant differences in cardiovascular disease events.

Thus, while the data are not perfect, they strongly suggest that improvement in glycemic control will reduce cardiovascular disease but that the reduction will be modest and will not lower the risk to levels observed in non-diabetic subjects. Of note is that in the UKPDS, metformin, while producing a similar improvement in glycemic control as insulin or sulfonylureas, markedly reduced cardiovascular disease by 40%. This indicates that effects of metformin other than improving glucose control, such as decreasing lipids, decreasing insulin resistance, preventing weight gain, etc., may account for the beneficial effects on cardiovascular disease. Additionally, this demonstrates that the method by which one improves glycemic control may be very important. Studies with thiazolidindiones, such as rosiglitazone and pioglitazone, have suggested that they may have anti-atherogenic properties (reduction of proatherogenic lipid profile, decreases in C-reactive protein, decreases in PIA-1, reduction in carotid intimal media thickness, etc). Moreover, a recent randomized controlled trial (ProActive Study) examined the effect of pioglitazone vs. placebo over a 3 year period in type 2 diabetics with pre-existing macrovascular disease. With regards to the primary endpoint ( a composite of all-cause mortality, non-fatal myocardial infarction including silent MI, stroke, acute coronary syndrome, endovascular or surgical intervention in the coronary or leg arteries, and amputation above the ankle) there was a 10% reduction in events in the pioglitazone group but this difference was not statistically significant (p=0.095), However, the pioglitazone treated group did demonstrate a 16% reduction in the main secondary endpoint (composite of all cause mortality, non-fatal myocardial infarction, and stroke) that was statistically significant (p=0.027). In the pioglitazone treated group the blood pressure, and HBA1c, triglyceride, and HDL levels were all improved compared to the placebo group making it very likely that the mechanism by which pioglitazone decreased vascular events was multifactorial.