Editor's note:An earlier version of the story erroneously linked rosiglitazone to excess deaths in ACCORD. This mistake has been corrected in this version.

Participants in the Veterans Affairs Diabetes Trial (VADT) were 97% male, with an average age of 60 years at the time of study entry. Participants were assigned to standard or intensive blood glucose treatment groups. Average period of follow-up for individual patients was 6.25 years. VADT lasted 7.5 years, closing in May 2008.

VADT focused on patients who had not been helped by simpler therapies. At the time of enrollment, more than 40% of the participants had experienced previous cardiovascular events, more than 50% had abnormal lipids, 80% had high blood pressure, and most were obese.

The study was designed to eliminate variables other than blood glucose levels in both treatment groups by controlling blood pressure and lipids, improving diet and exercise, and treating with aspirin. The primary focus of glycemic control was to reduce CVD events, including stroke, MI, severe congestive heart failure, bypass surgery, amputations resulting from vascular insufficiency, and death from CVD. Secondary end points (microvascular events) will be analyzed and reported later this year.

VADT participants were a higher-risk population than those enrolled in the two other major trials. "[Their A_1c levels] averaged 9.5%, when a normal A_1c is 6%," said Dr William C Duckworth (University of Arizona, Tucson), cochair of the trial, in an ADA press release.

The majority of patients received several oral drugs, and 90% of participants in the intensive group and 74% of the standard group took insulin during most of the trial. Rosiglitazone was used in 72% of the intensive group and 62% of the standard group by year 3 of the trial; metformin was more commonly used in obese patients, and glimepiride in nonobese patients. Drug use varied over the course of the trial to maintain the established A_1c levels. VADT found no increase in deaths associated with use of rosiglitazone.

Average A_1c was reduced to 6.9% in the VADT intensive group and 8.4% in the standard group. Once achieved, these A_1c levels were maintained for the duration of the trial, providing enough difference between the standard- and intensive-group A_1c that its effect should have been evident. Nevertheless, blood sugar reduction was not found to have a significant effect in reducing CVD events.

"For intensive glucose control to yield a significant benefit on cardiovascular risk reduction, you may have to do it early," Duckworth observed. In a population such as this, with prior CVD, many risk factors, and long-term glucose control problems, "you cannot expect benefits from glucose control in the short term. . . . You can't expect miracles," he said.

Despite the nonsignificant relationship between glycemic control and cardiovascular events, both treatment groups had far fewer CVD events than predicted: 263 in the standard-treatment group and 231 in the intensive-treatment group. The predicted total of CVD events was from 650 to 700. This decrease was attributed to careful control of factors other than blood glucose.

A less publicized but highly significant finding of the VADT was presented by Dr Peter Reaven (University of Arizona). In this ancillary study, vascular calcium levels provided a good estimate of baseline atherosclerosis in many VADT subjects, and the baseline level of atherosclerosis is important in determining future CVD events.

Reaven explained these findings in an email. "The baseline level of atherosclerosis also influenced the response to tight glycemic control. . . . Baseline level of coronary calcium was the strongest determinant of future CVD outcome," he said, "and interestingly, intensive glycemic treatment appeared more favorable in those with less advanced disease. This appears to reflect the results in the larger VADT population, where those with less duration of diabetes seemed to gain benefit from tight glycemic control."

Reaven noted that this finding supports the results of ACCORD, which suggested some benefit of intensive glycemic control in patients with no previous CVD event.

In an email, Dr John B Buse (North Carolina School of Medicine in Chapel Hill), president, medicine and science, of the ADA, stated: "The unique finding [of VADT] was reported by Peter Reaven—namely, that in those with the least burden of CVD, at least as assessed by coronary artery calcium, there did seem to be a suggestion of benefit to more intensive therapy."

Reaven summarized: "Overall, these data suggest that identifying those individuals with less advanced vascular disease (ie, intervening earlier in the course of diabetes) may help determine who may benefit most from more aggressive glucose lowering."