The VADT-F study

The Veterans Affairs Diabetes Trial (VADT) at 15 Years

Presented by:
Wyndy L Wiitala, PhD
VA Center for Clinical Management Research, Ann Arbor, MI; USA
Peter D. Reaven, MD
Phoenix VA Health Care System, University of Arizona, AR; USA
Nicholas V Emanuele, MD
Hines VA Hospital, Loyola University of Chicago, IL; USA

The primary objective of the Veterans Affairs Diabetes Trial (VADT) was to determine if a reduction in major cardiovascular (CV) events could be observed in an intensive glycemic treatment group compared to a standard care group in patients with type 2 diabetes. The VADT Follow-up study was designed to continue to follow patients enrolled in the trial after it ended. The secondary objectives of the VADT study were to assess differences between treatment groups on other CV endpoints. The patients included in the trial were generally older patients who had poorly controlled diabetes (nonresponsive to a maximal dose of at least one oral agent and/or daily insulin). A total of 1,655 patients were randomized to the study (818 in the standard treatment group and 837 in the intensive treatment group).

The VADT study design was such that the only planned difference between groups was the level of glycemic control (median HbA1c separation at the end of the study: 1.5%). The use of medications between treatment groups was similar, with differences in dose and insulin intensity only. All other treatment aspects (i.e., blood pressure, lipid control, aspirin therapy, education, diet, and exercise) were nearly identical between groups. The time to the first major CV event was a composite of myocardial infarction (MI), stroke, CV death, new/worsening congestive heart failure (CHF), amputation (ischemic gangrene), surgery (cardiac, cerebrovascular, peripheral vascular disease), and inoperable coronary artery disease (CAD). Secondary outcomes included CV mortality, major DM event, microvascular event, quality of life, and total mortality. Results from the VADT intervention study can be found in the New England Journal of Medicine.1

At the end of the VADT intervention, all patients returned to usual care without any further intervention. In the 15-year follow-up study, a subset of patients (n = 1,391) were followed through national data registries and surveys. The median follow-up for total mortality equaled 15 years and the median follow-up for primary outcome equaled 13.6 years. While starting HbA1c levels were >9%, an HbA1c separation of 1.5% was obtained early and maintained throughout the active treatment phase. Median HbA1c values were 6.9 in the intensive group and 8.4 in the control groups. However, other CVD risk factors were equal between the two treatment arms during this period. The VADT and VADT follow-up study represent the longest follow-up for patients with advanced type 2 diabetes at a high-risk for CVD.

At the end of the active phase of the VADT, there was a non-statistical 12% decline in the primary CVD outcomes. No statistical differences between treatment groups in death from CV causes or in death from any cause were observed. In the VADT-F study the goal was to evaluate if a CV benefit emerged, since there was an indication that group differences were increasing towards the end of the study and benefits in CVD outcomes take longer to appear.

Systolic/Diastolic blood pressure, triglycerides, and high-density lipoprotein-cholesterol (HDL-c)/non-HDL-c levels remained consistent over time. Body mass index (BMI) values were higher in the intensive treated group and were maintained through the trial, suggesting a possible negative consequence to intensive therapy. In the 10-year interim follow-up of primary CVD outcomes, patients who had been randomly assigned to intensive glucose control had fewer major CV events than those assigned to standard therapy, but no improvement was seen in the rate of overall survival.2 There was no evidence of heterogeneity for treatment interactions in results based on baseline for primary CVD outcome by baseline diabetes duration, prior CV event, and UKPDS score, nor any differences for any secondary outcomes (any major diabetes outcome, CVD death, or death from any cause).

When comparing the major CVD events during HbA1c separation between years 0-10 and 11-15, years 0-10 demonstrated a statistical reduction in major CVD events (HR, 0.83; 95% CI, 0.70-0.99). However, this equalized to the standard treatment group in years 11-15 showing no statistically significant difference.

Other observations, including hospitalization, median number of hospitalizations and health-related quality of life, did not demonstrate any statistical differences between groups. These results are relatively consistent with other recent glucose lowering trials that also examined post trial follow-up. In both the ACCORD and the ADVANCE follow-up studies there was also no evidence of reduced CVD or CVD death resulting from earlier intensive glucose control. Also, there was also no evidence of legacy benefits among any of the studies.3,4 Collectively, these results suggest there are modest long-term CVD benefits of therapies directed towards bringing glucose control to near normal range in high-risk type 2 diabetes, and that substantial and continuous glucose separation may be required to maintain these improvements. This highlights the importance of considering nonglycemic approaches to reducing CVD events and mortality in these patients.

In a meta-analysis of the VADT, UKPDS, ACCORD, and ADVANCE studies, the effect of intensive glucose control on microvascular outcomes was explored in patients with type 2 diabetes.5 The main focus was the review of kidney and eye complications. Composite kidney outcomes included end-stage renal disease (ESRD), renal death, eGFR <30 ml/min/1.73 m2, and sustained macroalbuminuria. Of the more than 1,600 primary kidney events in the 4 studies, there was a statistically significant 20% reduction in composite kidney outcomes in the intensely treated group compared to the standard treated group (HR 0.80, 95% CI, 0.72-0.88; p <0.0001), while there was an HbA1c separation between treatment groups.5

Statistical significant difference was evident at the VADT interim 10-year mark for eGFR <45 ml/min/1.73 m2 and sustained macroalbuminuria. However, in the 15-year VADT follow-up data, when HbA1c separation waned, no statistical significance persisted for any renal outcomes.

The composite eye outcomes included retinal photocoagulation therapy, vitrectomy, development of proliferative retinopathy, or progression of diabetic retinopathy ≥3 steps on the Early Treatment Diabetic Retinopathy Study (ETDRS) severity scale.5 Of the almost 800 primary eye events in the 4 studies, there was a statistically significant 13% reduction in composite eye outcomes in the intensely treated group compared to the standard treated group, while there was an HbA1c separation between treatment groups.5

In the 10-year time point and 15-year VADT follow-up data, there was no statistical significance for any cataract outcomes and borderline significance for the retinal event composite outcome.

Key messages

  • There was a delayed benefit after nearly 10 years of improved glucose control.
  • There was no demonstrated CVD benefit after 15 years of intensive glucose therapy.
  • There was a delayed beneficial effect of intensive glycemic control on the composite kidney outcome which waned as glycemic separation decreased.
  • There was a delayed beneficial trend of intensive glycemic control in reducing a composite of adverse retinal events with no effect on cataract extraction or on participants’ self-reported vision.


Present disclosure: All of the presenters, with the exclusion of those listed below, have reported no disclosures related to the VADT study. Peter D. Reaven, PhD: the presenter reported that he participated in an advisory panel for Sanofi and that he provided research support for Astra Zeneca and Novo Nordisk.

Written by: Debbie Anderson, PhD

Reviewed by: Marco Gallo, MD


Diabetic kidney disease & glycemic control

Pathophysiology of DKD and Glycemic Goals in Patients with Low GFR