Diabetes Reduction Assessment With Ramipril and Rosiglitazone Medication - DREAM - Rosiglitazone


The goal of the trial was to evaluate treatment with rosiglitazone compared with placebo among patients without cardiovascular disease or a history of diabetes but with impaired glucose.

Study Design

Study Design:

Patients Screened: 24,592
Patients Enrolled: 5,269
Mean Follow Up: Median 3 years
Mean Patient Age: Mean age 55 years
Female: 59

Patient Populations:

Age ≥30 years; impaired fasting plasma glucose levels (110 - <126 mg/dl) or impaired glucose tolerance (140 - <200 mg/dl 2 hours after an oral glucose load) but without a history of diabetes, cardiovascular disease, or intolerance of either angiotensin-converting enzyme inhibitors or thiazolidinediones

Primary Endpoints:

Newly diagnosed diabetes or death

Secondary Endpoints:

Composite of cardiac and renal events, defined as myocardial infarction, stroke, cardiovascular death, revascularization, heart failure, newly diagnosed angina with objective evidence of ischemia, or ventricular arrhythmia requiring resuscitation or renal event based on urine and blood analysis; glucose levels

Drug/Procedures Used:

Following a 17-day placebo run-in phase, patients were randomized in a double-blind manner to rosiglitazone (8 mg/day) (n = 2,635) or placebo (n = 2,634). An oral glucose tolerance was test performed at 2 years and at final study visit.

In a 2 x 2 factorial design, patients were also randomized to ramipril (up to 15 mg/day) (n = 2,623) or placebo (n = 2,646).

Principal Findings:

Medication compliance of at least 80% adherence at study end was 71.7% in the rosiglitazone group and 75.1% in the placebo group. Median fasting plasma glucose level at baseline was 5.8 mmol/L. Mean systolic blood pressure (SBP) at baseline was 136 mm Hg and mean diastolic blood pressure (DBP) was 83 mm Hg. Both SBP and DBP were reduced to a greater extent in the rosiglitazone group compared with placebo (by 1.7 mm Hg and 1.4 mm Hg, respectively, p < 0.0001).

The primary endpoint of death or new diabetes occurred less frequently in the rosiglitazone group compared with the placebo group (11.6% vs. 26.0%, hazard ratio [HR] 0.40, p < 0.001), driven by a reduction in diabetes (10.6% vs. 25.0%, HR 0.38, p < 0.001) with no difference in mortality (1.1% vs. 1.3%, p = 0.70). The diabetes reduction was evident in both physician-diagnosed diabetes (1.9% vs. 3.9%, p < 0.001) and glucose-diagnosed diabetes (8.8% vs. 21.1%, p < 0.001). Glucose levels returned to normal more frequently in the rosiglitazone group (50.5% vs. 30.3%, HR 1.71, p < 0.0001). At final study visit, median fasting plasma glucose levels were 0.5 mmol/L lower in the rosiglitazone group than in the placebo group (p < 0.0001); post-load glucose levels were 1.6 mmol/L lower in the rosiglitazone group than in the placebo group (p < 0.0001).

Cardiovascular events trended higher with rosiglitazone (2.9% vs. 2.1%, HR 1.37, p = 0.08). Chronic heart failure occurred significantly more frequently with rosiglitazone (0.5% vs. 0.1%, p = 0.01). Discontinuation due to edema was higher in the rosiglitazone group (4.8% vs. 1.6%), as was discontinuation due to weight gain (1.9% vs. 0.6%). Mean weight increase was 2.2 kg higher in the rosiglitazone group.


Among patients without cardiovascular disease or a history of diabetes but with impaired glucose, treatment with rosiglitazone was associated with a reduction in the primary endpoint of death or new diabetes compared with placebo at a median follow-up of 3 years.

The other agent studied in the present trial in a factorial design, ramipril, was not associated with a reduction in new diabetes or death compared with placebo. The reduction in diabetes development associated with rosiglitazone approached the levels seen with lifestyle modification. However, there were potential limitations to rosiglitazone therapy, including increased weight gain and a trend toward higher cardiovascular events, driven in part by an increase in heart failure. The increase in heart failure with a thiazolidinedione was also reported in the PROactive trial of pioglitazone in diabetic patients, possibly due to edema and fluid overload associated with thiazolidinediones. Data evaluating the effect of rosiglitazone on atherosclerosis was assessed on sequential carotid ultrasounds as a substudy of the DREAM trial and will be presented separately.


The DREAM Trial Investigators. Effect of rosiglitazone on the frequency of diabetes in patients with impaired glucose tolerance or impaired fasting glucose: a randomised controlled trial. Lancet 2006;368:1096-105.

Clinical Topics: Heart Failure and Cardiomyopathies, Prevention, Acute Heart Failure, Diet

Keywords: Follow-Up Studies, Atherosclerosis, Edema, Weight Gain, Ramipril, Systole, Medication Adherence, Glucose Intolerance, Glucose Tolerance Test, Heart Failure, Hypoglycemic Agents, Diastole, Thiazolidinediones, Diabetes Mellitus, Fasting

< Back to Listings