Aims: In an international prospective cohort study we assessed the relationship between glucose levels and incident cardiovascular events and death. Methods and results: 18,990 men and women were screened for entry into the DREAM clinical trial from 21 different countries. All had clinical and biochemical information collected at baseline, including an oral glucose tolerance test (OGTT), and were prospectively followed over a median (IQR) of 3.5 (3.0–4.0) years for incident cardiovascular (CV) events including coronary artery disease (CAD), stroke, congestive heart failure (CHF) requiring hospitalization, and death. After OGTT screening, 8000 subjects were classified as normoglycaemic, 8427 had impaired fasting glucose (IFG) or impaired glucose tolerance (IGT), and 2563 subjects had newly diagnosed type 2 diabetes mellitus (DM). There were incident events in 491 individuals: 282 CAD, 54 strokes, 19 CHF, and 164 died. The annualized CV or death event rate was 0.79/100 person-years in the overall cohort, 0.51/100 person-years in normoglycaemics, 0.92/100 person-years among subjects with IFG and/or IGT at baseline, and 1.27/100 person-years among those with DM ( p for trend <0.0001). Among all subjects, a 1 mmol/l increase in fasting plasma glucose (FPG) or a 2.52 mmol/l increase in the 2-h post-OGTT glucose was associated with a hazard ratio increase in the risk of CV events or death of 1.17 (95% CI 1.13–1.22). Conclusions: In this large multiethnic cohort, the risk of CV events or death increased progressively among individuals who were normoglycaemic, IFG or IGT, and newly diagnosed diabetics. A 1 mmol/l increase in FPG was associated with a 17% increase in the risk of future CV events or death. Therapeutic or behavioural interventions designed to either prevent glucose levels from rising, or lower glucose among individuals with dysglycaemia should be evaluated.