We examined the association between plasma insulin and cardiovascular mortality in non-diabetic European men and women based on data from eleven prospective studies.The study population comprised 6156 men and 5351 women aged 30-89 years. Baseline measurements included oral glucose tolerance test, fasting and 2-h plasma insulin, and conventional risk factors. Cox models were used to calculate hazard ratios (HRs) and their 95% confidence intervals, and overall HRs were assessed by meta-analyses.During the 8.8-year follow-up, 362 men and 70 women died from cardiovascular disease. The age- and smoking-adjusted overall HR of cardiovascular mortality for the highest vs the lower quartiles of fasting insulin was 1.58 (95% CI: 1.26-1.97) in men and 2.64 (1.54-4.51) in women. Adjusting for other risk factors in addition, the HR was 1.54 (1.16-2.03) in men and 2.66 (1.45-4.90) in women. For 2-h insulin these HRs were 1.28 (0.99-1.66), 1.87 (0.87-4.02), and 0.85 (0.60-1.21), 1.36 (0.53-3.45). The overall HRs for interquartile ranges for fasting and 2-h insulin, with full adjustment, were 1.13 (1.05-1.22) and 1.11 (1.01-1.23) in men, and 1.25 (1.08-1.45) and 1.11 (0.91-1.36) in women.Hyperinsulinaemia, defined by the highest quartile cut-off for fasting insulin, was significantly associated with cardiovascular mortality in both men and women independently of other risk factors. Associations between high 2-h insulin and cardiovascular mortality were weaker and non-significant. Weak positive associations of fasting and 2-h insulin with cardiovascular mortality over interquartile ranges were, however, more similar.