Aims: The relationship between 1- and 2-h glucose levels following an oral glucose tolerance test (OGTT) and long-term mortality was evaluated. Methods: Over a 33-year period, 2138 individuals were followed for all-cause mortality. Fasting and post-OGTT glucose parameters categorized the cohort according to baseline glycaemic status. Four categories were established according to 1- and 2-h glucose levels (in mmol/l): group A = 1 h ≤ 8.8 and 2 h < 7.8; group B = 1 h > 8.6 and 2 h < 7.8; group C = 1 h ≤ 8.6 and 2 h = 7.8–11.1 (impaired glucose tolerance); group D = 1 h > 8.6 and 2 h = 7.8–11.1 (impaired glucose tolerance). Individuals with diabetes at baseline were excluded from the cohort. Results: By August 2013, 51% of the study cohort had died. The worst prognosis occurred in group D (73.8% mortality), followed by groups C (67.5%), B and A (57.9% and 41.6%, respectively). When the 2-h glucose value is ‘normal’ (< 7.8 mmol/l), the 1-h glucose value > 8.6 mmol/l is an important predictor of mortality (28% increased risk) compared with group A, controlling for sex, age, smoking, BMI, systolic and diastolic blood pressures. A gradual increased hazard for mortality was seen by study group (hazard ratio = 1.28, 1.60 and 1.76, for groups B, C and D, respectively; group A = reference). Conclusions: A 1-h glucose value > 8.6 mmol/l predicts mortality even when the 2-h level is < 7.8 mmol/l. However, when the 2-h level is in the impaired glucose tolerance range, the hazard for mortality rises significantly independent of the 1-h value. Individuals at risk for developing diabetes could be identified earlier using the 1-h threshold value of 8.6 mmol/l, which could avert progression to diabetes and increased mortality..