We conducted the first prospective observational study in which we examined the association between incretin responses to an oral glucose tolerance test (OGTT) and mixed meal test (MMT) at baseline and changes in fasting glucose levels 7 years later, in individuals who were non-diabetic at baseline. We used data from the Hoorn Meal Study; a population-based cohort study among 121 subjects, aged 61.0±6.7y. GIP and GLP-1 responses were determined at baseline and expressed as total and incremental area under the curve (tAUC and iAUC). The association between incretin response at baseline and changes in fasting glucose levels was assessed using linear regression. The average change in glucose over 7 years was 0.43 ± 0.5 mmol/l. For GIP, no significant associations were observed with changes in fasting glucose levels. In contrast, participants within the middle and highest tertile of GLP-1 iAUC responses to OGTT had significantly smaller increases (actually decreases) in fasting glucose levels; -0.28 (95% confidence interval: -0.54;-0.01) mmol/l and -0.39 (-0.67;-0.10) mmol/l, respectively, compared to those in the lowest tertile. The same trend was observed for tAUC GLP-1 following OGTT (highest tertile: -0.32 (0.61;-0.04) mmol/l as compared to the lowest tertile). No significant associations were observed for GLP-1 responses following MMT. In conclusion, within our non-diabetic population-based cohort, a low GLP-1 response to OGTT was associated with a steeper increase in fasting glucose levels during 7 years of follow-up. This suggests that a reduced GLP-1 response precedes glucose deterioration and may play a role in the etiology of type 2 diabetes mellitus.