Background: Elevated fasting blood glucose(FBG) levels have been associated with poorer cognitive function. High FBG variability has also been associated with accelerated cognitive decline in late life. However, the impact of long term FBG variability from childhood on midlife cognitive function (CF) remains understudied. Hypothesis: We hypothesized that there would be an association of higher long-term FBG variability from childhood with poorer midlife CF measures. Aim: We aimed to identify the influence of higher long term FBG variability on midlife CF. Methods: We studied 1,005 midlife Bogalusa Heart Study participants (age at baseline: mean 9.4 years +/- SD 3.5 years; age at neuropsychological(NP) exam: mean 47.9 years+/- SD 5.2 years; 60.4% women(607 of 1005); 33.5 % Black(337 of 1005)) with ≥3 FBG measurements across follow up. Participants having type 2 diabetes mellitus(T2DM) were excluded. T2DM was defined as FBG ≥126 mg/dL or taking diabetes medication. Long-term visit-to visit FBG variability was measured as deviation from age predicted values (DEV) and residual SD (RSD). Using ≥3 FBG measurements, DEV and RSD were obtained from individual growth curves created by fitting random-effects models stratified by race and sex. Participants’ midlife CF was first measured with 10 NP tests, then were classified into 3 distinct NP profiles (optimal, average, and mixed-low) using cluster analysis. Multinomial logistic regression models were used to analyze the associations between long-term FBG variability parameters and NP profiles, adjusting for education status and mean of FBG measures from childhood to midlife . DEV and RSD were standardized to z-scores by race and sex before regression analysis. Tests for interaction according to race and sex were conducted. Results: After adjusting for the mean FBG level and education status, for a 1 SD increase in DEV, the odds of having a mixed-low CF increased by 57.3% (95% CI 1.244, 1.990) and the odds of an average CF increased by 36.9% (95% CI 1.144, 1.640) respectively, compared to optimal CF. After adjusting for the mean FBG level and education status, for a 1 SD increase in RSD, the odds of having a mixed-low CF increased by 56% (95% CI 1.232, 1.974) and the odds of an average CF increased by 37.7% (95% CI 1.147, 1.653), respectively, compared to optimal CF. No significant interactions between long-term FBG variability parameters and race and sex were found. Conclusions: Higher long-term FBG variability was associated with poorer midlife CF. Further studies are needed to identify the role of long-term FBG variability in the association of cardiometabolic risk factors and midlife CF.