Higher Fasting Plasma Glucose Levels,within the Normal Range,are Associated with Decreased Processing Speed in High Functioning Young Elderly

INTRODUCTION

Type 2 diabetes mellitus (T2DM) has been consistently associated with an increased risk of dementia and mild cognitive impairment (MCI) [1, 2]. Cognitivedeficits in non-demented diabetic adults include reduced Processing Speed (PS), impairments in verbal learning, and complex information processing [3, 4]. Circulating glucose levels are a consistent contributor to cognitive decline in T2D patients [5, 6].

Pre-diabetes is a state in which glucose level is higher than normal but not yet high enough to be classified as T2DM (fasting glucose <126 mg/dL but >110 mg/dL) [7]. Individuals diagnosed as pre diabetic have impaired fasting glucose (IFG) or impaired glucose tolerance (IGT) [7], and are at higher risk of developing T2D. Both IFG and IGT are found to be associated with impaired cognitive function [8]in several domains of cognitive function, e.g., verbal memory [1], attentional control, executive function [9], and PS [3].

Higher fasting plasma glucose levels in non-diabetic elderly were associated with an increased risk for incident dementia after a 5-year follow up [5]. However, considering the relatively old age of participants in this study (76 years on average), the possibility of reverse causation cannot be ruled out. Therefore, there is a need to examine this relationship in younger samples.

PS is a basic cognitive measure of mental efficiency which reflects how fast one can execute the mental operations needed to complete cognitive tasks [10]. When deficits in attention occur, some higher cognitive functions may remain intact, yet overall cognitive efficiency decreases, particularly in terms of PS [11]. Therefore, it is not surprising that PS is considered to be the first cognitive function to decline with age among healthy adults [10]. Furthermore, reduced PS was found to predict the progression from MCI to Alzheimer’s disease [12]. Taken together, these evidence suggest that PS is a sensitive cognitive function that enables detection of subtle changes in cognition.

Establishment of associations between Plasma Glucose levels (PGL) within the normal range and PS may enable early detection and preventive intervention in subjects at risk for cognitive decline and dementia. Therefore, the aim of the present study was to investigate whether higher fasting PGL, falling within the normal range as defined by the World Health Organization (WHO), in highly functioning young elderly individuals, free of T2DM, are associated with poorer PS, as measured by a computerized cognitive battery.

MATERIALS AND METHODS

Data was derived from the Sheba Medical Center Executive Screening Survey (ESS), and included senior executives from all parts of Israel, summoned for an annual medical examination. The examination included venous blood samples (following a fast of at least 10 hours) among other tests.

The experimental procedure was approved by the Sheba Medical Center Helsinki Committee and all participants signed an informed consent form.

RESULTS

Descriptive characteristics and Pearson correlation coefficients of all participants are presented in Table 1. A two step hierarchical regression was conducted with RT as the dependent variable (Table 2). PGL was entered at step 1, indicating significant contribution to the regression model, accounting for 16.3% of the variation in RT. ‘Age’ and ‘Body Mass Index’ (‘BMI’) were forcefully entered as potential confounding variables in step 2, indicating insignificant contribution to the model, while PGL remained a significant predictor for RT.

DISCUSSION

Findings showed a negative association between PGL and PS, indicating that higher fasting PGL within the normal range (<110 mg/dl), were associated with a decrease in PS. Our findings are in line with previous studies that have reported deficits in measurements of PS in subclinical [3, 16] and T2DM [3, 4] subjects. Three main factors make our results notable; First, as defined by the WHO’s cut off level (2013) for impaired fasting hyperglycemia (110 mg/dL) the research sample of relatively young elderly (M = 64.7 years) was all in range of normal PGL. In order to avoid the inclusion of participants with IFG or IGT; individuals with T2DM or pre-T2DM were excluded. Second, in the present study we used an index of PS from a computerized battery, found to be more accurate than paper and pencil tests [17] and more effective in detecting subtle changes in cognitive functions, especially among highly functioning individuals. Third, the study was performed among young elderly, where the risk of incipient dementia is minimal. It is important to note that although we used a composite score of all four CogState tests each CogState task alone was found significantly associated with glucose so none of the tests specifically influenced the association of PGL with cognition.

Normal aging, free of neurodegenerative disease, is associated with changes in brain structure and cognitive function. Evidence of decline in memory and verbal abilities were discovered among healthy individuals under the age of 60 [18, 19]. This is probably the result of a long-term subclinical pathological process that extends over several decades [20]. Such findings can have important implication for possible interventions at an early stage as to prevent or postpone cognitive decline later in life.

A study recently conducted by Nilsson et al. [6] indicates that maintaining a balanced diet influences glucose level and may influence cognitive performance, in healthy non-diabetic adults. It shows that a four-week diet including foods with anti-inflammatory potential action, improved cardiovascular risk variables (e.g., BMI, blood pressure, glucose level) and resulted in favorable effects on cognitive functions – attention abilities, working memory. These results illustrate the potential beneficial effects of early detection of high normal PGL in healthy non-diabetic adults in preventing and decreasing cognitive performance.

A few limitations are mentioned. First, future longitudinal studies should combine several glucose measurements in a larger and more representative sample. Second, cross-sectional analysis does not permit making temporal or causal inferences. Third, the number of days between measuring glucose level and cognitive test performance differed among subjects (mean = 80, SD = 80). Nonetheless, PGL in non-diabetic adults tends to remain relatively stable in the short term, and is usually 81 to 100 mg/dl with a deviation of 10–15% [21]. In any case we performed a linear regression controlling for this potential confounder and results remained essentially unchanged (r = 0.46, p <  0.01). Third, we do not have available information on the level of the participants’ education, a parameter which was found highly and consistently correlated with cognition [22]. However, all our participants hold executive positions in the Israeli industry thus constituting a relatively homogeneous group in terms of educational attainment. Finally, although a widely accepted WHO cut-off level for normal PGL (<110 mg/dl) was used [7], it must be noted that the cut-off for normality according to the American Diabetes Association (ADA), is stricter and stands on <100 mg/dl for normal fasting PGL [23]. We repeated the analyses using the ADA’s criteria and the effect size remained quite similar (r = 0.222) although the statistical significance was lost due to the smaller sample size (pv = 0.13).

Despite these limitations, the results of this study show a consistent pattern of association between high fasting PGL within the normal range in a young elderly population and decreased PS. This is further demonstrated by the robust results after controlling for additional potential confounders (mean arterial pressure [r = 0.38, pv = 0.007] and total cholesterol [r = 0.42, pv = 0.004]). These findings suggest that even in the subclinical range and in the absence of diabetes, monitoring and managing PGL could have an impact on cognitive function. Furthermore, the present findings suggest the need to reconsider the actual threshold for defining a pre-diabetic condition. Adopting a stricter cut off level of <100 mg/dl for normal PGL, rather than the WHOs <110 mg/dl cut off should be considered.

DISCLOSURE STATEMENT

Authors’ disclosures available online (http://j-alz.com/manuscript-disclosures/15-0433r2).