Fasting plasma glucose in young adults free of diabetes is associated with cognitive function in midlife

Fasting plasma glucose in young adults free of diabetes is associated with cognitive function in... Abstract Background Evidence for an association of fasting plasma glucose (FPG) with cognitive function in adults free of diabetes is scarce and based on middle-aged and older adults. We examined the association of FPG, measured at age 30, and of change in FPG from age 30 to 43, with cognitive function at age 50. Methods 505 nondiabetic participants of the population-based Jerusalem Lipid Research Clinic (LRC) cohort study had baseline FPG, 2-h post-oral challenge plasma glucose (OGTT) and insulin determined at ages 28–32, and FPG and OGTT again at ages 41–46. Subsequently at ages 48–52, global cognitive function and its five specific component domains were assessed with a NeuroTrax computerized test battery, using multiple linear regression and multivariable logistic models. Results Hyperglycemia (FPG ≥ 5.6 mmol/l vs. <5.6 mmol/l) at baseline was associated with poorer global cognitive function in midlife (predominantly in the visual spatial and attention domains), independent of socio-demographic characteristics, life style variables, body mass index (BMI), and inflammatory and biochemical variables (standardized Beta = −0.121, P = 0.002, plinear trend(FPG continuous) =0.016). Similarly, increased odds for low-ranked (lowest fifth) global cognition was evident (ORper mmol/l FPG=2.31, 95% CI = 1.30–4.13, P = 0.005). Baseline OGTT, insulin resistance (HOMA-IR) and change in FPG and OGTT over 13 years were not associated with cognition. Conclusion A higher FPG in young adults was associated with lower cognitive performance in midlife. Although we cannot dismiss the possibility of reverse causation, hyperglycemia at a young age may be a modifiable risk factor for low-ranked cognitive function in midlife. Introduction It is well established that type 2 diabetes is associated with an increased risk of incident dementia1 as well as with cognitive decline2 and poor cognitive function.3 Similarly, hyperglycemia and insulin resistance without clinically diagnosed diabetes have been related to poorer cognitive function in late-life.4–6 Nevertheless, data relating impaired glucose regulation to cognitive function in younger adults are scarce,7–10 are all based on cross-sectional assessment of glucose and cognition, and are based on middle-aged samples7,9,10 and type 2 diabetes patients7,8 or a small sample of convenience.11 We are unaware of reports relating either glucose levels or longitudinal change in plasma glucose with cognitive assessment in non-diabetic young adults. Accordingly, we aimed to evaluate the relationship of fasting plasma glucose (FPG), 2 h post-oral challenge plasma glucose (OGTT) and insulin resistance (HOMA-IR) measured at ages 28–32, as well as change in FPG and OGTT over a 13-year follow-up to ages 41–46, with global cognitive function, the primary outcome variable, in a population-based cohort aged 48–52. The specific multi-domain components that contribute to global function served as secondary outcomes. Methods Study population The study sample was derived from the Jerusalem Lipid Research Clinic (LRC) Study, a longitudinal, population-based cohort initiated in 1976–1978. Details regarding sampling and response rates have been reported.12 A flow diagram for Visits 1–6 from 1976 through 2011 is shown in figure 1. In brief, in 1976–1978 the Jerusalem LRC examined 8646 17-year-old Jewish residents of Jerusalem, representing full age cohorts (Visit 1), as part of a compulsory examination to determine their fitness for military service (irrespective of actual conscription). In 1989–1991 (Visit 4, at age 28–32, the baseline of the current study), 2 overlapping samples of Visit 1 were re-invited: a sex-stratified random sample consisting of 884 young adults who met the eligibility requirements (71% response), and offspring of parents (an additional 168 individuals) who had a documented acute myocardial infarction or sudden cardiac death over a 10 years follow-up period (65% response). A total of 1052 eligible subjects (70% response) were examined and interviewed at age 28–32y.13 Figure 1 View largeDownload slide LRC study design flow chart. *2 overlapping samples of Visit 1 were re-invited to Visit 4: a sex-stratified random subsample sample comprising 884 young adults with a male to female sampling ratio of 1.6: 1 (71% response), and offspring of parents (an additional 168 individuals) who had a documented acute myocardial infarction or sudden cardiac death over a 10y follow-up period (65% response). The sampling scheme was incorporated in the analysis—and did not affect the results. †Exclusion of ineligible participants who were not current Jerusalem residents, were pregnant or were within 3 months of delivery, had a serious incapacitating illness or had died Figure 1 View largeDownload slide LRC study design flow chart. *2 overlapping samples of Visit 1 were re-invited to Visit 4: a sex-stratified random subsample sample comprising 884 young adults with a male to female sampling ratio of 1.6: 1 (71% response), and offspring of parents (an additional 168 individuals) who had a documented acute myocardial infarction or sudden cardiac death over a 10y follow-up period (65% response). The sampling scheme was incorporated in the analysis—and did not affect the results. †Exclusion of ineligible participants who were not current Jerusalem residents, were pregnant or were within 3 months of delivery, had a serious incapacitating illness or had died In 2003–2006 (Visit 5, age 41–46), 631 of the Visit 4 participants were re-examined14 [71% response rate, after exclusion of 168 ineligible participants who were largely not current Jerusalem residents, were pregnant or were within 3 months of delivery, had a serious incapacitating illness or had died]. In 2009–11 (Visit 6, age 48–52),15–17 507 of the 631 Visit 5 participants (82% response rate among those eligible) underwent cognitive function assessment. We were unable to gain access to and thus adjust for a routine baseline cognitive measure that was taken as part of the compulsory military examination at age 17. FPG, OGTT, insulin, and insulin resistance were determined at Visit 4 (mean age 30 years); FPG and OGTT were repeated at Visit 5. The current analysis was undertaken on 505 Visit 6 non-diabetic individuals after excluding 2 patients with diabetes at baseline (0.4%). A comparison of Visit 1 characteristics between the Visit 6 participants (n = 505) and nonparticipants initially examined at Visit 1 (n = 8139) showed that the sex-specific distributions of body mass index (BMI), country of origin, education and plasma lipids and lipoproteins were very similar (Supplementary table S1). We also compared the Visit 6 participants who had a parent with a documented history of coronary heart disease (CHD) vs. those who did not for their sex-specific BMI, sociodemographic characteristics and lipid profile determined at Visit 4 to assess the effect of the v4 sampling scheme on generalizability (not shown). Other than cholesterol and triglyceride values that were higher in male participants and lower in female participants with parental CHD, distributions were similar. Ethical approval was obtained from the Hadassah Medical Center Helsinki Committee. Signed informed consent was obtained from all individual participants included in the study. Assessment of cognitive function Cognitive functions were assessed through a battery of NeuroTrax computerized cognitive tests (NeuroTrax Corp., Modiin, Israel). The battery was designed to evaluate performance in about 30 min (0: 32 ± 0: 04 h in our study) across an array of cognitive domains known to deteriorate during aging (including memory, executive function, visual spatial processing, attention, and information processing speed). It provides measurements of accuracy and response time in milliseconds and has been shown to be valid18,19 and reliable.18 Several of these tests are based on common neuropsychological paradigms [including the Benton Visual Retention Test, Brief Visuospatial Memory Test, Tova, Stroop, and subsets of WAIS-III (Wechsler Adult Intelligence Scale, 3rd ed.)] and have been previously used in clinical settings, as well as in studies of normal aging relating these neuropsychological measurements to genetic findings20 and brain imaging parameters,21 and have been used in middle-aged adults.20–24 A detailed description of the tests can be found in the online Supplementary data S1 and table S2, comparing individuals with hyperglycemia vs. those with normal glucose. Assessment of plasma glucose levels Plasma glucose was determined on both 12 h fasting samples and 2 h after a standard 75 g glucose challenge at Visit 4 (mean age 30 years) and Visit 5 (mean age 43 years) by standard enzymatic techniques. Diabetes at baseline (Visit 4), defined as the use of insulin or oral medications or a FPG ≥7.0 mmol/l (≥ 126 mg/dl) or OGTT ≥ 11.1 mmol/l (≥ 200 mg/dl), led to the exclusion of 2 participants. Hyperglycaemia was defined as FPG ≥5.6 mmol/l (≥ 100 mg/dl and <126 mg/dl). Assessment of insulin resistance 12-h fasting insulin was measured by radioimmunoassay. Insulin resistance at Visit 4 was estimated using the HOMA-IR. Assessment of covariates Socio-demographic characteristics consisted of age, sex, origin (father’s country of birth grouped by continent: Europe, Asia, North Africa, and Israel), highest educational attainment (university degree, high school graduate, incomplete high school, elementary), religiosity (ultraorthodox, religious, traditional, secular), childhood socio-economic position (SEP) (2 measures based on father’s occupation modified from the Israel Central Bureau of Statistics (ICBS)12 and the Vered Kraus occupational prestige based scale,12 both obtained at Visit 1, and adult SEP (2 measures based on the modified ICBS ranking and the MacArthur Scale of Subjective Social Status25). Depressive and anxiety symptoms were measured at Visit 6 using a translated Hebrew version of the Hospital Anxiety and Depression Scale (HADS) (two 7-item independent subscales).26 Measures of BMI (in kg/m2), blood pressure, health behaviors and biochemistry were obtained at baseline. Blood pressure was taken as the mean of the last 2 of 3 seated measurements using a standard mercury sphygmomanometer after 5 min of quiet rest. Health behaviors consisted of cigarette smoking (lifetime pack-years), alcohol intake (≤ once a week, and a median split of number of units per week as two dummy variables of low and high intake), and vigorous physical leisure-time activity for at least 20 min causing sweating and shortness of breath (yes/no). Plasma total cholesterol, HDL-C and triglycerides were measured on 12-h fasted samples by standard enzymatic techniques at Visits 1, 3, 4 and 5 at mean ages 17, 21, 30 and 43 years. LDL-C was computed by the Friedewald method. Inflammation markers assessed at baseline were plasma concentrations of C-reactive protein (CRP) (by ELISA), fibrinogen (Clauss method), and the white blood cell count (Beckman Coulter Counter). Plasma GlycA, a novel protein glycan inflammatory biomarker (determined by NMR), was quantified in both baseline (mean age 30 years) and follow-up (mean age 43 years) samples.15,27 Serum homocysteine was determined at baseline using HPLC with fluorometric detection. Statistical analysis Raw cognitive outcome measures (i.e. response time, accuracy and composite scores) were z-standardized to permit averaging across the different scales of measurement. Timed measures (response time and response time SD) were multiplied by −1 so that higher values indicate better performance. The z-standardized measures were then averaged to produce five scores, each indexing a different cognitive domain: memory, attention, executive function, visual spatial, and information processing speed. A summary global cognitive score, computed as the average of the 5 domain scores, was treated as the main dependent variable. Cognitive scores with negatively skewed distributions (global, memory and attention) were Box-Cox power transformed (λ = −0.5, i.e. inverse square-root transformation), to achieve an approximately Gaussian distribution, subsequent to reflection (computed by subtracting each value of a negatively skewed score from a constant). FPG at baseline was normally distributed. FPG was treated categorically as FPG ≥5.6 mmol/l (5.6–6.9 mmol/l) vs. FPG <5.6 mmol/l and as a continuous variable to test for trend. Secondarily, we also grouped FPG as <4.4, 4.4–4.9, 5.0–5.5, and ≥5.6. Multiple linear regression models were used to examine the associations between each glucose and insulin resistance measure (independent variables) and global cognitive function, the primary outcome. These models were repeated separately for each of the 5 cognitive domains to assess to which component(s) the association with global function can be attributed (secondary outcomes). Regression coefficients are reported as standardized Betas. Odds ratios (ORs) and 95% confidence intervals (CI) for the association of glucose and insulin resistance with poor cognitive function were computed from logistic models. A score in the lowest quintile was regarded as relatively poor cognitive performance, as has been previously defined by others,28 and was compared with the upper 4 quintiles grouped. Nominal two-sided p-values are reported. Sex interaction with hyperglycemia on the main global cognitive function outcome, tested in a separate regression model using a multiplicative term, showed no evidence of interaction (P = 0.77); consequently, we did not further stratify the statistical models by sex. Sensitivity analyses were performed on global cognitive function with fasting plasma glucose restricted to <6.1 mmol/l. In all analyses, we accounted for the sampling scheme at Visit 4 by the introduction of a dichotomous (0, 1) term representing the random sample vs. the parental CHD sample, despite that its incorporation did not affect the associations (not shown). Additional stratified analysis comparing cases with a documented positive family history of CHD vs. those without parental CHD, showed stronger associations in the former group of cases; however, the parental CHD interactions were not statistically significant (all P > 0.7). Consequently, all analyses were based on the full cohort and were adjusted for the sampling scheme. Analyses were adjusted for age, sex, educational level, origin, religiosity, SEP in childhood (ICBS ranking), adult SEP (ICBS ranking) and cigarette smoking, leisure-time vigorous activity, alcohol intake, BMI, total cholesterol, HDL-cholesterol and GlycA measured at baseline. These were selected on the basis of their confounding effect size on the Beta coefficient of the bivariate baseline FPG-global cognition association (≥5%). Depression, anxiety, systolic and diastolic blood pressure, LDL-cholesterol, triglycerides, C-reactive protein, white blood cell count, fibrinogen and homocysteine showed no material confounding effect and were not included. In a sensitivity analysis using backward stepwise regression, we reduced the selected covariate set to sex, educational level, adult SEP and BMI, as the excluded group of covariates did not affect the Beta coefficient of the full multivariable-adjusted model (i.e. the effect size for the excluded group was below 5%). The association of FPG with global cognitive function was also analysed in a sensitivity analysis to examine the effect of controlling for waist-to-hip ratio rather than BMI. To avoid loss of observations in the multivariable analyses, missing values were replaced with non-missing median values (adult SEP, n = 5; leisure-time vigorous activity, n = 1) or for GlycA (n = 20) by imputation using a backward stepwise linear regression procedure applied to variables predicting GlycA (P for removal >0.2). A complete case analysis slightly attenuated the associations. Statistical analyses were carried out using SPSS v21.0 (IBM Corp., Armonk, NY). Power Setting an α value of 0.05 (2-tailed), this sample had a power of 0.9 to detect a correlation of 0.15 between FPG and cognition, and given a 0.16 prevalence of hyperglycemia, it had a power of 0.8 to detect an OR of 2.1 in a logistic model predicting poor cognitive function (lowest fifth) and a difference of about a 1/3 SD in the global cognitive score in a linear regression model. Results Characteristics of the study sample at baseline are presented in table 1. Participants were aged 28–32 at baseline and 41–46 at follow-up with a range of 12–16 years of follow-up (13.1 ± .7); 33% were women, and 54% were high school or university graduates. Mean BMI was high and mean HDL-cholesterol was low as has been reported.14 Alcohol intake was low as was leisure time vigorous activity and about one-third of the sample smoked. Table 1 Characteristics of the study sample at baseline: the Jerusalem LRC longitudinal study, 1976–2011 Characteristics Total n 505 Socio-demographic variables Age (years) (range) 30.1± 0.8a (28.1–32.1) Female (%) 32.5 Country of birth (%)     Israel 22.4     Europe 22.6     Asia 29.5     N. Africa 25.5 Religiosity (%)     Ultra-orthodox 7.7     Religious 19.2     Traditional 31.9     Secular 41.2 Education, highest level (%)     University graduate 32.5     High school graduate 21.6     High school not graduated (9–12 years) 40.2     Elementary school (≤ 8 years) 5.7 Adult SEP (ICBS ranking)b 2.7 ± 1.2 Adult SEP (MacArthur Scale)c 7.2 ± 1.5 Childhood SEP (ICBS ranking)b 3.8 ± 1.6 Childhood SEP (Vered Kraus Scale)c 44.5 ± 30.1 Social mobilityd 1.1 ± 1.6 Glucose regulation FPG (mmol/l) (range) [median, IQR] 5.1 ± 0.5 (3.6–6.8) [5.1, 4.8–5.4]     FPG change (mmol/l)e 0.4 ± 0.8 OGTT (mmol/l) 5.0 ± 1.2     Change in OGTT (mmol/l)e 0.7 ± 1.7 Hyperglycemia (%)f 16.2     Fasting plasma insulin (mlU/l) 17.0 ± 2.0 Insulin resistance     HOMA-IR (mlU/l*mmol/l)g 3.9 ± 1.9 Anthropometric and blood pressure     BMI (kg/m2) 24.7 ± 3.6     Systolic blood pressure (mmHg) 112 ± 10     Diastolic blood pressure (mmHg) 68 ± 9 Psychosocial variables     Depressive symptoms score (HADS 0–21)h 3.6 ± 2.9     Anxiety symptoms score (HADS 0–21)h 5.6 ± 3.7 Lifestyle variables at baseline Leisure-time vigorous activity (%)i 21.6 Alcohol intake of ≥once/week (%) 39.4     Low intake (units/week)j 1.4 ± 0.5     High intake (units/week)j 5.8 ± 2.8 Pack-years (whole sample) 4.5 ± 6.5     Among ever smoked 9.5 ± 6.5 % current smokers 36.3 Biochemistry Plasma Lipids (mmol/l)     Total cholesterol 4.4 ± 0.8     HDL-cholesterol 1.0 ± 0.3     Non-HDL-cholesterol 3.4 ± 0.9     LDL-cholesterolk 2.7 ± 0.7     Triglycerides 1.4 ± 0.9 Homocysteine (µmol/l) 12.2 ± 8.5 Inflammatory variables     C-reactive protein (mg/l) 2.2 ± 3.3     White blood cell count 6841 ± 1694     Fibrinogen (mg/dl) 233.5 ± 54.8     GlycA (µmol/l) 265 ± 52 Characteristics Total n 505 Socio-demographic variables Age (years) (range) 30.1± 0.8a (28.1–32.1) Female (%) 32.5 Country of birth (%)     Israel 22.4     Europe 22.6     Asia 29.5     N. Africa 25.5 Religiosity (%)     Ultra-orthodox 7.7     Religious 19.2     Traditional 31.9     Secular 41.2 Education, highest level (%)     University graduate 32.5     High school graduate 21.6     High school not graduated (9–12 years) 40.2     Elementary school (≤ 8 years) 5.7 Adult SEP (ICBS ranking)b 2.7 ± 1.2 Adult SEP (MacArthur Scale)c 7.2 ± 1.5 Childhood SEP (ICBS ranking)b 3.8 ± 1.6 Childhood SEP (Vered Kraus Scale)c 44.5 ± 30.1 Social mobilityd 1.1 ± 1.6 Glucose regulation FPG (mmol/l) (range) [median, IQR] 5.1 ± 0.5 (3.6–6.8) [5.1, 4.8–5.4]     FPG change (mmol/l)e 0.4 ± 0.8 OGTT (mmol/l) 5.0 ± 1.2     Change in OGTT (mmol/l)e 0.7 ± 1.7 Hyperglycemia (%)f 16.2     Fasting plasma insulin (mlU/l) 17.0 ± 2.0 Insulin resistance     HOMA-IR (mlU/l*mmol/l)g 3.9 ± 1.9 Anthropometric and blood pressure     BMI (kg/m2) 24.7 ± 3.6     Systolic blood pressure (mmHg) 112 ± 10     Diastolic blood pressure (mmHg) 68 ± 9 Psychosocial variables     Depressive symptoms score (HADS 0–21)h 3.6 ± 2.9     Anxiety symptoms score (HADS 0–21)h 5.6 ± 3.7 Lifestyle variables at baseline Leisure-time vigorous activity (%)i 21.6 Alcohol intake of ≥once/week (%) 39.4     Low intake (units/week)j 1.4 ± 0.5     High intake (units/week)j 5.8 ± 2.8 Pack-years (whole sample) 4.5 ± 6.5     Among ever smoked 9.5 ± 6.5 % current smokers 36.3 Biochemistry Plasma Lipids (mmol/l)     Total cholesterol 4.4 ± 0.8     HDL-cholesterol 1.0 ± 0.3     Non-HDL-cholesterol 3.4 ± 0.9     LDL-cholesterolk 2.7 ± 0.7     Triglycerides 1.4 ± 0.9 Homocysteine (µmol/l) 12.2 ± 8.5 Inflammatory variables     C-reactive protein (mg/l) 2.2 ± 3.3     White blood cell count 6841 ± 1694     Fibrinogen (mg/dl) 233.5 ± 54.8     GlycA (µmol/l) 265 ± 52 Missing data: adult SEP (ICBS ranking) (n = 5), adult SEP (MacArthur Scale) (n = 11), childhood SEP (Vered Kraus Scale) (n = 4), social mobility (n = 5), hyperglycemia (n = 7), HOMA-IR (n = 1), depressive symptoms score (n = 4), anxiety symptoms score (n = 4), leisure-time vigorous activity (n = 1), LDL-cholesterol (n = 7), homocysteine (n = 16), C-reactive protein (n = 14), white blood cell count (n = 12), fibrinogen (n = 43), GlycA (n = 20). LRC, Lipid Research Clinic; FPG, fasting plasma glucose; OGTT, 2 h post-oral challenge plasma glucose; HOMA-IR, homeostatic model assessment for insulin resistance; BMI, body mass index. a Mean ± SD (continuous/ interval variables). b An higher value infers a lower SEP. Scale range from 1 to 6. c An higher value infers a higher SEP. Our Vered Kraus scores range from 2.60 to 98.96; MacArthur Scale range from 1 to 10. d Computed by subtracting ICBS-based SEP in adulthood from SEP in childhood (both with a range from 1 (upper) to 6 (lower)). Range of social mobility score was from −5 (maximal downward drifting) to +5 (maximal upward mobility). No change/upward mobility corresponds to scores ≥0, whereas downward drifting corresponds to scores <0. e Computed by subtracting baseline measurement at ages 28–32 from follow-up measurement at ages 41–46. f Defined as FPG ≥100 mg/dl. g Calculated as the product of fasting serum glucose (mmol/l) x fasting serum insulin (mlU/l) divided by 22.5. (Diabetes Care. 2004; 27: 1487). h 7-items each scored 0–3. Scale range from 0 to 21. Cronbach's alphas were adequate at .71 and .785 for the depression and the anxiety subscale, respectively. i Exercise for at least 20 min causing heavy breathing and sweating. j Low/high intake, according to median split of alcohol intake among consumers of ≥once/week. k Computed by the Friedewald method; not computed for 7 men at age 30 and 11 participants (10 men and 1 woman) at age 43 with triglycerides > 400 mg/dl. Table 1 Characteristics of the study sample at baseline: the Jerusalem LRC longitudinal study, 1976–2011 Characteristics Total n 505 Socio-demographic variables Age (years) (range) 30.1± 0.8a (28.1–32.1) Female (%) 32.5 Country of birth (%)     Israel 22.4     Europe 22.6     Asia 29.5     N. Africa 25.5 Religiosity (%)     Ultra-orthodox 7.7     Religious 19.2     Traditional 31.9     Secular 41.2 Education, highest level (%)     University graduate 32.5     High school graduate 21.6     High school not graduated (9–12 years) 40.2     Elementary school (≤ 8 years) 5.7 Adult SEP (ICBS ranking)b 2.7 ± 1.2 Adult SEP (MacArthur Scale)c 7.2 ± 1.5 Childhood SEP (ICBS ranking)b 3.8 ± 1.6 Childhood SEP (Vered Kraus Scale)c 44.5 ± 30.1 Social mobilityd 1.1 ± 1.6 Glucose regulation FPG (mmol/l) (range) [median, IQR] 5.1 ± 0.5 (3.6–6.8) [5.1, 4.8–5.4]     FPG change (mmol/l)e 0.4 ± 0.8 OGTT (mmol/l) 5.0 ± 1.2     Change in OGTT (mmol/l)e 0.7 ± 1.7 Hyperglycemia (%)f 16.2     Fasting plasma insulin (mlU/l) 17.0 ± 2.0 Insulin resistance     HOMA-IR (mlU/l*mmol/l)g 3.9 ± 1.9 Anthropometric and blood pressure     BMI (kg/m2) 24.7 ± 3.6     Systolic blood pressure (mmHg) 112 ± 10     Diastolic blood pressure (mmHg) 68 ± 9 Psychosocial variables     Depressive symptoms score (HADS 0–21)h 3.6 ± 2.9     Anxiety symptoms score (HADS 0–21)h 5.6 ± 3.7 Lifestyle variables at baseline Leisure-time vigorous activity (%)i 21.6 Alcohol intake of ≥once/week (%) 39.4     Low intake (units/week)j 1.4 ± 0.5     High intake (units/week)j 5.8 ± 2.8 Pack-years (whole sample) 4.5 ± 6.5     Among ever smoked 9.5 ± 6.5 % current smokers 36.3 Biochemistry Plasma Lipids (mmol/l)     Total cholesterol 4.4 ± 0.8     HDL-cholesterol 1.0 ± 0.3     Non-HDL-cholesterol 3.4 ± 0.9     LDL-cholesterolk 2.7 ± 0.7     Triglycerides 1.4 ± 0.9 Homocysteine (µmol/l) 12.2 ± 8.5 Inflammatory variables     C-reactive protein (mg/l) 2.2 ± 3.3     White blood cell count 6841 ± 1694     Fibrinogen (mg/dl) 233.5 ± 54.8     GlycA (µmol/l) 265 ± 52 Characteristics Total n 505 Socio-demographic variables Age (years) (range) 30.1± 0.8a (28.1–32.1) Female (%) 32.5 Country of birth (%)     Israel 22.4     Europe 22.6     Asia 29.5     N. Africa 25.5 Religiosity (%)     Ultra-orthodox 7.7     Religious 19.2     Traditional 31.9     Secular 41.2 Education, highest level (%)     University graduate 32.5     High school graduate 21.6     High school not graduated (9–12 years) 40.2     Elementary school (≤ 8 years) 5.7 Adult SEP (ICBS ranking)b 2.7 ± 1.2 Adult SEP (MacArthur Scale)c 7.2 ± 1.5 Childhood SEP (ICBS ranking)b 3.8 ± 1.6 Childhood SEP (Vered Kraus Scale)c 44.5 ± 30.1 Social mobilityd 1.1 ± 1.6 Glucose regulation FPG (mmol/l) (range) [median, IQR] 5.1 ± 0.5 (3.6–6.8) [5.1, 4.8–5.4]     FPG change (mmol/l)e 0.4 ± 0.8 OGTT (mmol/l) 5.0 ± 1.2     Change in OGTT (mmol/l)e 0.7 ± 1.7 Hyperglycemia (%)f 16.2     Fasting plasma insulin (mlU/l) 17.0 ± 2.0 Insulin resistance     HOMA-IR (mlU/l*mmol/l)g 3.9 ± 1.9 Anthropometric and blood pressure     BMI (kg/m2) 24.7 ± 3.6     Systolic blood pressure (mmHg) 112 ± 10     Diastolic blood pressure (mmHg) 68 ± 9 Psychosocial variables     Depressive symptoms score (HADS 0–21)h 3.6 ± 2.9     Anxiety symptoms score (HADS 0–21)h 5.6 ± 3.7 Lifestyle variables at baseline Leisure-time vigorous activity (%)i 21.6 Alcohol intake of ≥once/week (%) 39.4     Low intake (units/week)j 1.4 ± 0.5     High intake (units/week)j 5.8 ± 2.8 Pack-years (whole sample) 4.5 ± 6.5     Among ever smoked 9.5 ± 6.5 % current smokers 36.3 Biochemistry Plasma Lipids (mmol/l)     Total cholesterol 4.4 ± 0.8     HDL-cholesterol 1.0 ± 0.3     Non-HDL-cholesterol 3.4 ± 0.9     LDL-cholesterolk 2.7 ± 0.7     Triglycerides 1.4 ± 0.9 Homocysteine (µmol/l) 12.2 ± 8.5 Inflammatory variables     C-reactive protein (mg/l) 2.2 ± 3.3     White blood cell count 6841 ± 1694     Fibrinogen (mg/dl) 233.5 ± 54.8     GlycA (µmol/l) 265 ± 52 Missing data: adult SEP (ICBS ranking) (n = 5), adult SEP (MacArthur Scale) (n = 11), childhood SEP (Vered Kraus Scale) (n = 4), social mobility (n = 5), hyperglycemia (n = 7), HOMA-IR (n = 1), depressive symptoms score (n = 4), anxiety symptoms score (n = 4), leisure-time vigorous activity (n = 1), LDL-cholesterol (n = 7), homocysteine (n = 16), C-reactive protein (n = 14), white blood cell count (n = 12), fibrinogen (n = 43), GlycA (n = 20). LRC, Lipid Research Clinic; FPG, fasting plasma glucose; OGTT, 2 h post-oral challenge plasma glucose; HOMA-IR, homeostatic model assessment for insulin resistance; BMI, body mass index. a Mean ± SD (continuous/ interval variables). b An higher value infers a lower SEP. Scale range from 1 to 6. c An higher value infers a higher SEP. Our Vered Kraus scores range from 2.60 to 98.96; MacArthur Scale range from 1 to 10. d Computed by subtracting ICBS-based SEP in adulthood from SEP in childhood (both with a range from 1 (upper) to 6 (lower)). Range of social mobility score was from −5 (maximal downward drifting) to +5 (maximal upward mobility). No change/upward mobility corresponds to scores ≥0, whereas downward drifting corresponds to scores <0. e Computed by subtracting baseline measurement at ages 28–32 from follow-up measurement at ages 41–46. f Defined as FPG ≥100 mg/dl. g Calculated as the product of fasting serum glucose (mmol/l) x fasting serum insulin (mlU/l) divided by 22.5. (Diabetes Care. 2004; 27: 1487). h 7-items each scored 0–3. Scale range from 0 to 21. Cronbach's alphas were adequate at .71 and .785 for the depression and the anxiety subscale, respectively. i Exercise for at least 20 min causing heavy breathing and sweating. j Low/high intake, according to median split of alcohol intake among consumers of ≥once/week. k Computed by the Friedewald method; not computed for 7 men at age 30 and 11 participants (10 men and 1 woman) at age 43 with triglycerides > 400 mg/dl. Median values for FPG and post-challenge glucose at baseline (mean age 30 years) were 5.1 [interquartile range (IQR), 4.8–5.4) mmol/l] and 4.8 (IQR, 4.2–5.6) mmol/l, respectively, and increased over the 13-year follow-up period (mean age 43 years) to 5.4 (IQR, 5.1–5.7) mmol/l and 5.4 (IQR, 4.7–6.3) mmol/l, respectively. Sex-adjusted Spearman tracking correlations over the 13 year average follow-up were rho=.397 and rho=.311 for FPG and OGTT levels, respectively. 16.2% (n = 82) participants had hyperglycemia defined as FPG ≥5.6 mmo/l at baseline (see Supplementary figure S1). The mean FPG of 5.1 mmol/l in the LRC cohort members aged 28–32 (5.2 mmol/l and 4.9 mmol/l in men and women, respectively) is in accordance with the mean values of 5.1 mmol/l and 4.8 mmol/l among white men and women of the CARDIA Study at ages 25–37.29 Table 2 presents the standardized Betas of cognitive function in midlife according to FPG at baseline. Hyperglycemia was independently associated with poorer global cognitive function in midlife (FPG ≥5.6 mmol/l vs. FPG <5.6 mmol/l, multivariable-adjusted standardized Beta= −0.121, P = 0.002; and FPG ≥5.6 mmol/l vs. FPG <4.4 mmol/l, multivariable-adjusted standardized Beta= −0.176, P = 0.007). A test for a linear trend with FPG entered as a continuous variable yielded a P values of .016. The domains that contributed to these associations were visual spatial and attention (table 2, Model 2). Table 2 Linear regression of the association of fasting plasma glucose and insulin resistance in young adulthood (∼30 years) with cognitive function in midlife (∼50 years): the Jerusalem LRC longitudinal study Models Globala Attentiona Information processing speed Executive Visual spatial Memorya (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl)vs. <5.6 mmol/l (100 mg/dl) −0.148** −0.121** −0.104* −0.083* −0.079 −0.061 −0.055 −0.020 −0.199** −0.185** −0.098* −0.081 (0.001) (0.002) (0.021) (0.050) (0.092) (0.162) (0.223) (0.638) (<0.001) (<0.001) (0.031) (0.066) 2. FPG (mmol/l) (continuous) −0.122** −0.094* −0.109* −0.090* −0.053 −0.043 −0.092* 0.068 −0.141** −0.130** −0.095* −0.058 (0.007) (0.016) (0.016) (0.034) (0.255) (0.321) (0.043) (0.119) (0.002) (0.002) (0.039) (0.191) 3. FPG (3 df)…(<80 mg/dl)(n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.032)b (0.034)b (0.084)b (0.101)b (0.568)b (0.627)b (0.113)b (0.192)b (0.004)b (0.002)b (0.054)b (0.140)b  4.4-4.9 mmol/l (80–89 mg/dl)(n = 172) −0.104 −0.102 −0.085 −0.073 −0.026 −0.002 −0.116 −0.127 −0.132 −0.121 −0.068 −0.064 (0.220) (0.179) (0.341) (0.381) (0.784) (0.985) (0.194) (0.133) (0.128) (0.129) (0.455) (0.462) 5.0-5.5 mmol/l (90–99 mg/dl)(n = 218) −0.059 −0.063 −0.064 −0.066 0.026 0.030 −0.128 −0.149 −0.071 −0.079 −0.076 −0.056 (0.517) (0.420) (0.488) (0.442) (0.784) (0.733) (0.165) (0.089) (0.427) (0.336) (0.419) (0.532) 5.6+ mmol/l (100+ mg/dl)(n = 82) −0.204** −0.176** −0.155* −0.131 −0.076 −0.049 −0.142 −0.119 −0.267** −0.252** −0.150 −0.123 (0.007) (0.007) (0.043) (0.054) (0.329) (0.489) (0.064) (0.103) (<.001) (<.001) (0.056) (0.100) 4. Insulin resistanceHOMA (mlU/l*mmol/l)a,c (continuous) −0.065 −0.012 −0.073 −0.040 −0.045 −0.019 −0.046 0.033 −0.041 0.002 −0.012 0.038 (0.157) (0.787) (0.112) (0.419) (0.347) (0.707) (0.309) (0.509) (0.369) (0.966) (0.801) (0.459) Models Globala Attentiona Information processing speed Executive Visual spatial Memorya (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl)vs. <5.6 mmol/l (100 mg/dl) −0.148** −0.121** −0.104* −0.083* −0.079 −0.061 −0.055 −0.020 −0.199** −0.185** −0.098* −0.081 (0.001) (0.002) (0.021) (0.050) (0.092) (0.162) (0.223) (0.638) (<0.001) (<0.001) (0.031) (0.066) 2. FPG (mmol/l) (continuous) −0.122** −0.094* −0.109* −0.090* −0.053 −0.043 −0.092* 0.068 −0.141** −0.130** −0.095* −0.058 (0.007) (0.016) (0.016) (0.034) (0.255) (0.321) (0.043) (0.119) (0.002) (0.002) (0.039) (0.191) 3. FPG (3 df)…(<80 mg/dl)(n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.032)b (0.034)b (0.084)b (0.101)b (0.568)b (0.627)b (0.113)b (0.192)b (0.004)b (0.002)b (0.054)b (0.140)b  4.4-4.9 mmol/l (80–89 mg/dl)(n = 172) −0.104 −0.102 −0.085 −0.073 −0.026 −0.002 −0.116 −0.127 −0.132 −0.121 −0.068 −0.064 (0.220) (0.179) (0.341) (0.381) (0.784) (0.985) (0.194) (0.133) (0.128) (0.129) (0.455) (0.462) 5.0-5.5 mmol/l (90–99 mg/dl)(n = 218) −0.059 −0.063 −0.064 −0.066 0.026 0.030 −0.128 −0.149 −0.071 −0.079 −0.076 −0.056 (0.517) (0.420) (0.488) (0.442) (0.784) (0.733) (0.165) (0.089) (0.427) (0.336) (0.419) (0.532) 5.6+ mmol/l (100+ mg/dl)(n = 82) −0.204** −0.176** −0.155* −0.131 −0.076 −0.049 −0.142 −0.119 −0.267** −0.252** −0.150 −0.123 (0.007) (0.007) (0.043) (0.054) (0.329) (0.489) (0.064) (0.103) (<.001) (<.001) (0.056) (0.100) 4. Insulin resistanceHOMA (mlU/l*mmol/l)a,c (continuous) −0.065 −0.012 −0.073 −0.040 −0.045 −0.019 −0.046 0.033 −0.041 0.002 −0.012 0.038 (0.157) (0.787) (0.112) (0.419) (0.347) (0.707) (0.309) (0.509) (0.369) (0.966) (0.801) (0.459) Model 1: Adjusted for sex and age at cognitive assessment. Model 2: Additionally adjusted for educational level, origin, religiosity, childhood SEP (ICBS ranking), adult SEP (ICBS ranking), and cigarette pack-yrs, leisure-time vigorous activity, alcohol intake, BMI, total cholesterol, HDL-cholesterol and GlycA, measured at baseline (at ages 28–32), and for the sampling scheme. Missing data: HOMA-IR (n = 1), adult SEP (n = 5), leisure-time vigorous activity (n = 1) and GlycA (n = 20). Missing values were replaced with non-missing median values (adult SEP, leisure-time vigorous activity) or by multiple linear regression imputation (baseline GlycA). LRC, Lipid Research Clinic; FPG, fasting plasma glucose; HOMA, homeostatic model assessment for insulin resistance. Beta = standardized regression coefficient. Numbers represent Beta coefficients, i.e. change in cognitive score per SD of variable. a Box-Cox transformed cognitive z-scores (λ= - 0.5), HOMA-IR (λ = 0). b Test for trend for the FPG groups introduced as an ordinal variable (1 df). c The HOMA-IR formula was defined as fasting insulin (mlU/l) multiplied by fasting glucose (mmol/l) divided by 22.5. * P ≤ 0.05 ** P ≤ 0.01 Table 2 Linear regression of the association of fasting plasma glucose and insulin resistance in young adulthood (∼30 years) with cognitive function in midlife (∼50 years): the Jerusalem LRC longitudinal study Models Globala Attentiona Information processing speed Executive Visual spatial Memorya (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl)vs. <5.6 mmol/l (100 mg/dl) −0.148** −0.121** −0.104* −0.083* −0.079 −0.061 −0.055 −0.020 −0.199** −0.185** −0.098* −0.081 (0.001) (0.002) (0.021) (0.050) (0.092) (0.162) (0.223) (0.638) (<0.001) (<0.001) (0.031) (0.066) 2. FPG (mmol/l) (continuous) −0.122** −0.094* −0.109* −0.090* −0.053 −0.043 −0.092* 0.068 −0.141** −0.130** −0.095* −0.058 (0.007) (0.016) (0.016) (0.034) (0.255) (0.321) (0.043) (0.119) (0.002) (0.002) (0.039) (0.191) 3. FPG (3 df)…(<80 mg/dl)(n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.032)b (0.034)b (0.084)b (0.101)b (0.568)b (0.627)b (0.113)b (0.192)b (0.004)b (0.002)b (0.054)b (0.140)b  4.4-4.9 mmol/l (80–89 mg/dl)(n = 172) −0.104 −0.102 −0.085 −0.073 −0.026 −0.002 −0.116 −0.127 −0.132 −0.121 −0.068 −0.064 (0.220) (0.179) (0.341) (0.381) (0.784) (0.985) (0.194) (0.133) (0.128) (0.129) (0.455) (0.462) 5.0-5.5 mmol/l (90–99 mg/dl)(n = 218) −0.059 −0.063 −0.064 −0.066 0.026 0.030 −0.128 −0.149 −0.071 −0.079 −0.076 −0.056 (0.517) (0.420) (0.488) (0.442) (0.784) (0.733) (0.165) (0.089) (0.427) (0.336) (0.419) (0.532) 5.6+ mmol/l (100+ mg/dl)(n = 82) −0.204** −0.176** −0.155* −0.131 −0.076 −0.049 −0.142 −0.119 −0.267** −0.252** −0.150 −0.123 (0.007) (0.007) (0.043) (0.054) (0.329) (0.489) (0.064) (0.103) (<.001) (<.001) (0.056) (0.100) 4. Insulin resistanceHOMA (mlU/l*mmol/l)a,c (continuous) −0.065 −0.012 −0.073 −0.040 −0.045 −0.019 −0.046 0.033 −0.041 0.002 −0.012 0.038 (0.157) (0.787) (0.112) (0.419) (0.347) (0.707) (0.309) (0.509) (0.369) (0.966) (0.801) (0.459) Models Globala Attentiona Information processing speed Executive Visual spatial Memorya (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl)vs. <5.6 mmol/l (100 mg/dl) −0.148** −0.121** −0.104* −0.083* −0.079 −0.061 −0.055 −0.020 −0.199** −0.185** −0.098* −0.081 (0.001) (0.002) (0.021) (0.050) (0.092) (0.162) (0.223) (0.638) (<0.001) (<0.001) (0.031) (0.066) 2. FPG (mmol/l) (continuous) −0.122** −0.094* −0.109* −0.090* −0.053 −0.043 −0.092* 0.068 −0.141** −0.130** −0.095* −0.058 (0.007) (0.016) (0.016) (0.034) (0.255) (0.321) (0.043) (0.119) (0.002) (0.002) (0.039) (0.191) 3. FPG (3 df)…(<80 mg/dl)(n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.032)b (0.034)b (0.084)b (0.101)b (0.568)b (0.627)b (0.113)b (0.192)b (0.004)b (0.002)b (0.054)b (0.140)b  4.4-4.9 mmol/l (80–89 mg/dl)(n = 172) −0.104 −0.102 −0.085 −0.073 −0.026 −0.002 −0.116 −0.127 −0.132 −0.121 −0.068 −0.064 (0.220) (0.179) (0.341) (0.381) (0.784) (0.985) (0.194) (0.133) (0.128) (0.129) (0.455) (0.462) 5.0-5.5 mmol/l (90–99 mg/dl)(n = 218) −0.059 −0.063 −0.064 −0.066 0.026 0.030 −0.128 −0.149 −0.071 −0.079 −0.076 −0.056 (0.517) (0.420) (0.488) (0.442) (0.784) (0.733) (0.165) (0.089) (0.427) (0.336) (0.419) (0.532) 5.6+ mmol/l (100+ mg/dl)(n = 82) −0.204** −0.176** −0.155* −0.131 −0.076 −0.049 −0.142 −0.119 −0.267** −0.252** −0.150 −0.123 (0.007) (0.007) (0.043) (0.054) (0.329) (0.489) (0.064) (0.103) (<.001) (<.001) (0.056) (0.100) 4. Insulin resistanceHOMA (mlU/l*mmol/l)a,c (continuous) −0.065 −0.012 −0.073 −0.040 −0.045 −0.019 −0.046 0.033 −0.041 0.002 −0.012 0.038 (0.157) (0.787) (0.112) (0.419) (0.347) (0.707) (0.309) (0.509) (0.369) (0.966) (0.801) (0.459) Model 1: Adjusted for sex and age at cognitive assessment. Model 2: Additionally adjusted for educational level, origin, religiosity, childhood SEP (ICBS ranking), adult SEP (ICBS ranking), and cigarette pack-yrs, leisure-time vigorous activity, alcohol intake, BMI, total cholesterol, HDL-cholesterol and GlycA, measured at baseline (at ages 28–32), and for the sampling scheme. Missing data: HOMA-IR (n = 1), adult SEP (n = 5), leisure-time vigorous activity (n = 1) and GlycA (n = 20). Missing values were replaced with non-missing median values (adult SEP, leisure-time vigorous activity) or by multiple linear regression imputation (baseline GlycA). LRC, Lipid Research Clinic; FPG, fasting plasma glucose; HOMA, homeostatic model assessment for insulin resistance. Beta = standardized regression coefficient. Numbers represent Beta coefficients, i.e. change in cognitive score per SD of variable. a Box-Cox transformed cognitive z-scores (λ= - 0.5), HOMA-IR (λ = 0). b Test for trend for the FPG groups introduced as an ordinal variable (1 df). c The HOMA-IR formula was defined as fasting insulin (mlU/l) multiplied by fasting glucose (mmol/l) divided by 22.5. * P ≤ 0.05 ** P ≤ 0.01 Similarly, using logistic models to predict low ranked cognition (the lower fifth of the cognitive score distribution compared with the upper 4 quintiles grouped), an approximately 130% increased odds for low global cognitive function was evident per 1 mmol/l FPG increment at baseline (OR = 2.31, 95% CI = 1.30–4.13, P = 0.005). FPG ≥5.6 mmol/l (vs. <5.6 mmol/l) was associated with a 2-fold increased odds for low ranked global cognitive function (OR = 2.10, 95% CI = 1.09–4.06, P = 0.026). Compared with FPG <4.4 mmol/l, increasing levels of FPG conferred a graded increase in the odds for low ranked global cognitive function (p for trend=.006) (table 3, Model 2). Table 3 Logistic modeling of the association of fasting plasma glucose and insulin resistance in young adulthood [∼30 years) with low ranked cognitive function (lowest fifth) in midlife (∼50 years)]: the Jerusalem LRC longitudinal study Models Global Attention Information processing speed Executive Visual spatial Memory (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 OR OR OR OR OR OR OR OR OR OR OR OR [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl) vs. <5.6 mmol/l (100 mg/dl) 2.09** 2.10* 1.81* 1.76 1.25 0.966 1.35 1.00 2.44** 2.26* 1.82* 1.55 [1.21–3.61] [1.09–4.06] [1.04–3.16] [0.95–3.25] [0.68–2.31] [0.48–1.95] [0.75–2.42] [0.53–1.91] [1.37–4.34] [1.15–4.44] [1.04–3.17] [0.81–2.95] (0.008) (0.026) (0.036) (0.071) (0.473) (0.923) (0.318) (0.995) (0.002)* (0.019) (0.035) (0.184) 2. FPG (mmol/l) (continuous) 2.26** 2.31** 1.67* 1.60 1.006 0.911 1.39 1.16 1.89* 1.78* 1.63 124 [1.37–3.73] [1.30–4.13] [1.02–2.74 [0.94–2.72] [0.62–1.79] [0.51–1.64] [0.85–2.28] [0.69–1.96] [1.13–3.19] [1.00–3.17] [1.00–2.67] [0.72–2.16] (0.001) (0.005) (0.042) (0.084) (0.837) (0.755) (0.188) (0.576) (0.016) (0.050) (0.055) (0.437) 3. FPG (3 df)…(<80 mg/dl) (n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.007)b (0.006)b (0.097)b (0.121)b (0.913)b (0.789)b (0.456)b (0.894)b (0.021)b (0.034)b (0.050)b (0.244)b  4.4–4.9 mmol/l (80–89 mg/dl) (n = 172) 2.13 2.28 1.29 1.11 1.88 1.76 1.51 1.70 1.67 1.37 2.38 2.20 [0.61–7.43] [0.57–9.10] [0.46–3.59] [0.38–3.28] [0.62–5.75] [0.53–5.85] [0.54–4.21] [0.58–5.02] [0.54–5.17] [0.40–4.64] [0.68–8.32] [0.59–8.21] (0.238) (0.243) (0.633) (0.849) (0.268) (0.354) (0.433) (0.334) (0.370) (0.618) (0.173) (0.241) 5.0–5.5 mmol/l (90–99 mg/dl) n = 218) 2.42 3.13 1.27 1.20 1.31 1.31 1.33 1.45 1.54 1.52 2.23 1.89 [0.70–8.35] [0.80–12.28] [0.46–3.52] [0.41–3.50] [0.43–4.00] [0.40–4.33] [0.48–3.69] [0.50–4.23] [0.50–4.73] [0.45–5.11] [0.64–7.72] [0.51–7.01] (0.162) (0.102) (0.645) (0.736) (0.640) (0.658) (0.585) (0.495) (0.452) (0.498) (0.207) (0.339) 5.6+ mmol/l (100+ mg/dl) (n = 82) 4.60* 5.44* 2.28 2.03 1.87 1.40 1.85 1.51 3.78* 3.19 3.99* 3.00 [1.26–16.72] [1.29–22.87] [0.77–6.73] [0.65–6.35] [0.56–6.23] [0.38–5.11] [0.61–5.57] [0.47–4.81] [1.15–12.38] [0.88–11.64] [1.09–14.60] [0.75–11.97] (0.021) (0.021) (0.137) (0.226) (0.305) (0.613) (0.276) (0.491) (0.028) (0.079) (0.037) (0.121) 4. Insulin resistance HOMA (mlU/l*mmol/l)a,c (continuous) 1.32 0.79 1.02 0.78 1.23* 0.80 1.59 0.90 1.06 0.65 1.03 0.56 [0.78–2.26] [0.39–1.61] [0.60–1.74] [0.40–1.53] [0.71–2.15] [0.39–1.61] [0.92–2.72] [0.46–1.76] [0.61–1.85] [0.32–1.34] [0.60–1.76] [0.28–1.13] (0.303) (0.517) (0.94) (0.47) (0.46) (0.54) (0.90) (0.75) (0.83) (0.25) (0.92) (0.11) Models Global Attention Information processing speed Executive Visual spatial Memory (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 OR OR OR OR OR OR OR OR OR OR OR OR [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl) vs. <5.6 mmol/l (100 mg/dl) 2.09** 2.10* 1.81* 1.76 1.25 0.966 1.35 1.00 2.44** 2.26* 1.82* 1.55 [1.21–3.61] [1.09–4.06] [1.04–3.16] [0.95–3.25] [0.68–2.31] [0.48–1.95] [0.75–2.42] [0.53–1.91] [1.37–4.34] [1.15–4.44] [1.04–3.17] [0.81–2.95] (0.008) (0.026) (0.036) (0.071) (0.473) (0.923) (0.318) (0.995) (0.002)* (0.019) (0.035) (0.184) 2. FPG (mmol/l) (continuous) 2.26** 2.31** 1.67* 1.60 1.006 0.911 1.39 1.16 1.89* 1.78* 1.63 124 [1.37–3.73] [1.30–4.13] [1.02–2.74 [0.94–2.72] [0.62–1.79] [0.51–1.64] [0.85–2.28] [0.69–1.96] [1.13–3.19] [1.00–3.17] [1.00–2.67] [0.72–2.16] (0.001) (0.005) (0.042) (0.084) (0.837) (0.755) (0.188) (0.576) (0.016) (0.050) (0.055) (0.437) 3. FPG (3 df)…(<80 mg/dl) (n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.007)b (0.006)b (0.097)b (0.121)b (0.913)b (0.789)b (0.456)b (0.894)b (0.021)b (0.034)b (0.050)b (0.244)b  4.4–4.9 mmol/l (80–89 mg/dl) (n = 172) 2.13 2.28 1.29 1.11 1.88 1.76 1.51 1.70 1.67 1.37 2.38 2.20 [0.61–7.43] [0.57–9.10] [0.46–3.59] [0.38–3.28] [0.62–5.75] [0.53–5.85] [0.54–4.21] [0.58–5.02] [0.54–5.17] [0.40–4.64] [0.68–8.32] [0.59–8.21] (0.238) (0.243) (0.633) (0.849) (0.268) (0.354) (0.433) (0.334) (0.370) (0.618) (0.173) (0.241) 5.0–5.5 mmol/l (90–99 mg/dl) n = 218) 2.42 3.13 1.27 1.20 1.31 1.31 1.33 1.45 1.54 1.52 2.23 1.89 [0.70–8.35] [0.80–12.28] [0.46–3.52] [0.41–3.50] [0.43–4.00] [0.40–4.33] [0.48–3.69] [0.50–4.23] [0.50–4.73] [0.45–5.11] [0.64–7.72] [0.51–7.01] (0.162) (0.102) (0.645) (0.736) (0.640) (0.658) (0.585) (0.495) (0.452) (0.498) (0.207) (0.339) 5.6+ mmol/l (100+ mg/dl) (n = 82) 4.60* 5.44* 2.28 2.03 1.87 1.40 1.85 1.51 3.78* 3.19 3.99* 3.00 [1.26–16.72] [1.29–22.87] [0.77–6.73] [0.65–6.35] [0.56–6.23] [0.38–5.11] [0.61–5.57] [0.47–4.81] [1.15–12.38] [0.88–11.64] [1.09–14.60] [0.75–11.97] (0.021) (0.021) (0.137) (0.226) (0.305) (0.613) (0.276) (0.491) (0.028) (0.079) (0.037) (0.121) 4. Insulin resistance HOMA (mlU/l*mmol/l)a,c (continuous) 1.32 0.79 1.02 0.78 1.23* 0.80 1.59 0.90 1.06 0.65 1.03 0.56 [0.78–2.26] [0.39–1.61] [0.60–1.74] [0.40–1.53] [0.71–2.15] [0.39–1.61] [0.92–2.72] [0.46–1.76] [0.61–1.85] [0.32–1.34] [0.60–1.76] [0.28–1.13] (0.303) (0.517) (0.94) (0.47) (0.46) (0.54) (0.90) (0.75) (0.83) (0.25) (0.92) (0.11) Model 1: Adjusted for sex and age at cognitive assessment. Model 2: Additionally adjusted for educational level, origin, religiosity, childhood SEP, adult SEP (ICBS ranking) and cigarette pack-yrs, leisure-time vigorous activity, alcohol intake, BMI, total cholesterol, HDL-cholesterol and GlycA measured at baseline (at ages 28–32), and for the sampling scheme. Missing data: HOMA-IR (n = 1), adult SEP (n = 5), leisure-time vigorous activity (n = 1) and GlycA (n = 20). Missing values were replaced with non-missing median values (adult SEP, leisure-time vigorous activity) or by multiple linear regression imputation (baseline GlycA). LRC, Lipid Research Clinic; FPG, fasting plasma glucose; HOMA, homeostatic model assessment for insulin resistance. OR, odds ratio. CI, confidence interval. a Box-Cox transformed HOMA-IR (λ = 0). b Test for trend for the FPG groups introduced as an ordinal variable (1 df). c The HOMA-IR formula was defined as fasting insulin (mlU/l) multiplied by fasting glucose (mmol/l) divided by 22.5. * P ≤ 0.05 ** P ≤0 .01 Table 3 Logistic modeling of the association of fasting plasma glucose and insulin resistance in young adulthood [∼30 years) with low ranked cognitive function (lowest fifth) in midlife (∼50 years)]: the Jerusalem LRC longitudinal study Models Global Attention Information processing speed Executive Visual spatial Memory (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 OR OR OR OR OR OR OR OR OR OR OR OR [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl) vs. <5.6 mmol/l (100 mg/dl) 2.09** 2.10* 1.81* 1.76 1.25 0.966 1.35 1.00 2.44** 2.26* 1.82* 1.55 [1.21–3.61] [1.09–4.06] [1.04–3.16] [0.95–3.25] [0.68–2.31] [0.48–1.95] [0.75–2.42] [0.53–1.91] [1.37–4.34] [1.15–4.44] [1.04–3.17] [0.81–2.95] (0.008) (0.026) (0.036) (0.071) (0.473) (0.923) (0.318) (0.995) (0.002)* (0.019) (0.035) (0.184) 2. FPG (mmol/l) (continuous) 2.26** 2.31** 1.67* 1.60 1.006 0.911 1.39 1.16 1.89* 1.78* 1.63 124 [1.37–3.73] [1.30–4.13] [1.02–2.74 [0.94–2.72] [0.62–1.79] [0.51–1.64] [0.85–2.28] [0.69–1.96] [1.13–3.19] [1.00–3.17] [1.00–2.67] [0.72–2.16] (0.001) (0.005) (0.042) (0.084) (0.837) (0.755) (0.188) (0.576) (0.016) (0.050) (0.055) (0.437) 3. FPG (3 df)…(<80 mg/dl) (n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.007)b (0.006)b (0.097)b (0.121)b (0.913)b (0.789)b (0.456)b (0.894)b (0.021)b (0.034)b (0.050)b (0.244)b  4.4–4.9 mmol/l (80–89 mg/dl) (n = 172) 2.13 2.28 1.29 1.11 1.88 1.76 1.51 1.70 1.67 1.37 2.38 2.20 [0.61–7.43] [0.57–9.10] [0.46–3.59] [0.38–3.28] [0.62–5.75] [0.53–5.85] [0.54–4.21] [0.58–5.02] [0.54–5.17] [0.40–4.64] [0.68–8.32] [0.59–8.21] (0.238) (0.243) (0.633) (0.849) (0.268) (0.354) (0.433) (0.334) (0.370) (0.618) (0.173) (0.241) 5.0–5.5 mmol/l (90–99 mg/dl) n = 218) 2.42 3.13 1.27 1.20 1.31 1.31 1.33 1.45 1.54 1.52 2.23 1.89 [0.70–8.35] [0.80–12.28] [0.46–3.52] [0.41–3.50] [0.43–4.00] [0.40–4.33] [0.48–3.69] [0.50–4.23] [0.50–4.73] [0.45–5.11] [0.64–7.72] [0.51–7.01] (0.162) (0.102) (0.645) (0.736) (0.640) (0.658) (0.585) (0.495) (0.452) (0.498) (0.207) (0.339) 5.6+ mmol/l (100+ mg/dl) (n = 82) 4.60* 5.44* 2.28 2.03 1.87 1.40 1.85 1.51 3.78* 3.19 3.99* 3.00 [1.26–16.72] [1.29–22.87] [0.77–6.73] [0.65–6.35] [0.56–6.23] [0.38–5.11] [0.61–5.57] [0.47–4.81] [1.15–12.38] [0.88–11.64] [1.09–14.60] [0.75–11.97] (0.021) (0.021) (0.137) (0.226) (0.305) (0.613) (0.276) (0.491) (0.028) (0.079) (0.037) (0.121) 4. Insulin resistance HOMA (mlU/l*mmol/l)a,c (continuous) 1.32 0.79 1.02 0.78 1.23* 0.80 1.59 0.90 1.06 0.65 1.03 0.56 [0.78–2.26] [0.39–1.61] [0.60–1.74] [0.40–1.53] [0.71–2.15] [0.39–1.61] [0.92–2.72] [0.46–1.76] [0.61–1.85] [0.32–1.34] [0.60–1.76] [0.28–1.13] (0.303) (0.517) (0.94) (0.47) (0.46) (0.54) (0.90) (0.75) (0.83) (0.25) (0.92) (0.11) Models Global Attention Information processing speed Executive Visual spatial Memory (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 OR OR OR OR OR OR OR OR OR OR OR OR [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl) vs. <5.6 mmol/l (100 mg/dl) 2.09** 2.10* 1.81* 1.76 1.25 0.966 1.35 1.00 2.44** 2.26* 1.82* 1.55 [1.21–3.61] [1.09–4.06] [1.04–3.16] [0.95–3.25] [0.68–2.31] [0.48–1.95] [0.75–2.42] [0.53–1.91] [1.37–4.34] [1.15–4.44] [1.04–3.17] [0.81–2.95] (0.008) (0.026) (0.036) (0.071) (0.473) (0.923) (0.318) (0.995) (0.002)* (0.019) (0.035) (0.184) 2. FPG (mmol/l) (continuous) 2.26** 2.31** 1.67* 1.60 1.006 0.911 1.39 1.16 1.89* 1.78* 1.63 124 [1.37–3.73] [1.30–4.13] [1.02–2.74 [0.94–2.72] [0.62–1.79] [0.51–1.64] [0.85–2.28] [0.69–1.96] [1.13–3.19] [1.00–3.17] [1.00–2.67] [0.72–2.16] (0.001) (0.005) (0.042) (0.084) (0.837) (0.755) (0.188) (0.576) (0.016) (0.050) (0.055) (0.437) 3. FPG (3 df)…(<80 mg/dl) (n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.007)b (0.006)b (0.097)b (0.121)b (0.913)b (0.789)b (0.456)b (0.894)b (0.021)b (0.034)b (0.050)b (0.244)b  4.4–4.9 mmol/l (80–89 mg/dl) (n = 172) 2.13 2.28 1.29 1.11 1.88 1.76 1.51 1.70 1.67 1.37 2.38 2.20 [0.61–7.43] [0.57–9.10] [0.46–3.59] [0.38–3.28] [0.62–5.75] [0.53–5.85] [0.54–4.21] [0.58–5.02] [0.54–5.17] [0.40–4.64] [0.68–8.32] [0.59–8.21] (0.238) (0.243) (0.633) (0.849) (0.268) (0.354) (0.433) (0.334) (0.370) (0.618) (0.173) (0.241) 5.0–5.5 mmol/l (90–99 mg/dl) n = 218) 2.42 3.13 1.27 1.20 1.31 1.31 1.33 1.45 1.54 1.52 2.23 1.89 [0.70–8.35] [0.80–12.28] [0.46–3.52] [0.41–3.50] [0.43–4.00] [0.40–4.33] [0.48–3.69] [0.50–4.23] [0.50–4.73] [0.45–5.11] [0.64–7.72] [0.51–7.01] (0.162) (0.102) (0.645) (0.736) (0.640) (0.658) (0.585) (0.495) (0.452) (0.498) (0.207) (0.339) 5.6+ mmol/l (100+ mg/dl) (n = 82) 4.60* 5.44* 2.28 2.03 1.87 1.40 1.85 1.51 3.78* 3.19 3.99* 3.00 [1.26–16.72] [1.29–22.87] [0.77–6.73] [0.65–6.35] [0.56–6.23] [0.38–5.11] [0.61–5.57] [0.47–4.81] [1.15–12.38] [0.88–11.64] [1.09–14.60] [0.75–11.97] (0.021) (0.021) (0.137) (0.226) (0.305) (0.613) (0.276) (0.491) (0.028) (0.079) (0.037) (0.121) 4. Insulin resistance HOMA (mlU/l*mmol/l)a,c (continuous) 1.32 0.79 1.02 0.78 1.23* 0.80 1.59 0.90 1.06 0.65 1.03 0.56 [0.78–2.26] [0.39–1.61] [0.60–1.74] [0.40–1.53] [0.71–2.15] [0.39–1.61] [0.92–2.72] [0.46–1.76] [0.61–1.85] [0.32–1.34] [0.60–1.76] [0.28–1.13] (0.303) (0.517) (0.94) (0.47) (0.46) (0.54) (0.90) (0.75) (0.83) (0.25) (0.92) (0.11) Model 1: Adjusted for sex and age at cognitive assessment. Model 2: Additionally adjusted for educational level, origin, religiosity, childhood SEP, adult SEP (ICBS ranking) and cigarette pack-yrs, leisure-time vigorous activity, alcohol intake, BMI, total cholesterol, HDL-cholesterol and GlycA measured at baseline (at ages 28–32), and for the sampling scheme. Missing data: HOMA-IR (n = 1), adult SEP (n = 5), leisure-time vigorous activity (n = 1) and GlycA (n = 20). Missing values were replaced with non-missing median values (adult SEP, leisure-time vigorous activity) or by multiple linear regression imputation (baseline GlycA). LRC, Lipid Research Clinic; FPG, fasting plasma glucose; HOMA, homeostatic model assessment for insulin resistance. OR, odds ratio. CI, confidence interval. a Box-Cox transformed HOMA-IR (λ = 0). b Test for trend for the FPG groups introduced as an ordinal variable (1 df). c The HOMA-IR formula was defined as fasting insulin (mlU/l) multiplied by fasting glucose (mmol/l) divided by 22.5. * P ≤ 0.05 ** P ≤0 .01 The FPG-cognition associations were evident, albeit somewhat attenuated, in a sensitivity analysis restricted to FPG <6.1 mmol/l (see Supplementary table S3). The associations for FPG ≥5.6 vs. <5.6 mmol/l persisted in a sensitivity analysis of global cognitive function adjusting to a reduced covariate set (standardized Beta=−0.125, P = 0.001 and OR = 2.06, 95%CI = 1.11–3.83, P = 0.023) and for FPG treated as a continuous variable (standardized Beta=−0.091, P = 0.018 and OR = 1.04, 95%CI = 1.00–1.07, P = 0.010) (not shown). The association of FPG with global cognitive function was not affected in a sensitivity analysis substituting BMI with waist-to-hip ratio (not shown). There was no association of cognition in midlife with baseline HOMA-IR (tables 2 and 3) or with baseline OGTT or OGTT change or FPG change over the mean 13 years follow-up (Supplementary table S4). Discussion In our population-based cohort free of diabetes at ages 28–32, higher FPG at age 30 years was associated with lower global cognitive function in midlife, mainly attributable to its visual spatial and attention domains. These associations persisted after adjustment for BMI, demographics, childhood and adult SEP, physical activity, alcohol intake, smoking, inflammatory variables and plasma lipids. The associations remained evident when the analysis was restricted to those with FPG <6.1 mmol/l. Baseline insulin resistance and OGTT, and change in FPG and OGTT over the 13 years follow-up were not associated with cognition. Our study shows, in line with previous reports, that blood glucose levels are associated with cognitive function, and in particular extends this finding to young adults free of diabetes. Recent studies report an association between higher glucose levels (combined FPG and HbA1c values) and increased risk of dementia in older adults (n = 2067; aged 76 years) with and without diabetes,30 as well as associations between hyperglycemia (diabetes) and impaired memory, visual perception and attention (n = 2126; ages 19–72),3 between impaired glucose metabolism (composite score of FPG, HbA1c and HOMA) and lower executive function in a national sample of middle-aged and older adults (n = 1076; median age, 57 years),31 between chronically higher blood glucose levels (HbA1c) and lower memory scores in older adults without diabetes (n = 141; age 63.1 ± 6.9),4 between higher FPG levels, even within the normal range (<6.1 mmol/l), and longer reaction times in processing speed (n = 41; age 64.7 years)32 and between higher FPG in 35 obese and 35 normal weight university students free of diabetes (age 20.8 ± 2.4)11 on a test of inhibitory control among 3 cognitive domains assessed. Longitudinal data on the relationship between blood glucose and cognition are sparse, inconsistent and also limited to older adults. Longitudinal trajectories of glycated haemoglobin (HbA1c) in older adults with type 2 diabetes (n = 835, mean age, 72.8 years) predicted overall cognitive performance and its executive domain.33 A study that examined change in HbA1c and cognition among older adults with normal glucose tolerance (NGT) (≥75 years old, n = 101)6 found a decrease in the Mini-Mental State Examination with increasing HbA1c. The only study to our knowledge that has examined longitudinal change in FPG34 did so in a sample of adults aged 25–85 at baseline (n = 4547) and reported, in line with our study, no association of FPG change with cognition measured 12 years later. Consistent with our findings, HOMA was not associated with cognitive function in the MONA LISA study (age 50.0 ± 8.1, range 35–64 years) in which 4% and <6%, had hyperinsulinemia or an HOMA value >2.6 mlU/l*mmol/l.10 In the Framingham Heart Study third-generation cohort (n = 2126, age 40.4 ± 8.7, range 19–72),3 no association of HOMA with cognition or of HOMA with structural brain measures (fractional anisotropy or gray matter density) was found. On the other hand, in the Framingham Offspring study (mean age 61 ± 9),35 HOMA was related to poorer executive and visuospatial memory scores. In the ARIC cohort (age 53.7 years, range, 45–64),36 HOMA was associated with lower baseline cognitive test scores. In the Framingham Offspring study and the ARIC cohort mentioned above,35,36 adiposity was not accounted for, and in Framingham diabetes35 was included. Our study sample was younger, was free of diabetes and accounted for BMI. Mechanisms underlying the relationship between glycemia and cognitive function might involve glucose-related increases in inflammatory responses and blood coagulation activation, leading to subclinical strokes, gray matter atrophy, reduced white matter integrity and subsequent cerebral volume loss.4,35 Moreover, hyperglycemia (HBA1C ≥5.5% or FPG ≥5.8 mmol/l) is linked to an increase in prevalence of retinopathy.37 Retinal microvascular abnormalities appear to reflect small vessel damage in the brain and are independently associated with poor cognitive function in midlife.38 Direct ‘toxic’ effects of glucose on neuronal structures include imbalance in the generation and scavenging of reactive oxygen species, or advanced glycation of important functional and structural proteins in the brain.39 Ultimately, these structural changes may lead to decreased neurotransmitter signaling and loss of synaptic contacts with subsequent cognitive impairment.40 The main limitation of this study to be considered when interpreting our findings is that the measure of cognition was done at one point in midlife. Consequently, no direct inference about the role of FPG on cognitive decline can be drawn. However, adjustment for childhood SEP, education and ethnic origin partly controls for baseline cognition and partly overcomes this drawback. We did not undertake a 24-h ABPM (ambulatory blood pressure monitoring). This may be a limitation of our study, as accounting for our three consecutive measurements of blood pressure does not eliminate the influence of hemodynamic factors. Finally, the present study includes a relatively small portion of the original Visit 1 participants recruited at age 17 in 1976–78. However, as depicted in figure 1, the main reason for this apparent discrepancy in sample sizes was due to us drawing a small random sample of the Visit 1 participants for the detailed Visit 4 round of examinations (70% response). The Visit 5 and Visit 6 samples were a follow-up of the Visit 4 study sample, with response rates of 71% and 82%, respectively. We emphasize that the current study participants were shown to be representative of the Visit 1 source population. Therefore, we do not assume that the reported associations of FPG and cognition are likely to be overestimated due to nonresponse or selection bias. The main strength of this study is its focus on young adults free of diabetes at baseline. Additional strengths are the longitudinal data on FPG, the wide range of potential confounders evaluated, including detailed socio-demographic, psychosocial, health behavioral and biochemical variables, and the comprehensive objective computerized cognitive measures with millisecond precision. Finally, the generalizability of these findings in a heterogeneous Jewish population sample should be assessed in other populations. In summary, our results indicate that even in healthy young adults aged 30 without diabetes, fasting hyperglycemia was inversely associated with a global measure of cognition in midlife. Should this finding be confirmed by longitudinally measured cognitive data, testing whether interventions in young adults aimed at maintaining or reducing glucose levels well below 5.6 mmol/l would delay cognitive decline in nondiabetic middle-aged adults is indicated. Multidomain interventions of lifestyle, management of metabolic and vascular risk factors and cognitive training in non-demented older adults have already been demonstrated to prevent cognitive decline.41 Supplementary data Supplementary data are available at EURPUB online. Funding Chief Scientist of the Israel Ministry of Health [300000-5352]; the Israel Science Foundation [593/01]; and the US-Israel Binational Science Foundation [87-00419] all to JDK. Conflicts of interest: ESS and GMD are employees of NeuroTrax Corp. The remaining authors have no potential conflicts of interest to declare. Key points Data relating impaired glucose regulation to cognitive function in young adults are scarce, entail only cross-sectional assessment of glucose and cognition, and are based on middle-aged samples, type 2 diabetes patients or a small sample of convenience. We are unaware of reports relating either plasma glucose concentrations or longitudinal change in plasma glucose with cognitive assessment in population-based samples of young adults free of diabetes. A higher fasting plasma glucose concentration in healthy young adults aged 30 was associated with lower global cognitive function at age 50. Hyperglycemia in young adults may be a preventable and modifiable risk factor for low ranked cognitive function in midlife. References 1 Lu FP , Lin KP , Kuo HK . Diabetes and the risk of multi-system aging phenotypes: a systematic review and meta-analysis . PLoS One 2009 ; 4 : e4144 . Google Scholar CrossRef Search ADS PubMed 2 Yaffe K , Falvey C , Hamilton N , et al. Diabetes, glucose control, and 9-year cognitive decline among older adults without dementia . Arch Neurol 2012 ; 69 : 1170 – 5 . Google Scholar CrossRef Search ADS PubMed 3 Weinstein G , Maillard P , Himali JJ , et al. Glucose indices are associated with cognitive and structural brain measures in young adults . Neurology 2015 ; 84 : 2329 – 37 . Google Scholar CrossRef Search ADS PubMed 4 Kerti L , Witte AV , Winkler A , et al. Higher glucose levels associated with lower memory and reduced hippocampal microstructure . Neurology 2013 ; 81 : 1746 – 52 . Google Scholar CrossRef Search ADS PubMed 5 Convit A , Wolf OT , Tarshish C , de Leon MJ . Reduced glucose tolerance is associated with poor memory performance and hippocampal atrophy among normal elderly . Proc Natl Acad Sci U S A 2003 ; 100 : 2019 – 22 . Google Scholar CrossRef Search ADS PubMed 6 Ravona-Springer R , Moshier E , Schmeidler J , et al. Changes in glycemic control are associated with changes in cognition in non-diabetic elderly . J Alzheimers Dis 2012 ; 30 : 299 – 309 . Google Scholar CrossRef Search ADS PubMed 7 Gold SM , Dziobek I , Sweat V , et al. Hippocampal damage and memory impairments as possible early brain complications of type 2 diabetes . Diabetologia 2007 ; 50 : 711 – 9 . Google Scholar CrossRef Search ADS PubMed 8 Yau PL , Javier DC , Ryan CM , et al. Preliminary evidence for brain complications in obese adolescents with type 2 diabetes mellitus . Diabetologia 2010 ; 53 : 2298 – 306 . Google Scholar CrossRef Search ADS PubMed 9 Fuh JL , Wang SJ , Hwu CM , Lu SR . Glucose tolerance status and cognitive impairment in early middle-aged women . Diabet Med 2007 ; 24 : 788 – 91 . Google Scholar CrossRef Search ADS PubMed 10 Sanz CM , Ruidavets JB , Bongard V , et al. Relationship between markers of insulin resistance, markers of adiposity, HbA1c, and cognitive functions in a middle-aged population-based sample: the MONA LISA study . Diabetes Care 2013 ; 36 : 1512 – 21 . Google Scholar CrossRef Search ADS PubMed 11 Hawkins MA , Gunstad J , Calvo D , Spitznagel MB . Higher fasting glucose is associated with poorer cognition among healthy young adults . Health Psychol 2016 ; 35 : 199 – 202 . Google Scholar CrossRef Search ADS PubMed 12 Slater PE , Friedlander Y , Baras M , et al. The Jerusalem Lipid Research Clinic: sampling, response and selected methodological issues . Isr J Med Sci 1982 ; 18 : 1106 – 12 . Google Scholar PubMed 13 Kark JD , Sinnreich R , Leitersdorf E , et al. Taq1B CETP polymorphism, plasma CETP, lipoproteins, apolipoproteins and sex differences in a Jewish population sample characterized by low HDL-cholesterol . Atherosclerosis 2000 ; 151 : 509 – 18 . Google Scholar CrossRef Search ADS PubMed 14 Kark JD , Goldberger N , Kimura M , et al. Energy intake and leukocyte telomere length in young adults . Am J Clin Nutr 2012 ; 95 : 479 – 87 . Google Scholar CrossRef Search ADS PubMed 15 Cohen-Manheim I , Doniger GM , Sinnreich R , et al. Increase in the inflammatory marker GlycA over 13 years in young adults is associated with poorer cognitive function in midlife . PLoS One 2015 ; 10 : e0138036 . Google Scholar CrossRef Search ADS PubMed 16 Cohen-Manheim I , Doniger GM , Sinnreich R , et al. Increased attrition of leukocyte telomere length in young adults is associated with poorer cognitive function in midlife . Eur J Epidemiol 2016 ; 31 : 147 – 57 . Google Scholar CrossRef Search ADS PubMed 17 Cohen-Manheim I , Doniger GM , Sinnreich R , et al. Body mass index, height and socioeconomic position in adolescence, their trajectories into adulthood, and cognitive function in midlife . J Alzheimers Dis 2017 ; 55 : 1207 – 21 . Google Scholar CrossRef Search ADS PubMed 18 Dwolatzky T , Whitehead V , Doniger GM , et al. Validity of a novel computerized cognitive battery for mild cognitive impairment . BMC Geriatr 2003 ; 3 : 4 . Google Scholar CrossRef Search ADS PubMed 19 Melton J . ( 2005 ) Psychometric evaluation of the Mindstreams neuropsychological screening tool. NEDU Technical Report 06-10, Navy Experimental Diving Unit, Panama City, FL. 20 Thaler A , Mirelman A , Gurevich T , et al. Lower cognitive performance in healthy G2019S LRRK2 mutation carriers . Neurology 2012 ; 79 : 1027 – 32 . Google Scholar CrossRef Search ADS PubMed 21 Sasson E , Doniger GM , Pasternak O , Assaf Y . Structural correlates of memory performance with diffusion tensor imaging . Neuroimage 2010 ; 50 : 1231 – 42 . Google Scholar CrossRef Search ADS PubMed 22 Boussi-Gross R , Golan H , Volkov O , et al. Improvement of memory impairments in poststroke patients by hyperbaric oxygen therapy . Neuropsychology 2015 ; 29 : 610 – 21 . Google Scholar CrossRef Search ADS PubMed 23 Hartman SJ , Marinac CR , Natarajan L , Patterson RE . Lifestyle factors associated with cognitive functioning in breast cancer survivors . Psychooncology 2015 ; 24 : 669 – 75 . Google Scholar CrossRef Search ADS PubMed 24 Zur D , Naftaliev E , Kesler A . Evidence of multidomain mild cognitive impairment in idiopathic intracranial hypertension . J Neuroophthalmol 2015 ; 35 : 26 – 30 . Google Scholar CrossRef Search ADS PubMed 25 Adler NE , Epel ES , Castellazzo G , Ickovics JR . Relationship of subjective and objective social status with psychological and physiological functioning: preliminary data in healthy white women . Health Psychol 2000 ; 19 : 586 – 92 . Google Scholar CrossRef Search ADS PubMed 26 Zigmond AS , Snaith RP . The hospital anxiety and depression scale . Acta Psychiatr Scand 1983 ; 67 : 361 – 70 . Google Scholar CrossRef Search ADS PubMed 27 Otvos JD , Shalaurova I , Wolak-Dinsmore J , et al. GlycA: A composite nuclear magnetic resonance biomarker of systemic inflammation . Clin Chem 2015 ; 61 : 714 – 23 . Google Scholar CrossRef Search ADS PubMed 28 Gutierrez J , Marshall RS , Lazar RM . Indirect measures of arterial stiffness and cognitive performance in individuals without traditional vascular risk factors or disease . JAMA Neurol 2015 ; 72 : 309 – 15 . Google Scholar CrossRef Search ADS PubMed 29 Folsom AR , Jacobs DR Jr. , Wagenknecht LE , et al. Increase in fasting insulin and glucose over seven years with increasing weight and inactivity of young adults. The CARDIA Study. Coronary artery risk development in young adults . Am J Epidemiol 1996 ; 144 : 235 – 46 . Google Scholar CrossRef Search ADS PubMed 30 Crane PK , Walker R , Hubbard RA , et al. Glucose levels and risk of dementia . N Engl J Med 2013 ; 369 : 540 – 8 . Google Scholar CrossRef Search ADS PubMed 31 Karlamangla AS , Miller-Martinez D , Lachman ME , et al. Biological correlates of adult cognition: midlife in the United States (MIDUS) . Neurobiol Aging 2014 ; 35 : 387 – 94 . Google Scholar CrossRef Search ADS PubMed 32 Raizes M , Elkana O , Franko M , et al. Higher fasting plasma glucose levels, within the normal range, are associated with decreased processing speed in high functioning young elderly . J Alzheimers Dis 2015 ; 49 : 589 – 92 . Google Scholar CrossRef Search ADS 33 Ravona-Springer R , Heymann A , Schmeidler J , et al. Trajectories in glycemic control over time are associated with cognitive performance in elderly subjects with type 2 diabetes . PLoS One 2014 ; 9 : e97384 . Google Scholar CrossRef Search ADS PubMed 34 Anstey KJ , Sargent-Cox K , Eramudugolla R , et al. Association of cognitive function with glucose tolerance and trajectories of glucose tolerance over 12 years in the AusDiab study . Alzheimers Res Ther 2015 ; 7 : 48 . Google Scholar CrossRef Search ADS PubMed 35 Tan ZS , Beiser AS , Fox CS , et al. Association of metabolic dysregulation with volumetric brain magnetic resonance imaging and cognitive markers of subclinical brain aging in middle-aged adults: the Framingham Offspring Study . Diabetes Care 2011 ; 34 : 1766 – 70 . Google Scholar CrossRef Search ADS PubMed 36 Young SE , Mainous AG 3rd , Carnemolla M . Hyperinsulinemia and cognitive decline in a middle-aged cohort . Diabetes Care 2006 ; 29 : 2688 – 93 . Google Scholar CrossRef Search ADS PubMed 37 Cheng YJ , Gregg EW , Geiss LS , et al. Association of A1C and fasting plasma glucose levels with diabetic retinopathy prevalence in the U.S. population: Implications for diabetes diagnostic thresholds . Diabetes Care 2009 ; 32 : 2027 – 32 . Google Scholar CrossRef Search ADS PubMed 38 Wong TY , Klein R , Sharrett AR , et al. Retinal microvascular abnormalities and cognitive impairment in middle-aged persons: the Atherosclerosis Risk in Communities Study . Stroke 2002 ; 33 : 1487 – 92 . Google Scholar CrossRef Search ADS PubMed 39 Brownlee M . Biochemistry and molecular cell biology of diabetic complications . Nature 2001 ; 414 : 813 – 20 . Google Scholar CrossRef Search ADS PubMed 40 Lamport DJ , Lawton CL , Mansfield MW , Dye L . Impairments in glucose tolerance can have a negative impact on cognitive function: a systematic research review . Neurosci Biobehav Rev 2009 ; 33 : 394 – 413 . Google Scholar CrossRef Search ADS PubMed 41 Ngandu T , Lehtisalo J , Solomon A , et al. A 2 year multidomain intervention of diet, exercise, cognitive training, and vascular risk monitoring versus control to prevent cognitive decline in at-risk elderly people (FINGER): a randomised controlled trial . Lancet 2015 ; 385 : 2255 – 63 . Google Scholar CrossRef Search ADS PubMed © The Author 2017. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The European Journal of Public Health Oxford University Press

Fasting plasma glucose in young adults free of diabetes is associated with cognitive function in midlife

Loading next page...
 
/lp/ou_press/fasting-plasma-glucose-in-young-adults-free-of-diabetes-is-associated-mOqQPNlkes
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved.
ISSN
1101-1262
eISSN
1464-360X
D.O.I.
10.1093/eurpub/ckx194
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Evidence for an association of fasting plasma glucose (FPG) with cognitive function in adults free of diabetes is scarce and based on middle-aged and older adults. We examined the association of FPG, measured at age 30, and of change in FPG from age 30 to 43, with cognitive function at age 50. Methods 505 nondiabetic participants of the population-based Jerusalem Lipid Research Clinic (LRC) cohort study had baseline FPG, 2-h post-oral challenge plasma glucose (OGTT) and insulin determined at ages 28–32, and FPG and OGTT again at ages 41–46. Subsequently at ages 48–52, global cognitive function and its five specific component domains were assessed with a NeuroTrax computerized test battery, using multiple linear regression and multivariable logistic models. Results Hyperglycemia (FPG ≥ 5.6 mmol/l vs. <5.6 mmol/l) at baseline was associated with poorer global cognitive function in midlife (predominantly in the visual spatial and attention domains), independent of socio-demographic characteristics, life style variables, body mass index (BMI), and inflammatory and biochemical variables (standardized Beta = −0.121, P = 0.002, plinear trend(FPG continuous) =0.016). Similarly, increased odds for low-ranked (lowest fifth) global cognition was evident (ORper mmol/l FPG=2.31, 95% CI = 1.30–4.13, P = 0.005). Baseline OGTT, insulin resistance (HOMA-IR) and change in FPG and OGTT over 13 years were not associated with cognition. Conclusion A higher FPG in young adults was associated with lower cognitive performance in midlife. Although we cannot dismiss the possibility of reverse causation, hyperglycemia at a young age may be a modifiable risk factor for low-ranked cognitive function in midlife. Introduction It is well established that type 2 diabetes is associated with an increased risk of incident dementia1 as well as with cognitive decline2 and poor cognitive function.3 Similarly, hyperglycemia and insulin resistance without clinically diagnosed diabetes have been related to poorer cognitive function in late-life.4–6 Nevertheless, data relating impaired glucose regulation to cognitive function in younger adults are scarce,7–10 are all based on cross-sectional assessment of glucose and cognition, and are based on middle-aged samples7,9,10 and type 2 diabetes patients7,8 or a small sample of convenience.11 We are unaware of reports relating either glucose levels or longitudinal change in plasma glucose with cognitive assessment in non-diabetic young adults. Accordingly, we aimed to evaluate the relationship of fasting plasma glucose (FPG), 2 h post-oral challenge plasma glucose (OGTT) and insulin resistance (HOMA-IR) measured at ages 28–32, as well as change in FPG and OGTT over a 13-year follow-up to ages 41–46, with global cognitive function, the primary outcome variable, in a population-based cohort aged 48–52. The specific multi-domain components that contribute to global function served as secondary outcomes. Methods Study population The study sample was derived from the Jerusalem Lipid Research Clinic (LRC) Study, a longitudinal, population-based cohort initiated in 1976–1978. Details regarding sampling and response rates have been reported.12 A flow diagram for Visits 1–6 from 1976 through 2011 is shown in figure 1. In brief, in 1976–1978 the Jerusalem LRC examined 8646 17-year-old Jewish residents of Jerusalem, representing full age cohorts (Visit 1), as part of a compulsory examination to determine their fitness for military service (irrespective of actual conscription). In 1989–1991 (Visit 4, at age 28–32, the baseline of the current study), 2 overlapping samples of Visit 1 were re-invited: a sex-stratified random sample consisting of 884 young adults who met the eligibility requirements (71% response), and offspring of parents (an additional 168 individuals) who had a documented acute myocardial infarction or sudden cardiac death over a 10 years follow-up period (65% response). A total of 1052 eligible subjects (70% response) were examined and interviewed at age 28–32y.13 Figure 1 View largeDownload slide LRC study design flow chart. *2 overlapping samples of Visit 1 were re-invited to Visit 4: a sex-stratified random subsample sample comprising 884 young adults with a male to female sampling ratio of 1.6: 1 (71% response), and offspring of parents (an additional 168 individuals) who had a documented acute myocardial infarction or sudden cardiac death over a 10y follow-up period (65% response). The sampling scheme was incorporated in the analysis—and did not affect the results. †Exclusion of ineligible participants who were not current Jerusalem residents, were pregnant or were within 3 months of delivery, had a serious incapacitating illness or had died Figure 1 View largeDownload slide LRC study design flow chart. *2 overlapping samples of Visit 1 were re-invited to Visit 4: a sex-stratified random subsample sample comprising 884 young adults with a male to female sampling ratio of 1.6: 1 (71% response), and offspring of parents (an additional 168 individuals) who had a documented acute myocardial infarction or sudden cardiac death over a 10y follow-up period (65% response). The sampling scheme was incorporated in the analysis—and did not affect the results. †Exclusion of ineligible participants who were not current Jerusalem residents, were pregnant or were within 3 months of delivery, had a serious incapacitating illness or had died In 2003–2006 (Visit 5, age 41–46), 631 of the Visit 4 participants were re-examined14 [71% response rate, after exclusion of 168 ineligible participants who were largely not current Jerusalem residents, were pregnant or were within 3 months of delivery, had a serious incapacitating illness or had died]. In 2009–11 (Visit 6, age 48–52),15–17 507 of the 631 Visit 5 participants (82% response rate among those eligible) underwent cognitive function assessment. We were unable to gain access to and thus adjust for a routine baseline cognitive measure that was taken as part of the compulsory military examination at age 17. FPG, OGTT, insulin, and insulin resistance were determined at Visit 4 (mean age 30 years); FPG and OGTT were repeated at Visit 5. The current analysis was undertaken on 505 Visit 6 non-diabetic individuals after excluding 2 patients with diabetes at baseline (0.4%). A comparison of Visit 1 characteristics between the Visit 6 participants (n = 505) and nonparticipants initially examined at Visit 1 (n = 8139) showed that the sex-specific distributions of body mass index (BMI), country of origin, education and plasma lipids and lipoproteins were very similar (Supplementary table S1). We also compared the Visit 6 participants who had a parent with a documented history of coronary heart disease (CHD) vs. those who did not for their sex-specific BMI, sociodemographic characteristics and lipid profile determined at Visit 4 to assess the effect of the v4 sampling scheme on generalizability (not shown). Other than cholesterol and triglyceride values that were higher in male participants and lower in female participants with parental CHD, distributions were similar. Ethical approval was obtained from the Hadassah Medical Center Helsinki Committee. Signed informed consent was obtained from all individual participants included in the study. Assessment of cognitive function Cognitive functions were assessed through a battery of NeuroTrax computerized cognitive tests (NeuroTrax Corp., Modiin, Israel). The battery was designed to evaluate performance in about 30 min (0: 32 ± 0: 04 h in our study) across an array of cognitive domains known to deteriorate during aging (including memory, executive function, visual spatial processing, attention, and information processing speed). It provides measurements of accuracy and response time in milliseconds and has been shown to be valid18,19 and reliable.18 Several of these tests are based on common neuropsychological paradigms [including the Benton Visual Retention Test, Brief Visuospatial Memory Test, Tova, Stroop, and subsets of WAIS-III (Wechsler Adult Intelligence Scale, 3rd ed.)] and have been previously used in clinical settings, as well as in studies of normal aging relating these neuropsychological measurements to genetic findings20 and brain imaging parameters,21 and have been used in middle-aged adults.20–24 A detailed description of the tests can be found in the online Supplementary data S1 and table S2, comparing individuals with hyperglycemia vs. those with normal glucose. Assessment of plasma glucose levels Plasma glucose was determined on both 12 h fasting samples and 2 h after a standard 75 g glucose challenge at Visit 4 (mean age 30 years) and Visit 5 (mean age 43 years) by standard enzymatic techniques. Diabetes at baseline (Visit 4), defined as the use of insulin or oral medications or a FPG ≥7.0 mmol/l (≥ 126 mg/dl) or OGTT ≥ 11.1 mmol/l (≥ 200 mg/dl), led to the exclusion of 2 participants. Hyperglycaemia was defined as FPG ≥5.6 mmol/l (≥ 100 mg/dl and <126 mg/dl). Assessment of insulin resistance 12-h fasting insulin was measured by radioimmunoassay. Insulin resistance at Visit 4 was estimated using the HOMA-IR. Assessment of covariates Socio-demographic characteristics consisted of age, sex, origin (father’s country of birth grouped by continent: Europe, Asia, North Africa, and Israel), highest educational attainment (university degree, high school graduate, incomplete high school, elementary), religiosity (ultraorthodox, religious, traditional, secular), childhood socio-economic position (SEP) (2 measures based on father’s occupation modified from the Israel Central Bureau of Statistics (ICBS)12 and the Vered Kraus occupational prestige based scale,12 both obtained at Visit 1, and adult SEP (2 measures based on the modified ICBS ranking and the MacArthur Scale of Subjective Social Status25). Depressive and anxiety symptoms were measured at Visit 6 using a translated Hebrew version of the Hospital Anxiety and Depression Scale (HADS) (two 7-item independent subscales).26 Measures of BMI (in kg/m2), blood pressure, health behaviors and biochemistry were obtained at baseline. Blood pressure was taken as the mean of the last 2 of 3 seated measurements using a standard mercury sphygmomanometer after 5 min of quiet rest. Health behaviors consisted of cigarette smoking (lifetime pack-years), alcohol intake (≤ once a week, and a median split of number of units per week as two dummy variables of low and high intake), and vigorous physical leisure-time activity for at least 20 min causing sweating and shortness of breath (yes/no). Plasma total cholesterol, HDL-C and triglycerides were measured on 12-h fasted samples by standard enzymatic techniques at Visits 1, 3, 4 and 5 at mean ages 17, 21, 30 and 43 years. LDL-C was computed by the Friedewald method. Inflammation markers assessed at baseline were plasma concentrations of C-reactive protein (CRP) (by ELISA), fibrinogen (Clauss method), and the white blood cell count (Beckman Coulter Counter). Plasma GlycA, a novel protein glycan inflammatory biomarker (determined by NMR), was quantified in both baseline (mean age 30 years) and follow-up (mean age 43 years) samples.15,27 Serum homocysteine was determined at baseline using HPLC with fluorometric detection. Statistical analysis Raw cognitive outcome measures (i.e. response time, accuracy and composite scores) were z-standardized to permit averaging across the different scales of measurement. Timed measures (response time and response time SD) were multiplied by −1 so that higher values indicate better performance. The z-standardized measures were then averaged to produce five scores, each indexing a different cognitive domain: memory, attention, executive function, visual spatial, and information processing speed. A summary global cognitive score, computed as the average of the 5 domain scores, was treated as the main dependent variable. Cognitive scores with negatively skewed distributions (global, memory and attention) were Box-Cox power transformed (λ = −0.5, i.e. inverse square-root transformation), to achieve an approximately Gaussian distribution, subsequent to reflection (computed by subtracting each value of a negatively skewed score from a constant). FPG at baseline was normally distributed. FPG was treated categorically as FPG ≥5.6 mmol/l (5.6–6.9 mmol/l) vs. FPG <5.6 mmol/l and as a continuous variable to test for trend. Secondarily, we also grouped FPG as <4.4, 4.4–4.9, 5.0–5.5, and ≥5.6. Multiple linear regression models were used to examine the associations between each glucose and insulin resistance measure (independent variables) and global cognitive function, the primary outcome. These models were repeated separately for each of the 5 cognitive domains to assess to which component(s) the association with global function can be attributed (secondary outcomes). Regression coefficients are reported as standardized Betas. Odds ratios (ORs) and 95% confidence intervals (CI) for the association of glucose and insulin resistance with poor cognitive function were computed from logistic models. A score in the lowest quintile was regarded as relatively poor cognitive performance, as has been previously defined by others,28 and was compared with the upper 4 quintiles grouped. Nominal two-sided p-values are reported. Sex interaction with hyperglycemia on the main global cognitive function outcome, tested in a separate regression model using a multiplicative term, showed no evidence of interaction (P = 0.77); consequently, we did not further stratify the statistical models by sex. Sensitivity analyses were performed on global cognitive function with fasting plasma glucose restricted to <6.1 mmol/l. In all analyses, we accounted for the sampling scheme at Visit 4 by the introduction of a dichotomous (0, 1) term representing the random sample vs. the parental CHD sample, despite that its incorporation did not affect the associations (not shown). Additional stratified analysis comparing cases with a documented positive family history of CHD vs. those without parental CHD, showed stronger associations in the former group of cases; however, the parental CHD interactions were not statistically significant (all P > 0.7). Consequently, all analyses were based on the full cohort and were adjusted for the sampling scheme. Analyses were adjusted for age, sex, educational level, origin, religiosity, SEP in childhood (ICBS ranking), adult SEP (ICBS ranking) and cigarette smoking, leisure-time vigorous activity, alcohol intake, BMI, total cholesterol, HDL-cholesterol and GlycA measured at baseline. These were selected on the basis of their confounding effect size on the Beta coefficient of the bivariate baseline FPG-global cognition association (≥5%). Depression, anxiety, systolic and diastolic blood pressure, LDL-cholesterol, triglycerides, C-reactive protein, white blood cell count, fibrinogen and homocysteine showed no material confounding effect and were not included. In a sensitivity analysis using backward stepwise regression, we reduced the selected covariate set to sex, educational level, adult SEP and BMI, as the excluded group of covariates did not affect the Beta coefficient of the full multivariable-adjusted model (i.e. the effect size for the excluded group was below 5%). The association of FPG with global cognitive function was also analysed in a sensitivity analysis to examine the effect of controlling for waist-to-hip ratio rather than BMI. To avoid loss of observations in the multivariable analyses, missing values were replaced with non-missing median values (adult SEP, n = 5; leisure-time vigorous activity, n = 1) or for GlycA (n = 20) by imputation using a backward stepwise linear regression procedure applied to variables predicting GlycA (P for removal >0.2). A complete case analysis slightly attenuated the associations. Statistical analyses were carried out using SPSS v21.0 (IBM Corp., Armonk, NY). Power Setting an α value of 0.05 (2-tailed), this sample had a power of 0.9 to detect a correlation of 0.15 between FPG and cognition, and given a 0.16 prevalence of hyperglycemia, it had a power of 0.8 to detect an OR of 2.1 in a logistic model predicting poor cognitive function (lowest fifth) and a difference of about a 1/3 SD in the global cognitive score in a linear regression model. Results Characteristics of the study sample at baseline are presented in table 1. Participants were aged 28–32 at baseline and 41–46 at follow-up with a range of 12–16 years of follow-up (13.1 ± .7); 33% were women, and 54% were high school or university graduates. Mean BMI was high and mean HDL-cholesterol was low as has been reported.14 Alcohol intake was low as was leisure time vigorous activity and about one-third of the sample smoked. Table 1 Characteristics of the study sample at baseline: the Jerusalem LRC longitudinal study, 1976–2011 Characteristics Total n 505 Socio-demographic variables Age (years) (range) 30.1± 0.8a (28.1–32.1) Female (%) 32.5 Country of birth (%)     Israel 22.4     Europe 22.6     Asia 29.5     N. Africa 25.5 Religiosity (%)     Ultra-orthodox 7.7     Religious 19.2     Traditional 31.9     Secular 41.2 Education, highest level (%)     University graduate 32.5     High school graduate 21.6     High school not graduated (9–12 years) 40.2     Elementary school (≤ 8 years) 5.7 Adult SEP (ICBS ranking)b 2.7 ± 1.2 Adult SEP (MacArthur Scale)c 7.2 ± 1.5 Childhood SEP (ICBS ranking)b 3.8 ± 1.6 Childhood SEP (Vered Kraus Scale)c 44.5 ± 30.1 Social mobilityd 1.1 ± 1.6 Glucose regulation FPG (mmol/l) (range) [median, IQR] 5.1 ± 0.5 (3.6–6.8) [5.1, 4.8–5.4]     FPG change (mmol/l)e 0.4 ± 0.8 OGTT (mmol/l) 5.0 ± 1.2     Change in OGTT (mmol/l)e 0.7 ± 1.7 Hyperglycemia (%)f 16.2     Fasting plasma insulin (mlU/l) 17.0 ± 2.0 Insulin resistance     HOMA-IR (mlU/l*mmol/l)g 3.9 ± 1.9 Anthropometric and blood pressure     BMI (kg/m2) 24.7 ± 3.6     Systolic blood pressure (mmHg) 112 ± 10     Diastolic blood pressure (mmHg) 68 ± 9 Psychosocial variables     Depressive symptoms score (HADS 0–21)h 3.6 ± 2.9     Anxiety symptoms score (HADS 0–21)h 5.6 ± 3.7 Lifestyle variables at baseline Leisure-time vigorous activity (%)i 21.6 Alcohol intake of ≥once/week (%) 39.4     Low intake (units/week)j 1.4 ± 0.5     High intake (units/week)j 5.8 ± 2.8 Pack-years (whole sample) 4.5 ± 6.5     Among ever smoked 9.5 ± 6.5 % current smokers 36.3 Biochemistry Plasma Lipids (mmol/l)     Total cholesterol 4.4 ± 0.8     HDL-cholesterol 1.0 ± 0.3     Non-HDL-cholesterol 3.4 ± 0.9     LDL-cholesterolk 2.7 ± 0.7     Triglycerides 1.4 ± 0.9 Homocysteine (µmol/l) 12.2 ± 8.5 Inflammatory variables     C-reactive protein (mg/l) 2.2 ± 3.3     White blood cell count 6841 ± 1694     Fibrinogen (mg/dl) 233.5 ± 54.8     GlycA (µmol/l) 265 ± 52 Characteristics Total n 505 Socio-demographic variables Age (years) (range) 30.1± 0.8a (28.1–32.1) Female (%) 32.5 Country of birth (%)     Israel 22.4     Europe 22.6     Asia 29.5     N. Africa 25.5 Religiosity (%)     Ultra-orthodox 7.7     Religious 19.2     Traditional 31.9     Secular 41.2 Education, highest level (%)     University graduate 32.5     High school graduate 21.6     High school not graduated (9–12 years) 40.2     Elementary school (≤ 8 years) 5.7 Adult SEP (ICBS ranking)b 2.7 ± 1.2 Adult SEP (MacArthur Scale)c 7.2 ± 1.5 Childhood SEP (ICBS ranking)b 3.8 ± 1.6 Childhood SEP (Vered Kraus Scale)c 44.5 ± 30.1 Social mobilityd 1.1 ± 1.6 Glucose regulation FPG (mmol/l) (range) [median, IQR] 5.1 ± 0.5 (3.6–6.8) [5.1, 4.8–5.4]     FPG change (mmol/l)e 0.4 ± 0.8 OGTT (mmol/l) 5.0 ± 1.2     Change in OGTT (mmol/l)e 0.7 ± 1.7 Hyperglycemia (%)f 16.2     Fasting plasma insulin (mlU/l) 17.0 ± 2.0 Insulin resistance     HOMA-IR (mlU/l*mmol/l)g 3.9 ± 1.9 Anthropometric and blood pressure     BMI (kg/m2) 24.7 ± 3.6     Systolic blood pressure (mmHg) 112 ± 10     Diastolic blood pressure (mmHg) 68 ± 9 Psychosocial variables     Depressive symptoms score (HADS 0–21)h 3.6 ± 2.9     Anxiety symptoms score (HADS 0–21)h 5.6 ± 3.7 Lifestyle variables at baseline Leisure-time vigorous activity (%)i 21.6 Alcohol intake of ≥once/week (%) 39.4     Low intake (units/week)j 1.4 ± 0.5     High intake (units/week)j 5.8 ± 2.8 Pack-years (whole sample) 4.5 ± 6.5     Among ever smoked 9.5 ± 6.5 % current smokers 36.3 Biochemistry Plasma Lipids (mmol/l)     Total cholesterol 4.4 ± 0.8     HDL-cholesterol 1.0 ± 0.3     Non-HDL-cholesterol 3.4 ± 0.9     LDL-cholesterolk 2.7 ± 0.7     Triglycerides 1.4 ± 0.9 Homocysteine (µmol/l) 12.2 ± 8.5 Inflammatory variables     C-reactive protein (mg/l) 2.2 ± 3.3     White blood cell count 6841 ± 1694     Fibrinogen (mg/dl) 233.5 ± 54.8     GlycA (µmol/l) 265 ± 52 Missing data: adult SEP (ICBS ranking) (n = 5), adult SEP (MacArthur Scale) (n = 11), childhood SEP (Vered Kraus Scale) (n = 4), social mobility (n = 5), hyperglycemia (n = 7), HOMA-IR (n = 1), depressive symptoms score (n = 4), anxiety symptoms score (n = 4), leisure-time vigorous activity (n = 1), LDL-cholesterol (n = 7), homocysteine (n = 16), C-reactive protein (n = 14), white blood cell count (n = 12), fibrinogen (n = 43), GlycA (n = 20). LRC, Lipid Research Clinic; FPG, fasting plasma glucose; OGTT, 2 h post-oral challenge plasma glucose; HOMA-IR, homeostatic model assessment for insulin resistance; BMI, body mass index. a Mean ± SD (continuous/ interval variables). b An higher value infers a lower SEP. Scale range from 1 to 6. c An higher value infers a higher SEP. Our Vered Kraus scores range from 2.60 to 98.96; MacArthur Scale range from 1 to 10. d Computed by subtracting ICBS-based SEP in adulthood from SEP in childhood (both with a range from 1 (upper) to 6 (lower)). Range of social mobility score was from −5 (maximal downward drifting) to +5 (maximal upward mobility). No change/upward mobility corresponds to scores ≥0, whereas downward drifting corresponds to scores <0. e Computed by subtracting baseline measurement at ages 28–32 from follow-up measurement at ages 41–46. f Defined as FPG ≥100 mg/dl. g Calculated as the product of fasting serum glucose (mmol/l) x fasting serum insulin (mlU/l) divided by 22.5. (Diabetes Care. 2004; 27: 1487). h 7-items each scored 0–3. Scale range from 0 to 21. Cronbach's alphas were adequate at .71 and .785 for the depression and the anxiety subscale, respectively. i Exercise for at least 20 min causing heavy breathing and sweating. j Low/high intake, according to median split of alcohol intake among consumers of ≥once/week. k Computed by the Friedewald method; not computed for 7 men at age 30 and 11 participants (10 men and 1 woman) at age 43 with triglycerides > 400 mg/dl. Table 1 Characteristics of the study sample at baseline: the Jerusalem LRC longitudinal study, 1976–2011 Characteristics Total n 505 Socio-demographic variables Age (years) (range) 30.1± 0.8a (28.1–32.1) Female (%) 32.5 Country of birth (%)     Israel 22.4     Europe 22.6     Asia 29.5     N. Africa 25.5 Religiosity (%)     Ultra-orthodox 7.7     Religious 19.2     Traditional 31.9     Secular 41.2 Education, highest level (%)     University graduate 32.5     High school graduate 21.6     High school not graduated (9–12 years) 40.2     Elementary school (≤ 8 years) 5.7 Adult SEP (ICBS ranking)b 2.7 ± 1.2 Adult SEP (MacArthur Scale)c 7.2 ± 1.5 Childhood SEP (ICBS ranking)b 3.8 ± 1.6 Childhood SEP (Vered Kraus Scale)c 44.5 ± 30.1 Social mobilityd 1.1 ± 1.6 Glucose regulation FPG (mmol/l) (range) [median, IQR] 5.1 ± 0.5 (3.6–6.8) [5.1, 4.8–5.4]     FPG change (mmol/l)e 0.4 ± 0.8 OGTT (mmol/l) 5.0 ± 1.2     Change in OGTT (mmol/l)e 0.7 ± 1.7 Hyperglycemia (%)f 16.2     Fasting plasma insulin (mlU/l) 17.0 ± 2.0 Insulin resistance     HOMA-IR (mlU/l*mmol/l)g 3.9 ± 1.9 Anthropometric and blood pressure     BMI (kg/m2) 24.7 ± 3.6     Systolic blood pressure (mmHg) 112 ± 10     Diastolic blood pressure (mmHg) 68 ± 9 Psychosocial variables     Depressive symptoms score (HADS 0–21)h 3.6 ± 2.9     Anxiety symptoms score (HADS 0–21)h 5.6 ± 3.7 Lifestyle variables at baseline Leisure-time vigorous activity (%)i 21.6 Alcohol intake of ≥once/week (%) 39.4     Low intake (units/week)j 1.4 ± 0.5     High intake (units/week)j 5.8 ± 2.8 Pack-years (whole sample) 4.5 ± 6.5     Among ever smoked 9.5 ± 6.5 % current smokers 36.3 Biochemistry Plasma Lipids (mmol/l)     Total cholesterol 4.4 ± 0.8     HDL-cholesterol 1.0 ± 0.3     Non-HDL-cholesterol 3.4 ± 0.9     LDL-cholesterolk 2.7 ± 0.7     Triglycerides 1.4 ± 0.9 Homocysteine (µmol/l) 12.2 ± 8.5 Inflammatory variables     C-reactive protein (mg/l) 2.2 ± 3.3     White blood cell count 6841 ± 1694     Fibrinogen (mg/dl) 233.5 ± 54.8     GlycA (µmol/l) 265 ± 52 Characteristics Total n 505 Socio-demographic variables Age (years) (range) 30.1± 0.8a (28.1–32.1) Female (%) 32.5 Country of birth (%)     Israel 22.4     Europe 22.6     Asia 29.5     N. Africa 25.5 Religiosity (%)     Ultra-orthodox 7.7     Religious 19.2     Traditional 31.9     Secular 41.2 Education, highest level (%)     University graduate 32.5     High school graduate 21.6     High school not graduated (9–12 years) 40.2     Elementary school (≤ 8 years) 5.7 Adult SEP (ICBS ranking)b 2.7 ± 1.2 Adult SEP (MacArthur Scale)c 7.2 ± 1.5 Childhood SEP (ICBS ranking)b 3.8 ± 1.6 Childhood SEP (Vered Kraus Scale)c 44.5 ± 30.1 Social mobilityd 1.1 ± 1.6 Glucose regulation FPG (mmol/l) (range) [median, IQR] 5.1 ± 0.5 (3.6–6.8) [5.1, 4.8–5.4]     FPG change (mmol/l)e 0.4 ± 0.8 OGTT (mmol/l) 5.0 ± 1.2     Change in OGTT (mmol/l)e 0.7 ± 1.7 Hyperglycemia (%)f 16.2     Fasting plasma insulin (mlU/l) 17.0 ± 2.0 Insulin resistance     HOMA-IR (mlU/l*mmol/l)g 3.9 ± 1.9 Anthropometric and blood pressure     BMI (kg/m2) 24.7 ± 3.6     Systolic blood pressure (mmHg) 112 ± 10     Diastolic blood pressure (mmHg) 68 ± 9 Psychosocial variables     Depressive symptoms score (HADS 0–21)h 3.6 ± 2.9     Anxiety symptoms score (HADS 0–21)h 5.6 ± 3.7 Lifestyle variables at baseline Leisure-time vigorous activity (%)i 21.6 Alcohol intake of ≥once/week (%) 39.4     Low intake (units/week)j 1.4 ± 0.5     High intake (units/week)j 5.8 ± 2.8 Pack-years (whole sample) 4.5 ± 6.5     Among ever smoked 9.5 ± 6.5 % current smokers 36.3 Biochemistry Plasma Lipids (mmol/l)     Total cholesterol 4.4 ± 0.8     HDL-cholesterol 1.0 ± 0.3     Non-HDL-cholesterol 3.4 ± 0.9     LDL-cholesterolk 2.7 ± 0.7     Triglycerides 1.4 ± 0.9 Homocysteine (µmol/l) 12.2 ± 8.5 Inflammatory variables     C-reactive protein (mg/l) 2.2 ± 3.3     White blood cell count 6841 ± 1694     Fibrinogen (mg/dl) 233.5 ± 54.8     GlycA (µmol/l) 265 ± 52 Missing data: adult SEP (ICBS ranking) (n = 5), adult SEP (MacArthur Scale) (n = 11), childhood SEP (Vered Kraus Scale) (n = 4), social mobility (n = 5), hyperglycemia (n = 7), HOMA-IR (n = 1), depressive symptoms score (n = 4), anxiety symptoms score (n = 4), leisure-time vigorous activity (n = 1), LDL-cholesterol (n = 7), homocysteine (n = 16), C-reactive protein (n = 14), white blood cell count (n = 12), fibrinogen (n = 43), GlycA (n = 20). LRC, Lipid Research Clinic; FPG, fasting plasma glucose; OGTT, 2 h post-oral challenge plasma glucose; HOMA-IR, homeostatic model assessment for insulin resistance; BMI, body mass index. a Mean ± SD (continuous/ interval variables). b An higher value infers a lower SEP. Scale range from 1 to 6. c An higher value infers a higher SEP. Our Vered Kraus scores range from 2.60 to 98.96; MacArthur Scale range from 1 to 10. d Computed by subtracting ICBS-based SEP in adulthood from SEP in childhood (both with a range from 1 (upper) to 6 (lower)). Range of social mobility score was from −5 (maximal downward drifting) to +5 (maximal upward mobility). No change/upward mobility corresponds to scores ≥0, whereas downward drifting corresponds to scores <0. e Computed by subtracting baseline measurement at ages 28–32 from follow-up measurement at ages 41–46. f Defined as FPG ≥100 mg/dl. g Calculated as the product of fasting serum glucose (mmol/l) x fasting serum insulin (mlU/l) divided by 22.5. (Diabetes Care. 2004; 27: 1487). h 7-items each scored 0–3. Scale range from 0 to 21. Cronbach's alphas were adequate at .71 and .785 for the depression and the anxiety subscale, respectively. i Exercise for at least 20 min causing heavy breathing and sweating. j Low/high intake, according to median split of alcohol intake among consumers of ≥once/week. k Computed by the Friedewald method; not computed for 7 men at age 30 and 11 participants (10 men and 1 woman) at age 43 with triglycerides > 400 mg/dl. Median values for FPG and post-challenge glucose at baseline (mean age 30 years) were 5.1 [interquartile range (IQR), 4.8–5.4) mmol/l] and 4.8 (IQR, 4.2–5.6) mmol/l, respectively, and increased over the 13-year follow-up period (mean age 43 years) to 5.4 (IQR, 5.1–5.7) mmol/l and 5.4 (IQR, 4.7–6.3) mmol/l, respectively. Sex-adjusted Spearman tracking correlations over the 13 year average follow-up were rho=.397 and rho=.311 for FPG and OGTT levels, respectively. 16.2% (n = 82) participants had hyperglycemia defined as FPG ≥5.6 mmo/l at baseline (see Supplementary figure S1). The mean FPG of 5.1 mmol/l in the LRC cohort members aged 28–32 (5.2 mmol/l and 4.9 mmol/l in men and women, respectively) is in accordance with the mean values of 5.1 mmol/l and 4.8 mmol/l among white men and women of the CARDIA Study at ages 25–37.29 Table 2 presents the standardized Betas of cognitive function in midlife according to FPG at baseline. Hyperglycemia was independently associated with poorer global cognitive function in midlife (FPG ≥5.6 mmol/l vs. FPG <5.6 mmol/l, multivariable-adjusted standardized Beta= −0.121, P = 0.002; and FPG ≥5.6 mmol/l vs. FPG <4.4 mmol/l, multivariable-adjusted standardized Beta= −0.176, P = 0.007). A test for a linear trend with FPG entered as a continuous variable yielded a P values of .016. The domains that contributed to these associations were visual spatial and attention (table 2, Model 2). Table 2 Linear regression of the association of fasting plasma glucose and insulin resistance in young adulthood (∼30 years) with cognitive function in midlife (∼50 years): the Jerusalem LRC longitudinal study Models Globala Attentiona Information processing speed Executive Visual spatial Memorya (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl)vs. <5.6 mmol/l (100 mg/dl) −0.148** −0.121** −0.104* −0.083* −0.079 −0.061 −0.055 −0.020 −0.199** −0.185** −0.098* −0.081 (0.001) (0.002) (0.021) (0.050) (0.092) (0.162) (0.223) (0.638) (<0.001) (<0.001) (0.031) (0.066) 2. FPG (mmol/l) (continuous) −0.122** −0.094* −0.109* −0.090* −0.053 −0.043 −0.092* 0.068 −0.141** −0.130** −0.095* −0.058 (0.007) (0.016) (0.016) (0.034) (0.255) (0.321) (0.043) (0.119) (0.002) (0.002) (0.039) (0.191) 3. FPG (3 df)…(<80 mg/dl)(n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.032)b (0.034)b (0.084)b (0.101)b (0.568)b (0.627)b (0.113)b (0.192)b (0.004)b (0.002)b (0.054)b (0.140)b  4.4-4.9 mmol/l (80–89 mg/dl)(n = 172) −0.104 −0.102 −0.085 −0.073 −0.026 −0.002 −0.116 −0.127 −0.132 −0.121 −0.068 −0.064 (0.220) (0.179) (0.341) (0.381) (0.784) (0.985) (0.194) (0.133) (0.128) (0.129) (0.455) (0.462) 5.0-5.5 mmol/l (90–99 mg/dl)(n = 218) −0.059 −0.063 −0.064 −0.066 0.026 0.030 −0.128 −0.149 −0.071 −0.079 −0.076 −0.056 (0.517) (0.420) (0.488) (0.442) (0.784) (0.733) (0.165) (0.089) (0.427) (0.336) (0.419) (0.532) 5.6+ mmol/l (100+ mg/dl)(n = 82) −0.204** −0.176** −0.155* −0.131 −0.076 −0.049 −0.142 −0.119 −0.267** −0.252** −0.150 −0.123 (0.007) (0.007) (0.043) (0.054) (0.329) (0.489) (0.064) (0.103) (<.001) (<.001) (0.056) (0.100) 4. Insulin resistanceHOMA (mlU/l*mmol/l)a,c (continuous) −0.065 −0.012 −0.073 −0.040 −0.045 −0.019 −0.046 0.033 −0.041 0.002 −0.012 0.038 (0.157) (0.787) (0.112) (0.419) (0.347) (0.707) (0.309) (0.509) (0.369) (0.966) (0.801) (0.459) Models Globala Attentiona Information processing speed Executive Visual spatial Memorya (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl)vs. <5.6 mmol/l (100 mg/dl) −0.148** −0.121** −0.104* −0.083* −0.079 −0.061 −0.055 −0.020 −0.199** −0.185** −0.098* −0.081 (0.001) (0.002) (0.021) (0.050) (0.092) (0.162) (0.223) (0.638) (<0.001) (<0.001) (0.031) (0.066) 2. FPG (mmol/l) (continuous) −0.122** −0.094* −0.109* −0.090* −0.053 −0.043 −0.092* 0.068 −0.141** −0.130** −0.095* −0.058 (0.007) (0.016) (0.016) (0.034) (0.255) (0.321) (0.043) (0.119) (0.002) (0.002) (0.039) (0.191) 3. FPG (3 df)…(<80 mg/dl)(n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.032)b (0.034)b (0.084)b (0.101)b (0.568)b (0.627)b (0.113)b (0.192)b (0.004)b (0.002)b (0.054)b (0.140)b  4.4-4.9 mmol/l (80–89 mg/dl)(n = 172) −0.104 −0.102 −0.085 −0.073 −0.026 −0.002 −0.116 −0.127 −0.132 −0.121 −0.068 −0.064 (0.220) (0.179) (0.341) (0.381) (0.784) (0.985) (0.194) (0.133) (0.128) (0.129) (0.455) (0.462) 5.0-5.5 mmol/l (90–99 mg/dl)(n = 218) −0.059 −0.063 −0.064 −0.066 0.026 0.030 −0.128 −0.149 −0.071 −0.079 −0.076 −0.056 (0.517) (0.420) (0.488) (0.442) (0.784) (0.733) (0.165) (0.089) (0.427) (0.336) (0.419) (0.532) 5.6+ mmol/l (100+ mg/dl)(n = 82) −0.204** −0.176** −0.155* −0.131 −0.076 −0.049 −0.142 −0.119 −0.267** −0.252** −0.150 −0.123 (0.007) (0.007) (0.043) (0.054) (0.329) (0.489) (0.064) (0.103) (<.001) (<.001) (0.056) (0.100) 4. Insulin resistanceHOMA (mlU/l*mmol/l)a,c (continuous) −0.065 −0.012 −0.073 −0.040 −0.045 −0.019 −0.046 0.033 −0.041 0.002 −0.012 0.038 (0.157) (0.787) (0.112) (0.419) (0.347) (0.707) (0.309) (0.509) (0.369) (0.966) (0.801) (0.459) Model 1: Adjusted for sex and age at cognitive assessment. Model 2: Additionally adjusted for educational level, origin, religiosity, childhood SEP (ICBS ranking), adult SEP (ICBS ranking), and cigarette pack-yrs, leisure-time vigorous activity, alcohol intake, BMI, total cholesterol, HDL-cholesterol and GlycA, measured at baseline (at ages 28–32), and for the sampling scheme. Missing data: HOMA-IR (n = 1), adult SEP (n = 5), leisure-time vigorous activity (n = 1) and GlycA (n = 20). Missing values were replaced with non-missing median values (adult SEP, leisure-time vigorous activity) or by multiple linear regression imputation (baseline GlycA). LRC, Lipid Research Clinic; FPG, fasting plasma glucose; HOMA, homeostatic model assessment for insulin resistance. Beta = standardized regression coefficient. Numbers represent Beta coefficients, i.e. change in cognitive score per SD of variable. a Box-Cox transformed cognitive z-scores (λ= - 0.5), HOMA-IR (λ = 0). b Test for trend for the FPG groups introduced as an ordinal variable (1 df). c The HOMA-IR formula was defined as fasting insulin (mlU/l) multiplied by fasting glucose (mmol/l) divided by 22.5. * P ≤ 0.05 ** P ≤ 0.01 Table 2 Linear regression of the association of fasting plasma glucose and insulin resistance in young adulthood (∼30 years) with cognitive function in midlife (∼50 years): the Jerusalem LRC longitudinal study Models Globala Attentiona Information processing speed Executive Visual spatial Memorya (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl)vs. <5.6 mmol/l (100 mg/dl) −0.148** −0.121** −0.104* −0.083* −0.079 −0.061 −0.055 −0.020 −0.199** −0.185** −0.098* −0.081 (0.001) (0.002) (0.021) (0.050) (0.092) (0.162) (0.223) (0.638) (<0.001) (<0.001) (0.031) (0.066) 2. FPG (mmol/l) (continuous) −0.122** −0.094* −0.109* −0.090* −0.053 −0.043 −0.092* 0.068 −0.141** −0.130** −0.095* −0.058 (0.007) (0.016) (0.016) (0.034) (0.255) (0.321) (0.043) (0.119) (0.002) (0.002) (0.039) (0.191) 3. FPG (3 df)…(<80 mg/dl)(n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.032)b (0.034)b (0.084)b (0.101)b (0.568)b (0.627)b (0.113)b (0.192)b (0.004)b (0.002)b (0.054)b (0.140)b  4.4-4.9 mmol/l (80–89 mg/dl)(n = 172) −0.104 −0.102 −0.085 −0.073 −0.026 −0.002 −0.116 −0.127 −0.132 −0.121 −0.068 −0.064 (0.220) (0.179) (0.341) (0.381) (0.784) (0.985) (0.194) (0.133) (0.128) (0.129) (0.455) (0.462) 5.0-5.5 mmol/l (90–99 mg/dl)(n = 218) −0.059 −0.063 −0.064 −0.066 0.026 0.030 −0.128 −0.149 −0.071 −0.079 −0.076 −0.056 (0.517) (0.420) (0.488) (0.442) (0.784) (0.733) (0.165) (0.089) (0.427) (0.336) (0.419) (0.532) 5.6+ mmol/l (100+ mg/dl)(n = 82) −0.204** −0.176** −0.155* −0.131 −0.076 −0.049 −0.142 −0.119 −0.267** −0.252** −0.150 −0.123 (0.007) (0.007) (0.043) (0.054) (0.329) (0.489) (0.064) (0.103) (<.001) (<.001) (0.056) (0.100) 4. Insulin resistanceHOMA (mlU/l*mmol/l)a,c (continuous) −0.065 −0.012 −0.073 −0.040 −0.045 −0.019 −0.046 0.033 −0.041 0.002 −0.012 0.038 (0.157) (0.787) (0.112) (0.419) (0.347) (0.707) (0.309) (0.509) (0.369) (0.966) (0.801) (0.459) Models Globala Attentiona Information processing speed Executive Visual spatial Memorya (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl)vs. <5.6 mmol/l (100 mg/dl) −0.148** −0.121** −0.104* −0.083* −0.079 −0.061 −0.055 −0.020 −0.199** −0.185** −0.098* −0.081 (0.001) (0.002) (0.021) (0.050) (0.092) (0.162) (0.223) (0.638) (<0.001) (<0.001) (0.031) (0.066) 2. FPG (mmol/l) (continuous) −0.122** −0.094* −0.109* −0.090* −0.053 −0.043 −0.092* 0.068 −0.141** −0.130** −0.095* −0.058 (0.007) (0.016) (0.016) (0.034) (0.255) (0.321) (0.043) (0.119) (0.002) (0.002) (0.039) (0.191) 3. FPG (3 df)…(<80 mg/dl)(n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.032)b (0.034)b (0.084)b (0.101)b (0.568)b (0.627)b (0.113)b (0.192)b (0.004)b (0.002)b (0.054)b (0.140)b  4.4-4.9 mmol/l (80–89 mg/dl)(n = 172) −0.104 −0.102 −0.085 −0.073 −0.026 −0.002 −0.116 −0.127 −0.132 −0.121 −0.068 −0.064 (0.220) (0.179) (0.341) (0.381) (0.784) (0.985) (0.194) (0.133) (0.128) (0.129) (0.455) (0.462) 5.0-5.5 mmol/l (90–99 mg/dl)(n = 218) −0.059 −0.063 −0.064 −0.066 0.026 0.030 −0.128 −0.149 −0.071 −0.079 −0.076 −0.056 (0.517) (0.420) (0.488) (0.442) (0.784) (0.733) (0.165) (0.089) (0.427) (0.336) (0.419) (0.532) 5.6+ mmol/l (100+ mg/dl)(n = 82) −0.204** −0.176** −0.155* −0.131 −0.076 −0.049 −0.142 −0.119 −0.267** −0.252** −0.150 −0.123 (0.007) (0.007) (0.043) (0.054) (0.329) (0.489) (0.064) (0.103) (<.001) (<.001) (0.056) (0.100) 4. Insulin resistanceHOMA (mlU/l*mmol/l)a,c (continuous) −0.065 −0.012 −0.073 −0.040 −0.045 −0.019 −0.046 0.033 −0.041 0.002 −0.012 0.038 (0.157) (0.787) (0.112) (0.419) (0.347) (0.707) (0.309) (0.509) (0.369) (0.966) (0.801) (0.459) Model 1: Adjusted for sex and age at cognitive assessment. Model 2: Additionally adjusted for educational level, origin, religiosity, childhood SEP (ICBS ranking), adult SEP (ICBS ranking), and cigarette pack-yrs, leisure-time vigorous activity, alcohol intake, BMI, total cholesterol, HDL-cholesterol and GlycA, measured at baseline (at ages 28–32), and for the sampling scheme. Missing data: HOMA-IR (n = 1), adult SEP (n = 5), leisure-time vigorous activity (n = 1) and GlycA (n = 20). Missing values were replaced with non-missing median values (adult SEP, leisure-time vigorous activity) or by multiple linear regression imputation (baseline GlycA). LRC, Lipid Research Clinic; FPG, fasting plasma glucose; HOMA, homeostatic model assessment for insulin resistance. Beta = standardized regression coefficient. Numbers represent Beta coefficients, i.e. change in cognitive score per SD of variable. a Box-Cox transformed cognitive z-scores (λ= - 0.5), HOMA-IR (λ = 0). b Test for trend for the FPG groups introduced as an ordinal variable (1 df). c The HOMA-IR formula was defined as fasting insulin (mlU/l) multiplied by fasting glucose (mmol/l) divided by 22.5. * P ≤ 0.05 ** P ≤ 0.01 Similarly, using logistic models to predict low ranked cognition (the lower fifth of the cognitive score distribution compared with the upper 4 quintiles grouped), an approximately 130% increased odds for low global cognitive function was evident per 1 mmol/l FPG increment at baseline (OR = 2.31, 95% CI = 1.30–4.13, P = 0.005). FPG ≥5.6 mmol/l (vs. <5.6 mmol/l) was associated with a 2-fold increased odds for low ranked global cognitive function (OR = 2.10, 95% CI = 1.09–4.06, P = 0.026). Compared with FPG <4.4 mmol/l, increasing levels of FPG conferred a graded increase in the odds for low ranked global cognitive function (p for trend=.006) (table 3, Model 2). Table 3 Logistic modeling of the association of fasting plasma glucose and insulin resistance in young adulthood [∼30 years) with low ranked cognitive function (lowest fifth) in midlife (∼50 years)]: the Jerusalem LRC longitudinal study Models Global Attention Information processing speed Executive Visual spatial Memory (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 OR OR OR OR OR OR OR OR OR OR OR OR [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl) vs. <5.6 mmol/l (100 mg/dl) 2.09** 2.10* 1.81* 1.76 1.25 0.966 1.35 1.00 2.44** 2.26* 1.82* 1.55 [1.21–3.61] [1.09–4.06] [1.04–3.16] [0.95–3.25] [0.68–2.31] [0.48–1.95] [0.75–2.42] [0.53–1.91] [1.37–4.34] [1.15–4.44] [1.04–3.17] [0.81–2.95] (0.008) (0.026) (0.036) (0.071) (0.473) (0.923) (0.318) (0.995) (0.002)* (0.019) (0.035) (0.184) 2. FPG (mmol/l) (continuous) 2.26** 2.31** 1.67* 1.60 1.006 0.911 1.39 1.16 1.89* 1.78* 1.63 124 [1.37–3.73] [1.30–4.13] [1.02–2.74 [0.94–2.72] [0.62–1.79] [0.51–1.64] [0.85–2.28] [0.69–1.96] [1.13–3.19] [1.00–3.17] [1.00–2.67] [0.72–2.16] (0.001) (0.005) (0.042) (0.084) (0.837) (0.755) (0.188) (0.576) (0.016) (0.050) (0.055) (0.437) 3. FPG (3 df)…(<80 mg/dl) (n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.007)b (0.006)b (0.097)b (0.121)b (0.913)b (0.789)b (0.456)b (0.894)b (0.021)b (0.034)b (0.050)b (0.244)b  4.4–4.9 mmol/l (80–89 mg/dl) (n = 172) 2.13 2.28 1.29 1.11 1.88 1.76 1.51 1.70 1.67 1.37 2.38 2.20 [0.61–7.43] [0.57–9.10] [0.46–3.59] [0.38–3.28] [0.62–5.75] [0.53–5.85] [0.54–4.21] [0.58–5.02] [0.54–5.17] [0.40–4.64] [0.68–8.32] [0.59–8.21] (0.238) (0.243) (0.633) (0.849) (0.268) (0.354) (0.433) (0.334) (0.370) (0.618) (0.173) (0.241) 5.0–5.5 mmol/l (90–99 mg/dl) n = 218) 2.42 3.13 1.27 1.20 1.31 1.31 1.33 1.45 1.54 1.52 2.23 1.89 [0.70–8.35] [0.80–12.28] [0.46–3.52] [0.41–3.50] [0.43–4.00] [0.40–4.33] [0.48–3.69] [0.50–4.23] [0.50–4.73] [0.45–5.11] [0.64–7.72] [0.51–7.01] (0.162) (0.102) (0.645) (0.736) (0.640) (0.658) (0.585) (0.495) (0.452) (0.498) (0.207) (0.339) 5.6+ mmol/l (100+ mg/dl) (n = 82) 4.60* 5.44* 2.28 2.03 1.87 1.40 1.85 1.51 3.78* 3.19 3.99* 3.00 [1.26–16.72] [1.29–22.87] [0.77–6.73] [0.65–6.35] [0.56–6.23] [0.38–5.11] [0.61–5.57] [0.47–4.81] [1.15–12.38] [0.88–11.64] [1.09–14.60] [0.75–11.97] (0.021) (0.021) (0.137) (0.226) (0.305) (0.613) (0.276) (0.491) (0.028) (0.079) (0.037) (0.121) 4. Insulin resistance HOMA (mlU/l*mmol/l)a,c (continuous) 1.32 0.79 1.02 0.78 1.23* 0.80 1.59 0.90 1.06 0.65 1.03 0.56 [0.78–2.26] [0.39–1.61] [0.60–1.74] [0.40–1.53] [0.71–2.15] [0.39–1.61] [0.92–2.72] [0.46–1.76] [0.61–1.85] [0.32–1.34] [0.60–1.76] [0.28–1.13] (0.303) (0.517) (0.94) (0.47) (0.46) (0.54) (0.90) (0.75) (0.83) (0.25) (0.92) (0.11) Models Global Attention Information processing speed Executive Visual spatial Memory (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 OR OR OR OR OR OR OR OR OR OR OR OR [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl) vs. <5.6 mmol/l (100 mg/dl) 2.09** 2.10* 1.81* 1.76 1.25 0.966 1.35 1.00 2.44** 2.26* 1.82* 1.55 [1.21–3.61] [1.09–4.06] [1.04–3.16] [0.95–3.25] [0.68–2.31] [0.48–1.95] [0.75–2.42] [0.53–1.91] [1.37–4.34] [1.15–4.44] [1.04–3.17] [0.81–2.95] (0.008) (0.026) (0.036) (0.071) (0.473) (0.923) (0.318) (0.995) (0.002)* (0.019) (0.035) (0.184) 2. FPG (mmol/l) (continuous) 2.26** 2.31** 1.67* 1.60 1.006 0.911 1.39 1.16 1.89* 1.78* 1.63 124 [1.37–3.73] [1.30–4.13] [1.02–2.74 [0.94–2.72] [0.62–1.79] [0.51–1.64] [0.85–2.28] [0.69–1.96] [1.13–3.19] [1.00–3.17] [1.00–2.67] [0.72–2.16] (0.001) (0.005) (0.042) (0.084) (0.837) (0.755) (0.188) (0.576) (0.016) (0.050) (0.055) (0.437) 3. FPG (3 df)…(<80 mg/dl) (n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.007)b (0.006)b (0.097)b (0.121)b (0.913)b (0.789)b (0.456)b (0.894)b (0.021)b (0.034)b (0.050)b (0.244)b  4.4–4.9 mmol/l (80–89 mg/dl) (n = 172) 2.13 2.28 1.29 1.11 1.88 1.76 1.51 1.70 1.67 1.37 2.38 2.20 [0.61–7.43] [0.57–9.10] [0.46–3.59] [0.38–3.28] [0.62–5.75] [0.53–5.85] [0.54–4.21] [0.58–5.02] [0.54–5.17] [0.40–4.64] [0.68–8.32] [0.59–8.21] (0.238) (0.243) (0.633) (0.849) (0.268) (0.354) (0.433) (0.334) (0.370) (0.618) (0.173) (0.241) 5.0–5.5 mmol/l (90–99 mg/dl) n = 218) 2.42 3.13 1.27 1.20 1.31 1.31 1.33 1.45 1.54 1.52 2.23 1.89 [0.70–8.35] [0.80–12.28] [0.46–3.52] [0.41–3.50] [0.43–4.00] [0.40–4.33] [0.48–3.69] [0.50–4.23] [0.50–4.73] [0.45–5.11] [0.64–7.72] [0.51–7.01] (0.162) (0.102) (0.645) (0.736) (0.640) (0.658) (0.585) (0.495) (0.452) (0.498) (0.207) (0.339) 5.6+ mmol/l (100+ mg/dl) (n = 82) 4.60* 5.44* 2.28 2.03 1.87 1.40 1.85 1.51 3.78* 3.19 3.99* 3.00 [1.26–16.72] [1.29–22.87] [0.77–6.73] [0.65–6.35] [0.56–6.23] [0.38–5.11] [0.61–5.57] [0.47–4.81] [1.15–12.38] [0.88–11.64] [1.09–14.60] [0.75–11.97] (0.021) (0.021) (0.137) (0.226) (0.305) (0.613) (0.276) (0.491) (0.028) (0.079) (0.037) (0.121) 4. Insulin resistance HOMA (mlU/l*mmol/l)a,c (continuous) 1.32 0.79 1.02 0.78 1.23* 0.80 1.59 0.90 1.06 0.65 1.03 0.56 [0.78–2.26] [0.39–1.61] [0.60–1.74] [0.40–1.53] [0.71–2.15] [0.39–1.61] [0.92–2.72] [0.46–1.76] [0.61–1.85] [0.32–1.34] [0.60–1.76] [0.28–1.13] (0.303) (0.517) (0.94) (0.47) (0.46) (0.54) (0.90) (0.75) (0.83) (0.25) (0.92) (0.11) Model 1: Adjusted for sex and age at cognitive assessment. Model 2: Additionally adjusted for educational level, origin, religiosity, childhood SEP, adult SEP (ICBS ranking) and cigarette pack-yrs, leisure-time vigorous activity, alcohol intake, BMI, total cholesterol, HDL-cholesterol and GlycA measured at baseline (at ages 28–32), and for the sampling scheme. Missing data: HOMA-IR (n = 1), adult SEP (n = 5), leisure-time vigorous activity (n = 1) and GlycA (n = 20). Missing values were replaced with non-missing median values (adult SEP, leisure-time vigorous activity) or by multiple linear regression imputation (baseline GlycA). LRC, Lipid Research Clinic; FPG, fasting plasma glucose; HOMA, homeostatic model assessment for insulin resistance. OR, odds ratio. CI, confidence interval. a Box-Cox transformed HOMA-IR (λ = 0). b Test for trend for the FPG groups introduced as an ordinal variable (1 df). c The HOMA-IR formula was defined as fasting insulin (mlU/l) multiplied by fasting glucose (mmol/l) divided by 22.5. * P ≤ 0.05 ** P ≤0 .01 Table 3 Logistic modeling of the association of fasting plasma glucose and insulin resistance in young adulthood [∼30 years) with low ranked cognitive function (lowest fifth) in midlife (∼50 years)]: the Jerusalem LRC longitudinal study Models Global Attention Information processing speed Executive Visual spatial Memory (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 OR OR OR OR OR OR OR OR OR OR OR OR [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl) vs. <5.6 mmol/l (100 mg/dl) 2.09** 2.10* 1.81* 1.76 1.25 0.966 1.35 1.00 2.44** 2.26* 1.82* 1.55 [1.21–3.61] [1.09–4.06] [1.04–3.16] [0.95–3.25] [0.68–2.31] [0.48–1.95] [0.75–2.42] [0.53–1.91] [1.37–4.34] [1.15–4.44] [1.04–3.17] [0.81–2.95] (0.008) (0.026) (0.036) (0.071) (0.473) (0.923) (0.318) (0.995) (0.002)* (0.019) (0.035) (0.184) 2. FPG (mmol/l) (continuous) 2.26** 2.31** 1.67* 1.60 1.006 0.911 1.39 1.16 1.89* 1.78* 1.63 124 [1.37–3.73] [1.30–4.13] [1.02–2.74 [0.94–2.72] [0.62–1.79] [0.51–1.64] [0.85–2.28] [0.69–1.96] [1.13–3.19] [1.00–3.17] [1.00–2.67] [0.72–2.16] (0.001) (0.005) (0.042) (0.084) (0.837) (0.755) (0.188) (0.576) (0.016) (0.050) (0.055) (0.437) 3. FPG (3 df)…(<80 mg/dl) (n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.007)b (0.006)b (0.097)b (0.121)b (0.913)b (0.789)b (0.456)b (0.894)b (0.021)b (0.034)b (0.050)b (0.244)b  4.4–4.9 mmol/l (80–89 mg/dl) (n = 172) 2.13 2.28 1.29 1.11 1.88 1.76 1.51 1.70 1.67 1.37 2.38 2.20 [0.61–7.43] [0.57–9.10] [0.46–3.59] [0.38–3.28] [0.62–5.75] [0.53–5.85] [0.54–4.21] [0.58–5.02] [0.54–5.17] [0.40–4.64] [0.68–8.32] [0.59–8.21] (0.238) (0.243) (0.633) (0.849) (0.268) (0.354) (0.433) (0.334) (0.370) (0.618) (0.173) (0.241) 5.0–5.5 mmol/l (90–99 mg/dl) n = 218) 2.42 3.13 1.27 1.20 1.31 1.31 1.33 1.45 1.54 1.52 2.23 1.89 [0.70–8.35] [0.80–12.28] [0.46–3.52] [0.41–3.50] [0.43–4.00] [0.40–4.33] [0.48–3.69] [0.50–4.23] [0.50–4.73] [0.45–5.11] [0.64–7.72] [0.51–7.01] (0.162) (0.102) (0.645) (0.736) (0.640) (0.658) (0.585) (0.495) (0.452) (0.498) (0.207) (0.339) 5.6+ mmol/l (100+ mg/dl) (n = 82) 4.60* 5.44* 2.28 2.03 1.87 1.40 1.85 1.51 3.78* 3.19 3.99* 3.00 [1.26–16.72] [1.29–22.87] [0.77–6.73] [0.65–6.35] [0.56–6.23] [0.38–5.11] [0.61–5.57] [0.47–4.81] [1.15–12.38] [0.88–11.64] [1.09–14.60] [0.75–11.97] (0.021) (0.021) (0.137) (0.226) (0.305) (0.613) (0.276) (0.491) (0.028) (0.079) (0.037) (0.121) 4. Insulin resistance HOMA (mlU/l*mmol/l)a,c (continuous) 1.32 0.79 1.02 0.78 1.23* 0.80 1.59 0.90 1.06 0.65 1.03 0.56 [0.78–2.26] [0.39–1.61] [0.60–1.74] [0.40–1.53] [0.71–2.15] [0.39–1.61] [0.92–2.72] [0.46–1.76] [0.61–1.85] [0.32–1.34] [0.60–1.76] [0.28–1.13] (0.303) (0.517) (0.94) (0.47) (0.46) (0.54) (0.90) (0.75) (0.83) (0.25) (0.92) (0.11) Models Global Attention Information processing speed Executive Visual spatial Memory (n = 505) (n = 505) (n = 475) (n = 504) (n = 503) (n = 497) Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 OR OR OR OR OR OR OR OR OR OR OR OR [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] [95% CI] (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) (P value) 1. FPG ≥5.6 mmol/l (100 mg/dl) vs. <5.6 mmol/l (100 mg/dl) 2.09** 2.10* 1.81* 1.76 1.25 0.966 1.35 1.00 2.44** 2.26* 1.82* 1.55 [1.21–3.61] [1.09–4.06] [1.04–3.16] [0.95–3.25] [0.68–2.31] [0.48–1.95] [0.75–2.42] [0.53–1.91] [1.37–4.34] [1.15–4.44] [1.04–3.17] [0.81–2.95] (0.008) (0.026) (0.036) (0.071) (0.473) (0.923) (0.318) (0.995) (0.002)* (0.019) (0.035) (0.184) 2. FPG (mmol/l) (continuous) 2.26** 2.31** 1.67* 1.60 1.006 0.911 1.39 1.16 1.89* 1.78* 1.63 124 [1.37–3.73] [1.30–4.13] [1.02–2.74 [0.94–2.72] [0.62–1.79] [0.51–1.64] [0.85–2.28] [0.69–1.96] [1.13–3.19] [1.00–3.17] [1.00–2.67] [0.72–2.16] (0.001) (0.005) (0.042) (0.084) (0.837) (0.755) (0.188) (0.576) (0.016) (0.050) (0.055) (0.437) 3. FPG (3 df)…(<80 mg/dl) (n = 33) (reference) 0 0 0 0 0 0 0 0 0 0 0 0 (0.007)b (0.006)b (0.097)b (0.121)b (0.913)b (0.789)b (0.456)b (0.894)b (0.021)b (0.034)b (0.050)b (0.244)b  4.4–4.9 mmol/l (80–89 mg/dl) (n = 172) 2.13 2.28 1.29 1.11 1.88 1.76 1.51 1.70 1.67 1.37 2.38 2.20 [0.61–7.43] [0.57–9.10] [0.46–3.59] [0.38–3.28] [0.62–5.75] [0.53–5.85] [0.54–4.21] [0.58–5.02] [0.54–5.17] [0.40–4.64] [0.68–8.32] [0.59–8.21] (0.238) (0.243) (0.633) (0.849) (0.268) (0.354) (0.433) (0.334) (0.370) (0.618) (0.173) (0.241) 5.0–5.5 mmol/l (90–99 mg/dl) n = 218) 2.42 3.13 1.27 1.20 1.31 1.31 1.33 1.45 1.54 1.52 2.23 1.89 [0.70–8.35] [0.80–12.28] [0.46–3.52] [0.41–3.50] [0.43–4.00] [0.40–4.33] [0.48–3.69] [0.50–4.23] [0.50–4.73] [0.45–5.11] [0.64–7.72] [0.51–7.01] (0.162) (0.102) (0.645) (0.736) (0.640) (0.658) (0.585) (0.495) (0.452) (0.498) (0.207) (0.339) 5.6+ mmol/l (100+ mg/dl) (n = 82) 4.60* 5.44* 2.28 2.03 1.87 1.40 1.85 1.51 3.78* 3.19 3.99* 3.00 [1.26–16.72] [1.29–22.87] [0.77–6.73] [0.65–6.35] [0.56–6.23] [0.38–5.11] [0.61–5.57] [0.47–4.81] [1.15–12.38] [0.88–11.64] [1.09–14.60] [0.75–11.97] (0.021) (0.021) (0.137) (0.226) (0.305) (0.613) (0.276) (0.491) (0.028) (0.079) (0.037) (0.121) 4. Insulin resistance HOMA (mlU/l*mmol/l)a,c (continuous) 1.32 0.79 1.02 0.78 1.23* 0.80 1.59 0.90 1.06 0.65 1.03 0.56 [0.78–2.26] [0.39–1.61] [0.60–1.74] [0.40–1.53] [0.71–2.15] [0.39–1.61] [0.92–2.72] [0.46–1.76] [0.61–1.85] [0.32–1.34] [0.60–1.76] [0.28–1.13] (0.303) (0.517) (0.94) (0.47) (0.46) (0.54) (0.90) (0.75) (0.83) (0.25) (0.92) (0.11) Model 1: Adjusted for sex and age at cognitive assessment. Model 2: Additionally adjusted for educational level, origin, religiosity, childhood SEP, adult SEP (ICBS ranking) and cigarette pack-yrs, leisure-time vigorous activity, alcohol intake, BMI, total cholesterol, HDL-cholesterol and GlycA measured at baseline (at ages 28–32), and for the sampling scheme. Missing data: HOMA-IR (n = 1), adult SEP (n = 5), leisure-time vigorous activity (n = 1) and GlycA (n = 20). Missing values were replaced with non-missing median values (adult SEP, leisure-time vigorous activity) or by multiple linear regression imputation (baseline GlycA). LRC, Lipid Research Clinic; FPG, fasting plasma glucose; HOMA, homeostatic model assessment for insulin resistance. OR, odds ratio. CI, confidence interval. a Box-Cox transformed HOMA-IR (λ = 0). b Test for trend for the FPG groups introduced as an ordinal variable (1 df). c The HOMA-IR formula was defined as fasting insulin (mlU/l) multiplied by fasting glucose (mmol/l) divided by 22.5. * P ≤ 0.05 ** P ≤0 .01 The FPG-cognition associations were evident, albeit somewhat attenuated, in a sensitivity analysis restricted to FPG <6.1 mmol/l (see Supplementary table S3). The associations for FPG ≥5.6 vs. <5.6 mmol/l persisted in a sensitivity analysis of global cognitive function adjusting to a reduced covariate set (standardized Beta=−0.125, P = 0.001 and OR = 2.06, 95%CI = 1.11–3.83, P = 0.023) and for FPG treated as a continuous variable (standardized Beta=−0.091, P = 0.018 and OR = 1.04, 95%CI = 1.00–1.07, P = 0.010) (not shown). The association of FPG with global cognitive function was not affected in a sensitivity analysis substituting BMI with waist-to-hip ratio (not shown). There was no association of cognition in midlife with baseline HOMA-IR (tables 2 and 3) or with baseline OGTT or OGTT change or FPG change over the mean 13 years follow-up (Supplementary table S4). Discussion In our population-based cohort free of diabetes at ages 28–32, higher FPG at age 30 years was associated with lower global cognitive function in midlife, mainly attributable to its visual spatial and attention domains. These associations persisted after adjustment for BMI, demographics, childhood and adult SEP, physical activity, alcohol intake, smoking, inflammatory variables and plasma lipids. The associations remained evident when the analysis was restricted to those with FPG <6.1 mmol/l. Baseline insulin resistance and OGTT, and change in FPG and OGTT over the 13 years follow-up were not associated with cognition. Our study shows, in line with previous reports, that blood glucose levels are associated with cognitive function, and in particular extends this finding to young adults free of diabetes. Recent studies report an association between higher glucose levels (combined FPG and HbA1c values) and increased risk of dementia in older adults (n = 2067; aged 76 years) with and without diabetes,30 as well as associations between hyperglycemia (diabetes) and impaired memory, visual perception and attention (n = 2126; ages 19–72),3 between impaired glucose metabolism (composite score of FPG, HbA1c and HOMA) and lower executive function in a national sample of middle-aged and older adults (n = 1076; median age, 57 years),31 between chronically higher blood glucose levels (HbA1c) and lower memory scores in older adults without diabetes (n = 141; age 63.1 ± 6.9),4 between higher FPG levels, even within the normal range (<6.1 mmol/l), and longer reaction times in processing speed (n = 41; age 64.7 years)32 and between higher FPG in 35 obese and 35 normal weight university students free of diabetes (age 20.8 ± 2.4)11 on a test of inhibitory control among 3 cognitive domains assessed. Longitudinal data on the relationship between blood glucose and cognition are sparse, inconsistent and also limited to older adults. Longitudinal trajectories of glycated haemoglobin (HbA1c) in older adults with type 2 diabetes (n = 835, mean age, 72.8 years) predicted overall cognitive performance and its executive domain.33 A study that examined change in HbA1c and cognition among older adults with normal glucose tolerance (NGT) (≥75 years old, n = 101)6 found a decrease in the Mini-Mental State Examination with increasing HbA1c. The only study to our knowledge that has examined longitudinal change in FPG34 did so in a sample of adults aged 25–85 at baseline (n = 4547) and reported, in line with our study, no association of FPG change with cognition measured 12 years later. Consistent with our findings, HOMA was not associated with cognitive function in the MONA LISA study (age 50.0 ± 8.1, range 35–64 years) in which 4% and <6%, had hyperinsulinemia or an HOMA value >2.6 mlU/l*mmol/l.10 In the Framingham Heart Study third-generation cohort (n = 2126, age 40.4 ± 8.7, range 19–72),3 no association of HOMA with cognition or of HOMA with structural brain measures (fractional anisotropy or gray matter density) was found. On the other hand, in the Framingham Offspring study (mean age 61 ± 9),35 HOMA was related to poorer executive and visuospatial memory scores. In the ARIC cohort (age 53.7 years, range, 45–64),36 HOMA was associated with lower baseline cognitive test scores. In the Framingham Offspring study and the ARIC cohort mentioned above,35,36 adiposity was not accounted for, and in Framingham diabetes35 was included. Our study sample was younger, was free of diabetes and accounted for BMI. Mechanisms underlying the relationship between glycemia and cognitive function might involve glucose-related increases in inflammatory responses and blood coagulation activation, leading to subclinical strokes, gray matter atrophy, reduced white matter integrity and subsequent cerebral volume loss.4,35 Moreover, hyperglycemia (HBA1C ≥5.5% or FPG ≥5.8 mmol/l) is linked to an increase in prevalence of retinopathy.37 Retinal microvascular abnormalities appear to reflect small vessel damage in the brain and are independently associated with poor cognitive function in midlife.38 Direct ‘toxic’ effects of glucose on neuronal structures include imbalance in the generation and scavenging of reactive oxygen species, or advanced glycation of important functional and structural proteins in the brain.39 Ultimately, these structural changes may lead to decreased neurotransmitter signaling and loss of synaptic contacts with subsequent cognitive impairment.40 The main limitation of this study to be considered when interpreting our findings is that the measure of cognition was done at one point in midlife. Consequently, no direct inference about the role of FPG on cognitive decline can be drawn. However, adjustment for childhood SEP, education and ethnic origin partly controls for baseline cognition and partly overcomes this drawback. We did not undertake a 24-h ABPM (ambulatory blood pressure monitoring). This may be a limitation of our study, as accounting for our three consecutive measurements of blood pressure does not eliminate the influence of hemodynamic factors. Finally, the present study includes a relatively small portion of the original Visit 1 participants recruited at age 17 in 1976–78. However, as depicted in figure 1, the main reason for this apparent discrepancy in sample sizes was due to us drawing a small random sample of the Visit 1 participants for the detailed Visit 4 round of examinations (70% response). The Visit 5 and Visit 6 samples were a follow-up of the Visit 4 study sample, with response rates of 71% and 82%, respectively. We emphasize that the current study participants were shown to be representative of the Visit 1 source population. Therefore, we do not assume that the reported associations of FPG and cognition are likely to be overestimated due to nonresponse or selection bias. The main strength of this study is its focus on young adults free of diabetes at baseline. Additional strengths are the longitudinal data on FPG, the wide range of potential confounders evaluated, including detailed socio-demographic, psychosocial, health behavioral and biochemical variables, and the comprehensive objective computerized cognitive measures with millisecond precision. Finally, the generalizability of these findings in a heterogeneous Jewish population sample should be assessed in other populations. In summary, our results indicate that even in healthy young adults aged 30 without diabetes, fasting hyperglycemia was inversely associated with a global measure of cognition in midlife. Should this finding be confirmed by longitudinally measured cognitive data, testing whether interventions in young adults aimed at maintaining or reducing glucose levels well below 5.6 mmol/l would delay cognitive decline in nondiabetic middle-aged adults is indicated. Multidomain interventions of lifestyle, management of metabolic and vascular risk factors and cognitive training in non-demented older adults have already been demonstrated to prevent cognitive decline.41 Supplementary data Supplementary data are available at EURPUB online. Funding Chief Scientist of the Israel Ministry of Health [300000-5352]; the Israel Science Foundation [593/01]; and the US-Israel Binational Science Foundation [87-00419] all to JDK. Conflicts of interest: ESS and GMD are employees of NeuroTrax Corp. The remaining authors have no potential conflicts of interest to declare. Key points Data relating impaired glucose regulation to cognitive function in young adults are scarce, entail only cross-sectional assessment of glucose and cognition, and are based on middle-aged samples, type 2 diabetes patients or a small sample of convenience. We are unaware of reports relating either plasma glucose concentrations or longitudinal change in plasma glucose with cognitive assessment in population-based samples of young adults free of diabetes. A higher fasting plasma glucose concentration in healthy young adults aged 30 was associated with lower global cognitive function at age 50. Hyperglycemia in young adults may be a preventable and modifiable risk factor for low ranked cognitive function in midlife. References 1 Lu FP , Lin KP , Kuo HK . Diabetes and the risk of multi-system aging phenotypes: a systematic review and meta-analysis . PLoS One 2009 ; 4 : e4144 . Google Scholar CrossRef Search ADS PubMed 2 Yaffe K , Falvey C , Hamilton N , et al. Diabetes, glucose control, and 9-year cognitive decline among older adults without dementia . Arch Neurol 2012 ; 69 : 1170 – 5 . Google Scholar CrossRef Search ADS PubMed 3 Weinstein G , Maillard P , Himali JJ , et al. Glucose indices are associated with cognitive and structural brain measures in young adults . Neurology 2015 ; 84 : 2329 – 37 . Google Scholar CrossRef Search ADS PubMed 4 Kerti L , Witte AV , Winkler A , et al. Higher glucose levels associated with lower memory and reduced hippocampal microstructure . Neurology 2013 ; 81 : 1746 – 52 . Google Scholar CrossRef Search ADS PubMed 5 Convit A , Wolf OT , Tarshish C , de Leon MJ . Reduced glucose tolerance is associated with poor memory performance and hippocampal atrophy among normal elderly . Proc Natl Acad Sci U S A 2003 ; 100 : 2019 – 22 . Google Scholar CrossRef Search ADS PubMed 6 Ravona-Springer R , Moshier E , Schmeidler J , et al. Changes in glycemic control are associated with changes in cognition in non-diabetic elderly . J Alzheimers Dis 2012 ; 30 : 299 – 309 . Google Scholar CrossRef Search ADS PubMed 7 Gold SM , Dziobek I , Sweat V , et al. Hippocampal damage and memory impairments as possible early brain complications of type 2 diabetes . Diabetologia 2007 ; 50 : 711 – 9 . Google Scholar CrossRef Search ADS PubMed 8 Yau PL , Javier DC , Ryan CM , et al. Preliminary evidence for brain complications in obese adolescents with type 2 diabetes mellitus . Diabetologia 2010 ; 53 : 2298 – 306 . Google Scholar CrossRef Search ADS PubMed 9 Fuh JL , Wang SJ , Hwu CM , Lu SR . Glucose tolerance status and cognitive impairment in early middle-aged women . Diabet Med 2007 ; 24 : 788 – 91 . Google Scholar CrossRef Search ADS PubMed 10 Sanz CM , Ruidavets JB , Bongard V , et al. Relationship between markers of insulin resistance, markers of adiposity, HbA1c, and cognitive functions in a middle-aged population-based sample: the MONA LISA study . Diabetes Care 2013 ; 36 : 1512 – 21 . Google Scholar CrossRef Search ADS PubMed 11 Hawkins MA , Gunstad J , Calvo D , Spitznagel MB . Higher fasting glucose is associated with poorer cognition among healthy young adults . Health Psychol 2016 ; 35 : 199 – 202 . Google Scholar CrossRef Search ADS PubMed 12 Slater PE , Friedlander Y , Baras M , et al. The Jerusalem Lipid Research Clinic: sampling, response and selected methodological issues . Isr J Med Sci 1982 ; 18 : 1106 – 12 . Google Scholar PubMed 13 Kark JD , Sinnreich R , Leitersdorf E , et al. Taq1B CETP polymorphism, plasma CETP, lipoproteins, apolipoproteins and sex differences in a Jewish population sample characterized by low HDL-cholesterol . Atherosclerosis 2000 ; 151 : 509 – 18 . Google Scholar CrossRef Search ADS PubMed 14 Kark JD , Goldberger N , Kimura M , et al. Energy intake and leukocyte telomere length in young adults . Am J Clin Nutr 2012 ; 95 : 479 – 87 . Google Scholar CrossRef Search ADS PubMed 15 Cohen-Manheim I , Doniger GM , Sinnreich R , et al. Increase in the inflammatory marker GlycA over 13 years in young adults is associated with poorer cognitive function in midlife . PLoS One 2015 ; 10 : e0138036 . Google Scholar CrossRef Search ADS PubMed 16 Cohen-Manheim I , Doniger GM , Sinnreich R , et al. Increased attrition of leukocyte telomere length in young adults is associated with poorer cognitive function in midlife . Eur J Epidemiol 2016 ; 31 : 147 – 57 . Google Scholar CrossRef Search ADS PubMed 17 Cohen-Manheim I , Doniger GM , Sinnreich R , et al. Body mass index, height and socioeconomic position in adolescence, their trajectories into adulthood, and cognitive function in midlife . J Alzheimers Dis 2017 ; 55 : 1207 – 21 . Google Scholar CrossRef Search ADS PubMed 18 Dwolatzky T , Whitehead V , Doniger GM , et al. Validity of a novel computerized cognitive battery for mild cognitive impairment . BMC Geriatr 2003 ; 3 : 4 . Google Scholar CrossRef Search ADS PubMed 19 Melton J . ( 2005 ) Psychometric evaluation of the Mindstreams neuropsychological screening tool. NEDU Technical Report 06-10, Navy Experimental Diving Unit, Panama City, FL. 20 Thaler A , Mirelman A , Gurevich T , et al. Lower cognitive performance in healthy G2019S LRRK2 mutation carriers . Neurology 2012 ; 79 : 1027 – 32 . Google Scholar CrossRef Search ADS PubMed 21 Sasson E , Doniger GM , Pasternak O , Assaf Y . Structural correlates of memory performance with diffusion tensor imaging . Neuroimage 2010 ; 50 : 1231 – 42 . Google Scholar CrossRef Search ADS PubMed 22 Boussi-Gross R , Golan H , Volkov O , et al. Improvement of memory impairments in poststroke patients by hyperbaric oxygen therapy . Neuropsychology 2015 ; 29 : 610 – 21 . Google Scholar CrossRef Search ADS PubMed 23 Hartman SJ , Marinac CR , Natarajan L , Patterson RE . Lifestyle factors associated with cognitive functioning in breast cancer survivors . Psychooncology 2015 ; 24 : 669 – 75 . Google Scholar CrossRef Search ADS PubMed 24 Zur D , Naftaliev E , Kesler A . Evidence of multidomain mild cognitive impairment in idiopathic intracranial hypertension . J Neuroophthalmol 2015 ; 35 : 26 – 30 . Google Scholar CrossRef Search ADS PubMed 25 Adler NE , Epel ES , Castellazzo G , Ickovics JR . Relationship of subjective and objective social status with psychological and physiological functioning: preliminary data in healthy white women . Health Psychol 2000 ; 19 : 586 – 92 . Google Scholar CrossRef Search ADS PubMed 26 Zigmond AS , Snaith RP . The hospital anxiety and depression scale . Acta Psychiatr Scand 1983 ; 67 : 361 – 70 . Google Scholar CrossRef Search ADS PubMed 27 Otvos JD , Shalaurova I , Wolak-Dinsmore J , et al. GlycA: A composite nuclear magnetic resonance biomarker of systemic inflammation . Clin Chem 2015 ; 61 : 714 – 23 . Google Scholar CrossRef Search ADS PubMed 28 Gutierrez J , Marshall RS , Lazar RM . Indirect measures of arterial stiffness and cognitive performance in individuals without traditional vascular risk factors or disease . JAMA Neurol 2015 ; 72 : 309 – 15 . Google Scholar CrossRef Search ADS PubMed 29 Folsom AR , Jacobs DR Jr. , Wagenknecht LE , et al. Increase in fasting insulin and glucose over seven years with increasing weight and inactivity of young adults. The CARDIA Study. Coronary artery risk development in young adults . Am J Epidemiol 1996 ; 144 : 235 – 46 . Google Scholar CrossRef Search ADS PubMed 30 Crane PK , Walker R , Hubbard RA , et al. Glucose levels and risk of dementia . N Engl J Med 2013 ; 369 : 540 – 8 . Google Scholar CrossRef Search ADS PubMed 31 Karlamangla AS , Miller-Martinez D , Lachman ME , et al. Biological correlates of adult cognition: midlife in the United States (MIDUS) . Neurobiol Aging 2014 ; 35 : 387 – 94 . Google Scholar CrossRef Search ADS PubMed 32 Raizes M , Elkana O , Franko M , et al. Higher fasting plasma glucose levels, within the normal range, are associated with decreased processing speed in high functioning young elderly . J Alzheimers Dis 2015 ; 49 : 589 – 92 . Google Scholar CrossRef Search ADS 33 Ravona-Springer R , Heymann A , Schmeidler J , et al. Trajectories in glycemic control over time are associated with cognitive performance in elderly subjects with type 2 diabetes . PLoS One 2014 ; 9 : e97384 . Google Scholar CrossRef Search ADS PubMed 34 Anstey KJ , Sargent-Cox K , Eramudugolla R , et al. Association of cognitive function with glucose tolerance and trajectories of glucose tolerance over 12 years in the AusDiab study . Alzheimers Res Ther 2015 ; 7 : 48 . Google Scholar CrossRef Search ADS PubMed 35 Tan ZS , Beiser AS , Fox CS , et al. Association of metabolic dysregulation with volumetric brain magnetic resonance imaging and cognitive markers of subclinical brain aging in middle-aged adults: the Framingham Offspring Study . Diabetes Care 2011 ; 34 : 1766 – 70 . Google Scholar CrossRef Search ADS PubMed 36 Young SE , Mainous AG 3rd , Carnemolla M . Hyperinsulinemia and cognitive decline in a middle-aged cohort . Diabetes Care 2006 ; 29 : 2688 – 93 . Google Scholar CrossRef Search ADS PubMed 37 Cheng YJ , Gregg EW , Geiss LS , et al. Association of A1C and fasting plasma glucose levels with diabetic retinopathy prevalence in the U.S. population: Implications for diabetes diagnostic thresholds . Diabetes Care 2009 ; 32 : 2027 – 32 . Google Scholar CrossRef Search ADS PubMed 38 Wong TY , Klein R , Sharrett AR , et al. Retinal microvascular abnormalities and cognitive impairment in middle-aged persons: the Atherosclerosis Risk in Communities Study . Stroke 2002 ; 33 : 1487 – 92 . Google Scholar CrossRef Search ADS PubMed 39 Brownlee M . Biochemistry and molecular cell biology of diabetic complications . Nature 2001 ; 414 : 813 – 20 . Google Scholar CrossRef Search ADS PubMed 40 Lamport DJ , Lawton CL , Mansfield MW , Dye L . Impairments in glucose tolerance can have a negative impact on cognitive function: a systematic research review . Neurosci Biobehav Rev 2009 ; 33 : 394 – 413 . Google Scholar CrossRef Search ADS PubMed 41 Ngandu T , Lehtisalo J , Solomon A , et al. A 2 year multidomain intervention of diet, exercise, cognitive training, and vascular risk monitoring versus control to prevent cognitive decline in at-risk elderly people (FINGER): a randomised controlled trial . Lancet 2015 ; 385 : 2255 – 63 . Google Scholar CrossRef Search ADS PubMed © The Author 2017. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

The European Journal of Public HealthOxford University Press

Published: Nov 13, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off