Background: The relationship between long-term exposure to whole body or central obesity and cognitive function, as well as its potential determinants, remain controversial. In this study, we assessed (1) the potential impact of 30 years exposure to different patterns of whole body and central adiposity on cognitive function at 60–64 years, (2) whether trajectories of central adiposity can provide additional information on later cognitive function compared to trajectories of whole body adiposity, and (3) the influence of vascular phenotypes on these associations. Methods: The study included 1249 participants from the prospective cohort MRC National Survey of Health and Development. Body mass index (BMI), waist circumference (WC), and vascular (carotid intima-media thickness, carotid- femoral pulse wave velocity) and cognitive function (memory, processing speed, reaction time) data, at 60–64 years, were used to assess the associations between different patterns of adult WC or BMI (from 36 years of age) and late midlife cognitive performance, as well as the proportion of this association explained by cardiovascular phenotypes. Results: Longer exposure to elevated WC was related to lower memory performance (p < 0.001 for both) and longer choice reaction time (p = 0.003). A faster gain of WC between 36 and 43 years of age was associated with the largest change in reaction time and memory test (P < 0.05 for all). Similar associations were observed when patterns of WC were substituted with patterns of BMI, but when WC and BMI were included in the same model, only patterns of WC remained significantly associated with cognitive function. Participants who dropped one BMI category and maintained a lower BMI had similar memory performance to those of normal weight during the whole follow-up. Conversely, those who dropped and subsequently regained one BMI category had a memory function similar to those with 30 years exposure to elevated BMI. Adjustment for vascular phenotypes, levels of cardiovascular risk factors, physical activity, education, childhood cognition and socioeconomic position did not affect these associations. Conclusions: Longer exposure to elevated WC or BMI and faster WC or BMI gains between 36 and 43 years are related to lower cognitive function at 60–64 years. Patterns of WC in adulthood could provide additional information in predicting late midlife cognitive function than patterns of BMI. The acquisition of an adverse cardiovascular phenotype associated with adiposity is unlikely to account for these relationships. Keywords: Obesity, waist circumference, cognitive function, vascular phenotypes, lifetime risk * Correspondence: email@example.com Equal contributors National Centre for Cardiovascular Prevention and Outcomes, Institute of Cardiovascular Science, University College London, 1 St Martin le Grande, London EC1A 4NP, UK Department of Clinical and Experimental Medicine, University of Pisa, Pisa, Italy Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Masi et al. BMC Medicine (2018) 16:75 Page 2 of 12 Background Methods The increasing prevalence of obesity represents a major Population public health concern, as it is associated with an in- The MRC NSHD is a nationally representative sample of creased risk of several chronic diseases, including cardio- 5362 singleton births to married parents in England, vascular disease (CVD). Several reports have suggested Scotland and Wales, stratified by social class, during 1 week that exposure to whole body and abdominal obesity in March 1946 [25, 26]. The cohort has been followed-up could influence cognitive function and risk of dementia, 23 times from birth to age 69 years. The present study is although results are conflicting and limited to older co- based on the 1249 (74%) of 1690 participants who had adi- horts [1–7]. A recent study found that gene variants as- posity measures, vascular phenotype and cognitive data at sociated with greater body mass index (BMI) are also 60–64 years with a BMI > 18.5 kg/m . Further details on related to lower cognitive function , supporting the the sample invited at the 60–64 years assessment are hypothesis that shared biological pathways could in- provided in Additional file 1. crease the risk of obesity and cognitive dysfunction. However, whether the association of obesity with cogni- Cognitive assessment tive outcomes is related to the cumulative burden of ex- Cognitive function was assessed at age 60–64 years posure or vulnerability to the effects of rapid changes of using a validated verbal memory test, a letter search whole body or abdominal fat at specific stages of adult- speed test and two reaction time tests (simple reaction hood remains unclear. time, choice reaction time) . Details of each cognitive Several factors could account for the association be- test are provided in Additional file 1. tween amount and distribution of body fat and reduced cognitive function. A lifetime exposure to obesity is asso- ciated with the acquisition of an adverse cardiovascular Adiposity measures phenotype. Carotid-to-femoral pulse wave velocity (PWV) Weight, height and WC were measured during adult- and common carotid artery intima-media thickness hood at ages 36, 43, 53 and 60–64 years. BMI was calcu- (cIMT) are validated surrogate markers of arterial stiffness lated as weight (kg) divided by squared height (m ), and and atherosclerotic CVD [9, 10], and are known to be af- was used to define adiposity status according to World fected by exposure to whole body and abdominal obesity Health Organization criteria (BMI 18.5–25 kg/m [11–14]. In turn, increased cIMT and PWV are associated 2 2 normal weight, 25–29 kg/m overweight, and ≥ 30 kg/m with lower cognitive performance [15–20] and with a obese) at each age. Similarly, we identified three classes higher burden and rate of deposition of β-amyloid in the of cardiometabolic risk related to WC, namely (1) low brain [21, 22]. Therefore, an altered vascular phenotype risk = WC ≤ 94 cm for males and ≤ 80 cm for females; identified by greater cIMT or PWV could contribute to (2) increased risk = WC > 94 cm and ≤ 102 cm for males the association between adiposity and cognitive perform- and > 80 cm and ≤ 88 cm for females; (3) substantially ance, but this has not been studied. increased risk = WC > 102 cm for males and > 88 cm for The MRC National Survey of Health and Develop- females . Participants in the classes defined as ment (NSHD, also known as the 1946 British Birth Co- increased and substantially increased risk were hort) is the oldest of the British Birth Cohort studies combined into an elevated WC group. , and is unique in providing measures of BMI and waist circumference (WC) across the entire life course, together with a characterisation of vascular phenotypes, Vascular phenotypes cardiovascular risk factors and cognitive function at At age 60–64 years, PWV and cIMT were measured 60–64 years. Using this population, we have previously using validated devices and following standard protocols explored the impact of BMI and its change over time [28, 29], as reported in Additional file 1. on cognitive function at the age of 53 years . We now extend this work to investigate whether central adiposity has an effect over and above the effect of gen- Covariates eral adiposity on cognitive function at 60–64 years of Covariates were either selected a priori or were those age. We also assess whether rapid changes of BMI or variables that were associated with cognitive measures in WC over different periods of adult life can have a spe- univariable models. These included level of education, cific influence on cognitive function at 60–64 years of childhood cognition, socioeconomic position, heart rate, age. Finally, we explored what proportion, if any, of the systolic blood pressure, smoking, diabetes and its association between adult patterns of adiposity and duration, total cholesterol, and levels of physical activity. cognitive function could be explained by acquisition of Methods used for their assessment are reported in an adverse cardiovascular phenotype. Additional file 1. Masi et al. BMC Medicine (2018) 16:75 Page 3 of 12 Statistical analysis Cross-sectional analysis Mean (standard deviation), or median (IQR) for Association of BMI and WC with cognitive function at skewed variables, were used to describe continuous 60–64 years variables and percentages for binary variables. We We did not find any significant effect modification of tested for effect modification by sex of obesity indices obesity indices on cognitive outcomes by sex, apart from on cognitive outcomes by introducing relevant inter- the letter search speed test. In models adjusted for sex, action terms (BMI*sex) in multivariable regression education and childhood cognition, BMI was positively models. Level of statistical significance for interaction associated with cIMT (regression coefficient (β)= 0. terms was set at 0.1 and, when a significant inter- 003 mm per kg/m ; 95% confidence interval (CI) 0.001 action was found, results were stratified by sex. We to 0.005; p = 0.01) and PWV (β = 0.040 m/sec per kg/m ; fitted a series of linear multivariable regression 95% CI 0.012 to 0.068; p = 0.005) (Table 2). The models to establish the associations between (1) BMI association of BMI with cIMT remained significant in at 60–64 years, (2) patterns of overweight/obesity, the fully adjusted model, while that with PWV was and (3) conditional change in BMI from 36 to 43, 43 attenuated in MODEL 2 and further reduced in the final to 53, and 53 to 60–64 with each cognitive outcome model (Table 2). A higher BMI was associated with a and with each vascular phenotype (PWV, cIMT). In lower performance on the verbal memory test (β = each model, inverse probability weighting was imple- −0.195 number of words per kg/m ;95% CI –0.274 mented to account for dropout due to death. The to −0.116; p < 0.001) and the letter search speed test models were sequentially adjusted for covariates: (β = −0.005 number of targets hit per kg/m ; 95% CI MODEL 1 adjusted for sex, education and childhood –0.010 to −0.001; p = 0.018). When stratified by sex, cognition; MODEL 2 = MODEL 1 + socioeconomic the association between higher BMI and lower letter position at 53 years, systolic blood pressure and heart search speed test performance was stronger in females rate at age 60–64; MODEL 3 (fully adjusted) = than in males (Additional file 1: Table S1). These MODEL 2 + total cholesterol, smoking, diabetes and associations remained significant in the fully adjusted levels of physical activity at 60–64 years. In analyses models. No associations were observed between BMI and exploring the association between patterns or condi- choice reaction time or simple reaction time (Table 2). tional changes of BMI with cognitive outcomes (2 Additional file 1: Table S2 reports the associations and 3), MODEL 3 was further adjusted for duration between WC and the vascular and cognitive measures at of diabetes. We also tested for effect modification by 60–64 years. WC was associated with vascular socioeconomic position and education levels of the phenotypes, verbal memory test performance and, association between BMI changes and cognitive out- differently from BMI, with choice reaction time, but not comes. All multivariable regression models of cogni- with performance in the letter search speed test. When tive outcomes on cross-sectional measures or BMI and WC were included in the same model, the longitudinal patterns of BMI were further adjusted for associations of both with performance in the verbal PWV and, separately, cIMT to explore what propor- memory test and letter search speed test were strongly tion of these associations could be explained by car- attenuated, suggesting that the two measures of adiposity diovascular phenotype. Each analysis was repeated provided similar information. In turn, WC remained using WC rather than BMI as the exposure. To assess significantly associated with choice reaction time whether central adiposity had an effect over and after adjustment for BMI (β = 6.73; 95% CI 1.32 to above the effect of general adiposity, each cross- 12.1; p = 0.015). sectional and longitudinal analysis was repeated in- cluding BMI and WC in the same models. Further Associations of cardiovascular risk factors/phenotypes with details on the statistical methods are reported in cognitive function at 60–64 years Additional file 1. Statistical analyses were performed Associations between cardiovascular risk factors and using Stata 13.1. each cognitive outcome are reported in Additional file 1: Table S3. Higher PWV was associated with a lower per- Results formance in the verbal memory test (β = −0.355 number Table 1 reports the characteristics of the total sample of words per m/s; 95% CI –0.616 to −0.095; p = 0.008), used in these analyses (n = 1249) as well as the differ- with the strength of this association attenuated in ences between groups of normal weight, overweight MODEL 3, while no association was found between and obese at 60–64 years old. We have previously re- cIMT and verbal memory test performance (Table 3). ported that the sample attending the clinical research The letter search speed and both reaction time tests facility showed some differences in characteristics were not associated with any vascular phenotype. Substi- compared with those not attending . tution of pulse pressure for systolic blood pressure in Masi et al. BMC Medicine (2018) 16:75 Page 4 of 12 Table 1 Participant characteristics by BMI category at the age of 60–64 years N Entire sample Normal (30.5%) Overweight (42.5%) Obese (27.0%) p value trend Male sex (%) 1249 46.3 37.0 54.7 43.8 0.046 Height (m) 1249 1.68 (0.1) 1.68 (0.1) 1.69 (0.1) 1.67 (0.1) 0.44 Weight (kg) 1249 78.4 (14.9) 65.3 (8.2) 78.1 (9.2) 94.3 (12.5) < 0.001 BMI (kg/m ) 1249 27.6 (4.6) 23.0 (1.5) 27.2 (1.4) 33.6 (3.6) < 0.001 Waist circumference (cm) 1249 96.6 (12.6) 84.6 (8.2) 96.0 (7.6) 109.2 (9.8) < 0.001 Hip circumference (cm) 1249 105.8 (9.5) 97.7 (4.9) 104.5 (5.1) 115.7 (9.0) < 0.001 Waist-to-hip ratio 1249 0.91 (0.1) 0.87 (0.1) 0.92 (0.1) 0.95 (0.1) < 0.001 Heart rate (bpm) 1248 68.4 (11.3) 67.2 (11.4) 68.3 (10.6) 69.8 (12.1) 0.009 SBP (mmHg) 1248 134.8 (17.9) 129.3 (16.7) 136.6 (18.1) 138 (17.9) < 0.001 DBP (mmHg) 1248 77.2 (9.7) 74.5 (9.4) 78.2 (9.4) 78.7 (10.1) < 0.001 Education (% above A-level) 909 61.3 61.9 65 54.6 0.14 Smoking (% current smokers) 1171 34.9 37.7 33.7 34.3 0.34 Total cholesterol (mmol/L) 1186 5.69 (1.2) 5.91 (1.1) 5.7 (1.2) 5.4 (1.3) < 0.001 LDL cholesterol (mmol/L) 1156 3.52 (1.0) 3.66 (0.9) 3.55 (1.0) 3.3 (1.1) 0.001 HDL cholesterol (mmol/L) 1186 1.62 (0.41) 1.81 (0.4) 1.59 (0.4) 1.43 (0.3) < 0.001 Triglycerides (mmol/L) 1162 1.1 (0.8–1.5) 0.8 (0.6–1.1) 1.1 (0.8–1.5) 1.3 (1.0–1.9) < 0.001 HbA1c (%) 1157 5.8 (5.5–6) 5.7 (5.5–5.9) 5.7 (5.5–6.0) 5.9 (5.6–6.2) < 0.001 Adiponectin (μg/mL) 1183 12.9 (7.6–19.5) 16.6 (10.9–25.1) 12.1 (7.0–17.8) 10.2 (6.3–16.4) < 0.001 Leptin (ng/mL) 1183 12.1 (6.6–23.6) 7.8 (4.1–13.7) 11.6 (6.4–19.1) 26.1 (14.1–43.2) < 0.001 Levels of physical activity (in the last 4 weeks) � Not physical active 1223 703 (58.15) 205 (54.81) 297 (58.01) 201 (62.23) 0.375 � Moderate 189 (15.63) 61 (16.31) 80 (15.63) 48 (14.86) � Intense 317 (26.22) 108 (28.88) 135 (26.37) 74 (22.91) PWV (m/s) 1249 8.2 (1.51) 7.95 (1.58) 8.27 (1.5) 8.37 (1.42) < 0.001 cIMT (mm) 900 0.69 (0.12) 0.67 (0.12) 0.69 (0.13) 0.71 (0.11) <0.001 VMT (number of words) 1227 24.7 (6.1) 26.0 (6.0) 24.5 (6.2) 23.6 (5.7) < 0.001 LSST (targets) 1249 282 (231–307) 287 (231–329) 239 (231–329) 263 (174–288) 0.009 S-RT (s) 1234 281 (63.5) 279 (64.5) 278 (60.2) 289 (67.2) 0.038 C-RT (s) 1230 612 (76.5) 607 (77.4) 608 (72.9) 622 (80.6) 0.012 Values are presented as mean ± standard deviation, N (%) or median (IQR). Comparisons between normal weight, overweight and obese groups were performed by test for trend using linear regression Numbers in italic indicate statistical significance SBP systolic blood pressure, DBP diastolic blood pressure, cIMT common carotid artery intima-media thickness, PWV pulse wave velocity, VMT verbal memory test, LSST letter search speed test, C-RT choice reaction time test, S-RT simple reaction time test MODELS 2 and 3 did not affect the association between Longitudinal analysis PWV and verbal memory test. As PWV was associated Association of patterns of cumulative exposure to with both BMI and memory function, we tested whether overweight or obesity and elevated WC in adulthood with the association between BMI and verbal memory test per- cognitive function formance was attenuated by PWV, suggesting that PWV From age 36 to 60–64 years, the prevalence of over- may be a mediator of the association between adiposity weight/obesity increased from 29% to 69%. Of 1021 and memory performance. However, adjustment for PWV participants with vascular phenotypes at 60–64 years only slightly attenuated the association between BMI and and complete BMI records at all ages, 141 (14%) had verbal memory test in the fully adjusted model (unadjusted a reduction in BMI category during 27 years of for PWV: β = −0.172 number of words per kg/m ; 95% CI follow-up; 78 (8%) subsequently regained weight, leav- –0.258 to −0.086 vs. adjusted for PWV β = −0.170 number ing 63 (6%) with stable weight reduction. Earlier on- of words per kg/m ;95% CI –0.256 to −0.084; proportional set of overweight/obesity was associated with a worse difference in β =1.3%). cardiometabolic profile and higher PWV and cIMT Masi et al. BMC Medicine (2018) 16:75 Page 5 of 12 Table 2 Cross-sectional associations of BMI with carotid intima-media thickness and pulse wave velocity with verbal memory, letter search speed, choice and simple reaction test at age 60–64 years MODEL 1 MODEL 2 MODEL 3 β (95% CI) p β (95% CI) p β (95% CI) p VASCULAR MEASURES 1. cIMT (mm) 0.003 (0.0003 to 0.005) 0.025 0.002 (−0.0003 to 0.005) 0.085 0.003 (0.0004 to 0.006) 0.024 2. PWV (m/s) 0.040 (0.012 to 0.068) 0.005 0.018 (−0.009 to 0.045) 0. 192 0.014 (−0.015 to 0.043) 0.355 COGNITIVE MEASURES 1. VMT (n. of words) –0.195 (−0.274 to −0.116) < 0.001 −0.178 (−0.255 to −0.098) < 0.001 −0.162 (−0.248 to −0.076) < 0.001 2. LSST (targets) –0.005 (−0.010 to −0.001) 0.018 −0.005 (−0.010 to −0.001) 0.027 −0.004 (−0.009 to 0.001) 0.117 3. RT (s) a. Simple 0.485 (−0.373 to 1.344) 0.267 0.552 (−0.324 to 1.428) 0.217 0.689 (−0.300 to 1.679) 0.172 b. Choice 0.910 (−0.222 to 2.042) 0.115 0.983 (−0.157 to 2.123) 0.091 1.014 (−0.292 to 2.320) 0.128 Linear regression models were used to assess associations between variables MODEL 1 adjusted for sex, education, childhood cognition; MODEL 2 = MODEL 1 + adjustments for socioeconomic position at age 53, systolic blood pressure and heart rate at age 60–64; MODEL 3 (fully adjusted) = MODEL 2 + adjustments for total cholesterol, smoking, diabetes and levels of physical activity. In each model, inverse probability weighting was implemented to account for the probability of survival until the end of follow-up Numbers in italics indicate statistical significance cIMT common carotid artery intima-media thickness, PWV pulse wave velocity, VMT verbal memory test, LSST letter search speed test, RT reaction time test Indicates log transformed dependent variables (Additional file 1: Table S4). There was a graded rela- category had a verbal memory test score similar to those tionship between increasing length of time being over- with onset of overweight/obesity at 36 years old. There weight/obese and decreasing verbal memory test was no association between patterns of overweight/obesity performance (β = −0.752 words per category of increasing and letter search, simple and choice reaction time tests length of time overweight/obese; 95% CI –1.157 to −0. (Additional file 1: Figures S1 and S2). 346; p for trend < 0.001). Individuals who were classified Full adjustment for cardiovascular risk factors as overweight/obese at age 36 years recalled 2.3 (95% CI (MODEL 3) only slightly reduced the association be- −3.5 to −1.1) fewer words compared to those who were al- tween increased duration of overweight/obesity exposure ways normal (Fig. 1a). Participants who were able to drop and verbal memory test (β = −0.666 words per category one BMI category and maintain a lower BMI had a similar of increasing length of time being overweight/obese; verbal memory test performance to those who had never 95% CI –1.119 to −0.213; p for trend = 0.004). Further been overweight or obese. Conversely, those who dropped adjustment of this association by PWV resulted in a 0.9% one BMI category but who subsequently moved up a change in the β coefficient (β = −0.672 words per category Table 3 Cross-sectional associations of carotid intima-media thickness and pulse wave velocity with verbal memory test, letter search speed and reaction time at 60–64 years VMT (n. of words) LSST (targets) S-RT (s) C-RT (s) β (95% CI) p β (95% CI) p β (95% CI) p β (95% CI) p PWV Model 1 −0.355 (−0.616 to −0.095) 0.008 0.004 (−0.018 to 0.026) 0.701 −0.958 (−3.676 to 1.760) 0.489 −1.871 (−5.485 to 1.741) 0.309 Model 2 −0.297 (−0.578 to −0.015) 0.039 0.006 (−0.020 to 0.032) 0.629 −0.344 (−3.402 to 2.714) 0.825 −1.357 (−5.270 to 2.557) 0.496 Model 3 −0.166 (0.461 to 0.129) 0.270 0.007 (−0.022 to 0.036) 0.634 −0.029 (−3.417 to 3.359) 0.987 −1.148 (−5.710 to 2.749) 0.492 cIMT Model 1 0.220 (−3.470 to 3.912) 0.906 −0.036 (−0.218 to 0.145) 0.695 6.819 (−27.267 to 40.906) 0.694 16.606 (−32.294 to 65.507) 0.505 Model 2 0.487 (−3.162 to 4.136) 0.793 −0.049 (−0.235 to 0.138) 0.607 10.673 (−24.072 to 45.418) 0.546 21.282 (−27.069 to 69.635) 0.388 Model 3 1.287 (−2.340 to 4.913) 0.486 −0.015 (−0.222 to 0.192) 0.888 12.470 (−24.800 to 49.740) 0.511 25.184 (−25.895 to 76.264) 0.333 Linear regression models were used to assess associations between variables cIMT common carotid artery intima-media thickness, PWV pulse wave velocity, VMT verbal memory test, LSST letter search speed test, C-RT choice reaction time test, S-RT simple reaction time test MODEL 1 adjusted for sex, education and childhood cognition; MODEL 2 = MODEL 1 + adjustments for socioeconomic position at age 53, systolic blood pressure and heart rate at age 60–64; MODEL 3 (fully adjusted) = MODEL 2 + adjustments for total cholesterol, smoking, diabetes and levels of physical activity. In each model, inverse probability weighting was implemented to account for the probability of survival until the end of follow-up Indicates log transformed dependent variables Numbers in italics indicate statistical significance Masi et al. BMC Medicine (2018) 16:75 Page 6 of 12 Fig. 1 a Patterns of overweight/obesity change and performance in the verbal memory test (VMT) at age 60–64 years. Data points represent mean number of words recalled and vertical bars indicate 95% CI for each group. O/O@36, O/O@43, O/O@53 and O/O@60–64 = Overweight/ obesity since 36, 43, 53 and 60–64 years old, respectively; Never O/O = never overweight/obese; Lost/Non-regain = dropped and did not regain one category of BMI; Lost/Regain = dropped and regained one category of BMI. VMT = Verbal memory test. Results are adjusted for sex, childhood cognition and education, and inverse probability weighting was implemented to account for dropout due to death. Test for trend for O/O@36 to O/O@60–64: p < 0.0001. Pairwise comparison of Never O/O versus Lost/Non-regain (p = 0.9) and Lost/Regained (p = 0.002). Pairwise comparison of O/O@36 versus Lost/Non-regain (p = 0.007) and Lost/Regain (p = 0.9). Adjustment for cardiovascular risk factors, PWV and cIMT did not affect these differences. b Patterns of waist circumference (WC) change and performance in the verbal memory test (VMT) at age 60–64 years. Data points represent mean number of words recalled and vertical bars indicate 95% CI for each group. HiWC@36, HiWC@43, HiWC@53 and HiWC@60–64 = elevated WC since 36, 43, 53 and 60–64 years old, respectively; Never HiWC = WC always normal; Lost/Non-regain = dropped and did not regain one category of WC; Lost/Regain = dropped and regained one category of WC. Results are adjusted for sex, childhood cognition and education, and inverse probability weighting was implemented to account for the probability of survival until the end of follow-up. Test for trend for HiWC@36 to HiWC@60–64: p < 0.001. Adjustment for cardiovascular risk factors, PWV and cIMT did not affect these differences of increasing length of time being overweight/obese; 95% When added to the model, the interaction between medi- CI –1.125 to −0.220; p for trend = 0.0037), and a 6.9% cation use and overweight/obese groups was not signifi- change of the β coefficient was observed when the same cant, and exclusion of participants with previous CVD association was adjusted for cIMT (β = −0.620 words per history did not substantially affect results. category of increasing length of time being overweight/ Analyses were repeated using patterns of WC in place obese; 95% CI –1.176 to −0.063; p for trend = 0.029). of BMI. Similar to findings for BMI, longer exposure to Masi et al. BMC Medicine (2018) 16:75 Page 7 of 12 elevated WC was associated with a worse cardiometa- S8 and Figure S6S). Evidence of effect modification by so- bolic profile and higher PWV and cIMT (Additional file cial class was observed for the association between BMI 1: Table S5). Greater length of time with elevated WC change from 36 to 43 years and choice reaction time, such was associated with decreases in verbal memory test that this association was stronger in those from more score, so that those with elevated WC from age advantaged social classes (Additional file 1:Table S9). 36 years had the lowest mean scores (β = −0.980 A greater increase in WC between 36 and 43 years words per category of increasing length of exposure was associated with poorer performance in all cognitive to elevated WC; 95% CI –1.424 to −0.536; p for trend tests at 60–64 years (Fig. 2a–d). The progressive adjust- <0.001) (Fig. 1b); this association remained in the ments from MODEL 1 to MODEL 3 minimally attenu- fully adjusted model (β = −1.051; 95% CIs −1.532 to ated the strength of these associations (Table 5). As for −0.571, p for trend < 0.001). Individuals who were BMI, a faster increase in WC between 53 to 60–64 years able to drop and not regain one category of WC had was related to lower verbal memory test score at age a similar memory performance as the group who al- 60–64, but this association was attenuated in MODEL 3 ways had a normal WC. Further adjustment of (Table 5). There was no evidence of effect modification MODEL 3 for PWV resulted in a 0.8% decrease in by socioeconomic position or education level of the as- the regression coefficient (β = −1.043 words per cat- sociation between WC increase and cognitive outcomes. egory of increasing length of exposure to elevated When changes of BMI and WC were included in the WC; 95% CIs −1.529 to −0.556; p for trend < 0.001) same fully adjusted model, faster WC gain between 36 and a 13.8% decrease was observed when the same and 43 years remained significantly associated with association was adjusted for cIMT (β = −0.906 words choice reaction time (β = 8.664; 95% CI 2.054 to 15.275; per category of increasing length of exposure to elevated p = 0.010), and a weaker association was observed with WC; 95% CI –1.508 to −0.304; p for trend = 0.003). No verbal memory test performance (β = −0.530; 95% CI –1. clear trends were observed for the associations between 116 to 0.050; p = 0.070). In contrast, the association be- categories of WC and letter search speed or simple reac- tween gain of WC at 36–43 years with letter search tion time tests (Additional file 1: Figures S3 and S4A). For speed as well as between gain of BMI at 36–43 and letter choice reaction time, participants with high WC at search speed became non-significant. 53 years and before had slower times than those with raised WC at 60–64 and those never having had a raised Discussion WC (β = 14.115 words per category of increasing length of This study shows that different patterns of whole body exposure to elevated WC; 95% CI 6.453 to 21.776; p for and abdominal obesity are associated with cognitive trend = 0.003) (Additional file 1:Figure S4B). function at 60–64 years. Cumulative exposure to ele- When both BMI and WC were included in the same vated BMI and WC over 30 years was related to poorer fully adjusted model, the linear trend across categories memory function at 60–64 years. We identified a sensi- of exposure to elevated WC for verbal memory test (β tive period in early adulthood when a faster gain of = −0.871 per change in WC category; 95% CI −1.720 to BMI and WC might have a greater impact on cognitive −0.024, p for trend = 0.044) and choice reaction time capacities in late midlife compared to weight gain in test (β = 14.087 per change in WC category; 95% CI 3. other periods, and show that patterns of cumulative ex- 300 to 24.873; p for trend = 0.011) remained significant, posure or rapid changes in WC remain associated with while the association between categories of BMI and ver- cognition even after adjustment for BMI. Finally, we bal memory test score was considerably reduced and no found that the relationships between patterns of adi- longer significant (β = −0.257 per change in BMI cat- posity and cognitive function were not explained by egory; 95% CI –0.979 to 0.465, p =0.485). CVD risk factors and vascular phenotypes. The process of neurodegeneration leading to cognitive decline and Association of BMI and WC gains in adulthood with dementia is complex and likely to result from the inter- cognitive function action of multiple factors. Our findings support the A faster increase in BMI between 53 to 60–64 years was adoption of early interventions based on the prevention related to lower verbal memory test score at age 60–64 of central and whole body obesity as possible measures (Additional file 1: Figure S5A), but this association was to reduce the burden of cognitive decline in the general attenuated in MODELS 2 and 3 (Table 4). A greater population. BMI increase between 36 and 43 years was related to The negative association between measures of abdom- lower log (letter search speed) at age 60–64 (Additional inal and whole body obesity with cognitive function ob- file 1: Figure S5B). This association remained highly sig- served in our survey is supported by previous nificant in the fully adjusted model (Table 4), and was epidemiological and genetic studies [3, 4, 7, 8, 31, 32]. stronger in females than in males (Additional file 1:Table However, only limited data are available on the potential Masi et al. BMC Medicine (2018) 16:75 Page 8 of 12 Table 4 Relationship between cognitive tests at age 60–64 years and residual changes in BMI between three time periods: BMI =36–43 years, BMI =43–53 years, and BMI =53–60 to 64 years 36–43 43–53 53–60 to 64 MODEL 1 MODEL 2 MODEL 3 β (95% CI) p β (95% CI) p β (95% CI) p VMT BMI −0.216 (−0.629 to 0.198) 0.306 −0.161 (−0.553 to 0.231) 0.421 −0.199 (−0.632 to 0.234) 0.367 36–43 BMI 0.065 (−0.390 to 0.520) 0.780 0.068 (−0.378 to 0.514) 0.763 0.098 (−0.393 to 0.590) 0.694 43–53 BMI −0.524 (−0.923 to −0.124) 0.010 −0.515 (−0.915 to −0.115) 0.012 −0.380 (−0.839 to 0.078) 0.104 53–60 to 64 LSST BMI −0.023 (−0.044 to −0.002) 0.029 −0.022 (−0.044 to −0.002) 0.035 −0.024 (−0.047 to −0.002) 0.034 36–43 BMI −0.012 (−0.033 to 0.010) 0.280 −0.011 (−0.033 to 0.010) 0.305 −0.008 (−0.031 to 0.014) 0.462 43–53 BMI 0.005 (−0.019 to 0.030) 0.680 0.005 (−0.020 to 0.029) 0.704 0.013 (−0.016 to 0.043) 0.367 53–60 to 64 S-RT BMI 2.116 (−1.973 to 6.206) 0.310 2.205 (−1.903 to 6.312) 0.290 3.701 (−0.836 to 8.238) 0.110 36–43 BMI 1.638 (−2.437 to 5.714) 0.430 1.775 (−2.377 to 5.928) 0.402 2.387 (−2.079 to 6.853) 0.294 43–53 BMI 0.698 (−3.119 to 4.514) 0.720 1.142 (−2.672 to 4.955) 0.557 1.859 (−2.514 to 6.233) 0.404 53–60 to 64 C-RT BMI −1.610 (−7.038 to 3.818) 0.561 −1.433 (−6.968 to 4.101) 0.611 −0.798 (−6.749 to 5.152) 0.792 36–43 BMI 2.581 (−3.305 to 8.468) 0.389 3.145 (−2.879 to 9.170) 0.306 3.021 (−3.570 to 9.612) 0.368 43–53 BMI 2.146 (−3.157 to 7.449) 0.427 2.716 (−2.537 to 7.969) 0.310 4.230 (−1.835 to 10.295) 0.171 53–60 to 64 Linear regression models were used to assess associations between variables. Significant associations (p < 0.05) are highlighted in bold Indicates log transformed dependent variables. MODEL 1 adjusted for sex, education and childhood cognition; MODEL 2 = MODEL 1 + adjustments for socioeconomic position at age 53, systolic blood pressure and heart rate at age 60–64; MODEL 3 (fully adjusted) = MODEL 2 + adjustments for total cholesterol, smoking, diabetes, diabetes duration and levels of physical activity. In each model, inverse probability weighting was implemented to account for dropout due to death VMT verbal memory test, LSST letter search speed test, S-RT simple reaction time test, C-RT choice reaction time influence of adult patterns of BMI on cognition. Using show that patterns of WC in adulthood could provide data from the NSHD cohort, Albanese et al.  docu- additional information in predicting late midlife cogni- mented that weight gain during specific periods of life is tive functions than patterns of BMI. associated with cognitive capacities at 53 years, although In cross-sectional and longitudinal analyses, adjust- this association was attenuated by socioeconomic pos- ment for PWV or cIMT had little effect on the rela- ition and childhood cognitive capacities. Because our tionship between adult patterns of BMI or WC and analyses use cognitive function and adiposity measures verbal memory test performance. The influence of from a later assessment, comparisons are difficult to cardiovascular factors on the association between make. Albanese et al.  also reported results stratified whole body and abdominal obesity with cognitive by sex, as significant sex × BMI interactions were identi- function has been previously explored, albeit with fied. In our sample, we found no evidence of effect conflicting results [2, 38, 39]. In these studies, cardio- modification by sex. Our results are broadly consistent vascular risk factors were measured only at a single with those obtained in the Whitehall II study by Sabia et point, and no measures of subclinical CVD were al. , where a dose–response relationship was identified available. As cIMT and PWV are recognised markers between longer exposure to obesity and lower cognitive of end-organ damage and reflect the lifetime burden of function at 60 years old. Similarly, we also provide infor- cardiovascular risk factor exposure, the minimal attenu- mation on the importance of patterns of WC in addition ation of the association between patterns of WC and BMI to BMI for cognition, as well as on the different impact with memory function after adjustment for PWV and of rapid weight gain at different age of adult life on later cIMT suggests that obesity and vascular factors might cognition. Previous studies exploring the relationship be- affect cognitive function by different mechanisms and tween indices of central adiposity and cognitive function should be treated early and concomitantly to reduce the are based on samples of older adults (65+) [6, 33], were risk of cognitive impairment. This is supported by results cross-sectional [2, 34–36] or had small sample sizes . of recent clinical trials, wherein multidomain interven- We are the first to report the influence of cumulative ex- tions have been indicated as those likely to represent the posure to elevated WC and of rapid changes of WC dur- most effective strategies to improve cognitive function in ing adulthood on different cognitive outcomes, and to overweight populations . Masi et al. BMC Medicine (2018) 16:75 Page 9 of 12 Fig. 2 Association of rate of change in waist circumference (WC) in three time periods (36–43 years, 43–53 years and 53–60 to 64 years) according to MODEL 1 with (a) verbal memory test (VMT), (b) letter search speed test (LSST), (c) simple reaction time test and (d) choice reaction time (C-RT) at age 60–64 years. β represents the slope of the linear regression and indicates the difference in units of the cognitive outcome for 1 standard deviation increase in WC in each interval Our study has several strengths. The NSHD is the 60–64 years remained broadly representative of the longest-running longitudinal study in the UK, with mul- British-born population of that age. Finally, the results tiple measures of height, weight and WC available at dif- in relation to the groups who achieve a stable weight re- ferent ages. It includes individuals without cognitive duction should be interpreted with caution, as only a impairment and is generally representative of the limited number of participants had sustained weight British-born population of similar age. The availability of loss/re-gain during follow-up. multiple vascular and cognitive measures enabled ex- ploration of the association between different vascular Conclusion phenotypes and a wide range of cognitive domains in Increasing cumulative exposure to elevated BMI and WC late mid-life, with appropriate adjustment for environ- in adulthood is associated with lower memory function at mental and behavioural factors. 60–64 years, and a rapid gain of WC across the third and Nevertheless, the study also has limitations. First, we fourth decades is associated with a global reduction of examined associations in an observational study and cognitive capacities in later life. Cardiovascular risk factors therefore cannot reliably assign causality. Second, the and vascular phenotypes are unlikely to account for these outcome of our analysis was cognitive function and associations. Our findings suggest that lifelong prevention more studies are necessary to test the relevance of our of whole body and abdominal obesity, particularly in early findings against the risk of dementia. Third, attrition is midlife, might represent the most effective strategy to unavoidable in long-running studies such as NSHD, but prevent the burden of cognitive decline attributable to previous analyses have shown that the samples at 53 and obesity in the general population. Masi et al. BMC Medicine (2018) 16:75 Page 10 of 12 Table 5 Relationship between cognitive tests at age 60–64 years and residual changes in waist circumference between three time periods: Waist =36–43 years, Waist =43–53 years, and Waist =53–60 to 64 years 36–43 43–53 53–60 to 64 MODEL 1 MODEL 2 MODEL 3 β (95% CI) P β (95% CI) P β (95% CI) P VMT WC −0.500 (−0.930 to −0.071) 0.022 −0.447 (−0.875, −0.019) 0.041 −0.501 (−0.972 to −0.030) 0.037 36–43 WC −0.160 (−0.562 to 0.241) 0.434 −0.127 (−0.525 to 0.270) 0.530 −0.090 (−0.511 to 0.330) 0.672 43–53 WC −0.488 (−0.863 to −0.113) 0.011 −0.444 (−0.822 to −0.067) 0.021 −0.358 (−0.766 to 0.049) 0.085 53–60 to 64 LSST WC −0.026 (−0.051 to −0.002) 0.035 −0.025 (−0.050, −0.001) 0.045 −0.027 (−0.055 to 0.001) 0.058 36–43 WC −0.001 (−0.026 to 0.024) 0.932 0.0003 (−0.025 to 0.025) 0.983 0.001 (−0.026 to 0.029) 0.916 43–53 WC 0.012 (−0.016 to 0.040) 0.395 0.013 (−0.015 to 0.041) 0.372 0.018 (−0.014 to 0.051) 0.273 53–60 to 64 S-RT WC 4.011 (0.420 to 7.602) 0.029 4.140 (0.473 to 7.806) 0.027 4.435 (0.386 to 8.486) 0.032 36–43 WC 1.604 (−2.445 to 5.654) 0.437 1.718 (−2.408 to 5.844) 0.414 2.466 (−2.002 to 6.936) 0.279 43–53 WC 1.453 (−2.357 to 5.263) 0.454 1.753 (−2.125 to 5.631) 0.375 2.196 (−2.120 to 6.512) 0.318 53–60 to 64 C-RT WC 5.471 (.356 to 10.588) 0.036 0.787 (0.599 to 10.976) 0.029 60.086 (0.389 to 11.783) 0.036 36–43 WC 5.024 (−0.390 to 10.439) 0.069 5.530 (0.095 to 10.965) 0.046 50.251 (−0.752 to 11.254) 0.086 43–53 WC 1.670 (−3.389 to 6.728) 0.517 1.940 (−3.090 to 6.971) 0.449 30.589 (−1.963 to 9.141) 0.205 53–60 to 64 Linear regression models were used to assess associations between variables VMT verbal memory test, LSST letter search speed test, C-RT choice reaction time test, S-RT simple reaction time test Significant associations (p < 0.05) are highlighted in bold Indicates log transformed dependent variables MODEL 1 adjusted for sex, education and childhood cognition; MODEL 2 = MODEL 1 + adjustments for socioeconomic position at age 53, systolic blood pressure and heart rate at age 60–64; MODEL 3 (fully adjusted) = MODEL 2 + adjustments for total cholesterol, smoking, diabetes, diabetes duration and levels of physical activity. In each model, inverse probability weighting was implemented to account for dropout due to death Numbers in italics indicate statistical significance Additional file Funding The UK Medical Research Council provides core funding for the MRC National Survey of Health and Development and supports RH and MR (grant Additional file 1: Additional Methods and Results (including additional numbers MC_UU_12019/1, MC_UU_12019/2 and MC_UU_12019/3). WJ is Tables and Figures). (DOC 688 kb) supported by a UK Medical Research Council (MRC) New Investigator Research Grant (MR/P023347/1), and acknowledges support from the National Institute for Health Research (NIHR) Leicester Biomedical Research Centre, which is a partnership between University Hospitals of Leicester NHS Trust, Loughborough University, and the University of Leicester. JD and SM Abbreviations are supported by the British Heart Foundation. BMI: Body Mass Index; CVD: Cardiovascular Disease; cIMT: common carotid artery Intima-Media Thickness; C-RT: Choice Reaction Time Test; LSST: Letter Search Speed Test; MRC: Medical Research Council; NSHD: National Survey of Availability of data and materials Health and Development; PWV: Pulse Wave Velocity; S-RT: Simple Reaction NSHD data are made available to bona fide researchers who submit data Time Test; VMT: Verbal Memory Test; WC: Waist Circumference requests to firstname.lastname@example.org; see also the full policy documents at http://www.nshd.mrc.ac.uk/data.aspx. https://doi.org/10.5522/NSHD/Q101; https://doi.org/10.5522/NSHD/Q102; https://doi.org/10.5522/NSHD/S102A. Acknowledgements The authors are grateful to National Survey of Health and Development (NSHD) participants who took part in this latest data collection for their Authors’ contributions continuing support. We thank members of the NSHD scientific and data SM, JD and RH planned the study. PW, MR, RH, JD, AH and AW collected collection team at the following centres: MRC Unit for Lifelong Health and data. SM, GG, WJ and TK analysed the data. SM wrote the manuscript. JD, RH, Ageing; MRC Lifecourse Epidemiology Unit, University of Southampton; MRC MR, AH, PW, AW, WJ, MC and GG reviewed/edited the manuscript. SM, MR, Human Nutrition Research, Cambridge; Wellcome Trust (WT) Clinical JD, RH and AH contributed to the discussion and reviewed/edited the Research Facility (CRF) Manchester and the Department of Clinical Radiology manuscript. All authors read and approved the final manuscript. at the Central Manchester University Hospitals NHS Foundation Trust; WT CRF and Medical Physics at the Western General Hospital in Edinburgh; WT CRF and the Department of Nuclear Medicine at University Hospital Birmingham; WT CRF and the Department of Nuclear Medicine at University Ethics approval and consent to participate College London Hospital; the CRF and the Department of Medical Physics at All participants provided written informed consent, and the study received the University Hospital of Wales; the CRF and Twin Research Unit at St ethical approval from the Central Manchester Research Ethics Committee Thomas’ Hospital London. (07/H1008/245) and the Scottish A Research Ethics Committee (08/MRE00/12). Masi et al. BMC Medicine (2018) 16:75 Page 11 of 12 Consent for publication 14. Zebekakis PE, Nawrot T, Thijs L, Balkestein EJ, van der Heijden-Spek J, Van This manuscript does not contain details, images, or videos relating to an Bortel LM, Struijker-Boudier HA, Safar ME, Staessen JA. Obesity is associated individual person, and thus does not require written informed consent for with increased arterial stiffness from adolescence until old age. J Hypertens. the publication of these details. 2005;23(10):1839–46. 15. Buratti L, Balucani C, Viticchi G, Falsetti L, Altamura C, Avitabile E, Provinciali Competing interests L, Vernieri F, Silvestrini M. Cognitive deterioration in bilateral asymptomatic The authors declare that they have no competing interests. severe carotid stenosis. Stroke. 2014;45(7):2072–7. 16. Frazier DT, Seider T, Bettcher BM, Mack WJ, Jastrzab L, Chao L, Weiner MW, DeCarli C, Reed BR, Mungas D, et al. The role of carotid intima-media Publisher’sNote thickness in predicting longitudinal cognitive function in an older adult Springer Nature remains neutral with regard to jurisdictional claims in cohort. Cerebrovasc Dis. 2014;38(6):441–7. published maps and institutional affiliations. 17. Masley SC, Masley LV, Gualtieri CT. Cardiovascular biomarkers and carotid IMT scores as predictors of cognitive function. J Am Coll Nutr. 2014;33(1):63–9. Author details 18. Mitchell GF, van Buchem MA, Sigurdsson S, Gotal JD, Jonsdottir MK, National Centre for Cardiovascular Prevention and Outcomes, Institute of Kjartansson O, Garcia M, Aspelund T, Harris TB, Gudnason V, et al. Arterial Cardiovascular Science, University College London, 1 St Martin le Grande, stiffness, pressure and flow pulsatility and brain structure and function: the London EC1A 4NP, UK. Department of Clinical and Experimental Medicine, Age, Gene/Environment Susceptibility–Reykjavik study. Brain. 2011;134(Pt University of Pisa, Pisa, Italy. First Department of Cardiology|, Hippokration 11):3398–407. Hospital, University of Athens, Athens, Greece. School of Sport, Exercise and 19. Rogne SO, Solbu MD, Arntzen KA, Herder M, Mathiesen EB, Schirmer H. Health Sciences, Loughborough University, Loughborough, Leicestershire, UK. Albuminuria and carotid atherosclerosis as predictors of cognitive function MRC Unit for Lifelong Health and Ageing at UCL, Institute of Epidemiology in a general population. Eur Neurol. 2013;70(5–6):340–8. and Health Care, UCL, London, UK. Division of Population Health Sciences 20. Scuteri A, Wang H. Pulse wave velocity as a marker of cognitive impairment and Education, St George’s University of London, London, UK. in the elderly. J Alzheimers Dis. 2014;42(Suppl 4):S401–10. Cardiometabolic Phenotyping Group, Institute of Cardiovascular Science, 21. Hughes TM, Kuller LH, Barinas-Mitchell EJ, Mackey RH, McDade EM, Klunk UCL, London, UK. WE, Aizenstein HJ, Cohen AD, Snitz BE, Mathis CA, et al. Pulse wave velocity is associated with beta-amyloid deposition in the brains of very elderly Received: 16 December 2017 Accepted: 23 April 2018 adults. Neurology. 2013;81(19):1711–8. 22. Hughes TM, Kuller LH, Barinas-Mitchell EJ, McDade EM, Klunk WE, Cohen AD, Mathis CA, Dekosky ST, Price JC, Lopez OL. Arterial stiffness and beta-amyloid References progression in nondemented elderly adults. JAMA Neurol. 2014;71(5):562–8. 1. Beydoun MA, Beydoun HA, Wang Y. Obesity and central obesity as risk 23. Shetty P. Celebrating 65 years of the NSHD cohort. Lancet. 2011;377(9768):802. factors for incident dementia and its subtypes: a systematic review and 24. Albanese E, Hardy R, Wills A, Kuh D, Guralnik J, Richards M. No association meta-analysis. Obesity reviews: an official journal of the International between gain in body mass index across the life course and midlife Association for the Study of Obesity. 2008;9(3):204–18. cognitive function and cognitive reserve–the 1946 British Birth Cohort 2. Dore GA, Elias MF, Robbins MA, Budge MM, Elias PK. Relation between study. Alzheimers Dement. 2012;8(6):470–82. central adiposity and cognitive function in the Maine-Syracuse Study: 25. KuhD,Pierce M,Adams J,Deanfield J,Ekelund U, FribergP,Ghosh AK,Harwood attenuation by physical activity. Ann Behav Med. 2008;35(3):341–50. N, Hughes A, Macfarlane PW, et al. Cohort profile: updating the cohort profile for 3. Gustafson D, Rothenberg E, Blennow K, Steen B, Skoog I. An 18-year follow- the MRC National Survey of Health and Development: a new clinic-based data up of overweight and risk of Alzheimer disease. Arch Intern Med. 2003; collection for ageing research. Int J Epidemiol. 2011;40(1):e1–9. 163(13):1524–8. 26. Wadsworth M, Kuh D, Richards M, Hardy R. Cohort Profile: The 1946 4. Gustafson DR, Backman K, Waern M, Ostling S, Guo X, Zandi P, Mielke MM, National Birth Cohort (MRC National Survey of Health and Development). Bengtsson C, Skoog I. Adiposity indicators and dementia over 32 years in Int J Epidemiol. 2006;35(1):49–54. Sweden. Neurology. 2009;73(19):1559–66. 27. Department of Nutrition for Health and Development WHO: Waist 5. Qizilbash N, Gregson J, Johnson ME, Pearce N, Douglas I, Wing K, Evans SJ, Circumference and Waist–Hip Ratio: Report of a WHO Expert Consultation. Pocock SJ. BMI and risk of dementia in two million people over two decades: a Geneva: World Health Organization 2008. http://apps.who.int/iris/bitstream/ retrospective cohort study. Lancet Diabetes Endocrinol. 2015;3(6):431–6. handle/10665/44583/9789241501491_eng.pdf?sequence=1. 6. Rodriguez-Fernandez JM, Danies E, Martinez-Ortega J, Chen WC. Cognitive 28. Stein JH, Korcarz CE, Hurst RT, Lonn E, Kendall CB, Mohler ER, Najjar SS, decline, body mass index, and waist circumference in community-dwelling Rembold CM, Post WS, American Society of Echocardiography Carotid elderly participants. J Geriatr Psychiatry Neurol. 2017;30(2):67–76. Intima-Media Thickness Task Force. Use of carotid ultrasound to identify 7. Sabia S, Kivimaki M, Shipley MJ, Marmot MG, Singh-Manoux A. Body mass subclinical vascular disease and evaluate cardiovascular disease risk: a index over the adult life course and cognition in late midlife: the Whitehall consensus statement from the American Society of Echocardiography II Cohort Study. Am J Clin Nutr. 2009;89(2):601–7. Carotid Intima-Media Thickness Task Force. Endorsed by the Society for 8. Marioni RE, Yang J, Dykiert D, Mottus R, Campbell A, Group CCW, Davies G, Vascular Medicine. J Am Soc Echocardiogr. 2008;21(2):93–111. quiz 189–190 Hayward C, Porteous DJ, Visscher PM, et al. Assessing the genetic overlap 29. Weber T, Ammer M, Rammer M, Adji A, O'Rourke MF, Wassertheurer S, between BMI and cognitive function. Mol Psychiatry. 2016;21(10):1477–82. Rosenkranz S, Eber B. Noninvasive determination of carotid-femoral pulse 9. Laurent S, Cockcroft J, Van Bortel L, Boutouyrie P, Giannattasio C, Hayoz D, wave velocity depends critically on assessment of travel distance: a Pannier B, Vlachopoulos C, Wilkinson I, Struijker-Boudier H, et al. Expert comparison with invasive measurement. J Hypertens. 2009;27(8):1624–30. consensus document on arterial stiffness: methodological issues and clinical 30. Stafford M, Black S, Shah I, Hardy R, Pierce M, Richards M, Wong A, Kuh D. Using applications. Eur Heart J. 2006;27(21):2588–605. a birth cohort to study ageing: representativeness and response rates in the 10. O'Leary DH, Bots ML. Imaging of atherosclerosis: carotid intima-media National Survey of Health and Development. Eur J Ageing. 2013;10(2):145–57. thickness. Eur Heart J. 2010;31(14):1682–9. 31. Fitzpatrick AL, Kuller LH, Lopez OL, Diehr P, O'Meara ES, Longstreth WT Jr, 11. Charakida M, Khan T, Johnson W, Finer N, Woodside J, Whincup PH, Sattar N, Luchsinger JA. Midlife and late-life obesity and the risk of dementia: Kuh D, Hardy R, Deanfield J. Lifelong patterns of BMI and cardiovascular cardiovascular health study. Arch Neurol. 2009;66(3):336–42. phenotype in individuals aged 60-64 years in the 1946 British birth cohort 32. Kivipelto M, Ngandu T, Fratiglioni L, Viitanen M, Kareholt I, Winblad B, study: an epidemiological study. Lancet Diabetes Endocrinol. 2014;2(8):648–54. Helkala EL, Tuomilehto J, Soininen H, Nissinen A. Obesity and vascular risk 12. Safar ME, Czernichow S, Blacher J. Obesity, arterial stiffness, and factors at midlife and the risk of dementia and Alzheimer disease. Arch cardiovascular risk. J Am Soc Nephrol. 2006;17(4 Suppl 2):S109–11. Neurol. 2005;62(10):1556–60. 13. Recio-Rodriguez JI, Gomez-Marcos MA, Patino-Alonso MC, Agudo-Conde C, Rodriguez-Sanchez E, Garcia-Ortiz L. Vasorisk g: Abdominal obesity vs 33. Driscoll I, Espeland MA, Wassertheil-Smoller S, Gaussoin SA, Ding J, Granek general obesity for identifying arterial stiffness, subclinical atherosclerosis IA, Ockene JK, Phillips LS, Yaffe K, Resnick SM, et al. Weight change and and wave reflection in healthy, diabetics and hypertensive. BMC Cardiovasc cognitive function: findings from the Women's Health Initiative Study of Disord. 2012;12:3. Cognitive Aging. Obesity. 2011;19(8):1595–600. Masi et al. BMC Medicine (2018) 16:75 Page 12 of 12 34. Jagust W, Harvey D, Mungas D, Haan M. Central obesity and the aging brain. Arch Neurol. 2005;62(10):1545–8. 35. Kerwin DR, Zhang Y, Kotchen JM, Espeland MA, Van Horn L, McTigue KM, Robinson JG, Powell L, Kooperberg C, Coker LH, et al. The cross-sectional relationship between body mass index, waist-hip ratio, and cognitive performance in postmenopausal women enrolled in the Women's Health Initiative. J Am Geriatr Soc. 2010;58(8):1427–32. 36. Nourhashemi F, Andrieu S, Gillette-Guyonnet S, Reynish E, Albarede JL, Grandjean H, Vellas B. Is there a relationship between fat-free soft tissue mass and low cognitive function? Results from a study of 7,105 women. J Am Geriatr Soc. 2002;50(11):1796–801. 37. Yoon DH, Choi SH, Yu JH, Ha JH, Ryu SH, Park DH. The relationship between visceral adiposity and cognitive performance in older adults. Age Ageing. 2012;41(4):456–61. 38. Cournot M, Marquie JC, Ansiau D, Martinaud C, Fonds H, Ferrieres J, Ruidavets JB. Relation between body mass index and cognitive function in healthy middle-aged men and women. Neurology. 2006;67(7):1208–14. 39. Wolf PA, Beiser A, Elias MF, Au R, Vasan RS, Seshadri S. Relation of obesity to cognitive function: importance of central obesity and synergistic influence of concomitant hypertension. The Framingham Heart Study. Curr Alzheimer Res. 2007;4(2):111–6. 40. Ngandu T, Lehtisalo J, Solomon A, Levalahti E, Ahtiluoto S, Antikainen R, Backman L, Hanninen T, Jula A, Laatikainen T, 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(9984):2255–63.
BMC Medicine – Springer Journals
Published: May 28, 2018
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
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
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.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera