Association of Coffee Consumption with MRI Markers and Cognitive Function: A Population-Based Study

Abstract

Background: Coffee is one of the most widely consumed beverages worldwide and has been of considerable interest in research on cognition and dementia.

Objective: To investigate the effect of coffee on preclinical brain MRI markers of dementia and cognitive performance.

Methods: In 2,914 participants from the population-based Rotterdam Study (mean age: 59.3±7.2 years, 55% females), we assessed coffee consumption, performed brain MRI, and assessed cognition at baseline. To study cognitive change, cognitive assessment was repeated after 5 years of follow-up. Coffee consumption was analyzed continuously (per cup increase) and in categories (0–1, >1–3, >3 cups/day). Using logistic and linear regression, associations of coffee consumption with lacunar infarcts and brain tissue volumes on MRI, and cognitive performance (cross-sectional and longitudinal) were investigated, adjusting for relevant confounders.

Results: We found that higher coffee consumption was associated with a lower prevalence of lacunar infarcts [odds ratio per cup increase: 0.88 (95% CI:0.79;0.98)], and smaller hippocampal volume [difference: –0.01 (95% CI:–0.02;0.00)]. Also, we found that the highest category of coffee consumption was associated with better performance on the Letter Digit Substitution Task [difference: 1.13(95% CI:0.39;1.88)], Word Fluency test [0.74(95% CI:0.04,1.45)], Stroop interference task [1.82(95% CI:0.23;3.41)], and worse performance on the 15-Word Learning test delayed recall [–0.38(95% CI:–0.74;–0.02)]. These associations were not found when cognition was analyzed longitudinally.

Conclusion: We found complex associations between coffee consumption, brain structure, and cognition. Higher coffee consumption was cross-sectionally associated with a lower occurrence of lacunar infarcts and better executive function, but also with smaller hippocampal volume and worse memory function.

Keywords

Brain imaging brain tissue coffee consumption cognitive function epidemiology MRI

INTRODUCTION

Coffee is one of the most widely consumed beverages worldwide and has therefore been of considerable interest regarding its potential effects on health [1 –4]. Coffee is a complex mixture consisting of more than 1,000 different compounds with different physiological effects [5]. Compounds that are thought to be beneficial for health include caffeine, which is a strong neurostimulant, and polyphenols that have antioxidant properties [5, 6]. Several studies demonstrated that coffee consumption has an overall beneficial effect on various diseases, including Parkinson’s disease [7], diabetes [1, 2], cardiovascular disease [2, 3] and even on mortality, specifically stroke-related mortality [9 –11].

Importantly, more recent evidence now also links coffee consumption to a lower risk of dementia [4 , 12–14]. Especially in light of the huge societal burden of dementia, further research into mechanisms underlying the relation between coffee consumption and dementia risk is vital. Key features of dementia are its multifactorial etiology, and its long pre-clinical phase [15, 16]. Already during this phase, several changes in the structure of the brain take place, such as white matter lesions or early hippocampal atrophy, which can be visualized in-vivo using magnetic resonance imaging (MRI). In addition to these structural brain changes, subtle changes in cognitive performance may also already occur, which can be investigated using dedicated cognitive testing [17]. Elucidating the link between coffee consumption and these pre-clinical markers of neurodegenerative disease is vital to further establish the role of coffee in the etiology of neurodegenerative diseases, including dementia. Previous studies have investigated the association of coffee consumption with cognition or brain MRI-markers, but these had small samples, focused only one or two cognitive domains, or included persons not at risk of dementia (e.g., healthy young volunteers) [4 , 18]. Therefore, it remains largely unknown how coffee consumption affects pre-clinical markers of dementia.

Against this background, in a population-based setting of community-dwelling older adults, we investigated the association between habitual coffee consumption and structural brain-MRI-markers and cognitive function.

MATERIALS AND METHODS

Setting

This study was embedded within the population-based Rotterdam Study [19], which aims to investigate determinants of chronic diseases in middle aged and elderly persons. The study started in 1990 when all inhabitants of Ommoord (a district of Rotterdam, The Netherlands) that were aged 55 years and older were invited to participate. In total, 7,983 (87% of 10,215 invitees) were enrolled in the study. In 2000, the cohort was expanded with another 3,011 participants, with the same inclusion criteria. In a second expansion in 2006, the cohort was extended with 3,932 participants aged 45 years or older. Every 3–4 years, all participants undergo an extensive home interview and a physical examination at the research center. From 2005 onwards, all participants that came to the research center were invited for a brain MRI examination [22]. For the current study, we included all participants (n = 3,871) that visited the research center and underwent MRI between 2005 and 2009 (i.e., the second center visit for the first cohort expansion, and the first visit of the second cohort expansion). Coffee consumption and cognitive performance were assessed during a visit at the research center which was close to the MRI assessment (mean time interval in months: 6.2, 95% CI: 6.0;6.4). Additionally, a second assessment of cognitive performance was done during the following center visit to assess change in cognitive performance (mean time interval between cognitive tests in years: 5.5, 95% CI: 5.4;5.5).

The Rotterdam Study is approved by the Medical Ethics Committee of the Erasmus MC and by the Ministry of Health, Welfare and Sport of the Netherlands, implementing the “Wet Bevolkingsonderzoek: ERGO (Population Studies Act: Rotterdam Study)”. Written informed consent was obtained from all participants.

Sample for analyses

A total of 3,871 persons underwent MRI between 2005 and 2009. Of these, 13 MRI examinations had to be excluded due to image artifacts. Moreover, we excluded all persons with prevalent dementia (n = 16), cortical infarcts (n = 99), and prevalent clinical stroke (n = 55). Data on coffee consumption were not available for 774 participants because not all participants received a food frequency questionnaire. This was due to unavailability of trained personnel at times to administer the questionnaire. Therefore, information on coffee consumption was available for 2,914 persons, constituting the total sample available for the cross-sectional analyses. Importantly, not all persons underwent all cognitive tests, resulting in slightly different totals for each cognitive test (Fig. 1). The longitudinal analyses on cognition were performed on a total of 2,454 participants who completed both the cognitive examination at baseline and at follow-up.

Assessment of coffee consumption

Coffee consumption was assessed as part of a validated semi-quantitative food frequency questionnaire indicating all foods and drinks consumed more than once a month during the preceding year [20]. The questionnaire comprised food items and all relevant beverages, including coffee, and was administered by a trained dietician. Participants reported their habitual coffee intake as number of cups per day, week, or month. The dietary coffee consumption was converted into mL of coffee per day, and then into cups of coffee per day. Each cup corresponds to approximately 125 mL or 7 g of coffee [21].

Brain MRI and post-processing

Brain MRI-scanning was performed on a 1.5T-scanner with an eight-channel head coil (GE Healthcare, Milwaukee, Wisconsin, USA), and included a T1-weighted (T1 w) sequence, a proton-density (PDw) weighted sequence, and a fluid-attenuated-inversion-recovery (FLAIR) sequence [22]. Automated brain tissue classification based on a k-nearest-neighbor-classifier algorithm extended with white matter lesion segmentation was used to quantify brain volume, grey matter volume, white matter volume, white matter lesion volume, left and right hippocampal volume, and intracranial volume (in milliliters) [23 –25]. Lacunar and cortical infarcts were rated on the FLAIR, PDw, and T1w sequences [26]. Lacunar infarcts were dichotomized into absent versus present. Total hippocampal volume was defined as the sum of the left and right hippocampal volumes.

Cognitive test battery

Both at baseline and at follow-up, participants underwent a neuropsychological test battery that included the Letter Digit Substitution Task (LDST) [27], the Stroop test [28], Word Fluency test (WFT) [29], the 15-Word Learning test (WLT) [30], and the Purdue Pegboard (PPB) both hands [31]. The LDST was used to evaluate processing speed and executive function [27]. The Stroop test measured the speed of reading (subtask 1), the speed of color naming (subtask 2), and interference of automated processing and attention (subtask 3) [28]. The WFT was used to evaluate efficiency of verbal ability, lexical access ability, retrieval from semantic memory, and executive function [29]. The 15-WLT measured verbal learning (the immediate recall task), retrieval from verbal memory (the delayed recall task) and recognition of verbal memory (the recognition task) [30]. Finally, the PPB test was used to evaluate dexterity and fine motor skills [31]. Higher scores indicate better performance on all cognitive tests, except for the Stroop tests where a higher score indicates worse performance as it measures time taken to complete the task. In order to compare scores for the Stroop test directly with the other tests, these were inverted [32].

Covariates

Information on relevant covariates was gathered by interview, physical examination, and laboratory tests [19]. Educational attainment was assessed by interview and defined as university degree, higher vocational education, general secondary education, intermediate vocational education, lower secondary education, lower vocational education, and primary education. Body mass index was calculated as weight in kilograms/height in meters-squared (kg/m²). Hypertension was defined as a blood pressure≥140/90 mmHg or use of blood pressure lowering medication prescribed for the indication of hypertension [33]. Fasting serum glucose, total cholesterol, and high-density lipoprotein (HDL) cholesterol levels were acquired by an automated enzymatic procedure. Diabetes mellitus type 2 was defined as a fasting serum glucose level≥7.0 mmol/L, or use of anti-diabetic medication [34]. Alcohol consumption and smoking habits were assessed by interview and for alcohol use participants were categorized as users or nonusers, while for smoking, participants were categorized as current, former, or never smokers. Previous coronary heart disease status was defined as myocardial revascularization (as a proxy for significant coronary artery disease) and/or myocardial infarction. We assessed a depressive symptoms score using the Dutch version of the Center for Epidemiological Studies Depression scale (CES-D) [35] and the use of psychoanaleptic drugs by self report.

Statistical analysis

We compared characteristics between the participants included and excluded from the analyses using Analysis of Variance (ANOVA) for continuous variables, and chi-square tests for categorical variables adjusting for age and sex where applicable. White matter lesion-volume was natural log-transformed due to the positively skewed distribution.

We used logistic regression to investigate the association between coffee consumption and the presence of MRI-based lacunar infarcts, whilst linear regression was used to investigate relationships of coffee consumption with MRI-based brain tissue volumes (total brain volume, grey matter volume, white matter volume, hippocampal volume), white matter lesion volumes. These analyses were adjusted for age, sex, intracranial volume (as measure of head size), educational attainment, and time interval between assessment of coffee consumption and MRI measures in model 1. Additionally, we adjusted these associations for body mass index, hypertension, diabetes mellitus, total cholesterol levels, HDL-cholesterol levels, history of coronary heart disease, alcohol consumption, smoking, depressive symptoms, and psychoanaleptic drugs use in model 2.

Cognitive function was investigated cross-sectionally and longitudinally using linear regression models. These models were identical to model 1 and 2 as described above, but excluding intracranial volume. Moreover, we constructed a third model (model 3), in which we included the brain tissue volumes and lacunar infarcts that were associated with coffee consumption in order to test whether the association between coffee consumption and cognitive performance was influenced by these measures. For the longitudinal analyses we used two approaches. In the first approach we used the cognitive tests at follow-up as outcome and adjusted for the test-score at baseline. In the second approach we calculated the difference between the test score at follow-up and that at baseline, and used this difference as outcome in the regression model (Supplementary Table 2). We adjusted according to the models described above, and additionally for the time interval between the tests.

We repeated the analyses of model 2 and model 3 for the cross-sectional analyses, using coffee consumption with three categories as 0–1 cup/day, >1–3 cups/day, and >3 cups/day [21], to assess non-linear associations between coffee consumption and our outcomes. Also, we compared these groups on potential differences in baseline characteristics. Finally, we explored effect modification by sex for all associations. Analyses were conducted using Stata 13.0 (Stata Corporation, College Station, USA).

RESULTS

Characteristics of the study population are presented in Table 1. The mean age of the study population was 59.3±7.2 years, and 55% were women. As follows from Table 1, participants that were not in the analyses due to missing data on coffee consumption were more likely to be younger, less educated, currently smoking, non-users of alcohol, and to have more depressive symptoms. In addition, Supplementary Table 1 also demonstrates the population characteristics according to the different categories of coffee consumption.

Table 2 shows the associations of habitual coffee consumption with lacunar infarcts and brain tissue volumes. We found that higher coffee consumption was associated with a lower prevalence of lacunar infarcts [odds ratio (OR) per cup increase in coffee consumption 0.88 (95% confidence interval (CI):0.80;0.98)], even after additional adjustment for cardiovascular risk factors [OR 0.88 (95% CI:0.79;0.98)]. However, higher coffee consumption was also associated with a smaller hippocampal volume [difference in hippocampal volume per cup increase in coffee consumption: –0.01 (95% CI: –0.02;0.00)].

Table 3 shows the cross-sectional associations between coffee consumption and cognitive performance on the different cognitive tests. We found that coffee consumption was associated with better performance on the LDST, but with a worse performance on the 15-WLT delayed recall test. These associations persisted after adjustment for hippocampal volume and presence of lacunar infarcts [difference in LDST: 0.11 (95% CI:0.01;0.22); and for delayed recall: –0.06 (95% CI:–0.11;–0.01)]. When investigating the effect of coffee on cognitive function in a longitudinal fashion, we only found an association with decline on the LDST when using the difference-score as outcome (Table 4 and Supplementary Table 2).

When investigating categories of coffee consumption we found that consumption of more than 3 cups of coffee per day was related to a lower prevalence of lacunar infarcts as compared to consumption of 0-1 cups per day [OR 0.44 (95% CI:0.24;0.81)] (Fig. 2). Also, consumption of more than 3 cups per day was associated with a better performance on the LDST [difference as compared to consumption of 0-1 cups/day: 1.13 (95% CI:0.39;1.88)], the WFT [0.74 (95% CI:0.04;1.45)], the Stroop color naming subtask [0.68 (95% CI:0.16;1.20)], and the Stroop interference subtask [1.82 (95% CI:0.23;3.41)]. This category was also related to worse performance on the 15-WLT delayed recall [–0.38 (95% CI:–0.74;–0.02)] (Fig. 2).

After reanalyzing all associations, we found evidence for effect modification by sex only for the relation between coffee consumption and performance on the 15-WLT delay recall (p_interaction: 0.034). This relation was more pronounced in women that in men [difference in delayed recall in men: –0.03 (95% CI:–0.10;0.03), and in women: –0.11 (95% CI: –0.18;–0.03)].

DISCUSSION

In this large sample of community-dwelling middle-age and older adults, we found that higher habitual coffee consumption was associated with a lower prevalence of lacunar infarcts, and with smaller hippocampal volumes. Moreover, coffee consumption was cross-sectionally related to a better performance on cognitive tests of executive function, but worse performance on memory. We found that persons that consumed more coffee were less likely to have lacunar infarcts. This suggests that the beneficial effect of coffee on cardiovascular health may also extend to the brain [36, 37]. A potential underlying mechanism may be the anti-inflammatory effect of coffee. Given that inflammation is a key feature in the cascade of atherosclerosis, and atherosclerosis is the most important risk factor for cerebral infarcts, suppression of the inflammation component by coffee could be protect against the occurrence of these infarcts [36, 37]. Further evidence for a beneficial effect of coffee on cerebrovascular pathology, comes from studies demonstrating that coffee consumption is associated with a decreased risk of clinical stroke [2 , 37].

In contrast to the potential beneficial effect of coffee consumption on the occurrence of lacunar infarcts, we found that higher coffee consumption was also specifically associated with smaller hippocampal volumes. To our knowledge this is the first time this effect of coffee on the hippocampus is shown. In contrast, evidence from animal studies shows that chronic coffee consumption modulates the endogenous antioxidant system in the brain [38], and protects from cerebral degeneration and subsequent dementia. These contradictory results urge for the further investigation of the specific effect of coffee on the hippocampus.

With regard to cognitive performance, we also found both beneficial and detrimental effects of coffee consumption on different cognitive domains. Most prominently, we found that higher coffee consumption (>3 cups/day) was related to better performance on the LDST, WFT, and the Stroop test, all tests that center around executive function and processing speed. We did not find evidence that the most optimal coffee consumption was any other category than the highest consumption. Other studies have also reported on a potential beneficial effect of coffee on cognitive functioning [6 , 39]. Yet, especially in light of our finding on the relation between higher coffee consumption and a lower prevalence of lacunar infarcts, the effect on executive function is particularly interesting, because of all cognitive domains, executive function is thought to be primarily affected by vascular brain disease [17, 40]. Notably, the association between coffee consumption and executive function still persisted after adjustment for lacunar infarcts, suggesting that other underlying processes are also involved, warranting future research. Despite the positive effects on cognitive performance, and in contrast with others [14, 40], we found detrimental effects of coffee consumption on the 15-WLT delayed recall, which relates to memory function. Interestingly, this finding fits well with our finding of higher coffee consumption being related with a smaller volume of the hippocampus, which is an important structure for memory function [41]. Notably, previous studies [13 , 42], suggest that the effects of coffee might be stronger in women, especially with regard to memory function. Of all associations we investigated, we only found effect modification by sex for the relation between coffee consumption and the 15-WLT delayed recall. Interestingly, this was indeed more pronounced in women, underlining these findings. In contrast to the cross-sectional findings between coffee consumption and cognition, we did not find any association when investigating the effect of coffee on long-term cognitive change. Only when we used the change in cognitive score between follow-up and baseline as outcome, we found an association between coffee consumption and decline on the LDST. Yet, we should be careful not to over interpret this single finding in light of the phenomenon of regression to the mean [43]. That we did not find longitudinal associations is in line with our previous work on the relation between coffee consumption and the risk of dementia, in which we neither found an association with the long-term risk of dementia [21]. A possible explanation for this finding may be that there is only a short-term beneficial effect of coffee on cognitive performance. Given that coffee is a short-acting neurostimulant [5, 6], its effect may be beneficial in the beginning but diminishes with prolonged use. We found a similar short-term effect for the risk of dementia [21]. Yet, evidence for this hypothesis remains scarce and needs further extension during the coming years.

The strengths of our study include the population-based design, large sample size, extensive phenotyping, and the focus on both brain MRI biomarkers and cognitive function in the same population. Yet, there are also several considerations that should be taken into account. First, we could not account for the coffee preparation method (filtered, boiled, etc.), the type of coffee (espresso, ristretto, or lungo), or the use of additions (i.e., sugar or milk), which directly influences the amount of consumed coffee ingredients. It is, however, important to acknowledge that coffee is a very complex mixture of approximately 800–1000 chemical compounds that may all have different effects [44], directly highlighting the impossibility to study all compounds separately. Second, we did not have information on changes in coffee consumption habits during the follow-up period, which could have influenced performance to a certain extent. However, we note that there is evidence that coffee consumption habits remain relatively stable over time [14]. Third, although we adjusted for many important confounding factors in our study, there is a possibility of residual confounding due to other not measured factors, such as use of anticholinesterase and antiparkinsonian drugs. Fourth, although we assessed cognitive also in a longitudinal way, the effect of coffee consumption on brain MRI markers could only be assessed cross-sectionally. Nonetheless, the finding that associations with objective quantifiable structural MRI-markers correspond to those found for cognitive tests does add to the robustness of the findings.

In summary, we found complex associations between coffee consumption, brain structure, and cognition. Higher coffee consumption is associated with a lower occurrence of lacunar infarcts and better executive function, but also with smaller hippocampal volume and worse memory function. These intriguing results warrant further longitudinal studies on the effect of coffee on preclinical neurodegenerative disease.

Footnotes

ACKNOWLEDGMENTS

The Rotterdam Study is sponsored by the Erasmus Medical Center and Erasmus University Rotterdam, The Netherlands Organization for Scientific Research (NWO), The Netherlands Organization for Health Research and Development (ZonMW), The research Institute for Diseases in the Elderly (RIDE), The Netherlands Genomics Initiative, the Ministry of Education, Culture and Science, the Ministry of Health, Welfare and Sports, the European Commission (DG XII), and the Municipality of Rotterdam. Further support was obtained from the Institute for Scientific Information on Coffee (ISIC). LFA received a doctoral scholarship from Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG) and was a research fellow of Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), grant n° 99999.002279/2014-02, as a visiting scholar at Erasmus University Medical Center in Rotterdam-The Netherlands. LFA receive a postdoctoral scholarship from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), grant n° 150248/2015-6. None of the funding organizations or sponsors were involved in the study design; in collection, analysis, or interpretation of data; in writing the report; and in making decision to submit the article for publication.

Authors’ disclosures available online ().

Appendix

References

Ding

, Bhupathiraju

, Chen

, van Dam

, Hu

(2014) Caffeinated and decaffeinated coffee consumption and risk of type 2 diabetes: A systematic review and a dose-response meta-analysis. Diabetes Care 37, 569–586.

Ding

, Bhupathiraju

, Satija

, van Dam

, Hu

(2014) Long-term coffee consumption and risk of cardiovascular disease: A systematic review and a dose-response meta-analysis of prospective cohort studies. Circulation 129, 643–659.

Chrysant

(2015) Coffee consumption and cardiovascular health. Am J Cardiol 116, 818–821.

Santos

, Costa

, Santos

, Vaz-Carneiro

, Lunet

(2010) Caffeine intake and dementia: Systematic review and meta-analysis. J Alzheimers Dis 20, 187–204.

Caprioli

, Cortese

, Sagratini

, Vittori

(2015) The influence of different types of preparation (espresso and brew) on coffee aroma and main bioactive constituents. Int J Food Sci Nutr 66, 505–513.

Chu

, Brown

, Lyle

, Chen

, Black

, Williams

, Lin

, Hsu

, Cheng

(2009) Roasted coffees high in lipophilic antioxidants and chlorogenic acid lactones are more neuroprotective than green coffees. J Agric Food Chem 57, 9801–9808.

Ross

, Abbott

, Petrovitch

, Morens

, Grandinetti

, Tung

, Tanner

, Masaki

, Blanchette

, Curb

, Popper

, White

(2000) Association of coffee and caffeine intake with the risk of Parkinson disease. JAMA 283, 2674–2679.

Natella

, Scaccini

(2012) Role of coffee in modulation of diabetes risk. Nutr Rev 70, 207–217.

Freedman

, Park

, Abnet

, Hollenbeck

, Sinha

(2012) Association of coffee drinking with total and cause-specific mortality. N Engl J Med 366, 1891–1904.

10.

Malerba

, Turati

, Galeone

, Pelucchi

, Verga

, La Vecchia

, Tavani

(2013) A meta-analysis of prospective studies of coffee consumption and mortality for all causes, cancers and cardiovascular diseases. Eur J Epidemiol 28, 527–539.

11.

Saito

, Inoue

, Sawada

, Shimazu

, Yamaji

, Iwasaki

, Sasazuki

, Noda

, Iso

, Tsugane

(2015) Association of coffee intake with total and cause-specific mortality in a Japanese population: The Japan Public Health Center-based Prospective Study. Am J Clin Nutr 101, 1029–1037.

12.

Ribeiro

, Sebastiao

(2010) Caffeine and adenosine. J Alzheimers Dis 20, 3–15.

13.

Arab

, Khan

, Lam

(2013) Epidemiologic evidence of a relationship between tea, coffee, or caffeine consumption and cognitive decline. Am Soc Nutr 4, 115–122.

14.

Vercambre

, Berr

, Ritchie

, Kang

(2013) Caffeine and cognitive decline in elderly women at high vascular risk. J Alzheimers Dis 35, 413–421.

15.

Shukitt-Hale

(1999) The effects of aging and oxidative stress on psychomotor and cognitive behavior. Age 22, 9–17.

16.

Shukitt-Hale

, Miller

, Chu

, Lyle

, Joseph

(2013) Coffee, but not caffeine, has positive effects on cognition and psychomotor behavior in aging. Age 35, 2183–2192.

17.

Gelber

, Petrovitch

, Masaki

, Ross

, White

(2011) Coffee intake in midlife and risk of dementia and its neuropathologic correlates. J Alzheimers Dis 23, 607–615.

18.

Arab

, Biggs

, O’Meara

, Longstreth

, Crane

, Fitzpatrick

(2011) Gender differences in tea, coffee, and cognitive decline in the elderly: The Cardiovascular Health Study. J Alzheimers Dis 27, 553–566.

19.

Hofman

, Brusselle

GGO

, Murad

, van Duijn

, Franco

, Goedegebure

, Ikram

, Klaver

CCW

, Nijsten

TEC

, Peeters

, Stricker

, Tiemeier

, Uitterlinden

, Vernooij

(2015) The Rotterdam Study: 2016 objectives and design update. Eur J Epidemiol 30, 661–708.

20.

Klipstein-Grobusch

, den Breeijen

, Goldbohm

, Geleijnse

, Hofman

, Grobbee

, Witteman

(1998) Dietary assessment in the elderly: Validation of a semiquantitative food frequency questionnaire. Eur J Clin Nutr 52, 588–596.

21.

Mirza

, Tiemeier

, de Bruijn

RFAG

, Hofman

, Franco

, de Jong

, Koudstaal

, Ikram

(2014) Coffee consumption and incident dementia. Eur J Epidemiol 29, 735–741.

22.

Ikram

, van der Lugt

, Niessen

, Koudstaal

, Krestin

, Hofman

, Bos

, Vernooij

(2015) The Rotterdam Scan Study: Design update 2016 and main findings. Eur J Epidemiol 30, 1299–1315.

23.

de Boer

, Vrooman

, van der Lijn

, Vernooij

, Ikram

, van der Lugt

, Breteler

MMB

, Niessen

(2009) White matter lesions extension to automatic brain tissue segmentation on MRI. Neuroimage 45, 1151–1161.

24.

Vrooman

, Cocosco

, van der Lijn

, Stokking

, Ikram

, Vernooij

, Breteler

, Niessen

(2007) Multi-sprectral brain tissue segmentation using automatically trained k-Nearrest-Neighbor classification. Neuroimage 37, 71–81.

25.

van der Lijn

, den Heijer

, Breteler

, Niessen

(2008) Hippocampus segmentation in MR images using atlas registration, voxel classification, and graph cuts. Neuroimage 43, 708–720.

26.

Vernooij

, van der Lugt

, Ikram

, Wielopolski

, Niessen

, Hofman

, Krestin

, Breteler

(2008) Prevalence and risk factors of cerebral microbleeds: The Rotterdam Scan Study. Neurology 70, 1208–1214.

27.

Lezak

, Howieson

, Loring

(2004) Neuropsychological Assessment, Oxford University Press, New York.

28.

Houx

, Jolles

, Vreeling

(1993) Stroop interference: Aging effects assessed with the Stroop color-word test. Exp Aging Res 19, 209–224.

29.

Welsh

, Butters

, Mohs

, Beekly

, Edland

, Fillenbaum

, Heyman

(1994) The consortium to establish a registry for Alzheimer’s disease (CERAD). Part V. A normative study of the Neuropsychological battery. Neurology 44, 609–614.

30.

Bleecker

, Bolla-Wilson

, Agnew

, Meyers

(1988) Age related sex differences in verbal memory. J Clin Psychol 44, 403–411.

31.

Tiffin

, Asher

(1848) The Purdue pegboard; norms and studies of reliability and validity. J Appl Psychol 32, 234–247.

32.

Hoogendam

, Hofman

, van der Geest

, van der Lugt

, Ikram

(2014) Patterns of cognitive function in aging: The Rotterdam Study. Eur J Epidemiol 29, 133–140.

33.

Whitworth

(2003) World Health Organization (WHO)/International Society of Hypertension (ISH) statement on management of hypertension. J Hypertens 21, 1983–1992.

34.

Alberti

, Zimmet

(1998) Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: Diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med 15, 539–553.

35.

Sawyer-Radloff

(1977) The CES-D scale: A self-report depression scale for research in the general population. Appl Psychol Meas 1, 385–401.

36.

Andersen

, Jacobs

, Carlsen

, Blomhoff

(2006) Consumption of coffee is associated with reduced risk of death attributed to inflammatory and cardiovascular diseases in the Iowa Women’s health study. Am J Clin Nutr 83, 1039–1046.

37.

Broussard

, Mytar

, Li

, Klapstein

(2012) The role of inflammatory processes in Alzheimer’s disease. Inflammopharmacology 20, 109–126.

38.

Abreu

, Silva-Oliveira

, Moraes

MFD

, Pereira

, Moraes-Santos

(2011) Chronic coffee and caffeine ingestion effects on the cognitive function and antioxidant system of rat brains. Pharmacol Biochem Behav 99, 659–664.

39.

Ritchie

, Carriere

, de Ma Portet

, Dartigues

, Rouaud

, Barberger-Gateau

, Ancelin

(2007) The neuroprotective effects of caffeine: A prospective population study (The Three CitySstudy). Neurology 69, 536–545.

40.

Costenla

, Cunha

, de Mendonca

(2010) Caffeine, adenosine receptors, and synaptic plasticity. J Alzheimers Dis 20, 25–34.

41.

Duarte

, Agostinho

, Carvalho

, Cunha

(2012) Caffeine consumption prevents diabetes-induced memory impairment and synaptotoxicity in the hippocampus of NONcZNO10/LTJ mice. PLoS One 7, e21899.

42.

Perlaki

, Orsi

, Kovacs

, Schwarcz

, Pap

, Kalmar

, Plozer

, Csatho

, Gabriel

, Komoly

, Janszky

, József Janszky

(2011) Coffee consumption may influence hippocampal volume in young women. Brain Imaging Behav 20, 274–284.

43.

Vickers

, Altman

(2001) Statistics notes: Analysing controlled trials with baseline and follow up measurements. BMJ 323, 1123–1124.

44.

Carman

, Dacks

, Lane

, Shineman

, Fillit

(2014) Current evidence for the use of coffee and caffeine to prevemt age-related cognitive decline and Alzheimer’s disease. J Nutr Health Aging 18, 383–392.