The Proportion of APOE4 Carriers Among Non-Demented Individuals: A Pooled Analysis of 389,000 Community-Dwellers

Abstract

Background:

The apolipoprotein E epsilon 4 (APOE4) is the strongest genetic risk factor for sporadic Alzheimer’s disease (AD). Its carriage percentage in non-demented population varies across geographic regions and ethnic groups.

Objective:

To estimate the proportion of APOE4 (2/4, 3/4, or 4/4) carriers in non-demented community-dwellers.

Methods:

PubMed, EMBASE, and China National Knowledge Infrastructure were searched from inception to April 20, 2020. Community-based studies that reported APOE polymorphisms with a sample of≥500 non-demented participants were included. Random-effects models were used to pool the results. Meta-regression and subgroup analyses were performed to test the source of heterogeneity and stratified effects. Age-standardized pooled proportion estimates (ASPPE) were calculated by direct standardization method.

Results:

A total of 121 studies were included, with a pooled sample of 389,000 community-dwellers from 38 countries. The global average proportion of APOE4 carriers was 23.9% (age-standardized proportion: 26.3%; 2.1% for APOE4/4, 20.6% for APOE3/4 and 2.3% for APOE2/4), and varied significantly with geographical regions (from 19.3% to 30.0%) and ethnic groups (from 19.1% to 37.5%). The proportion was highest in Africa, followed by Europe, North America, Oceania, and lowest in South America and Asia (p < 0.0001). With respect to ethnicity, it was highest in Africans, followed by Caucasians, and was lowest in Hispanics/Latinos and Chinese (p < 0.0001).

Conclusion:

APOE4 carriers are common in communities, especially in Africans and Caucasians. Developing precision medicine strategies in this specific high-risk population is highly warranted in the future.

Keywords

Alzheimer’s disease proportion worldwide

INTRODUCTION

The apolipoprotein E (APOE) gene is a well-established risk factor for sporadic Alzheimer’s disease (AD) [1]. It has three three major allelic variants: E2, E3, and E4. Prior work has shown that possession of the APOE4 allele is associated with a greater risk of AD dementia [2 –5]. By contrast, APOE2 may exert protective effects on cognitive decline and delay the onset of AD [6, 7]. Genome-wide association studies discovered that the risk of AD was increased in APOE4 carriers compared with those who had the APO E3/3 genotype (odds ratio [OR] = 3.68) (http://www.alzgene.org/). Moreover, individuals who were homozygous for APOE4 had an extremely increased risk of Alzheimer’s disease, compared with those who were homozygous for APOE3 (OR: 31, 95% confidence interval [CI]: 17 to 59) [8] and those who were homozygous for APOE2 (OR: 250, 95% CI: 67 to 935) [1]. In addition, the causal relationship of APOE4 with AD was further enhanced by lines of evidence from longitudinal studies regarding cerebrospinal fluid (CSF) or brain positron-emission tomography (PET), such that APOE4 allele carriers developed cerebral amyloid-β (Aβ) deposition at a more rapid rate than non-carriers, even at an earlier age [9, 10]. All these findings highlight the central roles of APOE4 in AD genetic architecture and pathogenesis. More interestingly, AD environmental risk and protective factors were found to interact with the APOE4 gene, suggesting that prevention or intervention strategies may differ between APOE4 carriers and non-carriers [11 –13]. Accordingly, APOE4 carriers deserve particular concerns in developing precision medicine strategy of preventing and managing AD.

The critical question is how many APOE4 carriers exist and no systematic review in a global context has been conducted until now. Previous studies reporting polymorphisms at the APOE locus yielded inconsistent conclusions and APOE4 proportion vary substantially (ranging from 3% to 40%) [14 –18]. The heterogeneity might derive from geographic regions, ethnic groups, population source, subject diagnosis, age strata, and sample size. Herein, we summarized the global proportion of APOE4 carriers in non-demented community-dwellers and explored the allelic variation in different human groups and regions.

METHODS

Search strategy and selection criteria

We systematically searched PubMed, EMBASE, and China National Knowledge Infrastructure databases with the search term “APOE4” from inception to April 20, 2020 with no language restriction. The following inclusion criteria must be met simultaneously: 1) community-based studies; 2) participants without dementia; 3) sample size of at least 500; 4) data available on the proportion of APOE4 carriers. Studies were excluded if they met either of the following two criteria: 1) abstracts, editorials, or comments where the proportion of APOE4 was not accessible; 2) special recruitment sources such as donors, volunteers, staff members, or patients with specific comorbidities (due to uncertain generalizability). As for the studies from the same population, we only included the most comprehensive one with the largest sample size. The bibliographies of relevant research articles and systematic reviews were also hand-searched to identify additional potential data sources. Literature selection was performed independently by two investigators (YYW and YJG). When there was disagreement, the third investigator (WX) was involved until consensus was reached.

Data extraction

Information extracted included first author’s name, year of publication, cohort name, basic demographic information (including region, country, and continent), ethnicity, cognitive status, data sources, sample size, mean age or/and age range, proportion of females, proportion of APOE4 carriers, as well as APOE genotypes and alleles (Supplementary Table 1). Information about latitude and longitude were collected from OvitalMap software (https://en.ovital.com/). If any data mentioned above were unavailable, we attempted to obtain them via contacting the corresponding authors.

Quality assessment

The modified quality assessment tool developed by Hoy and his colleagues [19] was employed to evaluate the methodological quality of eligible studies. The tool assessed three aspects, including sample representativeness, the rationality of dementia assessment, and the rationality of the methods used to obtain the proportion of APOE4 carriers. We assigned each item a score of 1 (yes) or 0 (no), and the total score (ranged from 0 to 10) was regarded as a proxy for assessment of the overall risk of bias for each single study. According to the total scores, we classified studies as having low (>8), moderate (6–8), or high (≤5) risk of bias [20]. Two investigators (YYW and YJG) independently assessed study quality, and disagreements were resolved with consensus via discussion.

Statistical analyses

APOE4 carriers were defined as individuals who carry at least one APOE4 allele (2/4, 3/4, or 4/4). The proportion was counted manually with raw data when possible. The pooled proportion of APOE4 carriers and APOE genotypes with 95% CI was calculated using a random-effects meta-analysis model with Freeman-Tukey double arcsine transformation [21, 22]. Heterogeneity between the studies was assessed by Cochran’s Q test and quantified by the I² statistic. The source of heterogeneity was investigated via subgroup analyses according to multiple variables, including continent (Asia, Africa, Europe, North America, South America, and Oceania), ethnicity (Caucasians, Chinese, African Americans, Africans, and Hispanics/Latinos), gender, mean age (≤65 y versus > 65 y), mean sample size (MSS, the total number of participants divided by the total number of included studies; ≤MSS versus > MSS), cognitive status (normal cognition versus mild cognitive impairment), and risk of bias (high, moderate, and low). We also calculated age-standardized pooled proportion estimate (ASPPE) by direct standardization method. After subtracting the proportion of demented people, we used the population pyramid of the world (http://www.populationpyramid.net/) to adjust the structural age between different age groups. Univariable and multivariable meta-regression analyses were also performed to explore the source of heterogeneity. Variables with p-values of < 0.2 in the univariable analysis were entered into the multivariable model. The following variables were tested: continent, ethnicity, mean age, proportion of females, sample size, quality scores, year of publication, longitude and latitude. We assessed the publication bias using Egger test. A p-value less than 0.10 of Egger test indicated significant publication bias. Finally, the associations of the proportion of APOE4 carriers with latitude and longitude were assessed by Spearman’s correlation analyses.

R version 4.0.0 and statistical package for social studies SPSS version 25 were used for all these statistical analyses. The R package for meta-analysis was employed to generate the pooled estimates, forest plots, and bubble plots. The ‘metareg’ package was used to conduct meta-regression. Spearman’s correlation analyses were performed using SPSS version 25. Maps of the APOE allele distribution in the world were drawn by Smartdraw version 2013. A p-value less than 0.05 was considered statistically significant.

RESULTS

Search results and study characteristics

Figure 1 shows the flow diagram of study selection. The search yielded a total of 11,098 citations. After title and abstract screening, 10,665 articles were excluded, leaving 433 potentially eligible for full-text screening. We then assessed the full-texts, and 352 articles were excluded for specific reasons (Fig. 1). After further integrating with additional bibliographies of relevant studies and reviews, a total of 121 articles (388,894 samples) were finally included, consisting of 29 from Asia, 40 from Europe, 39 from North America, 4 from South America, 3 from Africa, 4 from Oceania, and 2 intercontinental. A total of 117 articles reported the proportion of APOE4 carriers. The mean ages of participants in 45% of included studies were greater than 65; and in the remaining 55% of the studies, the mean ages were not more than 65. The detailed characteristics of studies included in the meta-analysis can be seen in Supplementary Table 1. The study quality was rated as generally acceptable with a mean score of 8.7 (SD = 0.8). Specifically, 107 studies were rated with a low risk of bias; 14 studies were rated with a moderate risk of bias; and none with a high risk of bias (Supplementary Table 2).

Fig. 1

Flow diagram of study selection.

Proportion of APOE4 carriers

Among 374,356 participants from 117 studies in 37 countries, a total of 92,836 participants were found to carry at least one APOE4 allele. The global average proportion of APOE4 carriers was 23.9% (95% CI: 23.0% to 24.9%, p = 0.000, I² = 97.7%) (Fig. 3 or Supplementary Table 3–1), while the ASPPE of APOE4 carriers was 26.3% (95% CI: 25.7% to 26.9%) (Supplementary Figure 3 or Supplementary Table 3-4). In the analysis of APOE genotypes, the ASPPE was 2.1% (95% CI: 2.0% to 2.3%) for APOE4/4, 20.6% (95% CI: 20.1% to 21.2%) for APOE3/4, 2.3% (95% CI: 2.2% to 2.5%) for APOE2/4, 0.89% (95% CI: 0.82% to 0.98%) for APOE2/2, 11.1% (95% CI: 10.8% to 11.5%) for APOE2/3 and 62.8% (95% CI: 61.9% to 63.7%) for APOE3/3 (Supplementary Table 3-5). Additionally, the ASPPE of APOE2 was 7.4% (95% CI: 7.2% to 7.7%) and APOE3 was 79.1% (95% CI: 78.4% to 79.7%). More details about ASPPE of APOE2 and APOE3 can been seen in Supplementary Figure 3 or STable 3-4.

Fig. 2

Geographical gradients of APOE4 carriers in global populations. Worldwide distribution of the APOE4 carriers in included countries, symbol colors represented the frequency of APOE4 carriers and the size of the solid black circle represented the sample size for analysis (A); Proportion of APOE genotypes in included countries arranged from south to north by latitude (B); Bubble plot of the proportion of APOE4 carriers in global populations, drawn according to latitude and longitude (C). APOE, apolipoprotein E; Not included*, No data meeting inclusion criteria; the “Longitude 0” in Figure (C) represents the Greenwich prime meridian, while the “Latitude 0” represents the equator.

Fig. 3

Forest plot of the pooled proportion of APOE4 carriers stratified by demographical and methodological characteristics. The numbers of samples and the N estimates in subgroups vary because of the available data from the selected studies. Stratified analyses indicated that the significance of the primary result was related to continent, ethnicity, cognitive status and mean age, but not related to gender, mean sample size or quality scores. CI, confidence interval; N estimates, number of estimates; CN, cognitive normally; MCI, mild cognitive impairment.

The pooled results of stratified meta-analyses were shown in Fig. 3 and Supplementary Table 3-1. Accordingly, the ASPPE of APOE4 carriers was highest in Africa (0.308; 95% CI: 0.293 to 0.324), followed by Europe (0.265; 95% CI: 0.261 to 0.269), North America (0.254; 95% CI: 0.247 to 0.261), and Oceania (0.233; 95% CI: 0.216 to 0.251), and was lowest in South America (0.192; 95% CI: 0.177 to 0.208) and Asia (0.180; 95% CI: 0.172 to 0.189) (p < 0.0001). With respect to ethnicity, the ASPPE was highest in African Americans (0.381; 95% CI: 0.365 to 0.398), followed by Africans (0.308; 95% CI: 0.280 to 0.338), and Caucasians (0.262; 95% CI: 0.253 to 0.271), and was lowest in Hispanics/Latinos (0.204; 95% CI: 0.193 to 0.217) and Chinese (0.187; 95% CI: 0.173 to 0.201) (p < 0.0001). Participants with mild cognitive impairment (MCI) had a higher ASPPE than cognitively normal individuals (0.307, 95% CI: 0.240 to 0.388 versus 0.184, 95% CI: 0.173 to 0.195; p = 0.034) (Supplementary Figure 3 or Supplementary Table 3-4). Participants in their midlife tended to have a higher pooled proportion than older adults (0.249, 95% CI: 0.237 to 0.262 versus 0.227, 95% CI: 0.210 to 0.248; p = 0.042) (Fig. 3 and Supplementary Table 3–1). The pooled proportion showed no difference between subgroups stratified by gender, sample size, or quality scores. The results of the meta-regression analysis were shown in Supplementary Table 4–1. Ethnicity explained 13.7% of the observed heterogeneity (p < 0.05). None of the other demo-graphical and methodological characteristics was significantly related to the heterogeneity in the multivariable meta-regression analysis.

Geographical gradient of APOE allele

The correlation analyses showed that the proportion of APOE4 carriers was positively correlated with latitude (r = 0.366, p = 0.001, Fig. 4A) and negatively correlated with longitude east (r = –0.335, p = 0.024, Fig. 4B) globally. These correlations still persisted in the northern hemisphere (latitude: r = 0.465, p = 0.000; longitude: r = –0.240, p = 0.039; Supplementary Table 5). Europe exhibited a remarkable south-to-north gradient, with the proportion of APOE4 carriers increasing from southern to northern countries (r = 0.874, p = 0.000, Supplementary Figure 4). However, we did not observe a similar south-to-north gradient in Asia or North America (p > 0.05, Supplementary Table 5–1). As for Africa, South America, and Oceania, correlation analyses could not be performed due to the small numbers of included studies (N < 5). In addition, we also conducted correlation analyses between the APOE2 allele with latitude and longitude, but no apparent geographical gradient was found of APOE2 proportion (Supplementary Table 5).

Fig. 4

Correlation analyses of the proportion of APOE4 carriers with latitude (A) and longitude east (B) globally. The method we used is the Spearman’s correlation analysis. The proportion of APOE4 carriers is positively correlated with latitude (r = 0.366, p = 0.001, A) and negatively correlated with longitude east (r = –0.335, p = 0.024, B). The “Latitude 0” on the abscissa in Figure (A) represents the equator, while the “Longitude 0” in Figure (B) represents the Greenwich prime meridian.

DISCUSSION

Our meta-analysis suggested that the global distribution of APOE4 carriers varies with geographical areas and ethnicity. About one in four people carry at least one APOE4 allele in the world. The proportion of APOE4 carriers was higher in Africa, Europe, and North America than in other regions, such as Oceania, South America, and Asia. Africans and Caucasians had a higher proportion of APOE4 carriers than Hispanics/Latinos and Chinese.

The distributions of APOE4 carriers showed marked latitudinal and longitudinal diversity globally. Inconsistent with Eisenberg’s study [14], we could not detect the curvilinear relationship between the proportion of APOE4 carriers and latitude in the northern hemisphere. But when we restricted the analysis to latitudes greater than 35 degrees, there was a remarkable south-to-north gradient, in agreement with Eisenberg’s study. However, no negative correlation between proportion of APOE4 carriers and latitude was found in the regions below 35 degrees north in the current study. This may be due to our stringent inclusion criteria, resulting in the limited numbers of included studies in these regions. In subgroup analyses stratified by continents, only Europe exhibited a remarkable south-to-north gradient, with the proportion of APOE4 carriers increasing from south to north [15]. However, we did not find a significant south-to-north gradient in Asia and North America. The discrepancies may be partly due to different inclusion criteria for eligible studies or historical migration patterns.

The main mechanisms commonly invoked to explain geographic gradients in APOE4 allele proportion are migration events and natural selective advantages [14 , 23–27]. E4 is often considered to be the ancestral allele [28] and is higher in Africa [15, 27]. The ancestors of modern humans, which originated in Africa, probably migrated first to the Middle East and Central Asia and then to Asia and Europe [29]. The distribution of APOE4 allele might accompany migrations of anatomically modern humans out of Africa. A growing body of evidence also suggests that natural selective advantages may explain the distribution of APOE4 carriers [14 , 24–27], potentially including interactions of APOE4 with modifiable risk and protective factors, such as diet, lifestyle, and environment [24–27 , 30–32]. High latitudes with high metabolic rates due to extreme climates and the increased vitamin D demands in the geographical regions subjected to low solar radiation may play a role in the selection of north-to-south gradient of APOE4 allele distribution [14 , 33] in Europe. As for the detected negative correlation of APOE4 proportion with longitude, one possible explanation is that APOE3 allele might be oriented and prevailed in Asia, accompanied by the gradual decrease of APOE4 proportion in Eastern regions. While in the Western, the age-related disadvantages of APOE4 allele might be diluted on the basis of certain ethnic backgrounds, so the APOE4 proportion reduction was hampered, contributing to a relatively higher APOE4 proportion in Western regions [34].

Despite a higher proportion of APOE4 carriers, the age-standardized prevalence of AD in Europe was relatively low, as estimated by the Global Burden of Diseases, Injuries, and Risk Factors Study [35]. We assumed that the risk of carrying the APOE4 allele might be weakened by some modifiable protective factors in European samples, such as higher education, healthy lifestyles (e.g., Mediterranean diet), and health care systems [36]. The hypothesis could provide valuable insights for developing practical recommendations, especially for APOE4 carriers. Many clinical studies have reported that APOE4 allele carriers appeared to be more susceptible to the effects of some modifiable risk factors on AD [12, 13]. Considering the significant contribution of APOE4 to AD [36, 37], the development of APOE-targeted AD therapeutic strategies and behavioral interventions for those at-risk individuals is highly needed.

The strength of this study is the large sample of community-dwellers free of dementia, which represented a homogeneous population and thus could increase the power to reveal the genetic distribution. Several limitations exist. First, although our samples represent a more homogeneous population based on large sample size, caution is warranted due to the poor coverage of certain subgroups in our study, such as in Oceania or Africa, which may confound the average proportion we estimated. Second, we found substantial heterogeneity in estimation of the proportion of APOE allele across studies. Thus, we ran subgroup analysis and meta-regression to examine the source of heterogeneity. However, these variables could not thoroughly explain the source of heterogeneity, suggesting that other unknown confounding variables might be the source of heterogeneity, such as sampling methods, methods of cognitive assessment. Third, the assessment of AD was not confirmed by autopsy. Considering approximately one-third of cognitively intact older adults suffered from AD pathology [38, 39], some ascertainment bias might have been introduced due to the inclusion of preclinical AD. Last, some aged communities in our pooled analysis may already exclude the dementia cases caused by APOE4, which may inevitably lead to a relatively lower proportion of APOE4 carriers in these studies. Similarly, as APOE4 is associated with reduced human longevity [40, 41], the inclusion of aged participants will inevitably lead to the relatively lower proportion of APOE4 carriers in these studies. Future studies should be conducted to confirm our findings.

CONCLUSION

The distribution of APOE4 carriers exhibited geographical and ethnic differences, and approximately one in four people carry at least one APOE4 allele in non-demented community-dwellers globally. Future international investigations with large sample sizes and careful documentation of age are needed to further explore the geographic gradients and to elucidate the underlying mechanisms. Additionally, the development of precision medicine strategy, especially for AD prevention and therapy, in this specific high-risk population is highly warranted.

Footnotes

ACKNOWLEDGMENTS

This study was supported by grants from the National Natural Science Foundation of China (82001136 and 81901121).

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Serrano-Pozo

, Das

, Hyman

(2021) APOE and Alzheimer’s disease: Advances in genetics, pathophysiology, and therapeutic approaches. Lancet Neurol 20, 68–80.

Corder

, Saunders

, Strittmatter

, Schmechel

, Gaskell

, Small

, Roses

, Haines

, Pericak-Vance

(1993) Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer’s disease in late onset families. Science 261, 921–923.

Liu

, Liu

, Kanekiyo

, Xu

, Bu

(2013) Apolipoprotein E and Alzheimer disease: Risk, mechanisms and therapy. Nat Rev Neurol 9, 106–118.

Zhong

, Weisgraber

(2009) Understanding the association of apolipoprotein E4 with Alzheimer disease: Clues from its structure. J Biol Chem 284, 6027–6031.

, Tan

, Hardy

(2014) Apolipoprotein E in Alzheimer’s disease: An update. Annu Rev Neurosci 37, 79–100.

Suri

, Heise

, Trachtenberg

, Mackay

(2013) The forgotten APOE allele: A review of the evidence and suggested mechanisms for the protective effect of APOE ɛ2. Neurosci Biobehav Rev 37, 2878–2886.

Corder

, Saunders

, Risch

, Strittmatter

, Schmechel

, Gaskell

Jr. , Rimmler

, Locke

, Conneally

, Schmader

, Small

, Roses

, Haines

, Pericak-Vance

(1994) Protective effect of apolipoprotein E type 2 allele for late onset Alzheimer disease. Nat Genet 7, 180–184.

Reiman

, Arboleda-Velasquez

, Quiroz

, Huentelman

, Beach

, Caselli

, Chen

, Su

, Myers

, Hardy

, Paul Vonsattel

, Younkin

, Bennett

, De Jager

, Larson

, Crane

, Keene

, Kamboh

, Kofler

, Duque

, Gilbert

, Gwirtsman

, Buxbaum

, Dickson

, Frosch

, Ghetti

, Lunetta

, Wang

, Hyman

, Kukull

, Foroud

, Haines

, Mayeux

, Pericak-Vance

, Schneider

, Trojanowski

, Farrer

, Schellenberg

, Beecham

, Montine

, Jun

(2020) Exceptionally low likelihood of Alzheimer’s dementia in APOE2 homozygotes from a 5,000-person neuropathological study. Nat Commun 11, 667.

Bussy

, Snider

, Coble

, Xiong

, Fagan

, Cruchaga

, Benzinger

TLS

, Gordon

, Hassenstab

, Bateman

, Morris

(2019) Effect of apolipoprotein E4 on clinical, neuroimaging, and biomarker measures in noncarrier participants in the Dominantly Inherited Alzheimer Network. Neurobiol Aging 75, 42–50.

10.

Mishra

, Blazey

, Holtzman

, Cruchaga

, Su

, Morris

, Benzinger

TLS

, Gordon

(2018) Longitudinal brain imaging in preclinical Alzheimer disease: Impact of APOE ɛ4 genotype. Brain 141, 1828–1839.

11.

Wirth

, Villeneuve

, La Joie

, Marks

, Jagust

(2014) Gene-environment interactions: Lifetime cognitive activity, APOE genotype, and β-amyloid burden. J Neurosci 34, 8612–8617.

12.

Niti

, Yap

, Kua

, Tan

, Ng

(2008) Physical, social and productive leisure activities, cognitive decline and interaction with APOE-epsilon 4 genotype in Chinese older adults. Int Psychogeriatr 20, 237–251.

13.

Ferrari

, Xu

, Wang

, Winblad

, Sorbi

, Qiu

, Fratiglioni

(2013) How can elderly apolipoprotein E ɛ4 carriers remain free from dementia? Neurobiol Aging 34, 13–21.

14.

Eisenberg

, Kuzawa

, Hayes

(2010) Worldwide allele frequencies of the human apolipoprotein E gene: Climate, local adaptations, and evolutionary history. Am J Phys Anthropol 143, 100–111.

15.

Singh

, Singh

, Mastana

(2006) APOE distribution in world populations with new data from India and the UK. Ann Hum Biol 33, 279–308.

16.

Turney

, Chesebro

, Rentería

, Lao

, Beato

, Schupf

, Mayeux

, Manly

, Brickman

(2020) APOE ɛ4 and resting-state functional connectivity in racially/ethnically diverse older adults. Alzheimers Dement (Amst) 12, e12094.

17.

Fuzikawa

, Peixoto

, Taufer

, Moriguchi

, Lima-Costa

(2007) Apolipoprotein E polymorphism distribution in an elderly Brazilian population: The Bambuí Health and Aging Study. Braz J Med Biol Res 40, 1429–1434.

18.

Belloy

, Napolioni

, Greicius

(2019) A quarter century of APOE and Alzheimer’s disease: Progress to date and the path forward. Neuron 101, 820–838.

19.

Hoy

, Brooks

, Woolf

, Blyth

, March

, Bain

, Baker

, Smith

, Buchbinder

(2012) Assessing risk of bias in prevalence studies: Modification of an existing tool and evidence of interrater agreement. J Clin Epidemiol 65, 934–939.

20.

Noubiap

, Essouma

, Bigna

, Jingi

, Aminde

, Nansseu

(2017) Prevalence of elevated blood pressure in children and adolescents in Africa: A systematic review and meta-analysis. Lancet Public Health 2, e375–e386.

21.

Barendregt

, Doi

, Lee

, Norman

, Vos

(2017) Meta-analysis of prevalence. J Epidemiol Community Health 67, 974–978.

22.

Rucker

, Schwarzer

, Carpenter

, Olkin

(2009) Why add anything to nothing? The arcsine difference as a measure of treatment effect in meta-analysis with zero cells. Stat Med 28, 721–738.

23.

Fix

(1997) Gene frequency clines produced by kin-structured founder effects. Hum Biol 69, 663–673.

24.

Mahley

, Rall

Jr (2000) Apolipoprotein E: Far more than a lipid transport protein. Annu Rev Genomics Hum Genet 1, 507–537.

25.

Gerdes

, Gerdes

, Hansen

, Klausen

, Faergeman

, Dyerberg

(1996) The apolipoprotein E polymorphism in Greenland Inuit in its global perspective. Hum Genet 98, 546–550.

26.

Corbo

, Scacchi

, Rickards

, Martinez-Labarga

, De Stefano

(1999) An investigation of human apolipoproteins B and E polymorphisms in two African populations from Ethiopia and Benin. Am J Hum Biol 11, 297–304.

27.

Corbo

, Scacchi

(1999) Apolipoprotein E (APOE) allele distribution in the world. Is APOE^*4 a ‘thrifty’ allele? Ann Hum Genet 63, 301–310.

28.

Hanlon

, Rubinsztein

(1995) Arginine residues at codons 112 and 158 in the apolipoprotein E gene correspond to the ancestral state in humans. Atherosclerosis 112, 85–90.

29.

Bons

, Bauer

, Bocherens

, de Riese

, Drucker

, Francken

, Menéndez

, Uhl

, van Milligen

, Wißing

(2019) Out of Africa by spontaneous migration waves. PLoS One 14, e0201998.

30.

Livingston

, Huntley

, Sommerlad

, Ames

, Ballard

, Banerjee

, Brayne

, Burns

, Cohen-Mansfield

, Cooper

, Costafreda

, Dias

, Fox

, Gitlin

, Howard

, Kales

, Kivimäki

, Larson

, Ogunniyi

, Orgeta

, Ritchie

, Rockwood

, Sampson

, Samus

, Schneider

, Selbæk

, Teri

, Mukadam

(2020) Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet 396, 413–446.

31.

van Exel

, Koopman

JJE

, Bodegom

, Meij

, Knijff

, Ziem

, Finch

, Westendorp

RGJ

(2017) Effect of APOE epsilon4 allele on survival and fertility in an adverse environment. PLoS One 12, e0179497.

32.

Jofre-Monseny

, Minihane

, Rimbach

(2008) Impact of apoE genotype on oxidative stress, inflammation and disease risk. Mol Nutr Food Res 52, 131–145.

33.

Huebbe

, Rimbach

(2017) Evolution of human apolipoprotein E (APOE) isoforms: Gene structure, protein function and interaction with dietary factors. Ageing Res Rev 37, 146–161.

34.

Sierksma

, Escott-Price

, De Strooper

(2020) Translating genetic risk of Alzheimer’s disease into mechanistic insight and drug targets. Science 370, 61–66.

35.

(2019) Global, regional, and national burden of Alzheimer’s disease and other dementias, 1990-2016: A systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol 18, 88–106.

36.

Livingston

, Sommerlad

, Orgeta

, Costafreda

, Huntley

, Ames

, Ballard

, Banerjee

, Burns

, Cohen-Mansfield

, Cooper

, Fox

, Gitlin

, Howard

, Kales

, Larson

, Ritchie

, Rockwood

, Sampson

, Samus

, Schneider

, Selbæk

, Teri

, Mukadam

(2017) Dementia prevention, intervention, and care. Lancet 390, 2673–2734.

37.

Ritchie

, Carrière

, Ritchie

, Berr

, Artero

, Ancelin

(2010) Designing prevention programmes to reduce incidence of dementia: Prospective cohort study of modifiable risk factors. BMJ 341, c3885.

38.

Soldan

, Pettigrew

, Cai

, Wang

, Moghekar

, O’Brien

, Selnes

, Albert

(2016) Hypothetical preclinical Alzheimer disease groups and longitudinal cognitive change. JAMA Neurol 73, 698–705.

39.

Bennett

, Schneider

, Arvanitakis

, Kelly

, Aggarwal

, Shah

, Wilson

(2006) Neuropathology of older persons without cognitive impairment from two community-based studies. Neurology 66, 1837–1844.

40.

Schwanke

, da Cruz

, Leal

, Scheibe

, Moriguchi

(2002) Analysis of the association between apolipoprotein E polymorphism and cardiovascular risk factors in an elderly population with longevity. Arq Bras Cardiol 78, 561–579.

41.

Rebeck

, Perls

, West

, Sodhi

, Lipsitz

, Hyman

(1994) Reduced apolipoprotein epsilon 4 allele frequency in the oldest old Alzheimer’s patients and cognitively normal individuals. Neurology 44, 1513–1516.