[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure.  Age-Specific Dementia Incidence by Birth Cohort
Age-Specific Dementia Incidence by Birth Cohort

The graph includes participants with information available on birth cohort, age, sex, height, childhood financial stability, ability to afford basic needs during childhood, childhood household density, participants’ educational level, and parental educational levels. Age-specific dementia incidence rates were calculated by dividing the number of new dementia diagnoses in the age group by the number of person-years in the age group. Error bars indicate 95% CIs, which are shown for the pre–Great Depression birth cohort (n = 3174) as an example. WWI indicates World War I; WWII, World War II.

Table 1.  Selected Participant Characteristics by Birth Cohorta
Selected Participant Characteristics by Birth Cohorta
Table 2.  Association of Birth Cohort and Socioeconomic Status in Early Life With Incident Dementiaa
Association of Birth Cohort and Socioeconomic Status in Early Life With Incident Dementiaa
1.
Satizabal  CL, Beiser  AS, Chouraki  V, Chene  G, Dufouil  C, Seshadri  S.  Incidence of dementia over three decades in the Framingham Heart Study.   N Engl J Med. 2016;374(6):523-532. doi:10.1056/NEJMoa1504327 PubMedGoogle ScholarCrossref
2.
Rajan  KB, Weuve  J, Barnes  LL, Wilson  RS, Evans  DA.  Prevalence and incidence of clinically diagnosed Alzheimer’s disease dementia from 1994 to 2012 in a population study.   Alzheimers Dement. 2019;15(1):1-7. doi:10.1016/j.jalz.2018.07.216 PubMedGoogle ScholarCrossref
3.
Noble  JM, Schupf  N, Manly  JJ, Andrews  H, Tang  M-X, Mayeux  R.  Secular trends in the incidence of dementia in a multi-ethnic community.   J Alzheimers Dis. 2017;60(3):1065-1075. doi:10.3233/JAD-170300 PubMedGoogle ScholarCrossref
4.
Schrijvers  EMC, Verhaaren  BFJ, Koudstaal  PJ, Hofman  A, Ikram  MA, Breteler  MMB.  Is dementia incidence declining?: trends in dementia incidence since 1990 in the Rotterdam Study.   Neurology. 2012;78(19):1456-1463. doi:10.1212/WNL.0b013e3182553be6 PubMedGoogle ScholarCrossref
5.
Gao  S, Burney  HN, Callahan  CM, Purnell  CE, Hendrie  HC.  Incidence of dementia and Alzheimer disease over time: a meta-analysis.   J Am Geriatr Soc. 2019;67(7):1361-1369. doi:10.1111/jgs.16027 PubMedGoogle ScholarCrossref
6.
Livingston  G, Sommerlad  A, Orgeta  V,  et al.  Dementia prevention, intervention, and care.   Lancet. 2017;390(10113):2673-2734. doi:10.1016/S0140-6736(17)31363-6 PubMedGoogle ScholarCrossref
7.
Skoog  I.  Dementia: dementia incidence—the times, they are a-changing.   Nat Rev Neurol. 2016;12(6):316-318. doi:10.1038/nrneurol.2016.55 PubMedGoogle ScholarCrossref
8.
Rocca  WA.  Time, sex, gender, history, and dementia.   Alzheimer Dis Assoc Disord. 2017;31(1):76-79. doi:10.1097/WAD.0000000000000187 PubMedGoogle ScholarCrossref
9.
Cadar  D, Lassale  C, Davies  H, Llewellyn  DJ, Batty  GD, Steptoe  A.  Individual and area-based socioeconomic factors associated with dementia incidence in England: evidence from a 12-year follow-up in the English Longitudinal Study of Ageing.   JAMA Psychiatry. 2018;75(7):723-732. doi:10.1001/jamapsychiatry.2018.1012 PubMedGoogle ScholarCrossref
10.
Derby  CA, Katz  MJ, Lipton  RB, Hall  CB.  Trends in dementia incidence in a birth cohort analysis of the Einstein Aging Study.   JAMA Neurol. 2017;74(11):1345-1351. doi:10.1001/jamaneurol.2017.1964 PubMedGoogle ScholarCrossref
11.
Rocca  WA, Petersen  RC, Knopman  DS,  et al.  Trends in the incidence and prevalence of Alzheimer’s disease, dementia, and cognitive impairment in the United States.   Alzheimers Dement. 2011;7(1):80-93. doi:10.1016/j.jalz.2010.11.002 PubMedGoogle ScholarCrossref
12.
McKeown  T.  The Modern Rise of Population. Academic Press; 1976.
13.
Fogel  RW. Economic growth, population theory, and physiology: the bearing of long-term processes on the making of economic policy. National Bureau of Economic Research working paper 4638. February 1994. Accessed July 2019. https://www.nber.org/papers/w4638.pdf
14.
Knickmeyer  RC, Gouttard  S, Kang  C,  et al.  A structural MRI study of human brain development from birth to 2 years.   J Neurosci. 2008;28(47):12176-12182. doi:10.1523/JNEUROSCI.3479-08.2008 PubMedGoogle ScholarCrossref
15.
Levitt  P.  Structural and functional maturation of the developing primate brain.   J Pediatr. 2003;143(4)(suppl):S35-S45. doi:10.1067/S0022-3476(03)00400-1 PubMedGoogle ScholarCrossref
16.
National Scientific Council on the Developing Child. The timing and quality of early experiences combine to shape brain architecture: working paper no. 5. 2007. Accessed August 2018. https://developingchild.harvard.edu/resources/the-timing-and-quality-of-early-experiences-combine-to-shape-brain-architecture/
17.
Stern  Y.  What is cognitive reserve? theory and research application of the reserve concept.   J Int Neuropsychol Soc. 2002;8(3):448-460. doi:10.1017/S1355617702813248 PubMedGoogle ScholarCrossref
18.
Vandenbroucke  JP, von Elm  E, Altman  DG,  et al; STROBE Initiative.  Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration.   Ann Intern Med. 2007;147(8):W163-W194. doi:10.7326/0003-4819-147-8-200710160-00010-w1 PubMedGoogle Scholar
19.
Kukull  WA, Higdon  R, Bowen  JD,  et al.  Dementia and Alzheimer disease incidence: a prospective cohort study.   Arch Neurol. 2002;59(11):1737-1746. doi:10.1001/archneur.59.11.1737 PubMedGoogle ScholarCrossref
20.
Teng  EL, Hasegawa  K, Homma  A,  et al.  The Cognitive Abilities Screening Instrument (CASI): a practical test for cross-cultural epidemiological studies of dementia.   Int Psychogeriatr. 1994;6(1):45-58. doi:10.1017/S1041610294001602 PubMedGoogle ScholarCrossref
21.
American Psychiatric Association.  Diagnostic and Statistical Manual of Mental Disorders. 4th ed. American Psychiatric Association; 1994.
22.
Cole  TJ.  Secular trends in growth.   Proc Nutr Soc. 2000;59(2):317-324. doi:10.1017/S0029665100000355 PubMedGoogle ScholarCrossref
23.
Floud  R, Fogel  RW, Harris  B, Hong  SC.  The Changing Body: Health, Nutrition, and Human Development in the Western World Since 1700. Cambridge University Press; 2011. doi:10.1017/CBO9780511975912
24.
Maurer  J.  Height, education and later-life cognition in Latin America and the Caribbean.   Econ Hum Biol. 2010;8(2):168-176. doi:10.1016/j.ehb.2010.05.013 PubMedGoogle ScholarCrossref
25.
Abbott  RD, White  LR, Ross  GW,  et al.  Height as a marker of childhood development and late-life cognitive function: the Honolulu-Asia Aging Study.   Pediatrics. 1998;102(3 Pt 1):602-609. doi:10.1542/peds.102.3.602 PubMedGoogle ScholarCrossref
26.
Russ  TC, Kivimaki  M, Starr  JM, Stamatakis  E, Batty  GD.  Height in relation to dementia death: individual participant meta-analysis of 18 UK prospective cohort studies.   Br J Psychiatry. 2014;205(5):348-354. doi:10.1192/bjp.bp.113.142984 PubMedGoogle ScholarCrossref
27.
Hixson  JE, Vernier  DT.  Restriction isotyping of human apolipoprotein E by gene amplification and cleavage with HhaI.   J Lipid Res. 1990;31(3):545-548.PubMedGoogle Scholar
28.
Emi  M, Wu  LL, Robertson  MA,  et al.  Genotyping and sequence analysis of apolipoprotein E isoforms.   Genomics. 1988;3(4):373-379. doi:10.1016/0888-7543(88)90130-9 PubMedGoogle ScholarCrossref
29.
Billingsley  P.  Probability and Measure. John Wiley & Sons; 1995.
30.
Fine  JP, Gray  RJ.  A proportional hazards model for the subdistribution of a competing risk.   J Am Stat Assoc. 1999;94(446):496-509. doi:10.1080/01621459.1999.10474144 Google ScholarCrossref
31.
Tom  SE, Hubbard  RA, Crane  PK,  et al.  Characterization of dementia and Alzheimer’s disease in an older population: updated incidence and life expectancy with and without dementia.   Am J Public Health. 2015;105(2):408-413. doi:10.2105/AJPH.2014.301935 PubMedGoogle ScholarCrossref
32.
Skoog  I, Najar  J, Wetterberg  H.  The importance of birth year for the incidence of dementia.   J Am Geriatr Soc. 2019;67(7):1330-1332. doi:10.1111/jgs.16029 PubMedGoogle ScholarCrossref
33.
Cutler  D, Miller  G.  The role of public health improvements in health advances: the twentieth-century United States.   Demography. 2005;42(1):1-22. doi:10.1353/dem.2005.0002 PubMedGoogle ScholarCrossref
34.
Moceri  VM, Kukull  WA, Emanual  I,  et al.  Using census data and birth certificates to reconstruct the early-life socioeconomic environment and the relation to the development of Alzheimer’s disease.   Epidemiology. 2001;12(4):383-389. doi:10.1097/00001648-200107000-00007 PubMedGoogle ScholarCrossref
35.
Moceri  VM, Kukull  WA, Emanuel  I, van Belle  G, Larson  EB.  Early-life risk factors and the development of Alzheimer’s disease.   Neurology. 2000;54(2):415-420. doi:10.1212/WNL.54.2.415 PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Views 1,190
    Citations 0
    Original Investigation
    Public Health
    July 27, 2020

    Association of Demographic and Early-Life Socioeconomic Factors by Birth Cohort With Dementia Incidence Among US Adults Born Between 1893 and 1949

    Author Affiliations
    • 1Department of Neurology, Columbia University, New York, New York
    • 2Department of Epidemiology, Columbia University, New York, New York
    • 3Department of Biostatistics, Columbia University, New York, New York
    • 4Department of Biostatistics, Epidemiology, and Informatics, University of Pennsylvania, Philadelphia
    • 5Department of Medicine, University of Washington, Seattle
    • 6Kaiser Permanente Washington Health Research Institute, Seattle
    JAMA Netw Open. 2020;3(7):e2011094. doi:10.1001/jamanetworkopen.2020.11094
    Key Points español 中文 (chinese)

    Question  Are dementia incidence trends by birth cohort associated with early-life environment?

    Findings  In this cohort study of 4277 participants in the Adult Changes in Thought study who were born between 1893 and 1949 and were followed up for up to 20 years (1994-2015), the age- and sex-adjusted dementia incidence was lower among those born during the Great Depression (1929-1939) and the period during World War II and postwar (1940-1949) compared with those born in the period before the Great Depression (1921-1928). The association between birth cohort and dementia incidence remained when accounting for early-life socioeconomic environment, educational level, and late-life vascular risk factors.

    Meaning  The study’s findings indicate that dementia incidence is lower for individuals born after the mid-1920s compared with those born earlier, and this lower incidence is not associated with early-life environment as measured in this study.

    Abstract

    Importance  Early-life factors may be important for later dementia risk. The association between a more advantaged early-life environment, as reflected through an individual’s height and socioeconomic status indicators, and decreases in dementia incidence by birth cohort is unknown.

    Objectives  To examine the association of birth cohort and early-life environment with dementia incidence among participants in the Adult Changes in Thought study from 1994 to 2015.

    Design, Setting, and Participants  This prospective cohort study included 4277 participants from the Adult Changes in Thought study, an ongoing longitudinal population-based study of incident dementia in a random sample of adults 65 years and older who were born between 1893 and 1949 and are members of Kaiser Permanente Washington in the Seattle region. Participants in the present analysis were followed up from 1994 to 2015. At enrollment, all participants were dementia-free and completed a baseline evaluation. Subsequent study visits were held every 2 years until a diagnosis of dementia, death, or withdrawal from the study. Participants were categorized by birth period (defined by historically meaningful events) into 5 cohorts: pre–World War I (1893-1913), World War I and Spanish influenza (1914-1920), pre–Great Depression (1921-1928), Great Depression (1929-1939), and World War II and postwar (1940-1949). Participants’ height, educational level, childhood financial stability, and childhood household density were examined as indicators of early-life environment, and later-life vascular risk factors for dementia were assessed. Cox proportional hazards regression models, adjusted for competing survival risk, were used to analyze data. Data were analyzed from June 1, 2018, to April 29, 2020.

    Main Outcomes and Measures  Participants completed the Cognitive Abilities Screening Instrument every 2 years to assess global cognition. Those with scores indicative of cognitive impairment completed an evaluation for dementia, with dementia diagnoses determined during consensus conferences using criteria from the Diagnostic and Statistical Manual of Mental Disorders, 4th edition.

    Results  Among 4277 participants, the mean (SD) age was 74.5 (6.4) years, and 2519 participants (58.9%) were women. The median follow-up was 8 years (interquartile range, 4-12 years), with 730 participants developing dementia over 24 378 person-years. The age-specific dementia incidence was lower for those born in 1929 and later compared with those born earlier. Compared with participants born in the pre–Great Depression years (1921-1928), the age- and sex-adjusted hazard ratio was 0.67 (95% CI, 0.53-0.85) for those born in the Great Depression period (1929-1939) and 0.62 (95% CI, 0.29-1.31) for those born in the World War II and postwar period (1940-1949). Although indicators of a more advantaged early-life environment and higher educational level (college or higher) were associated with a lower incidence of dementia, these variables did not explain the association between birth cohort and dementia incidence, which remained when vascular risk factors were included and were similar by sex.

    Conclusions and Relevance  Age-specific dementia incidence was lower in participants born after the mid-1920s compared with those born earlier. In this population, the decrease in dementia incidence may reflect societal-level changes or individual differences over the life course rather than early-life environment, as reflected through recalled childhood socioeconomic status and measured height, educational level, and later-life vascular risk.

    Introduction

    Studies have found a stabilization or decrease in dementia incidence in the US and Europe beginning in the 1990s.1-5 Although low educational level, midlife hypertension, midlife obesity, hearing loss, late-life depression, diabetes, physical inactivity, smoking, and social isolation have been identified as risk factors for dementia,6 previous studies have specifically examined higher educational levels and lower prevalence of midlife vascular risk factors in association with secular trends. Confounding these chronological trends are birth cohort trends, given the major social, economic, and political changes and medical advances that have been associated with increases in standard of living across the life course during the 20th century.7,8 The few studies that have assessed dementia incidence trends by birth cohort have found decreasing dementia incidence beginning with cohorts born in the mid-1920s.9-11

    Improvements during the first half of the 20th century with regard to childhood standard of living,12 especially improvements in nutrition,13 may have implications for dementia risk. The brain develops most rapidly and is most plastic in the first 5 years of life.14,15 Strong early brain development supports more complex subsequent neuritic and intraneuronal connections and cognition, conferring lifelong advantage16 that may be associated with cognitive reserve.17 Previous studies of secular and birth cohort trends in dementia incidence have not considered indicators of neurocognitive development that occurred before education completion.

    Using a population-based cohort with up to 20 years of follow-up, we investigated the associations of birth cohort and early-life environment with dementia incidence. We tested the hypotheses that cohorts born more recently would have a lower dementia incidence and that this trend would be associated with higher early-life socioeconomic status and taller height.

    Methods

    The Adult Changes in Thought (ACT) study is an ongoing population-based cohort study of incident dementia in a random sample of adults 65 years and older who were born between 1893 and 1949 and are members of Kaiser Permanente Washington (formerly Group Health) in the Seattle region. The institutional review boards of Kaiser Permanente Washington, Columbia University, and the University of Washington approved this study. Written informed consent was obtained from all ACT study participants. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.18

    Between 1994 and 1996, 2581 people were enrolled in the ACT study, and between 2000 and 2002, an expansion cohort of 811 people was enrolled. In 2005, the study began ongoing enrollment to replace participants who died, developed dementia, or withdrew from the study, enrolling 120 to 180 participants annually. At study enrollment, participants were dementia-free and completed a baseline evaluation, with subsequent study visits every 2 years until the earliest of either a dementia diagnosis, death, or withdrawal from the study. The 5104 participants in the present study sample comprised those who were followed up from 1994 to 2015; 4888 participants in the sample provided their birth year. Of those, 4277 participants who provided information on their sex, age, height, educational level, and early-life socioeconomic status were included in the analysis.

    Methods used for dementia diagnosis have been previously described.19 In brief, participants received the Cognitive Abilities Screening Instrument, a global cognition assessment (score range, 0-100 points, with lower scores indicating worse cognition),20 every 2 years. A score of less than 86 points prompted a standardized diagnostic evaluation, which included physical and neurological examinations and a neuropsychological test battery. Dementia diagnoses were determined at consensus conferences based on criteria from the Diagnostic and Statistical Manual of Mental Disorders, 4th edition.21 The consensus conference has been previously described.19

    Height, an indicator of early-life nutrition22,23 that has been associated with dementia risk,24-26 was measured at baseline. Participants also reported their year of birth, sex, educational level, whether they had ever received antihypertensive medications or smoked, and whether they had prevalent cardiovascular disease (myocardial infarction, angina, coronary artery bypass grafting, or angioplasty), prevalent cerebrovascular disease (stroke, transient ischemic attack, or carotid endarterectomy), or prevalent diabetes. Body mass index (BMI; calculated as weight in kilograms divided by height in meters squared) was obtained,2 and apolipoprotein E (APOE) genotype was determined using published methods27,28 and categorized as the presence or absence of ε4 alleles.

    At baseline, participants also reported information about their early-life environment, including maternal and paternal educational levels, family ability to afford basic necessities (able to meet basic needs [food, housing, clothing, and medical care] vs not able to meet ≥1 basic need) and small luxuries in childhood, family financial stability up to a participant age of 15 years (financial stability score range, 1-10, with 1 indicating least stable and 10 indicating most stable), and family household density (number of inhabitants per number of bedrooms) up to a participant age of 6 years.

    Statistical Analysis

    Birth cohorts were classified as pre–World War I (1893-1913), World War I and Spanish influenza (1914-1920), pre–Great Depression (1921-1928), Great Depression (1929-1939), and World War II and postwar (1940-1949) based on historically meaningful events that were determined a priori. The pre–Great Depression group, which was the largest birth cohort, was selected as the reference group to provide statistical stability. To further describe the cohort, we created 5 categories for participant age at study entry (65-69 years, 70-74 years, 75-79 years, 80-84 years, 85-89 years, and ≥90 years). For models, we used quartiles of the distribution of participant age at study entry (65-69 years, 70-73 years, 74-79 years, and ≥80 years).

    We created a combined score for parental educational levels by computing z scores for maternal and paternal educational levels separately, then calculating z scores of the sum. Self-reported childhood financial stability was divided into 4 quartiles based on score ranges (quartile 1 included stability scores of 1-3 points, quartile 2 included scores of 4-5 points, quartile 3 included scores of 6-7 points, and quartile 4 included scores of 8-10 points), with higher quartiles indicating greater stability. Height was categorized into sex-specific quartiles (for women, quartile 1 was ≤155 cm, quartile 2 was >155 cm to ≤160 cm, quartile 3 was >160 cm to ≤165 cm, and quartile 4 was >165 cm; for men, quartile 1 was ≤170 cm, quartile 2 was >170 cm to ≤175 cm, quartile 3 was >175 cm to ≤180 cm, and quartile 4 was >180 cm). Because participants in the sample reported high educational levels, we dichotomized educational levels as high school and lower vs college and higher.

    We described age-specific dementia incidence rates (per 1000 person-years), overall and by birth cohort, by dividing the number of new dementia diagnoses in the age group by the number of person-years in the age group. Standard errors for log-transformed incidence rates were calculated using the delta method.29 We used the Fine and Gray subdistribution proportional hazards model,30 which provides a competing risks survival framework, with all-cause dementia as the outcome and death as a competing event. Models used baseline hazards that were stratified by quartiles of age group at study entry and time since study entry as the time axis.

    To account for the potential mediation between early-life environment and dementia, we investigated the following series of models (including nested sets of covariates): (1) model 1, which included birth cohort and sex; (2) model 2, which included all variables in model 1 plus height, childhood financial stability, family ability to afford basic needs and small luxuries during childhood, and childhood household density; and (3) model 3, which included all variables in model 2 plus participant educational level. We also assessed the potential interaction between sex and birth cohort.

    In sensitivity analyses, we separately included parental educational levels (for 3171 participants who provided data), APOE genotype (for 3554 participants who provided data), and vascular risk factors (for 4062 participants who provided data; vascular risk factors included prevalent cardiovascular disease, prevalent cerebrovascular disease, prevalent diabetes, and having ever received antihypertensive medication or smoked), and we added these factors to the model 3 variables. All analyses were conducted using R software, version 3.5.1 (R Project for Statistical Computing).

    Results

    Among 5104 participants in the present study sample, 4888 participants provided their birth year. Of those, 4277 participants (the mean [SD] age, 74.5 (6.4) years; 2519 women [58.9%]) who provided information on height, educational level, and early-life socioeconomic variables, were included in the analysis. Birth cohorts comprised 483 participants in the pre–World War I cohort, 985 participants in the World War I and Spanish influenza cohort, 1491 participants in the pre–Great Depression cohort, 906 participants in the Great Depression cohort, and 412 participants in the World War II and postwar cohort.

    During a total of 24 378 person-years, 730 participants developed dementia. Among the 611 participants who provided their birth year but were not included in the analysis sample, the follow-up period was shorter (median, 0 years; interquartile range [IQR], 0-5 years) compared with those in the analysis sample (median, 8 years; IQR, 4-12 years), but their characteristics were otherwise similar to participants included in the analysis. The maternal educational level in the pre–World War I birth cohort was 4 years lower (median, 8 years; IQR, 8-12 years) than that of the pre–Great Depression cohort (median, 12 years; IQR, 8-13 years) and later birth cohorts (Great Depression cohort, median, 12 years; IQR, 10-14 years; World War II and postwar cohort, median, 12 years; IQR, 12-15 years) (Table 1). The paternal educational level was 8 years (IQR, 7-13 years) for the pre–World War I birth cohort, 10 years (IQR, 8-13 years) for the pre–Great Depression birth cohort, and 12 years (IQR, 8-15 years and 11-16 years, respectively) for the Great Depression cohort and the World War II and postwar cohort.

    A total of 305 of 483 participants (63%) in the pre–World War I birth cohort reported that their families were able to afford basic needs and small luxuries during childhood compared with 841 of 1491 participants (56%) in the pre–Great Depression birth cohort. The proportion of those able to afford basic needs and small luxuries was greater in more recent birth cohorts (672 of 906 participants [74%] in the Great Depression cohort and 357 of 412 participants [87%] in the World War II and postwar cohort). A similar pattern was observed with regard to childhood financial stability across birth cohorts, with the proportion of participants in the most stable financial quartile increasing from 144 of 483 participants (30%) in the pre–World War I cohort to 178 of 412 participants (43%) in the World War II and postwar cohort.

    A comparison of earliest (pre-World War I) to most recent (World War II and postwar) birth cohorts indicated that median height increased by 8 cm for women (from 155 cm [IQR, 150-160 cm] to 163 cm [IQR, 158-168 cm]) and 2 cm for men (from 173 cm [IQR, 168-175 cm] to 175 cm [IQR, 173-180 cm]). Childhood household density was identical across birth cohorts, at 2 inhabitants per bedroom (IQR, 1-2 inhabitants). In the pre–World War I birth cohort, 150 of 483 participants (31%) completed college compared with 331 of 412 participants (80%) in the World War II and postwar birth cohort, with the largest difference in educational level between the pre–Great Depression cohort (666 of 1491 participants [45%] completed college) and the Great Depression cohort (648 of 906 participants [72%] completed college). Smoking was more prevalent among the pre–Great Depression cohort (809 of 1491 participants [56%] had ever smoked) and the Great Depression cohort (452 of 906 participants [54%] had ever smoked), while heart disease (114 of 483 participants [25%] in the pre–World War I cohort vs 32 of 412 participants [8%] in the World War II and postwar cohort) and cerebrovascular disease (59 of 483 participants [13%] in the pre–World War I cohort vs 11 of 412 participants [3%] in the World War II and postwar cohort) were less prevalent in the more recently born cohorts.

    Unadjusted dementia incidence increased with age for all birth cohorts (Figure; eTable 1 in the Supplement). Age-specific dementia incidence was similar for the pre–World War I, World War I and Spanish influenza, and pre–Great Depression cohorts and was lower in the Great Depression cohort, with the exception of participants aged 75 to 79 years. The World War II and postwar cohort had a similar dementia incidence for participants aged 65 to 69 years and a higher incidence for participants aged 70 to 74 years compared with those in the Great Depression cohort.

    After adjusting for age and sex, dementia risk was lower for the Great Depression cohort (hazard ratio [HR], 0.67; 95% CI, 0.53-0.85) (model 1 in Table 2) and remained lower for the World War II and postwar cohort (HR, 0.62; 95% CI, 0.29-1.31) compared with the referent pre–Great Depression birth cohort. The 2 shortest height quartiles were associated with higher dementia risk (quartile 1, HR, 1.71; 95% CI, 1.35-2.17; quartile 2, HR, 1.43; 95% CI, 1.12-1.83) compared with the tallest height quartile (quartile 4; referent group) (model 2 in Table 2). In model 2, participants in quartiles with the least childhood financial stability had a lower risk of dementia (quartile 1, HR, 0.73; 95% CI, 0.59-0.90; quartile 2, HR, 0.81; 95% CI, 0.67-0.98) compared with those in the quartile with the most financial stability (quartile 4; referent group). With regard to childhood household density, an increase of 1 inhabitant per bedroom was associated with an 8% decrease in dementia risk (HR, 0.92; 95% CI, 0.85-0.99). The association between more recent birth cohort and risk of dementia persisted after accounting for early-life socioeconomic variables (model 2 in Table 2).

    In model 3, which accounted for participant educational level, the associations of birth cohort and early-life socioeconomic status with dementia risk remained similar, with dementia risk lower for the Great Depression (HR, 0.71; 95% CI, 0.56-0.90) and World War II and postwar (HR, 0.69; 95% CI, 0.32-1.45) cohorts compared with the referent pre–Great Depression cohort and lower in quartiles with the least financial stability (quartile 1, HR, 0.71; 95% CI, 0.58-0.88; quartile 2, HR, 0.80; 0.67-0.97) compared with the most financial stability in childhood (model 3 in Table 2). The association between birth cohort and dementia incidence did not differ by sex (eTable 2 in the Supplement). In sensitivity analyses, the results were similar when parental educational level, APOE genotype, and vascular risk factors were individually included (eTable 3 in the Supplement).

    Discussion

    Among those born at the turn of the 20th century through the mid-20th century who participated in the ACT study, the age-specific dementia incidence was lower for participants born in 1929 and later compared with those born earlier. This trend was not explained by recalled childhood socioeconomic status and measured height, which reflect early-life environment, nor was it explained by educational level and vascular risk as an older adult. The literature on secular dementia trends reports a decrease in dementia incidence starting in the 1990s.1-5 This timing is consistent with participants in the 1929 to 1939 birth cohorts who are entering the eighth decade of life, when dementia risk increases.2,4,31 Political and economic changes during the first half of the 20th century may have had different implications for dementia risk based on the participant’s age during those experiences.32 Analysis by birth cohort captures this intersection of age and calendar time. Our results suggest that societal-level changes in the first half of the 20th century that were not captured by the individual early-life measures or the educational levels used in this study may have been associated with decreases in dementia incidence.

    The 40% decrease in the US mortality rate from 1900 to 1940 was likely owing to the decrease in infectious diseases,33 which disproportionately occur in the young. The decrease in dementia incidence observed in the ACT study began with birth cohorts who were born in the middle of this period. These early-life health gains may be factors in the decreased dementia incidence. Although we accounted for family-level socioeconomic status variables and height, these variables may not have captured all changes, such as economic innovation13 and nutritional improvement,12 that may have been associated with decreases in mortality. In addition, variables included in this study may not have captured public heath improvements during this period.33 It is possible that unmeasured differences were more important for assessing dementia risk by birth cohort than the socioeconomic factors we measured.

    Across birth cohorts, participants with lower financial status and greater household density in childhood had a lower risk of developing dementia, which is inconsistent with our hypothesis and the results of previous studies.34,35 While the Great Depression was a time of financial hardship, those in the pre–Great Depression and the World War I and Spanish influenza birth cohorts were the least likely to report the ability to afford both basic needs and small luxuries, and they had the smallest proportion of participants reporting the most stable childhood financial quartile. This pattern may reflect problems with measurement or sample selection. Participant responses may reflect experiences in later childhood and early adolescence, as recall of early-life experiences may be difficult. In contrast, parental educational levels, which were constant throughout childhood and adolescence for most of the birth cohorts, were higher for the World War I and Spanish influenza cohort and the pre–Great Depression cohort compared with cohorts born earlier. This pattern suggests a higher early-life standard of living in the more recent birth cohorts. Another possibility is that because these 2 birth cohorts were the oldest, those who survived to participate in the study were able to compensate for adverse early-life environments or had less accurate recall than younger participants.

    Our study considered death as a competing risk, while a previous case-control study did not.34,35 Most ACT participants were members of Kaiser Permanente Washington (formerly Group Health) when they were younger than 65 years, during which they primarily received health insurance through large employers. It is likely that those with lower financial status and higher household density during childhood survived adverse experiences to be able to participate the sample.

    Together with height, an individual’s parental educational level, childhood financial stability, and childhood household density are likely to reflect their early-life environment. These variables did not explain the decrease in dementia incidence among the more recent birth cohorts. In a minimally adjusted model, the decrease in dementia incidence began with the Great Depression birth cohort, suggesting that societal-level experiences during later childhood to adolescence may have been more important than those during the in-utero through early childhood phase. If this earliest stage of life were important for dementia incidence, we would expect those born in the Great Depression cohort to have the greatest dementia risk. The largest difference in college completion was found between the pre–Great Depression and Great Depression birth cohorts. This disruption to economic opportunity for those born in the pre–Great Depression years may have had implications for dementia risk. The inclusion of late-life vascular risk factors did not appreciably alter the association between a more recent birth cohort and a lower incidence of dementia, which is consistent with analyses of the Einstein Aging Cohort10 and the Framingham Heart Study, which considered the cohort of study entry.1

    We found similar associations between birth cohort and decreased dementia incidence in 2 previous studies. An analysis of the English Longitudinal Study of Aging examined 2 birth cohorts based on birth-year median (1902-1925 and 1926-1943),9 and an analysis of the Einstein Aging Study used a data-focused approach to detect a changing point in continuous birth years.10 Our birth cohort categories were based on historically meaningful events. Because the ACT study is larger than the Einstein Aging Study, we were able to separate participants born after 1928 into 2 groups. In the ACT study, the most recent birth cohort (1940–1949) had higher educational levels and childhood financial stability compared with cohorts born earlier. Such categorization also allowed for the separation of worldwide economic disruption from family-level financial stability.

    Our analysis may not have captured differences in adult social experiences. Educational level is associated with subsequent occupation and employment patterns. However, birth cohort may reflect experience of events during the 20th century that had broad implications, regardless of educational level. For example, men born in the first 2 decades of the 20th century are likely to have served in the armed forces during World War II and to have benefitted from the GI bill. Men and women from those birth cohorts would also have benefitted from the postwar economic expansion. Our analysis did not capture such adult experiences.

    Limitations

    Our study has several limitations. Participants in older cohorts necessarily had to survive longer to be included in the study. Because the greatest risk factor for dementia is age, the requirement of survival among the pre–World War I and World War I and Spanish influenza birth cohorts as a requirement to enter the ACT study may create differences in dementia risk that are difficult to detect in these groups. Our results suggest that the most recent birth cohorts may continue to experience lower age-specific dementia incidence. However, follow-up period is shorter in these birth cohorts. The ACT study participants are from 1 health system in the Pacific Northwest, and their educational level is high. The cohort is a random sample of age-eligible members of Kaiser Permanente Washington; results therefore reflect this specific population but may not be generalizable to the US population. Our results are consistent with a sample from the Bronx, New York,10 and a nationally representative sample from the United Kingdom,9 suggesting that the decrease in dementia incidence by birth cohort may be a widespread phenomenon. Because ACT study participants may be socioeconomically advantaged, the measures of early-life environment included in this study may not be sensitive enough to detect meaningful differences that have implications for dementia incidence trends by birth cohort.

    The study did not include key health variables from later in the life course that are associated with dementia risk, notably midlife hypertension, hearing loss, late-life depression, diabetes, physical inactivity, and social isolation.6 As the ACT study is currently collecting data on most of these variables, future studies will be able to more fully capture life-course dementia risk factors. As a long-standing study, the follow-up included substantial age overlap of multiple birth cohorts, which had been a limitation in previous studies.9 Dementia diagnosis procedures have been consistent throughout the study. The large size of the ACT study and the theoretical basis of the cohort groups allowed for the inclusion of 2 cohort groups born after 1928 that aligned with historically meaningful events, whereas previous studies have considered only 1 group born after the mid-1920s.9,10

    Dementia incidence has decreased in more recent birth cohorts. Our measures of early-life socioeconomic status and educational level do not account for these differences in this study population. Birth cohort may reflect other historical and social changes that occurred during childhood or adulthood.

    Conclusions

    The historical perspective of dementia risk highlights the need to consider social risk factors across the life course. Further follow-up is required to confirm dementia incidence patterns of those born during World War II and postwar and of those with the oldest ages who were born during the Great Depression period, which may illuminate differences by early-life environment not yet evident.

    Back to top
    Article Information

    Accepted for Publication: May 9, 2020.

    Published: July 27, 2020. doi:10.1001/jamanetworkopen.2020.11094

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Tom SE et al. JAMA Network Open.

    Corresponding Author: Sarah E. Tom, PhD, MPH, Department of Neurology, Columbia University, Neurological Institute of New York, 710 W 168th St, New York, NY 10032 (st3144@cumc.columbia.edu).

    Author Contributions: Dr Tom had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Tom, Crane, Larson.

    Acquisition, analysis, or interpretation of data: All authors.

    Drafting of the manuscript: Tom, Phadke, Larson.

    Critical revision of the manuscript for important intellectual content: Tom, Hubbard, Crane, Stern, Larson.

    Statistical analysis: Tom, Phadke, Hubbard.

    Obtained funding: Crane, Larson.

    Administrative, technical, or material support: Crane, Larson.

    Supervision: Crane, Stern.

    Conflict of Interest Disclosures: Dr Tom reported receiving grants from the National Institute on Aging and the National Institutes of Health during the conduct of the study. Dr Hubbard reported receiving grants from the National Institutes of Health during the conduct of the study and grants from Humana and Pfizer outside the submitted work. Dr Crane reported receiving grants from the National Institutes of Health during the conduct of the study. Dr Stern reported receiving grants from the National Institutes of Health during the conduct of the study. Dr Larson reported receiving grants from the National Institutes of Health and royalties from UpToDate during the conduct of the study. No other disclosures were reported.

    Funding/Support: This work was supported by grants U01AG006781 and K01AG050723 from the National Institute on Aging.

    Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    References
    1.
    Satizabal  CL, Beiser  AS, Chouraki  V, Chene  G, Dufouil  C, Seshadri  S.  Incidence of dementia over three decades in the Framingham Heart Study.   N Engl J Med. 2016;374(6):523-532. doi:10.1056/NEJMoa1504327 PubMedGoogle ScholarCrossref
    2.
    Rajan  KB, Weuve  J, Barnes  LL, Wilson  RS, Evans  DA.  Prevalence and incidence of clinically diagnosed Alzheimer’s disease dementia from 1994 to 2012 in a population study.   Alzheimers Dement. 2019;15(1):1-7. doi:10.1016/j.jalz.2018.07.216 PubMedGoogle ScholarCrossref
    3.
    Noble  JM, Schupf  N, Manly  JJ, Andrews  H, Tang  M-X, Mayeux  R.  Secular trends in the incidence of dementia in a multi-ethnic community.   J Alzheimers Dis. 2017;60(3):1065-1075. doi:10.3233/JAD-170300 PubMedGoogle ScholarCrossref
    4.
    Schrijvers  EMC, Verhaaren  BFJ, Koudstaal  PJ, Hofman  A, Ikram  MA, Breteler  MMB.  Is dementia incidence declining?: trends in dementia incidence since 1990 in the Rotterdam Study.   Neurology. 2012;78(19):1456-1463. doi:10.1212/WNL.0b013e3182553be6 PubMedGoogle ScholarCrossref
    5.
    Gao  S, Burney  HN, Callahan  CM, Purnell  CE, Hendrie  HC.  Incidence of dementia and Alzheimer disease over time: a meta-analysis.   J Am Geriatr Soc. 2019;67(7):1361-1369. doi:10.1111/jgs.16027 PubMedGoogle ScholarCrossref
    6.
    Livingston  G, Sommerlad  A, Orgeta  V,  et al.  Dementia prevention, intervention, and care.   Lancet. 2017;390(10113):2673-2734. doi:10.1016/S0140-6736(17)31363-6 PubMedGoogle ScholarCrossref
    7.
    Skoog  I.  Dementia: dementia incidence—the times, they are a-changing.   Nat Rev Neurol. 2016;12(6):316-318. doi:10.1038/nrneurol.2016.55 PubMedGoogle ScholarCrossref
    8.
    Rocca  WA.  Time, sex, gender, history, and dementia.   Alzheimer Dis Assoc Disord. 2017;31(1):76-79. doi:10.1097/WAD.0000000000000187 PubMedGoogle ScholarCrossref
    9.
    Cadar  D, Lassale  C, Davies  H, Llewellyn  DJ, Batty  GD, Steptoe  A.  Individual and area-based socioeconomic factors associated with dementia incidence in England: evidence from a 12-year follow-up in the English Longitudinal Study of Ageing.   JAMA Psychiatry. 2018;75(7):723-732. doi:10.1001/jamapsychiatry.2018.1012 PubMedGoogle ScholarCrossref
    10.
    Derby  CA, Katz  MJ, Lipton  RB, Hall  CB.  Trends in dementia incidence in a birth cohort analysis of the Einstein Aging Study.   JAMA Neurol. 2017;74(11):1345-1351. doi:10.1001/jamaneurol.2017.1964 PubMedGoogle ScholarCrossref
    11.
    Rocca  WA, Petersen  RC, Knopman  DS,  et al.  Trends in the incidence and prevalence of Alzheimer’s disease, dementia, and cognitive impairment in the United States.   Alzheimers Dement. 2011;7(1):80-93. doi:10.1016/j.jalz.2010.11.002 PubMedGoogle ScholarCrossref
    12.
    McKeown  T.  The Modern Rise of Population. Academic Press; 1976.
    13.
    Fogel  RW. Economic growth, population theory, and physiology: the bearing of long-term processes on the making of economic policy. National Bureau of Economic Research working paper 4638. February 1994. Accessed July 2019. https://www.nber.org/papers/w4638.pdf
    14.
    Knickmeyer  RC, Gouttard  S, Kang  C,  et al.  A structural MRI study of human brain development from birth to 2 years.   J Neurosci. 2008;28(47):12176-12182. doi:10.1523/JNEUROSCI.3479-08.2008 PubMedGoogle ScholarCrossref
    15.
    Levitt  P.  Structural and functional maturation of the developing primate brain.   J Pediatr. 2003;143(4)(suppl):S35-S45. doi:10.1067/S0022-3476(03)00400-1 PubMedGoogle ScholarCrossref
    16.
    National Scientific Council on the Developing Child. The timing and quality of early experiences combine to shape brain architecture: working paper no. 5. 2007. Accessed August 2018. https://developingchild.harvard.edu/resources/the-timing-and-quality-of-early-experiences-combine-to-shape-brain-architecture/
    17.
    Stern  Y.  What is cognitive reserve? theory and research application of the reserve concept.   J Int Neuropsychol Soc. 2002;8(3):448-460. doi:10.1017/S1355617702813248 PubMedGoogle ScholarCrossref
    18.
    Vandenbroucke  JP, von Elm  E, Altman  DG,  et al; STROBE Initiative.  Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration.   Ann Intern Med. 2007;147(8):W163-W194. doi:10.7326/0003-4819-147-8-200710160-00010-w1 PubMedGoogle Scholar
    19.
    Kukull  WA, Higdon  R, Bowen  JD,  et al.  Dementia and Alzheimer disease incidence: a prospective cohort study.   Arch Neurol. 2002;59(11):1737-1746. doi:10.1001/archneur.59.11.1737 PubMedGoogle ScholarCrossref
    20.
    Teng  EL, Hasegawa  K, Homma  A,  et al.  The Cognitive Abilities Screening Instrument (CASI): a practical test for cross-cultural epidemiological studies of dementia.   Int Psychogeriatr. 1994;6(1):45-58. doi:10.1017/S1041610294001602 PubMedGoogle ScholarCrossref
    21.
    American Psychiatric Association.  Diagnostic and Statistical Manual of Mental Disorders. 4th ed. American Psychiatric Association; 1994.
    22.
    Cole  TJ.  Secular trends in growth.   Proc Nutr Soc. 2000;59(2):317-324. doi:10.1017/S0029665100000355 PubMedGoogle ScholarCrossref
    23.
    Floud  R, Fogel  RW, Harris  B, Hong  SC.  The Changing Body: Health, Nutrition, and Human Development in the Western World Since 1700. Cambridge University Press; 2011. doi:10.1017/CBO9780511975912
    24.
    Maurer  J.  Height, education and later-life cognition in Latin America and the Caribbean.   Econ Hum Biol. 2010;8(2):168-176. doi:10.1016/j.ehb.2010.05.013 PubMedGoogle ScholarCrossref
    25.
    Abbott  RD, White  LR, Ross  GW,  et al.  Height as a marker of childhood development and late-life cognitive function: the Honolulu-Asia Aging Study.   Pediatrics. 1998;102(3 Pt 1):602-609. doi:10.1542/peds.102.3.602 PubMedGoogle ScholarCrossref
    26.
    Russ  TC, Kivimaki  M, Starr  JM, Stamatakis  E, Batty  GD.  Height in relation to dementia death: individual participant meta-analysis of 18 UK prospective cohort studies.   Br J Psychiatry. 2014;205(5):348-354. doi:10.1192/bjp.bp.113.142984 PubMedGoogle ScholarCrossref
    27.
    Hixson  JE, Vernier  DT.  Restriction isotyping of human apolipoprotein E by gene amplification and cleavage with HhaI.   J Lipid Res. 1990;31(3):545-548.PubMedGoogle Scholar
    28.
    Emi  M, Wu  LL, Robertson  MA,  et al.  Genotyping and sequence analysis of apolipoprotein E isoforms.   Genomics. 1988;3(4):373-379. doi:10.1016/0888-7543(88)90130-9 PubMedGoogle ScholarCrossref
    29.
    Billingsley  P.  Probability and Measure. John Wiley & Sons; 1995.
    30.
    Fine  JP, Gray  RJ.  A proportional hazards model for the subdistribution of a competing risk.   J Am Stat Assoc. 1999;94(446):496-509. doi:10.1080/01621459.1999.10474144 Google ScholarCrossref
    31.
    Tom  SE, Hubbard  RA, Crane  PK,  et al.  Characterization of dementia and Alzheimer’s disease in an older population: updated incidence and life expectancy with and without dementia.   Am J Public Health. 2015;105(2):408-413. doi:10.2105/AJPH.2014.301935 PubMedGoogle ScholarCrossref
    32.
    Skoog  I, Najar  J, Wetterberg  H.  The importance of birth year for the incidence of dementia.   J Am Geriatr Soc. 2019;67(7):1330-1332. doi:10.1111/jgs.16029 PubMedGoogle ScholarCrossref
    33.
    Cutler  D, Miller  G.  The role of public health improvements in health advances: the twentieth-century United States.   Demography. 2005;42(1):1-22. doi:10.1353/dem.2005.0002 PubMedGoogle ScholarCrossref
    34.
    Moceri  VM, Kukull  WA, Emanual  I,  et al.  Using census data and birth certificates to reconstruct the early-life socioeconomic environment and the relation to the development of Alzheimer’s disease.   Epidemiology. 2001;12(4):383-389. doi:10.1097/00001648-200107000-00007 PubMedGoogle ScholarCrossref
    35.
    Moceri  VM, Kukull  WA, Emanuel  I, van Belle  G, Larson  EB.  Early-life risk factors and the development of Alzheimer’s disease.   Neurology. 2000;54(2):415-420. doi:10.1212/WNL.54.2.415 PubMedGoogle ScholarCrossref
    ×