[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure.
Cumulative Incidence Dementia Adjusted for Death Rates by Birth Place
Cumulative Incidence Dementia Adjusted for Death Rates by Birth Place

HSMS indicates high stroke mortality state.

Table 1.  
Sample Characteristics by Birthplace and Overall
Sample Characteristics by Birthplace and Overall
Table 2.  
Adjusted Hazard Ratios (HRs) for Dementia by Birthplace From Cox Proportional Hazards Modelsa
Adjusted Hazard Ratios (HRs) for Dementia by Birthplace From Cox Proportional Hazards Modelsa
Table 3.  
Estimates of Cumulative Risk of Dementiaa
Estimates of Cumulative Risk of Dementiaa
Table 4.  
Combined Effects of Birth in the High Stroke Mortality States and Race on Risk of Dementiaa
Combined Effects of Birth in the High Stroke Mortality States and Race on Risk of Dementiaa
1.
Howard  VJ, Woolson  RF, Egan  BM,  et al.  Prevalence of hypertension by duration and age at exposure to the stroke belt.  J Am Soc Hypertens. 2010;4(1):32-41.PubMedGoogle ScholarCrossref
2.
Borhani  NO.  Changes and geographic distribution of mortality from cerebrovascular disease.  Am J Public Health Nations Health. 1965;55(5):673-681.PubMedGoogle ScholarCrossref
3.
Centers for Disease Control and Prevention.  Regional and racial differences in prevalence of stroke—23 states and District of Columbia, 2003.  MMWR Morb Mortal Wkly Rep. 2005;54(19):481-484.PubMedGoogle Scholar
4.
Wadley  VG, Unverzagt  FW, McGuire  LC,  et al.  Incident cognitive impairment is elevated in the stroke belt: the REGARDS study.  Ann Neurol. 2011;70(2):229-236.PubMedGoogle ScholarCrossref
5.
Hall  WD, Ferrario  CM, Moore  MA,  et al.  Hypertension-related morbidity and mortality in the southeastern United States.  Am J Med Sci. 1997;313(4):195-209.PubMedGoogle ScholarCrossref
6.
Glymour  MM, Kosheleva  A, Boden-Albala  B.  Birth and adult residence in the Stroke Belt independently predict stroke mortality.  Neurology. 2009;73(22):1858-1865.PubMedGoogle ScholarCrossref
7.
Osmond  C, Barker  DJ, Slattery  JM.  Risk of death from cardiovascular disease and chronic bronchitis determined by place of birth in England and Wales.  J Epidemiol Community Health. 1990;44(2):139-141.PubMedGoogle ScholarCrossref
8.
Strachan  DP, Leon  DA, Dodgeon  B.  Mortality from cardiovascular disease among interregional migrants in England and Wales.  BMJ. 1995;310(6977):423-427.PubMedGoogle ScholarCrossref
9.
Glymour  MM, Kosheleva  A, Wadley  VG, Weiss  C, Manly  JJ.  Geographic distribution of dementia mortality: elevated mortality rates for black and white Americans by place of birth.  Alzheimer Dis Assoc Disord. 2011;25(3):196-202.PubMedGoogle ScholarCrossref
10.
Glymour  MM, Avendaño  M, Berkman  LF.  Is the “stroke belt” worn from childhood? risk of first stroke and state of residence in childhood and adulthood.  Stroke. 2007;38(9):2415-2421.PubMedGoogle ScholarCrossref
11.
Howard  VJ, McClure  LA, Glymour  MM,  et al.  Effect of duration and age at exposure to the Stroke Belt on incident stroke in adulthood.  Neurology. 2013;80(18):1655-1661.PubMedGoogle ScholarCrossref
12.
Mayeda  ER, Glymour  MM, Quesenberry  CP, Whitmer  RA.  Inequalities in dementia incidence between six racial and ethnic groups over 14 years.  Alzheimers Dement. 2016;12(3):216-224.PubMedGoogle ScholarCrossref
13.
Glymour  MM, Manly  JJ.  Lifecourse social conditions and racial and ethnic patterns of cognitive aging.  Neuropsychol Rev. 2008;18(3):223-254.PubMedGoogle ScholarCrossref
14.
Gordon  NP, Kaplan  GA.  Some evidence refuting the HMO “favorable selection” hypothesis: the case of Kaiser Permanente.  Adv Health Econ Health Serv Res. 1991;12:19-39.PubMedGoogle Scholar
15.
Krieger  N.  Overcoming the absence of socioeconomic data in medical records: validation and application of a census-based methodology.  Am J Public Health. 1992;82(5):703-710.PubMedGoogle ScholarCrossref
16.
Gordon NP.  Similarity of the Kaiser Permanente Senior Member Population in Northern California to the Non-Kaiser Permanente Covered and General Population of Seniors in Northern California: Statistics From the 2009 California Health Interview Survey. Oakland, CA: Kaiser Permanente Northern California Division of Research; 2012.
17.
Centers for Disease Control and Prevention. Division for Heart Disease and Stroke Prevention. Interactive Atlas of Heart Disease and Stroke. http://nccd.cdc.gov/DHDSPAtlas. Accessed January 5, 2017.
18.
Exalto  LG, Biessels  GJ, Karter  AJ, Huang  ES, Quesenberry  CP  Jr, Whitmer  RA.  Severe diabetic retinal disease and dementia risk in type 2 diabetes.  J Alzheimers Dis. 2014;42(suppl 3)(suppl 3):S109-S117.PubMedGoogle Scholar
19.
Exalto  LG, Biessels  GJ, Karter  AJ,  et al.  Risk score for prediction of 10 year dementia risk in individuals with type 2 diabetes: a cohort study.  Lancet Diabetes Endocrinol. 2013;1(3):183-190.PubMedGoogle ScholarCrossref
20.
Mayeda  ER, Karter  AJ, Huang  ES, Moffet  HH, Haan  MN, Whitmer  RA.  Racial/ethnic differences in dementia risk among older type 2 diabetic patients: the diabetes and aging study.  Diabetes Care. 2014;37(4):1009-1015.PubMedGoogle ScholarCrossref
21.
Whitmer  RA, Karter  AJ, Yaffe  K, Quesenberry  CP  Jr, Selby  JV.  Hypoglycemic episodes and risk of dementia in older patients with type 2 diabetes mellitus.  JAMA. 2009;301(15):1565-1572.PubMedGoogle ScholarCrossref
22.
Whitmer  RA, Sidney  S, Selby  J, Johnston  SC, Yaffe  K.  Midlife cardiovascular risk factors and risk of dementia in late life.  Neurology. 2005;64(2):277-281.PubMedGoogle ScholarCrossref
23.
Whitmer  RA, Gustafson  DR, Barrett-Connor  E, Haan  MN, Gunderson  EP, Yaffe  K.  Central obesity and increased risk of dementia more than three decades later.  Neurology. 2008;71(14):1057-1064.PubMedGoogle ScholarCrossref
24.
Katon  WJ, Lin  EHB, Williams  LH,  et al.  Comorbid depression is associated with an increased risk of dementia diagnosis in patients with diabetes: a prospective cohort study.  J Gen Intern Med. 2010;25(5):423-429.PubMedGoogle ScholarCrossref
25.
Huang  ES, Laiteerapong  N, Liu  JY, John  PM, Moffet  HH, Karter  AJ.  Rates of complications and mortality in older patients with diabetes mellitus: the diabetes and aging study.  JAMA Intern Med. 2014;174(2):251-258.PubMedGoogle ScholarCrossref
26.
Beiser  A, D’Agostino  RB  Sr, Seshadri  S, Sullivan  LM, Wolf  PA.  Computing estimates of incidence, including lifetime risk: Alzheimer’s disease in the Framingham Study: the Practical Incidence Estimators (PIE) macro.  Stat Med. 2000;19(11-12):1495-1522.PubMedGoogle ScholarCrossref
27.
Liu  SY, Glymour  MM, Zahodne  LB, Weiss  C, Manly  JJ.  Role of place in explaining racial heterogeneity in cognitive outcomes among older adults.  J Int Neuropsychol Soc. 2015;21(9):677-687.PubMedGoogle ScholarCrossref
28.
United States Census Bureau. Historical County Level Poverty Estimates Tool. http://www.census.gov/library/visualizations/time-series/demo/census-poverty-tool.html. Updated May 16, 2016. Accessed January 10, 2017.
29.
Hackman  DA, Farah  MJ.  Socioeconomic status and the developing brain.  Trends Cogn Sci. 2009;13(2):65-73.PubMedGoogle ScholarCrossref
30.
Wadsworth  ME, Cripps  HA, Midwinter  RE, Colley  JR.  Blood pressure in a national birth cohort at the age of 36 related to social and familial factors, smoking, and body mass.  BMJ (Clin Res Ed). 1985;291(6508):1534-1538.PubMedGoogle ScholarCrossref
31.
Barker  DJP.  The developmental origins of chronic adult disease.  Acta Paediatr Suppl. 2004;93(446):26-33.PubMedGoogle Scholar
32.
Rich-Edwards  JW, Stampfer  MJ, Manson  JE,  et al.  Birth weight and risk of cardiovascular disease in a cohort of women followed up since 1976.  BMJ. 1997;315(7105):396-400.PubMedGoogle ScholarCrossref
33.
Osmond  C, Barker  DJ, Winter  PD, Fall  CH, Simmonds  SJ.  Early growth and death from cardiovascular disease in women.  BMJ. 1993;307(6918):1519-1524.PubMedGoogle ScholarCrossref
34.
Barker  DJP, Winter  PD, Osmond  C, Margetts  B, Simmonds  SJ.  Weight in infancy and death from ischaemic heart disease.  Lancet. 1989;2(8663):577-580.PubMedGoogle ScholarCrossref
35.
Patton  KK, Benjamin  EJ, Kosheleva  A, Curtis  LH, Glymour  MM.  Early-life antecedents of atrial fibrillation: place of birth and atrial fibrillation-related mortality.  Ann Epidemiol. 2011;21(10):732-738.PubMedGoogle ScholarCrossref
36.
Meng  X, D’Arcy  C.  Education and dementia in the context of the cognitive reserve hypothesis: a systematic review with meta-analyses and qualitative analyses.  PLoS One. 2012;7(6):e38268.PubMedGoogle ScholarCrossref
37.
Margo  RA.  Race and Schooling in the South, 1880-1950: An Economic History. Chicago, IL: University of Chicago Press; 1990.Crossref
Original Investigation
September 2017

Association Between Birth in a High Stroke Mortality State, Race, and Risk of Dementia

Author Affiliations
  • 1Division of Research, Kaiser Permanente, Oakland, California
  • 2Department of Epidemiology and Biostatistics, University of California, San Francisco
  • 3Department of Epidemiology, Fielding School of Public Health, University of California, Los Angeles
JAMA Neurol. 2017;74(9):1056-1062. doi:10.1001/jamaneurol.2017.1553
Key Points

Question  Is birth in a state with high stroke mortality associated with dementia risk in a cohort of individuals who all subsequently resided outside of those states?

Findings  In this cohort study, among 7423 Northern California residents with equal access to health care, dementia risk was approximately 27% higher among those born in high stroke mortality states compared with elsewhere. Compared with nonblack persons born outside of high stroke mortality states, black individuals born in a high stroke mortality state had the highest dementia risk, followed by black individuals not born in a high stroke mortality state, and lastly nonblack persons born in a high stroke mortality state.

Meaning  Racial inequalities in dementia may partially reflect geographic patterning of early-life exposures.

Abstract

Importance  Birth in a group of predominantly southern US states is robustly linked to increased stroke risk. Given the role of cerebrovascular disease in dementia risk, geographic patterning may also occur for dementia incidence.

Objective  To determine whether birth in 9 high stroke mortality states (HSMSs) is associated with dementia in a diverse cohort of individuals living in Northern California.

Design, Setting, and Participants  An observational cohort study included 7423 members of Kaiser Permanente Northern California (KPNC), an integrated health care delivery system, with health survey and clinical examination data available. Data were collected between 1964 and 1973 when the individuals were middle-aged and 1996 and 2015 when participants were in later life.

Exposures  Self-reported state of birth in an HSMS (top quintile of states for stroke mortality).

Main Outcomes and Measures  Dementia diagnoses obtained from electronic health records from January 1, 1996, to October 15, 2015. Place of birth, race, educational level, and midlife vascular risk factors data were collected between 1964 and 1973.

Results  Of the 7423 persons included in the analysis, 4049 (54.5%) were women; 1354 (18.2%) were black. The mean (SD) age of study participants at their first visit between 1963 and 1974 was 42.94 (1.73) years and mean (SD) age at the beginning of follow-up for dementia in 1996 was 71.14 (2.72) years. Dementia was diagnosed in 2254 (30.4%) of the participants and was more common among those born in an HSMS than those born outside of one (455 [39.0%] vs 1799 [28.8%]). Birth in an HSMS was 9.6 times more common for black participants (795 [58.7%]) than nonblack participants (371 [6.1%]). Overall, birth in an HSMS was associated with a 28% higher risk of dementia (adjusted hazard ratio [aHR], 1.28; 95% CI, 1.13-1.46) adjusted for age, sex, and race. Compared with nonblack persons born outside of an HSMS, black individuals born in an HSMS had the highest dementia risk (aHR, 1.67; 95% CI, 1.48-1.88), followed by black individuals not born in an HSMS (aHR, 1.48; 95% CI, 1.28-1.72), and nonblack persons born in an HSMS had a 46% increased risk (aHR, 1.46; 95% CI, 1.23-1.74). Cumulative 20-year dementia risks at age 65 years were 30.13% (95% CI, 26.87%-32.93%) and 21.80% (95% CI, 20.51%-22.91%) for individuals born in and outside an HSMS, respectively.

Conclusions and Relevance  To our knowledge, this is the first study to date of place of birth and incident dementia and shows increased risk for individuals born in an HSMS, even though all participants subsequently resided in California. Birth in an HSMS was common among black participants. Place of birth has enduring consequences for dementia risk and may be a major contributor to racial disparities in dementia.

Introduction

Quiz Ref IDAdult residence in a band of states in the southern United States known as the Stroke Belt is associated with increased risk of a plethora of conditions, including hypertension, diabetes, stroke, and cognitive impairment.1-6 A growing body of research demonstrates a link between birthplace and adult cardiovascular risk,6-10 and early-life residence in the Stroke Belt has been shown to increase the risk of poor health outcomes even for individuals who subsequently move out of the region.1,6,10,11

Although birth in the Stroke Belt has been associated with greater risk of dementia mortality, regardless of the state of residence at death,9 to our knowledge, no prior study has examined whether incident dementia is similarly elevated for individuals born in geographic regions with high stroke risk but who subsequently reside elsewhere. There is strong geographic patterning of place of birth and adult residence by race, which could play a role in racial or ethnic inequalities in rates of dementia.12,13 Geographic patterning of increased risk of disease can represent clustering of risk factors, some of which are modifiable. Stroke risk can reflect clustering of cerebrovascular risk factors, many of which are also robust predictors of dementia, so it is biologically plausible that early life residence in regions with elevated stroke risk may also increase dementia risk. Our objective was to determine whether birth in a high stroke mortality state (HSMS) is associated with increased dementia risk in a diverse cohort of individuals living in Northern California with equal access to medical care.

Methods
Study Participants

Kaiser Permanente Northern California (KPNC) is a large integrated health care delivery system that provides care to over 3 million members in Northern California (approximately 30% of the geographic region). The KPNC member population is generally representative of the overall population of the region, although individuals at extreme tails of the income distribution are underrepresented.14-16 We included members who participated in the Multiphasic Health Checkup (MHC), an optional checkup provided to health plan members in San Francisco and Oakland, California, in the 1960s and 1970s. During MHC visits, information was collected on demographics, lifestyle, and cardiometabolic health.

The study was approved by the Kaiser Internal Review Board and they approved a waiver of informed consent for this study. Data are deidentified.

Study Design

This was a cohort study of 7423 US-born individuals who participated in at least 1 MHC visit between 1964 and 1973, were aged 40 to 45 years at the time of participation, remained KPNC members when electronic health records became available in 1994, and reported state of birth. The eFigure in the Supplement shows the flowchart of exclusions. For members who participated in more than 1 MHC visit, data from the first visit were included. This cohort was followed-up from January 1, 1996, until the first diagnosis of dementia, death, lapse in health plan membership, or October 1, 2015, for up to 19 years of follow-up.

State of Birth

During MHC visits, state of birth was captured with an open-ended questionnaire item asking, “Where were you born?” To best capture the early life conditions of the population, we used overall Centers for Disease Control and Prevention stroke mortality rates from 2012 to 2014, which are the most current estimates of high stroke risk states. High stroke mortality states were those in the top quintiles of stroke mortality rates among US territories (ie, states with >83 stroke deaths per 100 000 people older than 35 years).17 Based on this information, we coded place of birth as inside or outside an HSMS. Quiz Ref IDThe following 9 states were considered HSMSs: Alabama, Alaska, Arkansas, Louisiana, Mississippi, Oklahoma, Tennessee, South Carolina, and West Virginia. Many of these states are part of what is commonly considered the Stroke Belt, although definitions of the Stroke Belt are not consistent across prior research and public health activities.

Dementia Diagnosis and Mortality

Consistent with previous publications in this population,12,18-23 dementia diagnoses were identified from KPNC electronic medical records. The following International Classification of Diseases, Ninth Revision (ICD-9) diagnosis codes were used to identify dementia cases between January 1, 1996, and October 1, 2015: Alzheimer disease (code 331.0), vascular dementia (code 290.4x), and other/nonspecific dementia (codes 290.0, 290.1x, 290.2x, 290.3, 294.2x, and 294.8). A previous study, also within a health care system, found a that a similar set of ICD-9 codes had a sensitivity of 77% and a specificity of 95%, compared with a consensus diagnosis of dementia utilizing medical records review, physical examination, structured interviews, and a neuropsychiatric battery.24

Death was captured through electronic medical records, the California State Mortality File, and Social Security Death Records. This method of death ascertainment has previously been used in studies of this population.12,25

Covariates

The 1964-1973 MHC questionnaire included self-report of the following covariates: age, sex, race (black, white, Asian, or other), educational attainment (high school or less, some college or technical school, at least college graduate), blood pressure medication, and smoking. The 1964-1973 MHC visit also included clinical measurements of height, weight, and systolic and diastolic blood pressure. Height and weight were combined to calculate midlife body mass index. Hypertension was defined as blood pressure measurements of systolic blood pressure 140 mm Hg or higher or diastolic pressure 90 mm Hg or higher. Hypertension was combined with self-reported blood pressure medication use to create 3 categories of midlife blood pressure: no hypertension, controlled hypertension, and uncontrolled hypertension. Information on ever smoking was connected with smoking duration to classify midlife smoking status as never smoked, smoked less than 10 years, smoked between 10 and 20 years, or smoked more than 20 years. Late-life cardiovascular risk (yes/no) was defined as having at least 1 of the following diagnoses between January 1, 1996, and October 1, 2015, in the electronic health records: diabetes, hypertension, heart failure, acute myocardial infarction, and stroke (eTable 1 in the Supplement). Missing indicators were used for missing data on smoking duration (n = 125), midlife body mass index (n = 432), and blood pressure medication (n = 8).

Statistical Analysis

First, we examined the distribution of sociodemographic variables, midlife risk factors, and late-life conditions stratified on birth in HSMSs (yes/no). We then estimated dementia incidence rates standardized to the 2000 US Census by region of birth. Next, we tested the assumption of proportional hazards and used a series of Cox proportional hazards models (with age as the timescale) to examine the association between birth in an HSMS and risk of dementia. We adjusted for covariates in sequential steps, mirroring exposure across the lifespan, in recognition that midlife and late-life risk factors may be partial mediators of birth place effects: (1) adjusted for sociodemographic variables, (2) further adjusted for midlife cardiovascular risk factors, and (3) with late-life cardiovascular health added. We tested for possible effect modification by race by using interaction terms in pooled models and then stratifying by race.

We concurrently examined the associations of birth in an HSMS and race with dementia risk by grouping individuals into 4 categories: (1) nonblack individuals born outside of HSMSs, (2) black individuals born outside of HSMSs, (3) nonblack individuals born in HSMSs, and (4) black individuals born in HSMSs. We examined the association of these categories with dementia risk in Cox proportional hazards models, adjusting for sociodemographics and midlife and late-life exposures sequentially.

We estimated and plotted the cumulative risk of dementia in 5-year increments from 10 to 25 years conditional on survival free of dementia to age 65 years by birth in an HSMS using the Practical Incidence Estimator macro, which incorporates information on death rates.26 All analyses were conducted using SAS, version 9.3 (SAS Institute Inc), and 2-sided P values <.05 were considered statistically significant.

Results

Overall, the sample was 18.2% black, 74.2% white, and 5.0% Asian; 15.7% were born in an HSMS (Table 1). Of those born in HSMSs, 68.2% were black and 31.8% were nonblack. Even though all sample members resided in Northern California by 1974, black participants were 9.6 times more likely to be born in an HSMS than nonblack participants (795 [58.7%] vs 371 [6.1%]) (eTable 2 and eTable 3 in the Supplement). A total of 30.3% of individuals born in an HSMS and 27.5% of those born outside of those states experienced a late-life stroke. The mean (SD) age of study participants at their first MHC visit between 1963 and 1974 was 42.94 (1.73) years (range, 39.96-45.96 years) and their mean age at the beginning of follow-up for dementia in 1996 was 71.14 (2.72) years (range, 63.80-76.63 years). The mean (SD) follow-up time for dementia (starting in 1996) was 11.61 (6.24) years (range, 0.02-19.75 years). By the end of follow-up, 2254 (30.4%) of the participants were diagnosed with dementia, 2473 (33.3%) had died without a dementia diagnosis, 1227 (16.5%) were censored owing to a lapse in membership, and 1469 (19.8%) were alive, members of KPNC, and dementia-free at the end of follow-up. The incidence rates for people born in an HSMS and people born outside of an HSMS were 23.56 (95% CI, 20.46-26.67) and 15.54 (95% CI, 14.42-16.66) per 1000 person-years, respectively.

Visual examination of the unadjusted survival plots by birth in HSMSs provided evidence that the proportional hazards assumption was met. Quiz Ref IDIn age- and sex-adjusted Cox proportional hazards models, birth in an HSMS was associated with a 54% increased risk of dementia (adjusted hazard ratio [aHR], 1.54; 95% CI, 1.39-1.71) (Table 2). The association remained significant after adjustment for black race (aHR, 1.28; 95% CI, 1.13-1.46). Further adjustment for education and midlife vascular risk factors did not attenuate the association between birth in an HSMS and dementia (aHR, 1.27; 95% CI, 1.11-1.44). Among nonblack individuals, birth in an HSMS was associated with a 46% increase in dementia risk (aHR,1.46; 95% CI, 1.23-1.75) adjusted for age and sex. In black individuals, birth in an HSMS was associated with a 13% increased dementia risk (aHR, 1.13; 95% CI, 0.94-1.35), although the 95% CI included the null. An interaction term for birth in an HSMS and black race was statistically significant (P = .04), indicating that the relative effects of being born in an HSMS differed for black and nonblack individuals.

The 10-, 15-, 20-, and 25-year cumulative risks of dementia conditional on survival to age 65 years without dementia (Table 3 and Figure) were each greater among individuals born in HSMSs compared with their counterparts born outside of those states. For example, the 20-year cumulative dementia risk was 30.13% (95% CI, 26.87%-32.93%) for people born in the HSMSs and was 21.80% (95% CI, 20.51%-22.91%) for those born outside those states.

Quiz Ref IDIn age- and sex-adjusted models examining joint effects of birth in an HSMS and race, with nonblack individuals born outside the HSMSs as the reference group, black individuals born in an HSMS had the greatest dementia risk (aHR,1.67; 95% CI, 1.48-1.88), followed by black individuals not born in an HSMS (aHR,1.48; 95% CI, 1.28-1.72) and nonblack persons born in an HSMS (aHR, 1.46; 95% CI, 1.23-1.74) (Table 4). After adjusting for education and midlife and late-life health exposures, compared with nonblack members born outside the HSMSs, both black and nonblack individuals born in an HSMS had more than a 40% higher dementia risk and black individuals born outside of the HSMSs had a 32% higher risk (aHR, 1.32; 95% CI, 1.13-1.54).

Discussion

In what we believe to be the first article examining place of birth, race, and dementia incidence, we found place of birth to be a robust risk factor for dementia even after accounting for race, educational level, and life-course vascular risk factors. Birth in an HSMS was associated with a 54% increased risk of dementia compared with birth outside of those states in age-adjusted models and persisted after adjustments. Estimates of cumulative risk, an absolute risk measure (as opposed to relative measures, such as HRs), further underscore the risk associated with birth in an HSMS. Individuals born in an HSMS had approximately 38% higher 20-year cumulative dementia risk compared with people born outside of an HSMS. Compared with nonblack individuals born outside of the HSMSs, black persons born in these states had a 67% higher incidence rate of dementia, whereas black persons born outside of these states had a 48% greater dementia incidence rate.

Even though the association was stronger among nonblack individuals, trends suggested an increased risk of dementia associated with birth in an HSMS for both the black and nonblack cohorts. Furthermore, dementia risk was greatest for individuals who were black and born in an HSMS. The excess dementia risk associated with geography may therefore contribute to racial disparities in dementia. Post hoc analyses showed that black race was associated with a 56% increased risk of dementia when controlling only for age. With adjustment for birth in an HSMS, the risk associated with black race was attenuated by more than a third. This has great public health relevance since many current black elderly individuals were born in the South and moved away during the “great migration.”13

To our knowledge, no study has previously examined birthplace and dementia incidence. However, our finding of elevated risk among people born in an HSMS is consistent with a previous death-records–based study reporting an elevated risk of dementia-related mortality among individuals born in the Stroke Belt.9 The present analysis is a valuable advance because studies based on mortality data are potentially biased owing to inaccurate reporting of place of birth and dementia diagnoses on death records and undercounts in census denominator populations.

Several of the states with high stroke mortality rates are part of the southern region of the United States, which has the largest racial disparities in cognitive outcomes between white and black individuals.27Quiz Ref ID In 1960, the South had the greatest proportion of people living in poverty compared with other census regions; 35.6% of people in the South lived below poverty level compared with 22.1% of the United States overall and only 14.4% of those in the Northeast.28 Poverty early in life can reflect a host of factors that could affect brain health and cognitive reserve, such as nutrition, lead exposure, chronic stress, and cognitive stimulation.29 Poverty is highly associated with low birthweight and it is very likely that low birthweight disproportionally affected southern black persons in the 1920s, placing them at greater risk for elevated blood pressure,30,31 stroke,31,32 and cardiovascular disease mortality.33,34 Relatedly, birth or childhood residence in the Stroke Belt specifically has also been associated with stroke,10,11 hypertension,1 and atrial fibrillation.35 One’s place of birth can have lasting effects by influencing individuals’ lifestyle and health behavior throughout adult-life, even when residence has changed and access to health care is uniform.

Birth in southern states may also be associated with poorer quality of education among older adults, which affects cognitive reserve. The cognitive reserve hypothesis posits that education (and other factors) during early life prompts protective biological changes in the brain that can delay against cognitive impairment.36 Southern state laws allowing school segregation were not overturned until 1954 when the Supreme Court’s Brown vs Board of Education ruled that these laws were unconstitutional. Segregated schools designated for black children received less funding, had higher student-teacher ratios, and had shorter term lengths than schools designated for white children.37 Although black children were most disadvantaged by these policies, average educational quality for segregated white schools was also lower than average quality in nonsegregated states.9 Even after the Brown vs Board of Education decision, de facto racial segregation of schools remained common. In our sample, 66.0% of individuals born in an HSMS had a high school or less level of education compared with 44.9% of individuals born outside of these states. While we know the highest grade of school participants completed, we can only extrapolate regarding educational quality. Additional research is needed to assess whether higher levels of formal education diminish the absolute risk of dementia associated with being born in an HSMS among black or nonblack persons.

Limitations and Strengths

Given that everyone in our cohort migrated to California, our results are susceptible to selection bias and probably underestimate the effect of birth in an HSMS due to a healthy migrant bias: migrants may be healthier or more highly educated than those who remained in an HSMS. It is reasonable to assume that it would be more challenging for black individuals born in the South to move to California than their white counterparts. Therefore, selection bias may especially lead to an underestimate of the effect among black individuals and the risk of dementia associated with birth in an HSMS may not differ by race.

Although we know both the birthplace and the midlife and current location of individuals in this study, we did not have complete residential history and could not determine how long they resided in an HSMS. Therefore, we cannot disentangle whether cumulative or longer time of residence was worse or whether the effect of birthplace varies by the age at which they left the HSMS. We also do not know childhood exposure to relevant residential risk factors, such as secondhand smoke, lead, or air pollution, although this is a goal of our future research. It is unclear whether these results are generalizable to all of Northern California and individuals who were born during a different time since early-life exposures are influenced by social norms and policies. We do not have brain imaging data in our sample and thus were unable to examine subclinical vascular brain injury; we therefore could not directly examine the association between birth in an HSMS and brain health. As with all observational studies, the possibility of residual confounding, unmeasured confounding, and measurement error persists. Strengths of this study include a long follow-up period during which an array of data regarding a large span of the life course were collected, a multiethnic cohort with equal access to health care, and granular health records with information about other health conditions. By evaluating birth in HSMSs in a group that had subsequently all moved to 1 geographic region, we had the unusual opportunity to determine the role of early life exposures and conditions that are captured in geographic differences even in individuals who did not remain in their state of birth. We found that black individuals born outside the HSMSs had very similar dementia risks as nonblack individuals born inside an HSMS.

Conclusions

To our knowledge, this is the first study of place of birth and risk of incident dementia. Our study confirms the overlap between race and place of birth as potential contributors to disparities in dementia risk, showing that black individuals are more likely to have been born in HSMSs than are those of other races. Furthermore, there was an influence of place of birth on dementia risk in this group of individuals who moved away from states with high stroke mortality rates decades before dementia diagnosis. Place of birth may reflect a host of social and environmental conditions in early life that could be some of the primary drivers of racial inequalities in rates of dementia. Future research should examine possible pathways through which early-life biological and psychosocial exposures affect dementia risk. In addition, further research needs to address the geographic patterning of dementia and elucidate the effect of place over the life course on health. It is also important to conduct such analyses in different cohorts of aging populations since the early life social and physical conditions to which individuals are exposed will vary by birth cohort. Our findings underscore the importance of examining early life factors when researching dementia risk as these may inform on the timing, populations, and mechanism of interventions.

Back to top
Article Information

Accepted for Publication: April 25, 2017.

Corresponding Author: Rachel A. Whitmer, PhD, Division of Research, Kaiser Permanente, 2000 Broadway, Oakland, CA 94612 (rachel.whitmer@kp.org).

Correction: This article was corrected to fix a numeric error in Key Points on August 7, 2017.

Published Online: July 31, 2017. doi:10.1001/jamaneurol.2017.1553

Author Contributions: Dr Whitmer 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.

Study concept and design: Gilsanz, Glymour, Whitmer.

Acquisition, analysis, or interpretation of data: Gilsanz, Mayeda, Quesenberry, Whitmer.

Drafting of the manuscript:Gilsanz, Whitmer.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Gilsanz, Glymour, Quesenberry, Whitmer.

Obtained funding: Whitmer.

Administrative, technical, or material support: Mayeda.

Study supervision: Whitmer.

Other–coding support and code review for statistical analysis: Mayeda.

Conflict of Interest Disclosures: None reported.

Funding/Support: This work was supported by grants RF1 AG052132 (principal investigator: Dr Whitmer), R01 AG050782 (principal investigator: Dr Whitmer), and K99 AG053410 (principal investigator: Dr Mayeda) from the National Institute on Aging, National Institutes of Health; Dr Gilsanz is supported by the University of California at San Francisco Training for Research on Aging and Chronic Disease (T32 AG049663).

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.
Howard  VJ, Woolson  RF, Egan  BM,  et al.  Prevalence of hypertension by duration and age at exposure to the stroke belt.  J Am Soc Hypertens. 2010;4(1):32-41.PubMedGoogle ScholarCrossref
2.
Borhani  NO.  Changes and geographic distribution of mortality from cerebrovascular disease.  Am J Public Health Nations Health. 1965;55(5):673-681.PubMedGoogle ScholarCrossref
3.
Centers for Disease Control and Prevention.  Regional and racial differences in prevalence of stroke—23 states and District of Columbia, 2003.  MMWR Morb Mortal Wkly Rep. 2005;54(19):481-484.PubMedGoogle Scholar
4.
Wadley  VG, Unverzagt  FW, McGuire  LC,  et al.  Incident cognitive impairment is elevated in the stroke belt: the REGARDS study.  Ann Neurol. 2011;70(2):229-236.PubMedGoogle ScholarCrossref
5.
Hall  WD, Ferrario  CM, Moore  MA,  et al.  Hypertension-related morbidity and mortality in the southeastern United States.  Am J Med Sci. 1997;313(4):195-209.PubMedGoogle ScholarCrossref
6.
Glymour  MM, Kosheleva  A, Boden-Albala  B.  Birth and adult residence in the Stroke Belt independently predict stroke mortality.  Neurology. 2009;73(22):1858-1865.PubMedGoogle ScholarCrossref
7.
Osmond  C, Barker  DJ, Slattery  JM.  Risk of death from cardiovascular disease and chronic bronchitis determined by place of birth in England and Wales.  J Epidemiol Community Health. 1990;44(2):139-141.PubMedGoogle ScholarCrossref
8.
Strachan  DP, Leon  DA, Dodgeon  B.  Mortality from cardiovascular disease among interregional migrants in England and Wales.  BMJ. 1995;310(6977):423-427.PubMedGoogle ScholarCrossref
9.
Glymour  MM, Kosheleva  A, Wadley  VG, Weiss  C, Manly  JJ.  Geographic distribution of dementia mortality: elevated mortality rates for black and white Americans by place of birth.  Alzheimer Dis Assoc Disord. 2011;25(3):196-202.PubMedGoogle ScholarCrossref
10.
Glymour  MM, Avendaño  M, Berkman  LF.  Is the “stroke belt” worn from childhood? risk of first stroke and state of residence in childhood and adulthood.  Stroke. 2007;38(9):2415-2421.PubMedGoogle ScholarCrossref
11.
Howard  VJ, McClure  LA, Glymour  MM,  et al.  Effect of duration and age at exposure to the Stroke Belt on incident stroke in adulthood.  Neurology. 2013;80(18):1655-1661.PubMedGoogle ScholarCrossref
12.
Mayeda  ER, Glymour  MM, Quesenberry  CP, Whitmer  RA.  Inequalities in dementia incidence between six racial and ethnic groups over 14 years.  Alzheimers Dement. 2016;12(3):216-224.PubMedGoogle ScholarCrossref
13.
Glymour  MM, Manly  JJ.  Lifecourse social conditions and racial and ethnic patterns of cognitive aging.  Neuropsychol Rev. 2008;18(3):223-254.PubMedGoogle ScholarCrossref
14.
Gordon  NP, Kaplan  GA.  Some evidence refuting the HMO “favorable selection” hypothesis: the case of Kaiser Permanente.  Adv Health Econ Health Serv Res. 1991;12:19-39.PubMedGoogle Scholar
15.
Krieger  N.  Overcoming the absence of socioeconomic data in medical records: validation and application of a census-based methodology.  Am J Public Health. 1992;82(5):703-710.PubMedGoogle ScholarCrossref
16.
Gordon NP.  Similarity of the Kaiser Permanente Senior Member Population in Northern California to the Non-Kaiser Permanente Covered and General Population of Seniors in Northern California: Statistics From the 2009 California Health Interview Survey. Oakland, CA: Kaiser Permanente Northern California Division of Research; 2012.
17.
Centers for Disease Control and Prevention. Division for Heart Disease and Stroke Prevention. Interactive Atlas of Heart Disease and Stroke. http://nccd.cdc.gov/DHDSPAtlas. Accessed January 5, 2017.
18.
Exalto  LG, Biessels  GJ, Karter  AJ, Huang  ES, Quesenberry  CP  Jr, Whitmer  RA.  Severe diabetic retinal disease and dementia risk in type 2 diabetes.  J Alzheimers Dis. 2014;42(suppl 3)(suppl 3):S109-S117.PubMedGoogle Scholar
19.
Exalto  LG, Biessels  GJ, Karter  AJ,  et al.  Risk score for prediction of 10 year dementia risk in individuals with type 2 diabetes: a cohort study.  Lancet Diabetes Endocrinol. 2013;1(3):183-190.PubMedGoogle ScholarCrossref
20.
Mayeda  ER, Karter  AJ, Huang  ES, Moffet  HH, Haan  MN, Whitmer  RA.  Racial/ethnic differences in dementia risk among older type 2 diabetic patients: the diabetes and aging study.  Diabetes Care. 2014;37(4):1009-1015.PubMedGoogle ScholarCrossref
21.
Whitmer  RA, Karter  AJ, Yaffe  K, Quesenberry  CP  Jr, Selby  JV.  Hypoglycemic episodes and risk of dementia in older patients with type 2 diabetes mellitus.  JAMA. 2009;301(15):1565-1572.PubMedGoogle ScholarCrossref
22.
Whitmer  RA, Sidney  S, Selby  J, Johnston  SC, Yaffe  K.  Midlife cardiovascular risk factors and risk of dementia in late life.  Neurology. 2005;64(2):277-281.PubMedGoogle ScholarCrossref
23.
Whitmer  RA, Gustafson  DR, Barrett-Connor  E, Haan  MN, Gunderson  EP, Yaffe  K.  Central obesity and increased risk of dementia more than three decades later.  Neurology. 2008;71(14):1057-1064.PubMedGoogle ScholarCrossref
24.
Katon  WJ, Lin  EHB, Williams  LH,  et al.  Comorbid depression is associated with an increased risk of dementia diagnosis in patients with diabetes: a prospective cohort study.  J Gen Intern Med. 2010;25(5):423-429.PubMedGoogle ScholarCrossref
25.
Huang  ES, Laiteerapong  N, Liu  JY, John  PM, Moffet  HH, Karter  AJ.  Rates of complications and mortality in older patients with diabetes mellitus: the diabetes and aging study.  JAMA Intern Med. 2014;174(2):251-258.PubMedGoogle ScholarCrossref
26.
Beiser  A, D’Agostino  RB  Sr, Seshadri  S, Sullivan  LM, Wolf  PA.  Computing estimates of incidence, including lifetime risk: Alzheimer’s disease in the Framingham Study: the Practical Incidence Estimators (PIE) macro.  Stat Med. 2000;19(11-12):1495-1522.PubMedGoogle ScholarCrossref
27.
Liu  SY, Glymour  MM, Zahodne  LB, Weiss  C, Manly  JJ.  Role of place in explaining racial heterogeneity in cognitive outcomes among older adults.  J Int Neuropsychol Soc. 2015;21(9):677-687.PubMedGoogle ScholarCrossref
28.
United States Census Bureau. Historical County Level Poverty Estimates Tool. http://www.census.gov/library/visualizations/time-series/demo/census-poverty-tool.html. Updated May 16, 2016. Accessed January 10, 2017.
29.
Hackman  DA, Farah  MJ.  Socioeconomic status and the developing brain.  Trends Cogn Sci. 2009;13(2):65-73.PubMedGoogle ScholarCrossref
30.
Wadsworth  ME, Cripps  HA, Midwinter  RE, Colley  JR.  Blood pressure in a national birth cohort at the age of 36 related to social and familial factors, smoking, and body mass.  BMJ (Clin Res Ed). 1985;291(6508):1534-1538.PubMedGoogle ScholarCrossref
31.
Barker  DJP.  The developmental origins of chronic adult disease.  Acta Paediatr Suppl. 2004;93(446):26-33.PubMedGoogle Scholar
32.
Rich-Edwards  JW, Stampfer  MJ, Manson  JE,  et al.  Birth weight and risk of cardiovascular disease in a cohort of women followed up since 1976.  BMJ. 1997;315(7105):396-400.PubMedGoogle ScholarCrossref
33.
Osmond  C, Barker  DJ, Winter  PD, Fall  CH, Simmonds  SJ.  Early growth and death from cardiovascular disease in women.  BMJ. 1993;307(6918):1519-1524.PubMedGoogle ScholarCrossref
34.
Barker  DJP, Winter  PD, Osmond  C, Margetts  B, Simmonds  SJ.  Weight in infancy and death from ischaemic heart disease.  Lancet. 1989;2(8663):577-580.PubMedGoogle ScholarCrossref
35.
Patton  KK, Benjamin  EJ, Kosheleva  A, Curtis  LH, Glymour  MM.  Early-life antecedents of atrial fibrillation: place of birth and atrial fibrillation-related mortality.  Ann Epidemiol. 2011;21(10):732-738.PubMedGoogle ScholarCrossref
36.
Meng  X, D’Arcy  C.  Education and dementia in the context of the cognitive reserve hypothesis: a systematic review with meta-analyses and qualitative analyses.  PLoS One. 2012;7(6):e38268.PubMedGoogle ScholarCrossref
37.
Margo  RA.  Race and Schooling in the South, 1880-1950: An Economic History. Chicago, IL: University of Chicago Press; 1990.Crossref
×