In Europe and the U.S., proportionately fewer people are developing dementia now than in the past. Is this driven by less-prevalent Alzheimer's disease pathology? No, say researchers led by Francine Grodstein at Rush University Medical Center in Chicago. In the February 20 JAMA Neurology, they reported that among 1,550 older Americans born over a 25-year period, all had similar amounts of amyloid plaques at death, which came at an average of 90 years.

  • Prevalence of amyloid plaques has remained steady over the past 25 years.
  • Neurofibrillary tangle burden may be on the rise.
  • Atherosclerosis in the brain is becoming less common.

If less pathology does not explain falling dementia incidence, then what does? People born in the 1920s had healthier blood vessels in their brains when they died than did those born in the 1900s, the authors found. They think better cardiovascular health among people born in more recent decades may make them more resilient to AD pathology.

“This research reinforces the importance of population-based efforts to improve vascular health and suggests that we may already be seeing dementia prevention in action,” wrote Jonathan Schott of University College London (comment below). Chengxuan Qiu, Karolinska Institutet, Sweden, called this an important study. “[It] could help explain the observation of declining incidence of dementia in Western societies,” Qiu told Alzforum. “Neurodegenerative pathology seems to be determined by age, but vascular pathology could be modified by reducing risk factors for cardiovascular disease,” he added.

Dementia incidence has steadily fallen by 20 to 25 percent over the past three decades in the U.S., U.K., Sweden, and the Netherlands (Jul 2013 conference news; Apr 2016 news; Sep 2017 news; Tom et al., 2020). Researchers suspect that this drop was due to better overall health—especially improvements in cardiovascular health—and higher levels of education (May 2013 news; Feb 2016 news). Might these factors stave off amyloid plaques and neurofibrillary tangles?

To find out, first author Grodstein quantified the extent of amyloid plaques, neurofibrillary tangles, Lewy bodies, TDP-43 aggregates, infarcts, and the severity of atherosclerosis and arteriosclerosis in cortical tissue from 1,554 participants in the Religious Orders Study and the Rush Memory and Aging Project (ROSMAP) cohort. About one-quarter were born in each of four periods: 1905 to 1914, 1915 to 1919, 1920 to 1924, and 1925 to 1930. All died between 1997 and 2022. Almost all participants were Caucasian; two-thirds were women.

The researchers compared the amount of each pathology, standardized by age at death, between birth cohorts. Between people born in the 1920s and those born in the 1900s, there were no differences in plaques or global AD pathology, as measured by the National Institute on Aging–Alzheimer Association criteria, The prevalence of postmortem AD diagnoses hovered between 64 and 68 percent in each birth epoch (see image below). The amount of Lewy bodies and TDP-43 inclusions remained the same across birth cohorts, as well.

Notably, people born later had higher densities of tau tangles than those born earlier. Grodstein wasn’t sure what to make of this and wanted to see if the trend held in a larger body of data, including other cohorts.

Steady AD Pathology. Over a 25-year span, the amount of global AD pathology, defined as both amyloid plaque and neurofibrillary tangle density, and amyloid plaques remained steady. Neurofibrillary tangle burden appears to tick up a tad in more recent birth cohorts. [Courtesy of Grodstein et al., JAMA Neurology, 2023.]

However, the extent of blood vessel damage, be it through athero- or arteriosclerosis, had decreased dramatically in participants born in the later cohorts. About half of those born from 1905 to 1914 had moderate to severe atherosclerosis, while only 22 percent of people born in the late 1920s, did (see image below).

Curiously though, people born in the later cohorts had more microinfarcts than their earlier cohort counterparts, with prevalence rising from 27 to 40 percent. Microinfarcts are poorly understood because they are most easily studied at autopsy, so Grodstein wasn’t sure how this increase may relate to cognition or AD pathology. The prevalence of larger infarcts, a sign of a previous stroke, remained steady.

Healthier Blood Vessels. The prevalence of atherosclerosis (left) and arteriosclerosis (right) was lower in later birth cohorts. [Courtesy of Grodstein et al., JAMA Neurology, 2023.]

All told, the researchers think that the declining prevalence of clinically diagnosed AD may be due to people having more cognitive reserve, which might be a product of better cardiovascular health and more education. “As less-prevalent AD pathology is unlikely to be the primary explanation for decreases in clinical dementia, resilience to pathology may be a more important factor,” Grodstein told Alzforum. “Understanding determinants of cognitive resilience could have a really important public health impact.”—Chelsea Weidman Burke


  1. This is the latest in a series of important papers by Julie Schneider, Dave Bennett, and the Rush team. It combines longitudinal community cohort studies with detailed pathological examination, in this case to understand the secular changes in brain pathologies of 1,554 individuals over a 25-year period.

    To me, the most striking finding is of a sharp decline in the extent of athero- and arteriosclerotic pathology, in contrast to relatively stable levels of neurodegenerative pathologies. Many Western countries are seeing a declining incidence of dementia in recent years—estimated to be 13 percent per decade over the last 25 years (Wolters et al., 2020). It is notable that, in the present study, the extent of athero- and arterio-sclerotic changes declined by 14-24 percent per decade.

    While care must be exercised when drawing causality from cross-sectional data, this is compatible with the hypothesis that improved cardiovascular risk control has led to decreased cerebrovascular disease without having major influences on Alzheimer’s and other neurodegenerative pathologies. This, in turn, is in keeping with in vivo studies that have found vascular risk factors, particularly in mid-life, relate to later-life vascular pathologies (as measured using MRI) (Lane et al., 2019) but not to fibrillar amyloid burden measured with PET (Lane et al., 2019Vemuri et al., 2017). 

    This reinforces the importance of population-based efforts to improve vascular health, and suggests that we may already be seeing dementia prevention in action. The corollary is that if Western countries (at least) are already seeing the benefit of having addressed some of the most “treatable” dementia risks, the extent to which dementia cases can be prevented in the future may be overstated—albeit there is much to be done to address vascular risks worldwide. And, if the incidence of dementia is being more driven by changes in vascular than neurodegenerative pathologies (many of which are highly heritable [Escott-Price et al., 2017]), and if the latter are responsible for the majority of cases of dementias, then while we must prevent what we can prevent, targeted treatments will still be required for the majority of people with, or at risk of, dementia.


    . Twenty-seven-year time trends in dementia incidence in Europe and the United States: The Alzheimer Cohorts Consortium. Neurology. 2020 Aug 4;95(5):e519-e531. Epub 2020 Jul 1 PubMed.

    . Associations Between Vascular Risk Across Adulthood and Brain Pathology in Late Life: Evidence From a British Birth Cohort. JAMA Neurol. 2019 Nov 4;:1-9. PubMed.

    . Age, vascular health, and Alzheimer disease biomarkers in an elderly sample. Ann Neurol. 2017 Nov;82(5):706-718. Epub 2017 Oct 26 PubMed.

    . Polygenic risk score analysis of pathologically confirmed Alzheimer disease. Ann Neurol. 2017 Aug;82(2):311-314. Epub 2017 Aug 9 PubMed.

  2. This new paper from the ROS/MAP team adds another jewel to their many prior valuable contributions regarding a better understanding of the complexity of neuropathological causes for cognitive decline among older adults in the United States. Their previous paper by Boyle et al., 2021, found that the 11 known neuropathologies they studied accounted for only 43 percent of the variation in cognitive decline in ROS/MAP, with AD pathologies accounting for only 30 to 36 percent of cognitive decline. As they concluded, “These findings highlight the complexity of cognitive aging and have important implications for the ongoing effort to develop effective therapeutics and identify novel treatment targets.”

    The key finding from this Grodstein et al. paper is that, in the ROS/MAP cohorts, secular trends across U.S. birth cohorts from 1905 through 1930 did not show significant change in AD neuropathologies at death, but did show large declines in “atherosclerosis and arteriolosclerosis in these individuals, consistent with documented national trends of decreasing vascular morbidity” (p. E5). These interesting and somewhat surprising findings suggest that declines in clinical dementia incidence and prevalence that have been reported in the U.S. and other high-income countries (and are hinted at in the ROS/MAP cohort in this paper) are most likely due to increase in “resilience factors,” such as increasing population levels of educational attainment leading to more “cognitive reserve,” and also to improved cardiovascular health that has led to “spillover” benefits in the brain.

    Regarding increasing educational attainment across the birth cohorts studied in this paper as a potential cause for decreasing incidence and prevalence of dementia in recent decades, the unique characteristics of the very well-educated ROS/MAP cohort may lead to underestimates of the impact of a secular trend in increased educational attainment on a decline in clinical dementia. The mean number of years of education in ROS/MAP remained high and relatively stable from the oldest cohort (16.2 years of education) to the youngest cohort (16.5 years); by contrast, our study of trends in dementia prevalence in the nationally representative Health and Retirement Study following individuals who were similar ages to those in the current ROS/MAP analysis, found an increase in educational attainment of nearly one full year of schooling for the 65+ U.S. population in 2000 (11.8 years) compared to the 65+ US population in 2012 (12.7 years, Langa et al., 2017).

    One additional implication of this eye-opening new study by Grodstein et al., as well as of the Boyle et al. paper noted above, is that societal investments in “resilience factors,” like expanded access to education across the life course, and cardiovascular risk factor control, may provide a bigger bang for the buck than treatments aimed narrowly at known AD neuropathologies.


    . To what degree is late life cognitive decline driven by age-related neuropathologies?. Brain. 2021 Aug 17;144(7):2166-2175. PubMed.

    . A Comparison of the Prevalence of Dementia in the United States in 2000 and 2012. JAMA Intern Med. 2017 Jan 1;177(1):51-58. PubMed.

  3. In this thought-provoking study, the authors combine data from two renowned datasets with exceptionally high autopsy rates and low loss-to-follow-up, to investigate generational changes in neuropathology at death, in an impressively large, well-delineated prospective cohort with post-mortem autopsy data. Accumulating evidence suggests that the age-standardized incidence of dementia in high-income countries may be declining, but the underlying causes remain elusive. This study provides some potentially important clues.

    A conceivable takeaway is that a clear and significant decrease in cerebral atherosclerosis occurs with later generations, which is accompanied by a less emphatic decrease in age-standardized cognitive decline and clinical dementia incidence. Global presence of AD pathology does not decrease with later generations, and cortical microinfarct and tau tangle density may actually increase. In the context of the literature, one may theorize that the age-standardized decrease in atherosclerosis with later generations, in line with decreases in age-adjusted cardiovascular mortality, may have led to lower age-standardized dementia rates. This would fit the important insights previously originating from this same research group that old-age dementia is often characterized by a mix of both cerebrovascular and traditional Alzheimer’s disease related neuropathology.

    Like many state-of-the-art studies, this investigation also raises many new questions. Might the fact that the decline in dementia incidence is less apparent than the decline in atherosclerosis be related to the potential increases in tau and cortical microinfarct burden in later generations? Has the average mix of Alzheimer’s and cerebrovascular neuropathology of individuals with clinical dementia changed with later generations? Is there a decline in amyloid burden from the second to the fourth generation? And what may be the underlying causes of the seemingly contradictive decrease in atherosclerosis and increases in these other neuropathological changes, given that they both share many (cardiovascular) risk factors?

    Ideally, the generations would have had the same age at inclusion—to exclude potential survival effects due to individuals having to be without dementia at inclusion—and the same maximum age at death, since none of the last generation’s participants died old enough to be included in the highest age-standardization group which might have influenced age adjustment. This could also provide more information on whether neuropathological profiles with clinical dementia actually change with later generations, or whether the same neuropathological profiles are only postponed to a later age. However, waiting for these data may not be an option, since it will likely be another 20-30 years until all participants in these birth cohorts have died.

    The current may represent an important steppingstone for such future endeavours.

  4. I see two main takeaways in this study:

    1. Over time there has been a marked decrease in cerebral atherosclerosis and arteriosclerosis that, surprisingly, has not translated into a reduction of ischemic lesions, Alzheimer-related pathology, or other neurodegenerative pathologies.
    2. While dementia prevalence has remained either flat or slightly declined, microinfarctions and tangle pathology have increased. This is an unexpected finding.

    The specific relationship between vascular disease—or more precisely atherosclerosis and arteriosclerosis—and AD pathology has been controversial. The surprising, and interesting, finding here is the increase of microinfarctions despite reduced conventionally defined vascular disease. Microinfarcts are more ill-defined in terms of mechanism. They could possibly reflect, or be related to, failures of the neurovascular unit, either as consequences of blood-brain barrier breakdown or failed neurovascular coupling.

    A tantalizing question that I did not see answered in this paper is whether or not there was an association in individual subjects between the increased tau pathology and microinfarcts.

    With regard to the falling clinical dementia incidence, this study does provide an argument for improved reserve capacity/resilience, the nature of which is not altogether clear. It would be interesting to see if synapse density was being better preserved in the face of increased tangle density and, if so, could provide a starting point for unraveling this.

  5. Smoking and passive smoking are known lifestyle risk factors for cardiovascular disease, cerebrovascular disease, and dementia, and smoking cessation decreases the risk (Pan et al., 2019Choi et al., 2018; Zhong et al., 2015). Tobacco smoking escalated steadily, along with higher rates of heart attack and cardiac deaths, from the early 20th century until the 1970s, when the connection between smoking and lung cancer led to a precipitous drop-off in tobacco smoking, followed several years later by a parallel reduction in heart attack rates beginning around 1980 (statins first became available around 1987). Per capita cigarette smoking and deaths from coronary artery disease each dropped roughly by half from about 1970 to 2000 (per U.S. Mortality Data, National Center for Health Statistics, CDC, USDA). It would follow that the incidence of microvascular disease in the brain and the related cognitive decline would drop off, as well.

    Another factor to consider is the rise and fall of dietary trans fat (partially hydrogenated vegetable oils, shortenings, margarines, etc.) consumption through the 20th century. It dropped off substantially after 2002, when the USDA began to require labeling of trans fat on nutrition labels in response to a connection being made between trans fat consumption and a higher risk of heart attacks (Willett et al., 1993). Trans fat has been nearly phased out of the U.S. food supply since 2018. Trans fatty acids can become incorporated into the cell membrane and affect transport through the cell membrane, the fluidity of the cell membrane, and shorten the life of the cell. There could be other dietary factors at play, as well, that could explain the decreasing incidence of dementia over the past two decades.

    Why have plaque and tangle burden not dropped off at the same time? Dozens of drugs that remove Aβ plaque (and also tangles in the case of semagacestat) have not improved cognition and, in some cases, accelerated progression of the disease (Espay et al., 2021). Maybe Aβ plaque is not the cause of Alzheimer's after all. Aβ is an antimicrobial peptide (Soscia et al., 2010) and there have been hundreds of studies suggesting a causative or at least contributory role of microbes like herpes simplex, herpes varicella-zoster, and borrelia burdorferi to Alzheimer's (Itzhaki et al., 2016). Conceivably, removing Aβ plaque would allow pathogenic microbes to thrive. However, this area of research has been largely ignored, despite the possibility that antimicrobial treatments could improve the outcome, and vaccinations for pneumococcal pneumonia, flu, and shingles have been associated with reduced risk of dementia (AAIC press release; Lophatananon et al., 2021; Scherrer et al., 2021). 

    Metabolic dysregulation is another area of Alzheimer's research that deserves more attention. The abnormal lipid deposits described along with plaques and tangles by Dr. Alois Alzheimer more than a century ago have only recently been investigated. Diabetes is a major risk factor for Alzheimer's and dementia, and glucose hypometabolism due to insulin resistance and other factors precedes Alzheimer's symptoms by at least one or two decades. Mitochondrial dysfunction is another related key pathology. PET imaging studies have shown that ketones are taken up normally in the areas of the brain affected by glucose hypometabolism, and a number of clinical trials of the ketogenic diet and/or MCT oil, some presented at AAIC over the past five years, have reported cognitive improvement (Cunnane et al., 2020). 

    Despite the lack of cognitive improvement claimed by any of the currently approved drugs for Alzheimer's, massive funding continues to flow into the study of the amyloid hypothesis. It is time to redirect more funding toward the above mentioned and other worthy but neglected areas of research.


    . Effect of smoking cessation on the risk of dementia: a longitudinal study. Ann Clin Transl Neurol. 2018 Oct;5(10):1192-1199. Epub 2018 Sep 5 PubMed.

    . Brain energy rescue: an emerging therapeutic concept for neurodegenerative disorders of ageing. Nat Rev Drug Discov. 2020 Sep;19(9):609-633. Epub 2020 Jul 24 PubMed.

    . Soluble Amyloid-β Consumption in Alzheimer's Disease. J Alzheimers Dis. 2021;82(4):1403-1415. PubMed.

    . Microbes and Alzheimer's Disease. J Alzheimers Dis. 2016;51(4):979-84. PubMed.

    . Shingles, Zostavax vaccination and risk of developing dementia: a nested case-control study-results from the UK Biobank cohort. BMJ Open. 2021 Oct 8;11(10):e045871. PubMed.

    . The relationship between smoking and stroke: A meta-analysis. Medicine (Baltimore). 2019 Mar;98(12):e14872. PubMed.

    . Impact of herpes zoster vaccination on incident dementia: A retrospective study in two patient cohorts. PLoS One. 2021;16(11):e0257405. Epub 2021 Nov 17 PubMed.

    . The Alzheimer's disease-associated amyloid beta-protein is an antimicrobial peptide. PLoS One. 2010 Mar 3;5(3):e9505. PubMed.

    . Intake of trans fatty acids and risk of coronary heart disease among women. Lancet. 1993 Mar 6;341(8845):581-5. PubMed.

    . Smoking is associated with an increased risk of dementia: a meta-analysis of prospective cohort studies with investigation of potential effect modifiers. PLoS One. 2015;10(3):e0118333. Epub 2015 Mar 12 PubMed.

    View all comments by Mary Newport

Make a Comment

To make a comment you must login or register.


News Citations

  1. Dementia Prevalence Falls in England
  2. Dementia Incidence in Britain Dropped, Mostly in Men
  3. More Evidence that Dementia Case Numbers Are Falling
  4. Dementia Incidence Said to Drop as Public Health Improves
  5. Falling Dementia Rates in U.S. and Europe Sharpen Focus on Lifestyle

Paper Citations

  1. . Association of Demographic and Early-Life Socioeconomic Factors by Birth Cohort With Dementia Incidence Among US Adults Born Between 1893 and 1949. JAMA Netw Open. 2020 Jul 1;3(7):e2011094. PubMed.

Further Reading


  1. . Tracing temporal trends in dementia incidence over 25 years in central Stockholm, Sweden. Alzheimers Dement. 2020 May;16(5):770-778. Epub 2020 Apr 9 PubMed.
  2. . A major role for cardiovascular burden in age-related cognitive decline. Nat Rev Cardiol. 2015 May;12(5):267-77. Epub 2015 Jan 13 PubMed.

Primary Papers

  1. . Trends in Postmortem Neurodegenerative and Cerebrovascular Neuropathologies Over 25 Years. JAMA Neurol. 2023 Apr 1;80(4):370-376. PubMed.