Dementia is a growing global challenge, with Alzheimer’s disease (AD) being the most common cause, characterized by ADNCs, encompassing brain deposits of amyloid-β (Aβ) plaque and neurofibrillary tau tangles. The prevalence of dementia and mild cognitive impairment (MCI) is well established2,3, but the prevalence of ADNCs in general populations remains uncertain. With the advent of drugs capable of reducing Aβ plaque pathology and slowing cognitive decline4,5, accurate knowledge of ADNC prevalence is essential for anticipating the number of individuals eligible for treatment and estimating future health-care demands and associated costs. A recent review has reported an overall prevalence of 22% ADNCs in all people 50 years of age and older globally2. However, studies examining the prevalence of ADNCs are typically enriched, including relatively small clinic-based samples, which tend to differ regarding important clinical and demographic features compared with general populations. Such studies may thus report inflated or deflated rates of AD pathology.
Until recently, ADNCs could only be verified in vivo using cerebrospinal fluid analysis or molecular positron emission tomography (PET), substantially hindering its evaluation in large population-based studies. Minimally invasive blood-based markers, particularly plasma phosphorylated tau at threonine 217 (pTau217), that have high accuracy for ADNCs have recently become available but have not yet been used in large community-based studies6. In this study, we capitalized on the large Norwegian population-based Trøndelag Health (HUNT) study7,8, with 11,486 blood samples of participants 58 years of age and older, to explore the following research questions: (1) what the prevalence of ADNCs in the population 58 years of age and older across age and sex groups is; (2) what the association between ADNCs and demographics, cognition, educational level, apolipoprotein E (APOE) ε2, ε3 or ε4 status and comorbidities is; and (3) what proportion of those 70 years of age or older is eligible for disease-modifying therapies (DMTs) according to current recommendations9,10.
The HUNT study has been ongoing for four decades, with a new wave taking place in the same population every 10 years, thus four waves exist so far. In this nested cross-sectional study, we included 2,537 individuals from HUNT3 (age range of 58–69.9 years, 51.2% women) and 8,949 from HUNT4 (age range of 70 years and older (hereafter the 70+ group), 53.6% women). Subsequent diagnostic history was considered when deciding who to approach for inclusion in HUNT3. Although a blood sample was provided in both surveys, the HUNT4 70+ cohort also underwent a standardized clinical assessment for a diagnosis of dementia and MCI11 (Extended Data Fig. 1 and Supplementary Information, ‘Assessment of cognition, physical performance, anxiety, depression, neuropsychiatric symptoms and activities of daily living’). The presence of ADNCs was established by measuring plasma pTau217 levels with a previously validated commercial kit (ALZpath p-Tau 217 Advantage PLUS, Quanterix)1. We used a two cut-off approach as recommended by the Global CEO Initiative on Alzheimer’s Disease12 to categorize individuals as ADNC negative (less than 0.40 pg ml−1), intermediate or positive (0.63 pg ml−1 or more), as previously described1. The agreement between elevated plasma pTau217 concentration and the presence of notable amounts of plaques and tangles at post-mortem examination has been previously found to be very high13. For terminological clarity, in the remainder of this article, the term ‘ADNC’ refers specifically to the presence of elevated plasma pTau217 concentration, used as a surrogate marker for ADNCs.
The demographic and clinical characteristics of the cohort are shown in Extended Data Table 1. The estimated proportion of people with and without ADNCs in different age groups is shown in Fig. 1 and Extended Data Table 2. There was a stepwise increase in the proportion of ADNCs across age groups; the proportion was 33.4% in the 70+ group. There was a significant association between ADNCs and cognitive diagnosis. Among individuals with dementia, 60% had ADNCs (that is, AD dementia), compared with 32.6% of those with MCI (that is, prodromal AD) and 23.5% in the cognitively unimpaired group (that is, preclinical AD). The proportion with ADNCs increased with age in each of the cognitive groups. ADNCs were ruled out based on plasma pTau217 concentrations being below the lower cut-off in 19.4% of the dementia group, 41% of the MCI group and 50.1% of the cognitively unimpaired group (Fig. 2 and Extended Data Table 3). Depending on age, 13.5–27.6% of participants had plasma pTau217 concentrations in the intermediate range, with only minor differences between the cognitively unimpaired, MCI and dementia groups (Extended Data Tables 2 and 3). Weighted estimated proportions of ADNCs in the respective clinical cognitive subgroups are shown in Extended Data Table 4.
Fig. 1: Plasma pTau217 concentrations in different age groups. Individual dots represent plasma pTau217 concentrations (n = 2,537 participants from HUNT3 and n = 8,949 participants from HUNT4 70+). Percentages (95% confidence interval) are estimates of how many in each age group have AD neuropathology, defined by plasma pTau217 concentration of 0.63 pg ml−1 or more. The lower cut-off of 0.40 pg ml−1 is also shown. The horizontal line in each box represents the median, and bottom and top edges delineate the second and third quartiles. The bottom whisker represents the first quartile, and the top whisker denotes the fourth quartile. Concentrations above 3 pg ml−1 are not shown. Full size image
Fig. 2: Plasma pTau217 concentrations in people 70 years of age or older who are cognitively unimpaired, have MCI or have dementia. Percentages (with 95% confidence intervals; colour-coded to match the box plots) are estimates of how many in each cognitive group have AD neuropathology, defined by plasma pTau217 concentration ≥ 0.63 pg ml−1. The lower cut-off of 0.40 pg ml−1 is also shown. n = 8,949 participants from HUNT4 70+. The horizontal line in each box represents the median, and top and bottom edges delineate the second and third quartiles. The bottom whisker represents the first quartile, and the top whisker denotes the fourth quartile. CU, cognitively unimpaired. Full size image
The estimated prevalence of preclinical AD, prodromal AD and AD dementia in the 70+ study population is shown in Fig. 3 and Extended Data Table 5. The estimated prevalence of AD dementia consistently increased with age, whereas the prevalence of preclinical AD increased from the 70–74 year age group to the 80–84 year age group before decreasing in the oldest old (those aged 85 years or older (the 85+ group)). The prevalence of prodromal AD remained stable after 80 years of age (Extended Data Fig. 2).
Fig. 3: Proportions and frequencies of ADNCs across the AD continuum in the 70+ population. Left, the percentage of participants with ADNCs, defined as plasma pTau217 concentration ≥ 0.63 pg ml−1. Stacked bars represent the estimated proportions of ADNCs. Right, absolute numbers of study participants with ADNCs. Colours represent different levels of cognitive effects. The values are stratified, on the x axis, by sex. The numbers displayed at the top of the graphs are age groups in years. Full size image
In the 80–89 year age group, men had a slightly higher estimated prevalence of ADNCs than women, mainly due to a higher prevalence of early-stage AD (preclinical and prodromal AD), although cognitive subgroup differences were not statistically significant. There was no sex difference in the estimated prevalence of AD dementia in any of the age groups. Details can be found in Extended Data Table 5 and Extended Data Fig. 2.
Estimated ADNC prevalence was higher among individuals with one (46.4%) or two (64.6%) APOE ε4 alleles than in those with none (27.1%; Extended Data Table 6). Estimated glomerular filtration rate (eGFR) was inversely associated with plasma pTau217 concentration in individuals with eGFR < 51 ml min−1 per 1.73 m2 (Extended Data Fig. 3). There was no significant association between ADNCs and self-reported cardiovascular and cerebrovascular disease, chronic obstructive pulmonary disease, diabetes, cancer, migraine, psoriasis, kidney disease, rheumatoid arthritis and gout, when adjusting for age, sex, APOE ε4 allele count, cognition, serum creatinine levels and education level (Extended Data Table 7).
... continue reading