Article Text
PDF
Statistics from Altmetric.com
Question
Question
Are plasma β-amyloid (Aβ) levels associated with cognitive decline among older people without dementia?
People
Before exclusions, 3075 community-dwelling adults aged 70–79 years were recruited to the Health Aging and Body Composition (Health ABC study) in 1997–1998. They were selected from a random sample of people eligible for Medicare who were living in the study area. A subgroup of 997 adults with repeated cognitive testing and measurements of Aβ 40 and 42 was selected for analysis. Eligibility criteria: no difficulties with activities of daily living, able to walk a quarter mile or climb 10 steps without resting, free from life-threatening cancer and planning to remain within the study area for at least 3 years. Those with measurements were more likely to be female (55.2% of the study sample) of Black ethnicity (54.0%) and to have lower mean educational level compared with the rest of the cohort.
Setting
Community setting – Memphis, Tennessee, USA and Pittsburgh, Pennsylvania, USA – recruitment between 1997 and 1998.
Risk factors
Plasma levels of Aβ 42 and Aβ 40 levels. Plasma samples were stored at the first follow-up assessment (median 53.4 weeks from baseline).
Outcomes
Cognitive function and incidence of dementia. The Modified Mini-Mental State Examination (3MS) was conducted at baseline, year 3 (n=932), year 5 (n=854), year 8 (n=642) and year 10 (n=568). Incident dementia was further determined by hospital records and use of cholinesterase inhibitors and memantine. Cognitive reserve was determined by self-reported years of education (less than high school diploma or at least up to high school diploma) and literacy (measured by the Rapid Estimate of Adult Literacy in Medicine: at least sixth grade (score ≤44) or higher than sixth grade (score ≥45)). Genetic vulnerability or reserve was determined by assessment of apolipoprotein E (APOE) e4 allele status.
Methods
Design
Prospective cohort study.
Follow-up period
Average 9 years (following baseline assessment to final follow-up 2006–2007).
Main results
Lower baseline Aβ 42/40 ratio was associated with greater cognitive decline during follow-up. In models adjusted for age, ethnicity, education, diabetes, smoking and APOE e4 status, mean decline in 3MS was 6.38 points (95% CI −5.15 to −7.61) for those in the lowest tertile versus 6.09 decline for those in the middle tertile (95% CI −4.89 to −7.29) versus 3.44 decline for those in the highest tertile (95% CI −2.21 to −4.67) (p=0.02 for between groups difference). A similar association was observed between Aβ 42 level and 3MS score decline: in fully adjusted models, mean decline was 6.70 (95% CI −5.45 to −7.95) for those in the lowest tertile versus 5.03 decline for those in the middle tertile (95% CI −3.81 to −6.25) versus 4.29 decline for those in the highest tertile (95% CI −3.06 to −5.52). Interactions were modified by measures of cognitive reserve: those in the lowest Aβ 42/40 tertile with less than a high school diploma had 8.94 score decline (95% CI −6.94 to −10.94) compared with those with at least a high school diploma who had a 4.60 score decline (95% CI −3.07 to −6.13); respective declines according to education were 4.45 (95% CI −2.31 to −6.59) and 2.88 (95% CI −1.41 to −4.35) in the highest tertile.
Conclusions
Older people with lower levels of Aβ 42/40 and Aβ 40 have a moderately greater rate of cognitive decline over 9 years compared with those with higher levels. This relationship is modified by measures of cognitive reserve.
Abstracted from
Commentary
Alzheimer's disease (AD) may be easily misdiagnosed by reliance on clinical evaluations alone, which creates a need for biomarkers that increase the accuracy of diagnoses. Cerebrospinal fluid (CSF) levels of β-amyloid (Aβ), tτ and pτ – either alone or preferably in combination with Aβ PET amyloid tracers – hold promise as AD biomarkers. However, a number of technical issues, in addition to patients' apprehension and cost, has prevented these procedures from being implemented for routine medical use.1 In contrast, all that is needed to measure plasma Aβ, as in the study by Yaffe and colleagues, is a blood draw.
So far, results from studies that have attempted to use plasma Aβ as biomarker of cognitive decline and conversion to AD have, in general, been either negative or inconsistent while showing no correlation between plasma Aβ and soluble Aβ CSF levels or with brain amyloid deposits. Nevertheless, this line of research continues.
The strengths of the present study are the relatively large sample size (997), the long follow-up (9 years) and the ethnic variety of the sample. The results showed that low plasma Aβ42/Aβ40 ratio at baseline correlated with greater cognitive decline during longitudinal follow-up, especially in individuals with a lower level of education. Thus, Yaffe and colleagues concluded that a greater brain reserve might be protective against the deleterious effects of increased brain amyloid.
The main concerns with this study are: 1) given the advanced age of their sample, who were aged 80 years by the time of the last evaluation, the difference in drop across groups was very small, being only ∼3 points on a 100-point MMSE scale between the highest- and lower-risk groups; and 2) the plasma Aβ determination was limited to a single time point at baseline, in a study that extended over nine years; finally, it is not clear whether this marker predicts conversion to AD in this sample. These points limit the clinical and theoretical significance of this study.
Reference
Footnotes
-
Source of funding National Institute on Aging.
Footnotes
-
Competing interests None.