Changes in Daily Functioning in Association With Tau and Amyloid Among Unimpaired Older Adults With and Without Elevated Amyloid

Abstract

Background and Objectives

Everyday functioning declines gradually over time in Alzheimer disease (AD), with the earliest changes potentially occurring at the preclinical stage. We investigated how changes in everyday functioning relate to (changes in) amyloid and tau in a large sample of cognitively unimpaired older adults, most of whom had elevated amyloid levels at the start of the study.

Methods

This prospective study included participants from a 240-week randomized controlled trial of an anti-amyloid drug, solanezumab, and individuals who screen-failed because of a negative amyloid PET scan. A subset (n = 434) underwent repeated tau PET scans. Participants and their study partners completed the Alzheimer’s Disease Cooperative Study Activities of Daily Living Prevention Instrument multiple times during the double-blind phase of the trial. Using generalized least-squares models, fit by restricted maximum likelihood, we analyzed how changes in everyday functioning related to amyloid and tau PET. We also correlated changes in amyloid and tau PET with changes in everyday functioning.

Results

A total of 1,707 participants (71.5 ± 4.7 years, 60% female) showed a marginal decline in everyday functioning. Among individuals with elevated amyloid, those with the highest levels of neocortical and medial temporal tau showed the largest decline in everyday functioning, as reported by the participants and their study partners. Increases in neocortical and medial temporal tau correlated moderately (correlation coefficient ranging from −0.2 to −0.5) with a decline in ADCS ADL-PI scores. Functional changes were not evident with amyloid alone. At the item level, participants and their study partners were most likely to report increased difficulty at the last visit with completing complex activities and selecting and paying for items when shopping.

Discussion

Higher tau levels are associated with the fastest decline in everyday functioning in the presence of elevated amyloid, and those accumulating more tau show a faster decline in everyday functioning. These findings demonstrate the utility of including sensitive measures of everyday functioning in clinical practice and clinical trials at the stage of preclinical AD.

Trial Registration Information

The A4 study, ClinicalTrials.gov ID: NCT02008357; and Longitudinal Evaluation of Amyloid Risk and Neurodegeneration, ClinicalTrials.gov ID: NCT02488720.

Introduction

The preclinical stage of Alzheimer disease (AD) is defined by the presence of elevated levels of cerebral amyloid in the absence of objective cognitive impairment.¹ Stemming from the belief that therapeutic strategies may be optimal before the onset of cognitive impairment, many studies now focus on preclinical AD. While there is, by definition, no cognitive or functional impairment in the preclinical stage, research has shown that subtle changes in cognition can be captured, even in early-stage AD, with the proper instruments.^2,3

The performance of cognitively complex “instrumental activities of daily living” (IADL), such as managing finances or preparing meals, relies on intact cognitive functioning. Therefore, IADL performance may start to change along with cognition in the preclinical stage, rendering it a potentially useful measure of cognitive changes in daily life.⁴ Indeed, IADL constitute a clinically meaningful outcome because they relate to a person's independent everyday functioning. Consequently, IADL performance often serves as an outcome measure in clinical trials in preclinical AD, albeit usually secondary to cognitive performance.

Several studies have shown that amyloid-positive but cognitively unimpaired individuals seem to have somewhat more difficulty performing IADL compared with their amyloid-negative counterparts.^5,6 A longitudinal decline in IADL among cognitively unimpaired older adults with elevated amyloid and tau has also been found.⁷ However, the association between amyloid alone and functional changes is unclear because not all studies report such an association.⁸ This may suggest either that the effect of amyloid alone is insufficient or that study design factors such as the sample size and characteristics, follow-up duration, and selected IADL measures play a role in demonstrating the association.

Recently, the first large secondary prevention trial in preclinical AD, the Anti-Amyloid Treatment in Asymptomatic Alzheimer's Disease (A4) study, published its results.⁹ The A4 study was designed to investigate the effects of solanezumab on cognitive change in preclinical AD¹⁰ and was accompanied by the observational Longitudinal Evaluation of Amyloid Risk and Neurodegeneration (LEARN) study, which used the same cognitive and functional assessments and visit schedule as the A4 study but included only individuals who did not have elevated amyloid and did not receive an intervention (solanezumab). Solanezumab was unsuccessful in slowing cognitive decline in the A4 study.⁹ However, the data from the A4 and LEARN cohorts can still be analyzed to investigate AD-related changes in cognition. Several studies using the longitudinal data from A4 and LEARN have shown how amyloid and tau may predict a decline in performance on cognitive tests^11-13 and self-perceived cognitive functioning.¹⁴

In this study, we sought to investigate the changes over time in daily functioning, as measured by participants and their study partners. We had the following aims: (1) to characterize daily functioning changes over time in the active and placebo groups of A4 and LEARN; (2) to relate daily functioning changes to longitudinal amyloid and tau measures; and (3) to identify specific activities that showed more change over time than others.

Methods

Participants and Procedures

The procedures of the A4 study have been previously described in detail.⁹ In brief, A4 was a 240-week double-blind, placebo-controlled phase 3 trial of an anti-amyloid monoclonal antibody, solanezumab. The A4 study recruited participants from 67 sites across North America, Japan, and Australia. Inclusion criteria were age 65–85 years and being cognitively unimpaired, defined as having a Mini-Mental State Examination score of 25–30, a Clinical Dementia Rating global score of 0, and a Logical Memory delayed recall score of 6–18. For inclusion in the A4 study, participants had to be amyloid positive as determined using [¹⁸F]-florbetapir PET scans. The companion study, LEARN, included participants at select A4 study sites who screen-failed for the A4 study based on their amyloid PET scan but would otherwise be eligible for inclusion in A4.

Standard Protocol Approvals, Registrations, and Patient Consents

The protocols of the A4 and LEARN studies were approved by an institutional review board or ethics committee at each site. All participants and their study partners provided written informed consent before data collection, which included consent for data sharing. Both studies are registered on ClinicalTrials.gov under IDs NCT02008357 (A4) and NCT02488720 (LEARN).

Materials

Daily Functioning

Participants and their study partners concurrently completed the Alzheimer’s Disease Cooperative Study Activities of Daily Living Prevention Instrument (ADCS ADL-PI) to assess IADL performance.¹⁵ The ADCS ADL-PI was administered at baseline and at weeks 48, 108, 168, 204, and 240, with the timing of some visits shifting because of the impact of the COVID-19 pandemic.⁹ Each item was rated on a 4-point scale, ranging from “did not do the activity” (0) to “did the activity as well as usual, with no difficulty” (3). The scores of the 15 items are summed (range 0–45), with higher scores representing better everyday functioning. We dichotomized item scores as “performed the activity without difficulty” (item score of 3) or “had some level of difficulty performing the activity” (item score of ≤2).

PET Imaging

As previously described,⁹ all participants underwent brain amyloid PET scans using the [¹⁸F]-florbetapir tracer before inclusion and at 240 weeks. A global cortical standardized uptake volume ratio (SUVr) was calculated with the cerebellar gray as a reference region. Amyloid SUVr was additionally converted to the centiloid (CL) scale for comparability with other cohorts. Tau PET scans were available in a subset of participants (n = 434 at baseline) using the [¹⁸F]-flortaucipir tracer, acquired at baseline and at 240 weeks. We selected 2 aggregates of regions of interest, in line with our earlier work with the A4/LEARN baseline data¹⁶: the neocortex (comprising the inferior temporal, fusiform, middle temporal, and inferior parietal regions) and medial temporal lobe (consisting of the amygdala, entorhinal, and parahippocampal regions).

Statistical Analyses

All analyses were performed with R version 4.4.0. We used generalized least-squares models, fit by restricted maximum likelihood and assuming heterogeneous Toeplitz covariance, to model ADCS ADL-PI scores from participant self-reports and study partner reports. All models were adjusted for age, education, APOE ε4 carrier status, baseline amyloid, group, and time by group. For these analyses, the LEARN sample was considered a separate group in addition to the randomized groups of the A4 study. Time was modeled using natural cubic splines with 2 or 3 degrees of freedom, as guided by the Akaike Information Criterion. The models did not include separate intercepts for the A4 randomized groups, which constrained the treatment and placebo groups to have the same mean at baseline. Within the A4 study sample, amyloid PET levels were split into tertiles (low: 46.1 centiloids; mid: 46.1–77.2 centiloids; high: >77.2 centiloids). Similarly, neocortical tau (low: <1.13 SUVr; mid: 1.13–1.20 SUVr; high: >1.20 SUVr) and medial temporal tau (low: <1.13 SUVr; mid: 1.13–1.24 SUVr; high: >1.24 SUVr) were also split into tertiles within the A4 study sample. The LEARN sample was considered a separate group, with amyloid and tau burden lower than the lowest tertiles from A4.

For analyses of change vs change, we computed the changes in amyloid, tau, and ADCS ADL-PI scores by subtracting the baseline value from the endpoint (240-week) value so that higher values for amyloid and tau represent an increase in amyloid or tau deposition and that lower (more negative) values represent a decline in everyday functioning. We subsequently correlated the change scores with each other using the Pearson correlation coefficient.

Last, we used a longitudinal binomial logistic regression of dichotomized item scores (3 vs <3) to estimate the odds of endorsing difficulty at the 240-week follow-up visit compared with baseline within the A4 study sample only. Months from baseline, age, APOE ε4 carrier status, and education were included as covariates. The models were fit by generalized estimating equations. We report odds ratios and accompanying 95% CIs.

Data Availability

The data used in this study are available through a4studydata.org/after registration and approval of a data access request submitted by any qualified researcher.

Results

We included 1,707 participants, of whom 564 (33%) received solanezumab and 583 (34%) received the placebo. The remaining 560 (33%) were participants in LEARN. The baseline demographic characteristics are summarized in the Table. The LEARN participants had significantly lower amyloid than the solanezumab and placebo participants by design. LEARN participants were also younger and less often APOE ε4 carriers than participants from both A4 study arms.

Table.

Baseline Demographic Characteristics

	Whole sample	Solanezumab	Placebo	Learn
N (%)	1,707 (100.0)	564 (33.0)	583 (34.2)	560 (32.8)
Age	71.5 ± 4.7	72.0 ± 4.7	71.9 ± 5.0	70.5 ± 4.3
Female, n (%)	1,024 (60.0)	329 (58.3)	352 (60.4)	343 (61.2)
Education in y	16.6 ± 2.8	16.6 ± 2.7	16.6 ± 2.9	16.8 ± 2.7
Race, n (%)
American Indian/Alaskan Native	6 (0.4)	1 (0.2)	0 (0.0)	5 (0.9)
Asian	36 (2.1)	11 (2.0)	13 (2.2)	12 (2.1)
Black/African American	42 (2.5)	12 (2.1)	15 (2.6)	15 (2.7)
White	1,601 (93.8)	531 (94.1)	549 (94.2)	521 (93.0)
More than one race	14 (0.8)	5 (0.9)	3 (0.5)	6 (1.1)
Not reported	8 (0.5)	4 (0.7)	3 (0.5)	1 (0.2)
Ethnicity, n (%)
Hispanic/Latino	52 (3.0)	16 (2.8)	18 (3.1)	18 (3.2)
Non-Hispanic/Latino	1,640 (96.1)	542 (96.1)	560 (96.1)	538 (96.1)
Not reported	15 (0.9)	6 (1.1)	5 (0.9)	4 (0.7)
MMSEa	28.9 ± 1.2	28.8 ± 1.3	28.8 ± 1.2	29.0 ± 1.2
ADCS ADL-PIb
Self-report	43.4 ± 2.5	43.3 ± 2.6	43.1 ± 2.7	43.7 ± 2.0
Study partner report	43.6 ± 2.4	43.4 ± 2.7	43.5 ± 2.6	44.0 ± 1.9
Amyloid PET
SUVr	1.2 ± 0.2	1.3 ± 0.2	1.3 ± 0.2	1.0 ± 0.1
Centiloid	45.8 ± 40.2	66.2 ± 33.5	65.9 ± 32.1	4.3 ± 12.6
Tau PET
Available, n (%)	434 (25.4)	187 (33.2)	193 (33.1)	54 (9.6)
Neocortical SUVr	1.2 ± 0.1	1.2 ± 0.1	1.2 ± 0.1	1.1 ± 0.1
Medial temporal SUVr	1.2 ± 0.1	1.2 ± 0.1	1.2 ± 0.2	1.1 ± 0.1
APOE ε4 carriership, n (%)
Noncarrier	901 (52.8)	231 (41.0)	241 (41.3)	429 (76.6)
Heterozygous	709 (41.5)	286 (50.7)	295 (50.6)	128 (22.9)
Homozygous	97 (5.7)	47 (8.3)	47 (8.1)	3 (0.5)

Abbreviations: ADCS ADL-PI = Alzheimer's Disease Cooperative Study Activities of Daily Living Prevention Instrument; LEARN = Longitudinal Evaluation of Amyloid Risk and Neurodegeneration; MMSE = Mini-Mental State Examination; SUVr = standardized uptake value ratio.

Data are shown as mean ± SD unless otherwise specified.

^aData missing for n = 7 (0.4%), all in LEARN.

^bData missing for n = 1 (<0.1%), in the solanezumab group.

Tau PET was available for 434 participants (25%). The subgroup of LEARN participants with available tau PET did not differ in demographic characteristics from those without tau PET (data shown in eTable 1). The subgroup of A4 participants with available tau PET had fewer years of education (mean 16.2 vs 16.8 years, p < 0.001), less amyloid PET burden (63.2 vs 67.4 CL, p = 0.039), and a smaller proportion of White individuals (91.1% vs 95.7% White, p < 0.001). eTable 2 lists the demographic characteristics of the tau subsample.

Change Over Time

In the solanezumab and placebo groups, self-reported everyday functioning declined marginally over time by approximately 1.5 points: after 4.6 years (240 weeks), the mean self-reported ADCS ADL-PI score was 41.6 (95% CI [41.3–42.0]) in the solanezumab group and 41.9 (95% CI = [41.5–42.2]) in the placebo group. Among LEARN participants, the decline was even more subtle (less than a point), with the mean score at 240 weeks being 42.8 (95% CI [42.4–43.1]). We observed a similar pattern in study partner–reported everyday functioning: at 240 weeks, the mean study partner–reported ADCS ADL-PI score was 41.2 (95% CI [40.8–41.6]) in the solanezumab group, 41.8 (95% CI [41.4–42.2]) in the placebo group, and 43.2 (95% CI [42.8–43.6]) in LEARN. Figure 1 shows the change over time in both self-reported and study partner–reported ADCS ADL-PI scores in the 3 groups.

Figure 1. Change Over Time in Everyday Functioning, as Reported by the Participants (Panel A) and Their Study Partners (Panel B).

Estimates are separated into 3 groups: solanezumab (red), placebo (grey), and LEARN (yellow).

Change by Amyloid and Tau

When change over time in everyday functioning was stratified by baseline amyloid PET tertiles and LEARN, we observed a separation between the lower tertile and LEARN participants on the one hand and the middle and higher tertiles on the other, where those with more amyloid showed a steeper decline in self-reported ADCS ADL-PI scores. Estimates for the tertile groups of A4 participants represent the trajectory of the participants in the placebo group. At week 240, the estimated self-reported ADCS ADL-PI score in LEARN had decreased by 0.72 (95% CI [0.50–0.93]) from baseline, whereas it decreased by 2.00 (95% CI [1.48–2.53]) in the high amyloid tertile group. In study partner–reported ADCS ADL-PI scores, a stronger separation was visible between all 3 tertiles among individuals with elevated amyloid (estimated decline from baseline to week 240, low: 0.42, 95% CI [0.06–0.89]; medium: 1.51, 95% CI [0.85–2.16]; high: 3.07, 95% CI [2.35–3.79]), where the lower tertile overlapped with the LEARN participants (estimate: 0.58, 95% CI [0.35–0.82]).

The decline in self-reported ADCS ADL-PI scores seemed more pronounced in relation to both neocortical and medial temporal tau, where participants in the highest tertiles showed the most rapid decline (240-week estimated decline for the high tertile, neocortical: 2.70, 95% CI [1.55–3.85]; medial temporal: 3.23, 95% CI [2.27–4.19]), distinguishably faster than participants in the middle and lower tertiles (estimated decline for the lowest tertile, neocortical: 0.85, 95% CI [0.21–1.50]; medial temporal: 0.83, 95% CI [0.05–1.72]). A similar pattern was visible for study partner–reported ADCS ADL-PI scores. Figure 2 displays the trajectories of change in ADCS ADL-PI scores by the amyloid and tau tertiles. Interactions between amyloid and neocortical tau (visualized in eFigure 1) and between amyloid and medial temporal tau (visualized in eFigure 2) revealed that the effects of tau were most evident among participants with the highest amyloid burden.

Figure 2. Change in Self-Reported (A) and Study Partner–Reported (B) ADCS ADL-PI Total Scores Over Time by Amyloid (n = 1,707; Top), Neocortical Tau (n = 434; Middle), and Medial Temporal Tau (n = 434; Bottom) Tertiles.

Trends for the A4 tertiles show the placebo group effects. ADCS ADL-PI = Alzheimer's Disease Cooperative Study Activities of Daily Living Prevention Instrument; LEARN = Longitudinal Evaluation of Amyloid Risk and Neurodegeneration; SUVr = standardized uptake value ratio.

Change vs Change

We found no correlation between the change in amyloid and the change in self-reported ADCS ADL-PI scores. There was a modest correlation (r = −0.1) between the change in amyloid and the change in study partner–reported ADCS ADL-PI scores in participants who received solanezumab and placebo (p = 0.009 and p = 0.003, respectively).

Among individuals with elevated amyloid (in the A4 study), changes in self-reported ADCS ADL-PI scores correlated modestly with changes in neocortical tau (r = −0.2, p = 0.002, in participants who received solanezumab and r = −0.3, p < 0.001, in participants who received placebo). We did not find a correlation between changes in self-reported ADCS ADL-PI scores and changes in medial temporal tau in these same participants. The correlations between changes in study partner–reported ADCS ADL-PI scores and changes in neocortical tau were stronger than with self-reported ADCS ADL-PI scores (r = −0.4, p < 0.001, in participants receiving solanezumab and r = −0.5, p < 0.001, in participants receiving the placebo). Changes in medial temporal tau were only correlated with changes in study partner–reported ADCS ADL-PI scores in the placebo group (r = −0.2, p = 0.003) but not in participants receiving solanezumab (r = −0.1, p = 0.100). We did not observe any correlations between changes in tau and changes in ADCS ADL-PI scores among participants who did not have elevated amyloid (in LEARN; all p > 0.05). All correlations are shown in eFigure 3 (amyloid), eFigure 4 (neocortical tau), and eFigure 5 (medial temporal tau).

Item-Level Analyses

Item endorsement by participants and their study partners over the duration of the study is displayed in eFigure 6. A4 study participants were more likely to self-report difficulties performing various activities by the 240-week follow-up visit, most notably in paying when shopping (odds ratio [OR] = 5.72, 95% CI [3.27–10.01], p < 0.001) and using appliances or performing household repairs (OR = 3.75, 95% CI [2.16–6.52], p < 0.001) compared with baseline. Study partners generally reported numerically similar but slightly larger increased likelihoods in the same activities as participants themselves, except for paying when shopping (OR = 3.78, 95% CI [2.54–5.62], p < 0.001). The odds ratios of participants and study partners reporting difficulty performing each activity are shown in Figure 3.

Figure 3. Odds Ratios for Endorsing Difficulty Performing Each Activity at the 240-Week Follow-Up Visit, Compared With Baseline.

Items were dichotomized, and the data shown are the odds ratios for endorsing an item response indicating any level of difficulty (item score <3) at the 240-week follow-up visit, compared with baseline. Data are shown as odds ratio (95% CI). OR = odds ratio.

Discussion

In this study, we demonstrated that among initially cognitively unimpaired individuals participating in the A4/LEARN studies, daily functioning declines slowly over time, as reported by the individuals themselves and their study partners. Within the group of amyloid-positive individuals, those with higher baseline levels of amyloid and tau showed a steeper decline in everyday functioning than those with relatively lower levels. Changes in neocortical and medial temporal tau, more so than in amyloid alone, were related to changes in everyday functioning but only among individuals with both elevated amyloid and tau.

In this study, we had the opportunity to delve deeper into the longitudinal assessment of cognitively complex everyday functioning in the A4 and LEARN studies. Previous publications have already shown that functional decline is more prominent among those with higher levels of amyloid and tau.^11,12 In this study, we found that self-reported everyday functioning did not change faster over time in those who were amyloid-positive with a centiloid value of less than 46.1 (representing the lower third of amyloid-positive participants) compared with amyloid-negative participants. The same was true for changes in study partner–reported everyday functioning. On the contrary, participants with a middle or high level of amyloid (i.e., more than 46.1 centiloids) showed significantly more self-reported and study partner–reported decline than those with lower levels of amyloid. While having a middle vs high level of amyloid did not seem to matter much for the observed rate of decline in self-reported everyday functioning, study partner–reported functioning declined even more substantially among those with the highest levels of amyloid (more than 77.2 centiloids). Thus, the impact of amyloid on daily functioning seems strongest at the highest levels of amyloid burden. By contrast, having a relatively low—but nonetheless elevated—amount of amyloid seems to affect daily functioning minimally within cognitively unimpaired individuals.

The cross-sectional relationship between amyloid and cognition is controversial,¹⁷ partly because not all studies unequivocally show that amyloid relates to cognitive performance in cognitively unimpaired individuals. This debate may also be extended to daily functioning: there is some evidence that amyloid levels relate to changes in daily functioning, even among older adults who are cognitively unimpaired,^18-20 yet not all studies have shown such an association.⁸ The findings from this study suggest that the detrimental effect of amyloid may not arise until the amyloid burden passes a certain threshold. Furthermore, it is possible that elevated amyloid by itself is insufficient to cause a detectable change in daily functioning.

It has been suggested previously that tau is more strongly related to clinical outcomes. Several studies have shown associations between functional change and tau burden, particularly in the temporal lobe.^8,21 At the same time, not all studies have shown such associations.²² In this study, we found that participants with higher medial temporal or neocortical tau levels showed a more pronounced decline in self-reported and study partner–reported daily functioning. Furthermore, even participants with a relatively lower tau burden showed substantially more decline, especially on the study partner report, compared with the change shown in relation to amyloid burden.

Interactions between amyloid and tau seem to exist in the relationship with IADL.^{8,16,21,23,24} In this sample, we observed that the differences between the lower, middle, and higher levels of neocortical and medial temporal tau in rates of decline were more pronounced among those with the highest levels of amyloid burden. This supports the hypothesis that amyloid by itself may not lead to functional changes but that these changes only emerge when the amyloid burden is heightened and accompanied by temporal tau.

We also noted that changes in study partner–reported everyday functioning were moderately associated with changes in neocortical tau. Over the four-and-a-half-year observation period from the baseline to the closeout visit, those who accumulated more neocortical tau showed a larger study partner–reported decline in everyday functioning. Evidence on how longitudinal changes in tau relate to changes in everyday functioning, to date, is scarce.

We reported several differences between how self-reported and how study partner–reported daily functioning changed over time in relation to AD biomarkers. In the whole sample, the change in functioning reported by participants was similar to that reported by their study partners. However, when stratified by the level of amyloid, it seemed that participants with the highest levels of amyloid were less perceptive of their decline than their study partners. Of course, we do not know who is “right” about the amount of change in IADL that objectively occurred. While the study partner's report has traditionally been considered the gold standard, it seems logical that a cognitively unimpaired person can adequately reflect on their everyday functioning. At the same time, a lack of self-awareness about functional difficulty may emerge even among cognitively unimpaired individuals. In an analysis of a questionnaire about cognitive concerns in this same study sample, others¹⁴ found a divergence over time between self-reported and study partner–reported cognitive concerns among those with the highest levels of amyloid. Using objective or performance-based assessments of IADL targeting early changes in cognitively unimpaired individuals in addition to the questionnaires may help avoid this reporter dilemma.^4,25

When looking at how endorsement of difficulties changed from baseline to the end of follow-up for specific activities, we found that participants were most likely to self-report having difficulty “selecting and paying for items when shopping” at the end of the double-blind phase of the study. Study partners were most likely to report that the participant would have difficulty “completing complex activities,” whereas participants were less likely to self-report difficulty with this activity. Other activities that may be especially sensitive to change over time in a group of cognitively unimpaired older adults were “making telephone calls” and “planning and organizing complex activities for themselves or groups of people.” In the baseline assessments from both A4 and LEARN, we previously noted cross-sectional relationships between difficulty performing several activities and amyloid and tau.^5,16 Of interest, the 4 items that were endorsed with most difficulty at baseline (“remembering appointments,” “finding belongings,” “following TV programs,” and “remembering current events”) were not among the top items to show progression over time. This may suggest that the concerns that people present with initially may not be the most indicative of what will be most relevant later. It also underscores the value of repeated assessment of problems in daily functioning.

The AD course is believed to last 15 years or more before apparent clinical symptoms emerge,²⁶ and detecting changes in the preclinical stage is not easy with the standard tests that have been in use for decades.²⁷ Although we did observe significant changes and find associations between function and biomarkers here, studies with smaller samples or shorter follow-up periods might not have the power to overcome the ceiling effects and restricted movement that the ADCS ADL-PI shows in cognitively unimpaired individuals. It must also be noted that all individuals were aware of their amyloid status because their inclusion in either the A4 or the LEARN study depended on it. Consequently, this knowledge may have affected how participants and their study partners monitored and reported changes in daily functioning. As suggested before,^28,29 outcome measures in (clinical trial) research should be tailored to the sample being studied and capable of showing both cross-sectional differences and longitudinal changes. To that end, performance-based IADL measures, where participants are given a task to complete, may hold additional value in assessing everyday functioning that can complement self-report or study partner–related reports. Several studies using performance-based IADL instruments have shown promising results, including sensitivity to change over time,³⁰ associations with other IADL and cognitive measures,³¹ and both cross-sectional and longitudinal relationships with amyloid and tau.^21,23,32

Strengths of this study included the large sample, elaboration of clinical assessments, repeated PET scans, and relatively long follow-up duration. A significant limitation is the underrepresentation of people from diverse ethnoracial and educational attainment backgrounds, limiting the generalizability of these findings to the entire population at risk of AD and related dementias. Another limitation, as discussed above, was the ceiling effect of the IADL measure used, the ADCS ADL-PI. Despite that, we demonstrated various longitudinal associations with AD biomarkers.

In conclusion, this study shows that a more significant pathologic burden in preclinical AD is associated with a steeper decline in daily functioning over time and that increases in disease markers lead to more functional decline. Thus, it is feasible and essential to include sensitive IADL measures in preclinical AD clinical trials and clinical practice to demonstrate clinically meaningful outcomes.

Glossary

A4

Anti-Amyloid Treatment in Asymptomatic Alzheimer's Disease

AD

Alzheimer disease

ADCS ADL-PI

Alzheimer's Disease Cooperative Study Activities of Daily Living - Prevention Instrument

CL

centiloid

IADL

instrumental activities of daily living

LEARN

Longitudinal Evaluation of Amyloid Risk and Neurodegeneration

OR

odds ratio

SUVr

standardized uptake volume ratio

Author Contributions

M.A. Dubbelman: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. A. Liu: drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data. M.C. Donohue: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. O. Langford: drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data. R. Raman: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design. D.M. Rentz: drafting/revision of the manuscript for content, including medical writing for content; study concept or design. R. Amariglio: drafting/revision of the manuscript for content, including medical writing for content. R.A. Sperling: drafting/revision of the manuscript for content, including medical writing for content; study concept or design. P.S. Aisen: drafting/revision of the manuscript for content, including medical writing for content; study concept or design. G.A. Marshall: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data.

Study Funding

The A4 and LEARN studies were supported by a public-private–philanthropic partnership, which included funding from the National Institute of Aging of the NIH (R01 AG063689, U19AG010483, and U24AG057437), Eli Lilly (also the supplier of active medication and placebo), the Alzheimer's Association, the Accelerating Medicines Partnership through the Foundation for the NIH, the Gerald and Henrietta Rauenhorst Foundation, the Davis Alzheimer Prevention Program, the Yugilbar Foundation, an anonymous foundation, and additional private donors to Brigham and Women's Hospital, with in-kind support from Avid Radiopharmaceuticals, Cogstate, Albert Einstein College of Medicine and the Foundation for Neurologic Diseases. M.A. Dubbelman received salary support from Alzheimer Nederland (WE.06-2023-02), paid to his institution.

Disclosure

M.A. Dubbelman reports no disclosures relevant to the manuscript. A. Liu has received research support from the NIH, the Alzheimer's Association, the American Heart Association, Eli Lilly, and Eisai. M.C. Donohue reports that his spouse is a full-time employee of Janssen; and that he has served as a consultant to Roche. O. Langford and R. Raman have received research support from the NIH, the Alzheimer's Association, the American Heart Association, Eli Lilly, and Eisai. D.M. Rentz received salary support from the A4 study (R01 AG063689, U19AG010483 and U24AG057437) and has received payment or honoraria from USC Institute on Methods and Protocols for Advancement of Clinical Trials in ADRD (IMPACT-AD) course, Grand Rounds and External Advisory Boards from the University of California, Washington University, Boston University and Northwestern; as well as travel support to Alzheimer's Clinical Trials Consortium meetings, to the University of California Advisory Board Meeting, and to the Washington University Advisory Board Meeting. R. Amariglio received salary support from the A4 study (R01AG063689, U19AG010483 and U24AG057437). R.A. Sperling reports grant support from Eisai and Eli Lilly and reports serving as a consultant for AbbVie, AC Immune, Alector, Bristol-Myers-Squibb, Ionis, Janssen, Genentech, Merck, Prothena, Roche, and Vaxxinity. P.S. Aisen has received grants or contracts from the NIH, Alzheimer's Association, the Foundation for the NIH, Lilly, Janssen, and Eisai; and has received consulting fees from Merck, Biogen, AbbVie, Roche, and Immunobrain Checkpoint. G.A. Marshall was a site principal investigator for A4; has received salary support from the A4 study (R01 AG063689, U19AG010483, and U24AG057437); has received salary support from Eisai Inc. and Eli Lilly and Company for serving as a site principal investigator for clinical trials; and has received payments for serving as a consultant for Ono Pharma USA Inc. Go to Neurology.org/N for full disclosures.