Impact of learning APOE genotype on cognitively unimpaired adults: a pre-screening cohort study of the Alzheimer’s Prevention Initiative Generation Study 1 Authors: Jessica B Langbaum, Angela R Bradbury, Brian L Egleston, Elisabeth McCarty Wood, Carolyn M Langlois, Emily A Largent, Kristin Harkins, Claire M Erickson, Shana D Stites, Emma Oyen, Marie-Emmanuelle Riviere, Fonda Liu, Ana Graf, Scott Y H Kim, Joshua D Grill, Eric M Reiman, Pierre N Tariot, J Scott Roberts, Jason Karlawish Published: 2025 ABSTRACT SummaryBackgroundThe apolipoprotein E (APOE) gene is the best established genetic risk factor for Alzheimer’s disease in later life, with the ε4 allele conferring higher risk. APOE disclosure is becoming increasingly common in the clinical care of people with Alzheimer’s disease and in cognitively unimpaired adults. In this study, we aimed to describe changes in measures of genetic disease knowledge and psychiatric symptoms following APOE disclosure to cognitively unimpaired adults.MethodsData were collected as part of the screening phase of the global, multicentre, Alzheimer’s Prevention Initiative Generation Study 1 (NCT02565511). Eligible individuals were cognitively unimpaired (Mini-Mental State Exam total score ≥24), aged 60–75 years, and psychologically pre-screened for readiness (by measures of depressive symptoms and anxiety) to receive their APOE genotype from a health-care provider. Participants were assessed before disclosure, and 2–7 days, 6 weeks, 6 months, and 12 months after disclosure. Multivariable linear and ordinal logistic regressions were used to compare changes in genetic disease knowledge, anxiety, depression, and distress by APOE4 genotype status, adjusting for key covariates, with a focus on 2–7 days after disclosure. Multiple imputation by chained equations methods was used to account for missing outcome data.FindingsThe trial took place between Nov 30, 2015, and Sept 23, 2019. In total, 9496 participants (including 790 APOE4 homozygotes, 4869 heterozygotes, and 3837 non-carriers) learned their APOE genotype from a health-care provider as part of Generation Study 1 screening. 4038 (42·5%) participants were in the 65–69-year age group, 5790 (61·0%) were female, 3706 (39·0%) were male, and 8862 (93·3%) self-identified as White. Increase in genetic disease knowledge 2–7 days after disclosure was greater in APOE4 homozygotes (mean 1·19 [SD 3·95]) than in heterozygotes (0·78 [3·95], p=0·042) and non-carriers (0·29 [3·96], p=0·0002). Disease-specific distress 2–7 days after disclosure increased more in homozygotes (2·25 [6·42]) than in heterozygotes (0·53 [5·08], p<0·0001) and non-carriers (0·79 [4·95], p<0·0001). Levels of anxiety 2–7 days after disclosure increased in homozygotes (0·17 [2·95]) but decreased in heterozygotes (−0·67 [2·68], p<0·0001) and non-carriers (−0·66 [2·67], p<0·0001). There were no significant changes in depressive symptoms following disclosure for any APOE4 group. Notably, for all APOE4 groups, increases in distress and anxiety were small and did not reach predefined levels of clinical concern.InterpretationIn cognitively unimpaired, psychologically pre-screened adults, APOE disclosure by a trained health-care provider was generally safe and well tolerated, consistent with results from previous studies. To our knowledge, this is the largest study experience of APOE disclosure to date, especially for homozygotes, and is notable for the older age of participants compared with previous research. These results are timely and important given anticipated increases in APOE disclosure to guide clinical decision making once an Alzheimer’s disease prevention treatment is approved for cognitively unimpaired adults or if patients’ family members are interested in genetic testing. Scalable approaches for returning Alzheimer’s disease risk information are critical to meeting anticipated demand. Results from this study may be useful to bolster clinical translatability of disclosure programmes.FundingThe National Institute on Aging, Alzheimer’s Association, Banner Alzheimer’s Foundation, GHR Foundation, F-Prime Biomedical Research Initiative (FBRI), and Novartis Pharma. Introduction The apolipoprotein E (APOE) ε4 allele is the strongest genetic risk factor for late-onset Alzheimer’s disease.1,2 Disclosing APOE genotype is increasingly part of clinical care for people with cognitive impairment.3 With recent approvals of lecanemab and donanemab by the US Food and Drug Administration (FDA), there will be a large increase in APOE disclosure to guide clinical decision making, as Appropriate Use Recommendations4,5 include informing patients of their risk of developing symptomatic amyloid-related imaging abnormalities if they have one or two copies of the APOE ε4 allele.6 Moreover, APOE disclosure is common in clinical research in people with or without cognitive impairment. Observational studies and clinical trials can also target or limit enrollment based on APOE genotype, requiring disclosure of APOE results and associated risk of developing mild cognitive impairment or dementia due to Alzheimer’s disease. There will be an even greater demand for APOE disclosure if the ongoing prevention trials of lecanemab (NCT04468659) or donanemab (NCT05026866) are positive. In USA, the 21st Century Cures Act requires the immediate return of most medical test results, including APOE results to patients. This requirement means that patients might learn their results through their electronic medical record (EMR) before speaking with their health-care provider. Thus, there is mounting need to develop scalable approaches for returning genetic results and to understand the impact of APOE disclosure, particularly in cognitively unimpaired adults. In a series of successive randomised clinical trials, the Risk Evaluation and Education for Alzheimer’s Disease (REVEAL) programme tested the impact of APOE disclosure in cognitively unimpaired adults with a family history of Alzheimer’s disease and in patients with mild cognitive impairment. The REVEAL data suggest that disclosing APOE results does not lead to serious psychological distress, although some individuals might experience short-term distress.7–9 Overall, about 80% (n=24) of 30 REVEAL participants who learned their APOE results reported that the information had a positive impact on them, 17% (n=5) perceived it as neutral, and only 3% (n=1) perceived it as negative.10 However, cognitively unimpaired participants enrolled in REVEAL were more than a decade younger than the average age for developing symptoms of Alzheimer’s disease. There were also few APOE4 homozygotes in the REVEAL sample. Although the REVEAL results are encouraging, it remains important to understand the short-term and long-term psychological effects of APOE disclosure for those at highest risk for Alzheimer’s disease due to genetic susceptibility or age. The Alzheimer’s Prevention Initiative (API) Generation Program conducted the first National Institutes of Health (NIH) and industry-supported Alzheimer’s disease prevention trials of putative amyloid-modifying treatments in cognitively unimpaired persons at genetic risk for late-onset Alzheimer’s disease.11 API Generation Study 1 was conducted in cognitively unimpaired APOE4 homozygotes. Trial screening occurred in two steps11,12 (appendix p 3). Screening step one included the API Genetic Counseling and Disclosure Process, in which experienced health-care providers returned APOE results to individuals, and participants were followed up for up to 12 months to assess the impact of APOE disclosure.13 Only participants who carried two copies of the APOE4 ε4 allele (APOE4 homozygotes) were invited to continue trial screening in step two. Aims of the present study included describing changes in measures of genetic disease knowledge and psychological symptoms (eg, depression, anxiety) after APOE disclosure in cognitively unimpaired adults aged 60–75 years enrolled in step one. Methods Study design and participants This manuscript reports on data from screening step one of the phase 2/3, multicentre, randomised, double-blind, placebo-controlled Generation Study 1 trial of CAD106 (cohort 1) and umibecestat (cohort 2) in cognitively unimpaired APOE ε4 homozygotes aged 60–75 years (NCT02565511; appendix pp 2–3). The trial began on Nov 30, 2015, and was terminated early due to safety concerns with umibecestat (Sept 23, 2019, for cohort 1, July 11, 2019, for cohort 2). Results from the two trial cohorts were published separately.14,15 Implementation of the European Union General Data Protection Regulations in May 2018 delayed the release of data for analysis of the secondary endpoints reported in this paper. The Generation Study 1 trial was conducted in accordance with the International Council for Harmonisation Good Clinical Practice guidelines and local regulatory requirements. The study protocol was approved by the governing institutional review board or independent ethics committee for each investigational site or country. All participants provided written informed consent. Participants were enrolled by investigators at 139 clinical trial sites in ten countries (appendix pp 4–9). Recruitment strategies were described elsewhere.12 Key inclusion criteria were age 60–75 years inclusive and a Mini-Mental State Exam total score of 24 or higher.16 Study clinicians screened participants for psychological readiness to receive APOE results; participants were eligible if they had (1) a Geriatric Depression Scale (GDS) score of 6 or lower (participants with scores 7–10 could be included based on investigator judgement)17 and (2) a Modified State-Trait Anxiety Inventory for adults (STAI-AD) score of 17 or lower18 (participants with scores of 18 or 19 could be included based on investigator judgement; appendix pp 2, 10). The API Genetic Counseling and Disclosure Process was completed in a single visit.13 An optional educational video and an informational brochure were made available to participants before providing consent; these materials provided an introduction to the APOE gene, its role as a risk factor for Alzheimer’s disease, and potential considerations of learning APOE results. Participants had the right to opt out from disclosure of their APOE results and stop trial screening. Genetic counselling and disclosure sessions were conducted by appropriately licensed clinicians and genetic counsellors with experience communicating sensitive and complex health information either in person, by telephone, or by videoconference, based on local laws and regulations that varied by country. These experienced health-care providers received training at investigator meetings on how to conduct the disclosure sessions, including guidance regarding the lifetime risk estimates provided by the study, calculated for each APOE genotype, and talking points to guide the disclosure session (appendix pp 11–13).13,19 Participants were required to complete their disclosure session if they wanted to learn their APOE results. After APOE disclosure, the provider reviewed visual aids comparing risk estimates for each APOE genotype (range 10–55%), discussed inheritance implications for biological relatives, talked about potential risk modifiers, and offered resources for more information. In many instances, people with one copy of the APOE4 ε4 allele (heterozygotes) were offered screening in API Generation Study 2 (NCT03131453), and both heterozygotes and non-carriers could be offered screening in other, non-API led trials (eg, NCT04468659). Assessments Psychological impact was assessed via disclosure-related adverse events with a focus on GDS,17 STAI-AD,18 and disease-specific distress (DSD; table 1).20 Assessments were administered to participants in-person or by telephone by trained study staff. Participants with scores on the GDS higher than 10, STAI-AD higher than 19, or DSD higher than 13 had additional follow-up assessments with the site clinician. Scores and disclosure related adverse events were reviewed by the independent Disclosure Monitoring Advisory Group (DMAG; appendix p 14).13 In December 2018, the trial Data Monitoring Committee endorsed the recommendation from the DMAG to amend the protocol to reduce disclosure follow-up visits for APOE4 heterozygotes and non-carriers, requiring only the 2-to-7-days post-disclosure follow-up if no clinical concern for the participant was identified on the GDS, STAI-AD, or DSD. The 12-month follow-up period remained for homozygotes. The timing of implementation of this amendment varied among sites. Other measures collected to characterise the impact of APOE disclosure included knowledge of genetic disease,21 perceived risk (rated on a scale of 0–100% believed chance of developing Alzheimer’s disease),10,22 perceived threat of Alzheimer’s disease,23 and test-related distress using the Impact of Genetic Testing in Alzheimer’s Disease (IGT-AD) scale24 (appendix pp 15–17). Statistical analysis We excluded participants who had previous knowledge of their APOE status from the analysis to focus only on those who learned their results for the first time. We used χ2 tests and ANOVAs to compare demographic characteristics among APOE groups, excluding participants who knew their APOE results before step one screening. Reported means and percentages use unstandardised scales. We used multivariable linear and ordinal logistic regressions to compare unstandardised change from screening responses among APOE groups, adjusting for age, sex, race, years of education, country (USA vs non-USA), and family history of Alzheimer’s disease. We accounted for outcome missing data using multiple imputation by chained equations methods. Bonferroni adjusted p values of less than 0·0004 defined statistical significance. Analyses were performed using STATA (version 15/18). Due to the early termination, missing data structurally increased at later timepoints (appendix p 18). Role of the funding source The National Institute on Aging, Alzheimer’s Association, Banner Alzheimer’s Foundation, GHR Foundation, and F-Prime Biomedical Research Initiative had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. Novartis Pharma was a collaborator on the study design, data collection, and data interpretation, and several employees are co-authors of this report. Results During the conduct of this study, 9496 participants learned their APOE genotype for the first time (figure). Across all participants, the median age group was 65–69 years, 5790 (61·0%) were female, 8862 (93·3%) reported being White (ethnicity collection was restricted in some countries and not included), 3569 (37·6%) had 17 or more years of education, and 5279 (55·6%) had a parent with Alzheimer’s disease or other dementia (table 2). Multivariable longitudinal associations of baseline variables with missing data (with non-missing data as the reference) by APOE4 group are described in the appendix (p 19). At screening, APOE4 groups were similar on assessments of genetic disease knowledge and wellbeing (distress, depression, anxiety, perceived threat; table 3). Following APOE disclosure, there were differences on several assessments by group (table 3; appendix pp 20). 2–7 days after disclosure, there was an increase in scores for knowledge of genetic disease, and this increase varied by group. APOE4 homozygotes had the greatest observed increase in mean knowledge scores (adjusted mean change score difference [β]=0·59, 95% CI 0·28–0·91, compared with non-carriers; table 3 and appendix pp 21–22). There was an increase in DSD scores 2–7 days after disclosure in all APOE groups, with homozygotes having significantly greater change from baseline than heterozygotes (β=1·76, 95% CI 1·28–2·24) and non-carriers (β =1·55, 1·06–2·05), although unstandardised effect sizes were less than 0·5 baseline standard deviations (table 3; appendix pp 21–22). The mean scores did not reach levels of clinical concern (table 1). At 2–7 days after disclosure, APOE4 homozygotes reported greater levels of test-related distress (IGT-AD) compared with heterozygotes (β=6·45, 95% CI 5·65–7·24) and non-carriers (β=12·14, 11·34–12·94; appendix pp 21–22). At 6 weeks post-disclosure, homozygotes IGT-AD scores were lower than at 2–7 days but higher than heterozygotes (cross-sectional β=3·61, 2·90–4·31) and non-carriers (β=6·56, 5·83–7·29). At 2–7 days after disclosure, APOE4 homozygotes perceived their risk of developing Alzheimer’s disease as increased, on average, compared with baseline; heterozygotes and non-carriers perceived their risk as decreased (appendix pp 21–22). The differences were significant between homozygotes and heterozygotes (cross-sectional β=18·00, 95% CI 16·66–19·34), between homozygotes and non-carriers (β=24·61, 29·19–36·03), and between heterozygotes and non-carriers (p<0·0001). At 6 weeks, homozygotes’ perceived risk remained higher than at baseline and greater than heterozygotes (cross sectional β=15·74, 13·92–17·57) and non-carriers (β=21·03, 18·79–23·26). The perceived risk at 6 weeks of heterozygotes and non-carriers was lower than at baseline, but higher than at 2–7 days. We observed an APOE4 dose-dependent change in perceived threat (ie, worry about developing Alzheimer’s disease) at 2–7 days after disclosure, with non-carriers having greater score decrease than homozygotes (β=−1·22, 95% CI −0·92 to −1·52) and heterozygotes (β=−0·69, −0·86 to −0·52), and heterozygotes having a greater decrease than homozygotes (β=−0·53, −0·81 to −0·24; appendix pp 21–22). This pattern continued at 6 weeks (non-carriers vs homozygotes β=−0·84, −1·17 to −0·51, and non-carriers vs heterozygotes β=−0·45, −0·26 to −0·64). There were no significant changes in depressive symptom scores at any timepoint after disclosure for any group (table 3). Few participants (<0·50% or n<36 at each period) had GDS scores indicating moderate severity of depressive symptoms (scores 9–11), and even fewer had scores suggestive of severe symptoms (scores 12–15, n<11 at each period; table 4; appendix pp 38–40). Severe depressive symptoms were observed in all three genotype groups (table 4; appendix pp 38–40). Changes in state anxiety were observed 2–7 days after disclosure (table 4). Homozygotes had higher STAI-AD scores at 2–7 days than baseline, whereas scores for heterozygotes and non-carriers decreased at 2–7 days compared with baseline, with homozygotes having greater increases at 2–7 days than heterozygotes (β=0·86, 95% CI 0·63–1·08) and non-carriers (β=0·88, 0·65–1·12). After disclosure, about 1% of homozygotes had a borderline score (score of 18 or 19) on the modified STAI-AD. Even fewer had a score greater than 20, although homozygotes were more likely than heterozygotes to have a borderline or severe score on the modified STAI-AD 2–7 days after disclosure (1·6% vs 0·4%). Discussion In cognitively unimpaired adults screened for an Alzheimer’s prevention trial and deemed psychologically ready and willing to learn their results from an experienced health-care provider, disclosure of APOE genotype was well tolerated in the short-term and did not cause persistent, clinically significant psychological symptoms based on established clinical thresholds. Findings from this study are consistent with results from the REVEAL programme7 and other related studies of APOE disclosure in cognitively unimpaired adults,25 but noteworthy in view of the large number of APOE4 homozygotes and older age of the sample. To our knowledge, this is the first study to examine differences in outcomes between APOE4 homozygotes, heterozygotes, and non-carriers, as previous studies focused on studying the differences between carriers and non-carriers.26 Until recently, APOE disclosure was neither common in the clinical care of cognitively impaired persons with Alzheimer’s disease nor recommended for predicting the risk of Alzheimer’s disease in cognitively unimpaired people,3 because of limited clinical utility. There were concerns that disclosing APOE results to cognitively unimpaired adults would result in significant short-term or long-term psychological harm.27 In addition to expected increases in APOE disclosure mentioned in the introduction, there is mounting demand to return genetic, biomarker, and clinical research data to study participants.28,29 Given the growing number of people who will learn their APOE results, it is imperative to understand the psychological impact of receiving genetic information and best practices for communicating results, particularly in cognitively unimpaired adults. We found that APOE4 homozygotes reported greater levels of state anxiety and disease-specific distress immediately following disclosure; however, these were generally small (mild) changes, of less than half the magnitude of the baseline standard deviation and below the threshold for clinical concern. These results are consistent with previous studies examining differences in outcomes between APOE4 carriers and non-carriers.7,9,25 We observed an APOE4 gene-dose effect on the IGT-AD, a measure of test-related distress, immediately following disclosure.24 The clinical significance is unclear, given there are no established clinical cutoffs for this assessment, although we will note that even immediately after disclosure, the mean score of APOE4 homozygotes was far below the maximum score of 80. Similarly, although we found that APOE4 homozygotes were more likely than heterozygotes or non-carriers to cross a clinically defined threshold on measures of anxiety and depressive symptoms, this occurred in less than 1% of homozygotes. There were no differences among APOE4 groups in levels of depressive symptoms after disclosure. However, it is important to reinforce that, in the present study, participants were screened for psychological readiness (though only 34 were deemed not ready), and APOE disclosure was led by experienced and trained health-care providers. Our findings might therefore not generalise to contexts where disclosure is offered by less experienced providers or to more psychologically vulnerable populations. Risk estimates for developing mild cognitive impairment or Alzheimer’s disease dementia by age 85 years for each APOE genotype were calculated19 and used in the study genetic counselling materials.13 Following APOE disclosure, perceived risk increased for homozygotes and continued to increase at each subsequent timepoint, which could be related to participants nearing the age where cognitive impairment is more likely to manifest. In contrast, perceived risk decreased for heterozygotes and non-carriers compared with baseline at most timepoints. It is noteworthy that, after disclosure, heterozygotes and non-carriers believed their chances of developing Alzheimer’s disease were far greater than the provided risk estimates13,19 (appendix p 12), even after adjusting for key variables. For homozygotes, mean perceived risk was near the high end of the provided risk range of 30–55%. These data are consistent with REVEAL results, which found that nearly half of participants believed their risk of Alzheimer’s disease was more than 5% different from what they were told, with the majority believing their risk was higher than they were told.22 In REVEAL, greater levels of perceived personal risk at baseline were predictive of discordant beliefs, perhaps due to anchoring-and-adjustment bias.30 Opportunities to mitigate discordant perceptions in perceived risk warrant further exploration, particularly if APOE disclosure becomes a routine part of clinical practice. Knowledge of genetic disease increased for all groups 2–7 days after disclosure, with APOE4 homozygotes having larger increases compared with heterozygotes and non-carriers. These results are consistent with REVEAL9 and suggest the condensed, single-session process was effective at communicating pertinent information. With the expected increase in genetic testing for Alzheimer’s disease, the shortage of dementia specialists and limited staffing of genetic counsellors in memory clinics to disclose results, and the 21st Century Cures Act in the USA, novel approaches are needed to responsibly communicate test results to patients without unduly burdening a limited workforce. Condensed disclosure materials such as those used here should be adapted and evaluated for self-directed use among cognitively unimpaired adults and those with symptoms (eg, mild cognitive impairment and early Alzheimer’s disease), leveraging artificial intelligence and should be embedded into the EMR before disclosing results. This approach can help patients make sense of and contextualise their results before and after speaking with their clinician.31,32 Building off the present study, the Evaluation of Self-Mediated Alternatives for Risk Testing and Return of Results (eSMARTER) trial (NCT06459583) is a randomised, non-inferiority study evaluating self-directed scalable eHealth methods for communicating genetic and biomarker results for Alzheimer’s disease and characterising the clinical impacts of learning this information.33 Results from the present study, eSMARTER, and others will help guide the use of condensed disclosure protocols with and without a health-care provider in clinical settings for different patient populations (ie, cognitively unimpaired, mild cognitive impairment, or early Alzheimer’s disease) and among people with certain comorbidities that have been exclusionary criteria in disclosure trials (eg, depressive symptoms, anxiety, etc). This study has some limitations. Individuals with moderate-to-severe symptoms of depression or anxiety were ineligible (n=34). Those learning their APOE results in clinical settings or through direct-to-consumer testing might not undergo the same level of screening, resulting in more adverse psychological reactions compared with those observed here.34 Additionally, when APOE disclosure is conducted to determine appropriateness for amyloid-modifying therapies, it will be one of many pieces of health-related information shared with the patient during the multistep process for starting treatment. As such, discussions related to APOE might be more limited than in the present study, meaning that patients might not remember the information shared, particularly if patients have cognitive impairment that could impede information processing, leading to potentially incorrect perceptions about Alzheimer’s disease risk. Future studies should examine disclosure of multiple results (eg, APOE and brain amyloid measured by PET or plasma) in a single counselling session. Most participants were White and had 12 or more years of education. Care should be taken when extrapolating the results to other groups, and future studies should examine APOE disclosure in broader populations at high risk for Alzheimer’s disease (eg, the IDEAL programme studying APOE disclosure among Latinos in northern Manhattan;35 NCT04471779). Moreover, following APOE disclosure, homozygotes had the opportunity to continue screening for the trial, perhaps with the expectation that the trial treatments could reduce their risk of developing cognitive impairment, which might have unintended effects on the assessments. We cannot track whether heterozygotes or non-carriers participated in other trials (eg, Generation Study 2). All participants learned their APOE results from an experienced health-care provider (genetic counsellor or other licensed clinician) who followed the study’s genetic counselling protocol with talking points and visual aid materials.13 This might not be the case for future patients as APOE testing increases, with requirements for test results to be shared with patients in real-time via their EMR. We originally intended to assess all participants up to 12 months following disclosure to understand the longer-term impact of APOE disclosure. However, while enrolment was ongoing, the protocol was amended to require only the immediate follow-up (2–7 days after disclosure) for heterozygotes and non-carriers if no clinical concern was identified. Additionally, the trial was terminated early, while enrolment was ongoing. Taken together, data are not missing at random, thus limiting our ability to compare changes between APOE groups, especially beyond the 2–7-day follow-up. Lastly, there have been considerable advances in the field since the trial concluded (eg, FDA approvals of lecanemab and donanemab) which, if the trial were taking place today, might have influenced how participants respond after learning their APOE results. In conclusion, APOE disclosure will become increasingly common in the clinical care of people with or at increased risk of Alzheimer’s disease, either due to having access to amyloid-modifying therapies or out of interest because a biological relative (eg, parent) learned their APOE genotype. In this study, APOE disclosure conducted by experienced providers to cognitively unimpaired adults was shown to be generally well tolerated. This is consistent with results from previous studies, but noteworthy given the large number of APOE4 homozygotes and older age of those in the current study. Results from this study can be used to inform the development and implementation of scalable programmes for disclosing APOE and other Alzheimer’s disease biomarker results to cognitively unimpaired adults, particularly once an Alzheimer’s disease prevention treatment becomes available.36 Supplementary Material