치매 치료제 레카네맙 2023년 NEJM

[문형철]

https://www.youtube.com/watch?v=UKXOp9S-yAE

Lecanemab in Early Alzheimer’s Disease

Authors: Christopher H. van Dyck, M.D., Chad J. Swanson, Ph.D., Paul Aisen, M.D., Randall J. Bateman, M.D., Christopher Chen, B.M., B.Ch., Michelle Gee, Ph.D., Michio Kanekiyo, M.S., +11, and Takeshi Iwatsubo, M.D.Author Info & Affiliations

Published November 29, 2022

N Engl J Med 2023;388:9-21

DOI: 10.1056/NEJMoa2212948

VOL. 388 NO. 1

Copyright © 2022

• • Abstract
  • Methods
  • Results
  • Discussion
  • Notes
  • Supplementary Material
  • References
• • Information & Authors
  • Metrics & Citations
  • View Options
  • References
  • Media
  • Tables
  • Share

Abstract

Background

The accumulation of soluble and insoluble aggregated amyloid-beta (Aβ) may initiate or potentiate pathologic processes in Alzheimer’s disease. Lecanemab, a humanized IgG1 monoclonal antibody that binds with high affinity to Aβ soluble protofibrils, is being tested in persons with early Alzheimer’s disease.

Methods

We conducted an 18-month, multicenter, double-blind, phase 3 trial involving persons 50 to 90 years of age with early Alzheimer’s disease (mild cognitive impairment or mild dementia due to Alzheimer’s disease) with evidence of amyloid on positron-emission tomography (PET) or by cerebrospinal fluid testing. Participants were randomly assigned in a 1:1 ratio to receive intravenous lecanemab (10 mg per kilogram of body weight every 2 weeks) or placebo. The primary end point was the change from baseline at 18 months in the score on the Clinical Dementia Rating–Sum of Boxes (CDR-SB; range, 0 to 18, with higher scores indicating greater impairment). Key secondary end points were the change in amyloid burden on PET, the score on the 14-item cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-cog14; range, 0 to 90; higher scores indicate greater impairment), the Alzheimer’s Disease Composite Score (ADCOMS; range, 0 to 1.97; higher scores indicate greater impairment), and the score on the Alzheimer’s Disease Cooperative Study–Activities of Daily Living Scale for Mild Cognitive Impairment (ADCS-MCI-ADL; range, 0 to 53; lower scores indicate greater impairment).

Download a PDF of the Research Summary.

Results

A total of 1795 participants were enrolled, with 898 assigned to receive lecanemab and 897 to receive placebo. The mean CDR-SB score at baseline was approximately 3.2 in both groups. The adjusted least-squares mean change from baseline at 18 months was 1.21 with lecanemab and 1.66 with placebo (difference, −0.45; 95% confidence interval [CI], −0.67 to −0.23; P<0.001). In a substudy involving 698 participants, there were greater reductions in brain amyloid burden with lecanemab than with placebo (difference, −59.1 centiloids; 95% CI, −62.6 to −55.6). Other mean differences between the two groups in the change from baseline favoring lecanemab were as follows: for the ADAS-cog14 score, −1.44 (95% CI, −2.27 to −0.61; P<0.001); for the ADCOMS, −0.050 (95% CI, −0.074 to −0.027; P<0.001); and for the ADCS-MCI-ADL score, 2.0 (95% CI, 1.2 to 2.8; P<0.001). Lecanemab resulted in infusion-related reactions in 26.4% of the participants and amyloid-related imaging abnormalities with edema or effusions in 12.6%.

Conclusions

Lecanemab reduced markers of amyloid in early Alzheimer’s disease and resulted in moderately less decline on measures of cognition and function than placebo at 18 months but was associated with adverse events. Longer trials are warranted to determine the efficacy and safety of lecanemab in early Alzheimer’s disease. (Funded by Eisai and Biogen; Clarity AD ClinicalTrials.gov number, NCT03887455.)

배경

용해성 및 불용성 응집된 아밀로이드 베타(Aβ)의 축적은

알츠하이머병의 병리학적 과정을 시작하거나 강화할 수 있습니다.

Aβ 용해성 프로토피브릴에 높은 친화도로 결합하는

인간화 IgG1 단일클론 항체인 레카네맙은

초기 알츠하이머병 환자를 대상으로 시험되고 있습니다.

방법

우리는 50세에서 90세 사이의 초기 알츠하이머병(양성자 방출 단층촬영(PET) 또는

뇌척수액 검사를 통해 아밀로이드가 발견된 경도 인지장애 또는 경도 치매) 환자를 대상으로

18개월 동안 다기관 이중맹검 3상 시험을 실시했습니다.

참가자들은 1:1 비율로 무작위로 배정되어

레카네맙 정맥 주사(체중 1kg당 10mg을 2주마다 1회 투여) 또는

위약을 투여받았습니다.

1차 평가 지표는 18개월에 기준치로부터의 변화로,

임상 치매 평가 척도-상자 합계(CDR-SB; 범위 0~18, 점수가 높을수록 장애가 심함을 나타냄) 점수였습니다.

주요 2차 평가 지표는

PET에서 아밀로이드 부담의 변화,

알츠하이머병 평가 척도(ADAS-cog14)의 14개 항목 인지 하위 척도 점수(범위: 0~90, 점수가 높을수록 장애가 심함을 나타냄),

알츠하이머병 종합 점수(ADCOMS)였습니다. (범위: 0~1.97, 점수가 높을수록 장애가 심함),

알츠하이머병 협력 연구-경미한 인지 장애를 위한 일상생활 활동 척도(ADCS-MCI-ADL; 범위: 0~53, 점수가 낮을수록 장애가 심함)에 대한 점수.

연구 요약 PDF를 다운로드하세요.

결과

총 1795명의 참가자가 등록되었고,

그 중 898명은 레카네맙을 투여받고

897명은 위약을 투여받았습니다.

기준선에서 두 그룹의 평균 CDR-SB 점수는 약 3.2였습니다. 18개월에 기준선으로부터의 조정된 최소 제곱 평균 변화는 레카네맙의 경우 1.21, 위약의 경우 1.66이었습니다(차이, -0.45; 95% 신뢰 구간 [CI], -0.67~-0.23; P<0.001). 698명의 참가자를 대상으로 한 하위 연구에서, 레카네맙을 투여한 경우 위약보다 뇌 아밀로이드 부담이 더 많이 감소했습니다(차이, -59.1 센틸로이드; 95% CI, -62.6~-55.6). 두 그룹의 기준치로부터의 변화에서 레카네맙이 유리한 다른 평균 차이는 다음과 같습니다: ADAS-cog14 점수, -1.44 (95% CI, -2.27 ~ -0.61; P<0.001); ADCOMS, -0. 0.050(95% CI, -0.074~-0.027; P<0.001); ADCS-MCI-ADL 점수 2.0(95% CI, 1.2~2.8; P<0.001). 레카네맙은 참가자의 26.4%에게서 주입 관련 반응을 일으켰고, 12.6%에게서 부종이나 삼출을 동반한 아밀로이드 관련 영상 이상이 나타났습니다.

결론

레카네맙은

초기 알츠하이머병 환자의 아밀로이드 표지자를 감소시켰고,

18개월 동안 인지 및 기능 측정에서 위약보다 약간 더 적은 감소를 가져왔지만

부작용과 관련이 있었습니다.

초기 알츠하이머병에서

레카네맙의 효능과 안전성을 확인하기 위해서는

더 긴 임상시험이 필요합니다.

(아이사와 바이오젠이 자금을 지원; Clarity AD ClinicalTrials.gov 번호, NCT03887455.)

Quick Take

Lecanemab in Alzheimer’s Disease

1m 47s

Current therapeutic agents for Alzheimer’s disease–related dementia temporarily improve symptoms but do not alter the underlying disease course.1,2 Some evidence suggests that amyloid removal slows the progression of disease.3 One anti-amyloid antibody (aducanumab) has received accelerated approval from the Food and Drug Administration.

Lecanemab is a humanized monoclonal antibody that binds with high affinity to soluble amyloid-beta (Aβ) protofibrils, which have been shown to be more toxic to neurons than monomers or insoluble fibrils.4-14 A phase 2b, dose-finding trial involving 854 participants with early Alzheimer’s disease did not show a significant difference between lecanemab and placebo in a Bayesian analysis of 12-month change in a composite score (primary end point). However, analyses at 18 months showed dose- and time-dependent clearance of amyloid with lecanemab, and the drug was associated with less clinical decline on some measures than placebo. In that trial, intravenous administration of 10 mg of lecanemab per kilogram of body weight every 2 weeks was identified as an appropriate dose, with a 9.9% incidence (<3% symptomatic) of amyloid-related imaging abnormalities (ARIA) with edema or effusions (ARIA-E).15 We conducted a phase 3 trial (Clarity AD) to determine the safety and efficacy of lecanemab in participants with early Alzheimer’s disease.

현재 알츠하이머병 관련 치매에 사용되는 치료제는

일시적으로 증상을 개선할 뿐

근본적인 질병의 진행을 바꾸지는 못합니다.1,2

일부 증거에 따르면,

아밀로이드 제거가 질병의 진행을 늦춘다고 합니다.3

식품의약국(FDA)은

한 가지 항아밀로이드 항체(아두카누맙)에 대해

신속 승인을 내렸습니다.

레카네맙은

수용성 아밀로이드 베타(Aβ) 원섬유에 높은 친화도로 결합하는 인간화 단일 클론 항체로서,

단량체나 불용성 피브릴보다 뉴런에 더 독성이 있는 것으로 나타났습니다.4-14

A 초기 알츠하이머병 환자 854명을 대상으로 한 2b상,

용량 결정 시험에서 12개월 동안의 복합 점수(1차 평가 지표) 변화에 대한 베이지안 분석에서 레카네맙과 위약 사이에 유의미한 차이가 나타나지 않았습니다.

그러나

18개월에 대한 분석에서는

레카네맙을 투여한 경우 용량과 시간에 따라 아밀로이드가 제거되는 것으로 나타났으며,

이 약물은 위약보다 일부 측정에서 임상적 감소가 적었습니다.

이 임상시험에서 2주마다 체중 1kg당 10mg의 레카네맙을 정맥 투여하는 것이 적절한 용량으로 확인되었으며, 부종 또는 삼출액(ARIA-E)을 동반한 아밀로이드 관련 영상 이상(ARIA)의 발생률(증상이 있는 경우 3% 미만)은 9.9%였습니다.15 우리는 초기 알츠하이머병 환자를 대상으로 레카네맙의 안전성과 효능을 확인하기 위해 3상 임상시험(Clarity AD)을 실시했습니다.

MethodsTrial Design and Oversight

Clarity AD was an 18-month, multicenter, double-blind, placebo-controlled, parallel-group trial involving persons with early Alzheimer’s disease. Eligible participants were randomly assigned in a 1:1 ratio to receive intravenous lecanemab (10 mg per kilogram every 2 weeks) or placebo. The randomization was stratified according to clinical subgroup (mild cognitive impairment due to Alzheimer’s disease or mild Alzheimer’s disease–related dementia on the basis of the criteria noted below), the presence or absence of concomitant approved medication for symptoms of Alzheimer’s disease at baseline (e.g., acetylcholinesterase inhibitors, memantine, or both), apolipoprotein E (ApoE) ε4 carriers or noncarriers, and geographic region. During the trial, participants underwent serial blood testing for plasma biomarkers and could participate in three optional substudies that eval‎uated longitudinal changes in brain amyloid burden as measured by positron-emission tomography (PET), brain tau pathologic features as measured by PET, and cerebrospinal fluid (CSF) biomarkers of Alzheimer’s disease.

The trial was conducted in accordance with International Council for Harmonisation guidelines and the ethical principles of the Declaration of Helsinki. The trial was approved by the institutional review board or independent ethics committee at each center, and all the participants provided written informed consent. The sponsor Eisai designed the trial and analyzed the data in collaboration with the academic authors, provided lecanemab and placebo, provided funding for medical writing, and aided in drafting the manuscript. The sponsor could not delay or interdict publication. The first, second, and sixteenth authors wrote the first draft of the manuscript, with professional medical writing assistance funded by Eisai, and all the authors contributed to subsequent drafts. Confidentiality agreements were in place between the sponsor and the authors and site investigators. Biogen provided partial funding for the trial.

An independent data and safety monitoring board consisting of experts in Alzheimer’s disease and statistics reviewed unblinded safety data during the trial. An independent medical monitoring team, whose members were unaware of the trial-group assignments, reviewed ARIA, infusion-related reactions, and hypersensitivity reactions. Clinical assessment raters were unaware of the safety assessments and the trial-group assignments. All the authors vouch for the completeness and accuracy of the data, the fidelity of the trial to the protocol (available with the full text of this article at NEJM.org), and the full reporting of adverse events.

Eligibility Criteria

The trial included participants 50 to 90 years of age, with either mild cognitive impairment due to Alzheimer’s disease or mild Alzheimer’s disease–related dementia on the basis of National Institute on Aging–Alzheimer’s Association criteria.16,17 Amyloid positivity was determined by PET or CSF measurement of Aβ1–42. All the participants had objective impairment in episodic memory as indicated by at least 1 standard deviation below the age-adjusted mean in the Wechsler Memory Scale IV–Logical Memory II.

End Points

The primary efficacy end point was the change in the score on the Clinical Dementia Rating (CDR)–Sum of Boxes (CDR-SB)18 from baseline at 18 months. The CDR-SB score is a validated outcome measure used in clinical trials of Alzheimer’s disease that is obtained by interviewing patients and their care partners and captures cognition and function. It assesses six domains that patients and caregivers identify as important (Memory, Orientation, Judgment and Problem Solving, Community Affairs, Home and Hobbies, and Personal Care). Scores for each domain range from 0 to 3, with higher scores indicating greater impairment. Total scores range from 0 to 18, with a score of 0.5 to 6 indicating early Alzheimer’s disease.

Key secondary end points were the change from baseline at 18 months in the following: amyloid burden on PET as measured in centiloids (with either florbetaben, florbetapir, or flutemetamol tracers) in a substudy, the score on the 14-item cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-cog14; range, 0 to 90, with higher scores indicating greater impairment),19 the Alzheimer’s Disease Composite Score (ADCOMS; range, 0 to 1.97, with higher scores indicating greater impairment),20 and the score on the Alzheimer’s Disease Cooperative Study–Activities of Daily Living Scale for Mild Cognitive Impairment (ADCS-MCI-ADL; range, 0 to 53, with lower scores indicating greater impairment).21 Biomarker assessments included CSF biomarkers (Aβ1–40, Aβ1–42, total tau, phosphorylated tau 181 [p-tau181], neurogranin, and neurofilament light chain [NfL]) and plasma biomarkers (Aβ42/40 ratio, p-tau181, glial fibrillary acidic protein [GFAP], and NfL). Tau PET and volumetric magnetic resonance imaging (MRI) results have not been fully analyzed.

A prespecified exploratory and multiplicity-unadjusted analysis examined the time to worsening of the global CDR score (range, 0 to 3, with higher scores indicating greater impairment). This end point was defined as the time to the first increase of at least 0.5 points in the global CDR score on two consecutive visits.

Statistical Analysis

The sample size for this trial was estimated on the basis of comparison of lecanemab and placebo with respect to the primary efficacy end point, the change from baseline at 18 months in the CDR-SB score. On the basis of data from the phase 2b trial of lecanemab,15 the estimated standard deviation of the change from baseline at 18 months in the CDR-SB score with placebo was 2.031 points, and the estimated treatment difference between lecanemab and placebo in all the participants was 0.373 points. This estimation corresponds to 25% less decline in cognitive function with lecanemab than with placebo and is consistent with a clinically meaningful difference on the basis of the Alzheimer’s disease literature, statistical principles, and agreements with regulatory authorities.15,22–24 Therefore, under the assumption of an estimated 20% dropout rate at 18 months in this trial, a total sample size of 1566 participants, including 783 participants receiving lecanemab and 783 participants receiving placebo, would provide the trial with 90% power to detect the treatment difference with the use of a two-sample t-test at a two-sided alpha level of 0.05. The sample size was increased by 200 to account for participants who missed three or more consecutive doses during the initial 6-month peak period of coronavirus disease 2019 (Covid-19), in accordance with previous agreement with regulatory authorities. No interim analyses for futility or efficacy were planned or performed.

Efficacy analyses were performed in the modified intention-to-treat population, which was defined as the group of randomly assigned participants who received at least one dose of lecanemab or placebo and who had a baseline assessment and at least one postdose primary efficacy (CDR-SB) measurement. Sensitivity analyses across efficacy end points to assess the robustness of the primary analysis to missing data included rank analysis of covariance with imputation of missing values. Additional sensitivity analyses were performed to eval‎uate potential effects of functional unblinding due to ARIA and effects of missed doses due to Covid-19–related absences (see the Supplementary Appendix, available at NEJM.org). Safety was eval‎uated in the safety population, which was defined as the group of participants who received at least one dose of lecanemab or placebo. Safety eval‎uations included monitoring of adverse events, vital signs, physical examinations, clinical laboratory variables, and 12-lead electrocardiograms. Occurrences of ARIA were monitored throughout the trial by central reading of MRI performed at weeks 9, 13, 27, 53, and 79 as well as at the 3-month follow-up visit (week 91) for safety monitoring. In addition, the populations for the substudies of amyloid burden on PET, tau pathologic features on PET, and CSF biomarkers of Alzheimer’s disease were the groups of participants who received at least one dose of lecanemab or placebo and who underwent a baseline PET or CSF eval‎uation and at least one postdose eval‎uation.

The primary analysis was performed without imputation of missing values. The primary analysis of the change from baseline at 18 months in the CDR-SB score was performed to compare lecanemab and placebo with the use of a mixed model for repeated measures that included the baseline CDR-SB score as a covariate, with trial group, visit, stratification variables (i.e., clinical subgroup, use of medication for symptoms of Alzheimer’s disease at baseline [yes or no], ApoE ε4 carrier status [carriers or noncarriers], and geographic region [North America, Europe, and Asia–Pacific]), baseline CDR-SB score–by–visit interaction, and trial group–by–visit interaction as fixed effects. If the between-group difference in primary end-point results was significant, then key secondary end points were to be tested hierarchically in the following order: change from baseline at 18 months in amyloid burden on PET as measured in centiloids in the subgroup tested and change from baseline at 18 months in the ADAS-cog14 score, change from baseline at 18 months in the ADCOMS, and change from baseline at 18 months in the ADCS-MCI-ADL score, all in the modified intention-to-treat population. Each test was performed at an alpha level of 0.05 (two-sided) and was to be performed only if the preceding test was significant at a two-sided level of 0.05. Additional details on the design and analysis methods are provided in the Supplementary Appendix and protocol.

ResultsParticipants

A total of 5967 persons were screened and 1795 underwent randomization; 898 were assigned to receive lecanemab and 897 to receive placebo at 235 sites in North America, Europe, and Asia from March 2019 through March 2021. Of these participants, 729 (81.2%) in the lecanemab group and 757 (84.4%) in the placebo group completed the trial and had data available on the primary end point (Figure 1). The modified intention-to-treat population included 1734 participants (859 in the lecanemab group and 875 in the placebo group), and the safety population included all 1795 randomly assigned participants. Enrollment in three longitudinal substudies included 698 participants in the substudy of amyloid burden on PET, 257 in the study of tau pathologic features on PET, and 281 in the substudy of CSF biomarkers of Alzheimer’s disease. The baseline characteristics of the substudy groups were generally similar to those in the main analysis. This trial made efforts to enhance global enrollment of a diverse group of participants (20% non-White), including in the United States, where 6.1% and 28.1% of the 3638 screened participants and 4.5% and 22.5% of randomly assigned participants were Black and Hispanic, respectively. The characteristics of the participants at baseline were generally similar in the two trial groups (Table 1). These characteristics were similar to what has been observed in population studies involving persons with early Alzheimer’s disease, although there was an underrepresentation of Black persons in the United States and an overrepresentation of Hispanic persons in the United States. The representativeness of the trial population is shown in Table S1 in the Supplementary Appendix.

Figure 1

Screening, Randomization, and Follow-up.

Table 1

Characteristics of the Participants at Baseline (Modified Intention-to-Treat Population).

End-Point Results

The mean CDR-SB score at baseline was approximately 3.2 in both the lecanemab and placebo groups, findings consistent with early Alzheimer’s disease (score of 0.5 to 6). The adjusted mean change from baseline at 18 months in the CDR-SB score was 1.21 in the lecanemab group and 1.66 in the placebo group (difference, −0.45; 95% confidence interval [CI], −0.67 to −0.23; P<0.001) (Figure 2A and Table 2).

Figure 2

Primary and Key Secondary End Points.

Table 2

Primary and Secondary End Points (Modified Intention-to-Treat Population).

In the substudy of amyloid burden on PET (a key secondary end point) involving 698 participants, the mean amyloid level at baseline was 77.92 centiloids in the lecanemab group and 75.03 centiloids in the placebo group. The adjusted mean change from baseline at 18 months was −55.48 centiloids in the lecanemab group and 3.64 centiloids in the placebo group (difference, −59.12 centiloids; 95% CI, −62.64 to −55.60; P<0.001) (Figure 2B and Table 2). In the modified intention-to-treat population, the mean ADAS-cog14 scores at baseline were 24.45 in the lecanemab group and 24.37 in the placebo group. The adjusted mean change from baseline at 18 months in the ADAS-cog14 score was 4.14 in the lecanemab group and 5.58 in the placebo group (difference, −1.44; 95% CI, −2.27 to −0.61; P<0.001) (Figure 2C and Table 2). The mean ADCOMS in the modified intention-to-treat population at baseline was 0.398 in the lecanemab group and 0.400 in the placebo group. The adjusted mean change from baseline at 18 months in the ADCOMS was 0.164 in the lecanemab group and 0.214 in the placebo group (difference, −0.050; 95% CI, −0.074 to −0.027; P<0.001) (Figure 2D and Table 2). In the modified intention-to-treat population, the mean ADCS-MCI-ADL scores at baseline were 41.2 for lecanemab and 40.9 for placebo. The adjusted mean change from baseline at 18 months in the ADCS-MCI-ADL score was −3.5 in the lecanemab group and −5.5 in the placebo group (difference, 2.0; 95% CI, 1.2 to 2.8; P<0.001) (Figure 2E and Table 2).

For each of these assessments, separation of the trial groups was apparent by visual inspection of graphs at 3 months. However, no conclusions can be drawn because there was no prespecified plan for analysis that included adjustment of confidence intervals for multiple comparisons at any intermediate time point.

Sensitivity analyses of the CDR-SB score that eval‎uated the effect of Covid-19 (missed doses) and potential for bias from functional unblinding due to ARIA were generally consistent with the primary analysis (Table S2). Results were also consistent across key randomization strata, as well as for other factors that affect Alzheimer’s disease (Figs. S1 through S4). The exploratory subgroup analysis involving ApoE ε4 homozygotes (15% of the trial population) numerically favored lecanemab for the ADAS-cog14 and ADCS-MCI-ADL scores but not for the CDR-SB score and the ADCOMS. Results of prespecified analyses of end points involving CSF and plasma biomarkers showed numerical improvements for all assessments comparing lecanemab with placebo, with the exception of CSF NfL (Fig. S5). In a prespecified, multiplicity-unadjusted analysis of the time to worsening of the global CDR score, the hazard ratio for progression to the next stage of dementia (0.69) numerically favored lecanemab over placebo (Fig. S6).

Safety

Deaths occurred in 0.7% of the participants in the lecanemab group and 0.8% of those in the placebo group (Table 3). No deaths were considered by the investigators to be related to lecanemab or occurred with ARIA. Serious adverse events occurred in 14.0% of the participants in the lecanemab group and 11.3% of those in the placebo group. The most commonly reported serious adverse events were infusion-related reactions (in 1.2% of the participants in the lecanemab group and 0 participants in the placebo group), ARIA-E (in 0.8% and 0, respectively), atrial fibrillation (in 0.7% and 0.3%), syncope (in 0.7% and 0.1%), and angina pectoris (in 0.7% and 0). The overall incidence of adverse events was similar in the two groups (Table 3). Adverse events leading to discontinuation of the trial agent occurred in 6.9% of the participants in the lecanemab group and 2.9% of those in the placebo group. The most common adverse events (affecting >10% of the participants) in the lecanemab group were infusion-related reactions (26.4% with lecanemab and 7.4% with placebo); ARIA with cerebral microhemorrhages, cerebral macrohemorrhages, or superficial siderosis (ARIA-H; 17.3% with lecanemab and 9.0% with placebo); ARIA-E (12.6% with lecanemab and 1.7% with placebo); headache (11.1% with lecanemab and 8.1% with placebo); and falls (10.4% with lecanemab and 9.6% with placebo). Infusion-related reactions were largely mild to moderate (grade 1 or 2, 96%) and occurred with the first dose (75%). A total of 56% of the participants did not take preventative medications (i.e., nonsteroidal antiinflammatory drugs, antihistamines, or glucocorticoids) for infusion-related reactions. Of those who took preventative medications for subsequent doses, 63% did not have additional reactions.

Table 3

Adverse Events.

Events of ARIA-E with lecanemab were mostly mild to moderate (91%) on the basis of central reading of imaging with the use of protocol definitions. These events were mostly asymptomatic (78%), occurred during the first 3 months of the treatment period (71%), and resolved within 4 months after detection (81%). A total of 2.8% of the participants in the lecanemab group had symptomatic ARIA-E; commonly reported symptoms were headache, visual disturbance, and confusion. The incidence of isolated ARIA-H (i.e., ARIA-H in participants who did not also have ARIA-E) was 8.9% in the lecanemab group and 7.8% in the placebo group. The incidence of isolated symptomatic ARIA-H was 0.7% in the lecanemab group and 0.2% in the placebo group. The most common symptom associated with isolated symptomatic ARIA-H was dizziness. Macrohemorrhage occurred in 5 of 898 participants (0.6%) in the lecanemab group and 1 of 897 participants (0.1%) in the placebo group. ARIA-H that occurred with ARIA-E tended to occur early (within 6 months). Isolated ARIA-H occurred throughout the trial. ARIA-E and ARIA-H were numerically less common among ApoE ε4 noncarriers than among carriers, with higher frequency among ApoE ε4 homozygotes than among ApoE ε4 heterozygotes (Table 3).

Discussion

In this phase 3 trial, the change from baseline at 18 months in the CDR-SB score (primary end point) was less with lecanemab than with placebo, favoring lecanemab. Results for secondary clinical end points were in the same direction as those for the primary end point. Lecanemab has high selectivity for soluble aggregated species of Aβ as compared with monomeric amyloid, with moderate selectivity for fibrillar amyloid; this profile is considered to target the most toxic pathologic amyloid species.4,7,8,13,14 After 18 months of treatment in the amyloid substudy, the mean amyloid level of 22.99 centiloids in the lecanemab group was below the threshold for amyloid positivity of approximately 30 centiloids, above which participants are considered to have elevated brain amyloid levels.25 In the CSF substudy and in plasma analyses involving the overall population, markers of amyloid, tau, neurodegeneration, and neuroinflammation (plasma GFAP) were reduced to a greater extent with lecanemab than with placebo, with the exception of NfL, which is less sensitive to neurodegeneration than the other markers and has a slower time course for change than the others.

A definition of clinically meaningful effects in the primary end point of the CDR-SB score has not been established; however, this trial exceeded the prospectively defined target, with an estimated treatment difference of 0.373 points on a scale range of 18, a baseline value of 3.2, and early Alzheimer’s disease typically characterized by a score of 0.5 to 6. In a prespecified exploratory and multiplicity-unadjusted analysis of the time to worsening (increase) of the global CDR score of at least 0.5 points on two consecutive visits, the hazard ratio for progression to the next stage of dementia numerically favored lecanemab over placebo. An open-label extension study of Clarity AD is ongoing to provide additional safety and efficacy data beyond 18 months.

In the lecanemab group, the incidence of ARIA-E was 12.6%, and the incidence of ARIA-H was 17.3%. These incidences compare with 9.9% and 10.7%, respectively, in the phase 2b trial of lecanemab, in which ApoE ε4 carriers were underrepresented in the group that received 10 mg per kilogram every 2 weeks.15 The incidence of ARIA, including symptomatic ARIA, was numerically lower than in similar clinical trials, but differences in the drugs used and in trial design do not allow direct comparisons.26,27 ARIA-E generally occurred in the first 3 months, was mild and asymptomatic, did not lead to discontinuation of lecanemab or placebo if mild, and resolved within 4 months. The incidences of both overall and symptomatic ARIA-E were highest among ApoE ε4 homozygotes.

Among the limitations of this trial is that it includes data for only 18 months of treatment; an open-label extension study is ongoing. The Clarity AD trial was conducted during the Covid-19 pandemic and encountered obstacles including missed doses, delayed assessments, and intercurrent illnesses. The dropout rate was 17.2%, and a sensitivity analysis that eval‎uated the effect of missed doses was consistent with the primary end-point analysis. An additional potential limitation was the use of modified intention-to-treat analysis without imputation of missing values. However, a sensitivity analysis that was conducted with the use of a standard intention-to-treat population with imputation yielded similar results. Finally, occurrences of ARIA may have caused participants and investigators to be aware of the trial-group assignments. We attempted to minimize this bias by making clinical raters unaware of the safety assessments and the trial-group assignments, and sensitivity analyses that were performed to examine the effect of ARIA on clinical outcomes showed that ARIA had no effect on the results. Additional trials of lecanemab include a 5-year phase 2 long-term extension trial (ClinicalTrials.gov number, NCT01767311) and a 4-year phase 3 long-term extension trial (NCT03887455) in early Alzheimer’s disease, the 4-year AHEAD 3-45 trial (NCT04468659) in preclinical Alzheimer’s disease, and the 4-year DIAN-TU (Dominantly Inherited Alzheimer Network Trials Unit) Next Generation trial (NCT05269394) in dominantly inherited Alzheimer’s disease.

In persons with early Alzheimer’s disease, lecanemab reduced brain amyloid levels and was associated with moderately less decline on clinical measures of cognition and function than placebo at 18 months but was associated with adverse events. Longer trials are warranted to determine the efficacy and safety of lecanemab in early Alzheimer’s disease.

Notes

This article was published on November 29, 2022, at NEJM.org.

A data sharing statement provided by the authors is available with the full text of this article at NEJM.org.

Supported by Eisai (regulatory sponsor), with partial funding by Biogen.

Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

We thank the trial participants and their families, as well as all the investigators and site staff who made the trial possible (see the Supplementary Appendix for a list of collaborators); the members of the data and safety monitoring board and the raters; Lars Lannfelt and the staff of BioArctic for their early research on lecanemab; the staff of the clinical research organization Worldwide Clinical Trials for their ongoing support in conducting the trial; and J. David Cox (Mayville Medical Communications) and Lisa Yarenis (Eisai) for writing and editing assistance with an earlier version of the manuscript, in accordance with Good Publication Practice 4 ethical guidelines.

Supplementary Material

Research Summary (nejmoa2212948_research-summary.pdf)

• Download
• 260.42 KB

Protocol (nejmoa2212948_protocol.pdf)

• Download
• 4.75 MB

Supplementary Appendix (nejmoa2212948_appendix.pdf)

• Download
• 783.04 KB

Disclosure Forms (nejmoa2212948_disclosures.pdf)

• Download
• 724.97 KB

Data Sharing Statement (nejmoa2212948_data-sharing.pdf)

• Download
• 69.01 KB

References

1.

Marasco RA. Current and evolving treatment strategies for the Alzheimer disease continuum. Am J Manag Care 2020;26:Suppl:S167-S176.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

2.

Long JM, Holtzman DM. Alzheimer disease: an update on pathobiology and treatment strategies. Cell 2019;179:312-339.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

3.

Avgerinos KI, Ferrucci L, Kapogiannis D. Effects of monoclonal antibodies against amyloid-β on clinical and biomarker outcomes and adverse event risks: a systematic review and meta-analysis of phase III RCTs in Alzheimer’s disease. Ageing Res Rev 2021;68:101339-101339.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

4.

Tucker S, Möller C, Tegerstedt K, et al. The murine version of BAN2401 (mAb158) selectively reduces amyloid-β protofibrils in brain and cerebrospinal fluid of tg-ArcSwe mice. J Alzheimers Dis 2015;43:575-588.

Crossref

PubMed

Web of Science

Google Scholar

5.

Sehlin D, Englund H, Simu B, et al. Large aggregates are the major soluble Aβ species in AD brain fractionated with density gradient ultracentrifugation. PLoS One 2012;7(2):e32014-e32014.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

Lecanemab in Early Alzheimer’s Disease

Authors: Christopher H. van Dyck, M.D., Chad J. Swanson, Ph.D., Paul Aisen, M.D., Randall J. Bateman, M.D., Christopher Chen, B.M., B.Ch., Michelle Gee, Ph.D., Michio Kanekiyo, M.S., +11, and Takeshi Iwatsubo, M.D.Author Info & Affiliations

Published November 29, 2022

N Engl J Med 2023;388:9-21

DOI: 10.1056/NEJMoa2212948

VOL. 388 NO. 1

Copyright © 2022

• • Abstract
  • Methods
  • Results
  • Discussion
  • Notes
  • Supplementary Material
  • References
• • Information & Authors
  • Metrics & Citations
  • View Options
  • References
  • Media
  • Tables
  • Share

AbstractBackground

The accumulation of soluble and insoluble aggregated amyloid-beta (Aβ) may initiate or potentiate pathologic processes in Alzheimer’s disease. Lecanemab, a humanized IgG1 monoclonal antibody that binds with high affinity to Aβ soluble protofibrils, is being tested in persons with early Alzheimer’s disease.

Methods

We conducted an 18-month, multicenter, double-blind, phase 3 trial involving persons 50 to 90 years of age with early Alzheimer’s disease (mild cognitive impairment or mild dementia due to Alzheimer’s disease) with evidence of amyloid on positron-emission tomography (PET) or by cerebrospinal fluid testing. Participants were randomly assigned in a 1:1 ratio to receive intravenous lecanemab (10 mg per kilogram of body weight every 2 weeks) or placebo. The primary end point was the change from baseline at 18 months in the score on the Clinical Dementia Rating–Sum of Boxes (CDR-SB; range, 0 to 18, with higher scores indicating greater impairment). Key secondary end points were the change in amyloid burden on PET, the score on the 14-item cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-cog14; range, 0 to 90; higher scores indicate greater impairment), the Alzheimer’s Disease Composite Score (ADCOMS; range, 0 to 1.97; higher scores indicate greater impairment), and the score on the Alzheimer’s Disease Cooperative Study–Activities of Daily Living Scale for Mild Cognitive Impairment (ADCS-MCI-ADL; range, 0 to 53; lower scores indicate greater impairment).

Download a PDF of the Research Summary.

Results

A total of 1795 participants were enrolled, with 898 assigned to receive lecanemab and 897 to receive placebo. The mean CDR-SB score at baseline was approximately 3.2 in both groups. The adjusted least-squares mean change from baseline at 18 months was 1.21 with lecanemab and 1.66 with placebo (difference, −0.45; 95% confidence interval [CI], −0.67 to −0.23; P<0.001). In a substudy involving 698 participants, there were greater reductions in brain amyloid burden with lecanemab than with placebo (difference, −59.1 centiloids; 95% CI, −62.6 to −55.6). Other mean differences between the two groups in the change from baseline favoring lecanemab were as follows: for the ADAS-cog14 score, −1.44 (95% CI, −2.27 to −0.61; P<0.001); for the ADCOMS, −0.050 (95% CI, −0.074 to −0.027; P<0.001); and for the ADCS-MCI-ADL score, 2.0 (95% CI, 1.2 to 2.8; P<0.001). Lecanemab resulted in infusion-related reactions in 26.4% of the participants and amyloid-related imaging abnormalities with edema or effusions in 12.6%.

Conclusions

Lecanemab reduced markers of amyloid in early Alzheimer’s disease and resulted in moderately less decline on measures of cognition and function than placebo at 18 months but was associated with adverse events. Longer trials are warranted to determine the efficacy and safety of lecanemab in early Alzheimer’s disease. (Funded by Eisai and Biogen; Clarity AD ClinicalTrials.gov number, NCT03887455.)

Quick Take

Lecanemab in Alzheimer’s Disease

1m 47s

Current therapeutic agents for Alzheimer’s disease–related dementia temporarily improve symptoms but do not alter the underlying disease course.1,2 Some evidence suggests that amyloid removal slows the progression of disease.3 One anti-amyloid antibody (aducanumab) has received accelerated approval from the Food and Drug Administration.

Lecanemab is a humanized monoclonal antibody that binds with high affinity to soluble amyloid-beta (Aβ) protofibrils, which have been shown to be more toxic to neurons than monomers or insoluble fibrils.4-14 A phase 2b, dose-finding trial involving 854 participants with early Alzheimer’s disease did not show a significant difference between lecanemab and placebo in a Bayesian analysis of 12-month change in a composite score (primary end point). However, analyses at 18 months showed dose- and time-dependent clearance of amyloid with lecanemab, and the drug was associated with less clinical decline on some measures than placebo. In that trial, intravenous administration of 10 mg of lecanemab per kilogram of body weight every 2 weeks was identified as an appropriate dose, with a 9.9% incidence (<3% symptomatic) of amyloid-related imaging abnormalities (ARIA) with edema or effusions (ARIA-E).15 We conducted a phase 3 trial (Clarity AD) to determine the safety and efficacy of lecanemab in participants with early Alzheimer’s disease.

MethodsTrial Design and Oversight

Clarity AD was an 18-month, multicenter, double-blind, placebo-controlled, parallel-group trial involving persons with early Alzheimer’s disease. Eligible participants were randomly assigned in a 1:1 ratio to receive intravenous lecanemab (10 mg per kilogram every 2 weeks) or placebo. The randomization was stratified according to clinical subgroup (mild cognitive impairment due to Alzheimer’s disease or mild Alzheimer’s disease–related dementia on the basis of the criteria noted below), the presence or absence of concomitant approved medication for symptoms of Alzheimer’s disease at baseline (e.g., acetylcholinesterase inhibitors, memantine, or both), apolipoprotein E (ApoE) ε4 carriers or noncarriers, and geographic region. During the trial, participants underwent serial blood testing for plasma biomarkers and could participate in three optional substudies that eval‎uated longitudinal changes in brain amyloid burden as measured by positron-emission tomography (PET), brain tau pathologic features as measured by PET, and cerebrospinal fluid (CSF) biomarkers of Alzheimer’s disease.

The trial was conducted in accordance with International Council for Harmonisation guidelines and the ethical principles of the Declaration of Helsinki. The trial was approved by the institutional review board or independent ethics committee at each center, and all the participants provided written informed consent. The sponsor Eisai designed the trial and analyzed the data in collaboration with the academic authors, provided lecanemab and placebo, provided funding for medical writing, and aided in drafting the manuscript. The sponsor could not delay or interdict publication. The first, second, and sixteenth authors wrote the first draft of the manuscript, with professional medical writing assistance funded by Eisai, and all the authors contributed to subsequent drafts. Confidentiality agreements were in place between the sponsor and the authors and site investigators. Biogen provided partial funding for the trial.

An independent data and safety monitoring board consisting of experts in Alzheimer’s disease and statistics reviewed unblinded safety data during the trial. An independent medical monitoring team, whose members were unaware of the trial-group assignments, reviewed ARIA, infusion-related reactions, and hypersensitivity reactions. Clinical assessment raters were unaware of the safety assessments and the trial-group assignments. All the authors vouch for the completeness and accuracy of the data, the fidelity of the trial to the protocol (available with the full text of this article at NEJM.org), and the full reporting of adverse events.

Eligibility Criteria

The trial included participants 50 to 90 years of age, with either mild cognitive impairment due to Alzheimer’s disease or mild Alzheimer’s disease–related dementia on the basis of National Institute on Aging–Alzheimer’s Association criteria.16,17 Amyloid positivity was determined by PET or CSF measurement of Aβ1–42. All the participants had objective impairment in episodic memory as indicated by at least 1 standard deviation below the age-adjusted mean in the Wechsler Memory Scale IV–Logical Memory II.

End Points

The primary efficacy end point was the change in the score on the Clinical Dementia Rating (CDR)–Sum of Boxes (CDR-SB)18 from baseline at 18 months. The CDR-SB score is a validated outcome measure used in clinical trials of Alzheimer’s disease that is obtained by interviewing patients and their care partners and captures cognition and function. It assesses six domains that patients and caregivers identify as important (Memory, Orientation, Judgment and Problem Solving, Community Affairs, Home and Hobbies, and Personal Care). Scores for each domain range from 0 to 3, with higher scores indicating greater impairment. Total scores range from 0 to 18, with a score of 0.5 to 6 indicating early Alzheimer’s disease.

Key secondary end points were the change from baseline at 18 months in the following: amyloid burden on PET as measured in centiloids (with either florbetaben, florbetapir, or flutemetamol tracers) in a substudy, the score on the 14-item cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-cog14; range, 0 to 90, with higher scores indicating greater impairment),19 the Alzheimer’s Disease Composite Score (ADCOMS; range, 0 to 1.97, with higher scores indicating greater impairment),20 and the score on the Alzheimer’s Disease Cooperative Study–Activities of Daily Living Scale for Mild Cognitive Impairment (ADCS-MCI-ADL; range, 0 to 53, with lower scores indicating greater impairment).21 Biomarker assessments included CSF biomarkers (Aβ1–40, Aβ1–42, total tau, phosphorylated tau 181 [p-tau181], neurogranin, and neurofilament light chain [NfL]) and plasma biomarkers (Aβ42/40 ratio, p-tau181, glial fibrillary acidic protein [GFAP], and NfL). Tau PET and volumetric magnetic resonance imaging (MRI) results have not been fully analyzed.

A prespecified exploratory and multiplicity-unadjusted analysis examined the time to worsening of the global CDR score (range, 0 to 3, with higher scores indicating greater impairment). This end point was defined as the time to the first increase of at least 0.5 points in the global CDR score on two consecutive visits.

Statistical Analysis

The sample size for this trial was estimated on the basis of comparison of lecanemab and placebo with respect to the primary efficacy end point, the change from baseline at 18 months in the CDR-SB score. On the basis of data from the phase 2b trial of lecanemab,15 the estimated standard deviation of the change from baseline at 18 months in the CDR-SB score with placebo was 2.031 points, and the estimated treatment difference between lecanemab and placebo in all the participants was 0.373 points. This estimation corresponds to 25% less decline in cognitive function with lecanemab than with placebo and is consistent with a clinically meaningful difference on the basis of the Alzheimer’s disease literature, statistical principles, and agreements with regulatory authorities.15,22–24 Therefore, under the assumption of an estimated 20% dropout rate at 18 months in this trial, a total sample size of 1566 participants, including 783 participants receiving lecanemab and 783 participants receiving placebo, would provide the trial with 90% power to detect the treatment difference with the use of a two-sample t-test at a two-sided alpha level of 0.05. The sample size was increased by 200 to account for participants who missed three or more consecutive doses during the initial 6-month peak period of coronavirus disease 2019 (Covid-19), in accordance with previous agreement with regulatory authorities. No interim analyses for futility or efficacy were planned or performed.

Efficacy analyses were performed in the modified intention-to-treat population, which was defined as the group of randomly assigned participants who received at least one dose of lecanemab or placebo and who had a baseline assessment and at least one postdose primary efficacy (CDR-SB) measurement. Sensitivity analyses across efficacy end points to assess the robustness of the primary analysis to missing data included rank analysis of covariance with imputation of missing values. Additional sensitivity analyses were performed to eval‎uate potential effects of functional unblinding due to ARIA and effects of missed doses due to Covid-19–related absences (see the Supplementary Appendix, available at NEJM.org). Safety was eval‎uated in the safety population, which was defined as the group of participants who received at least one dose of lecanemab or placebo. Safety eval‎uations included monitoring of adverse events, vital signs, physical examinations, clinical laboratory variables, and 12-lead electrocardiograms. Occurrences of ARIA were monitored throughout the trial by central reading of MRI performed at weeks 9, 13, 27, 53, and 79 as well as at the 3-month follow-up visit (week 91) for safety monitoring. In addition, the populations for the substudies of amyloid burden on PET, tau pathologic features on PET, and CSF biomarkers of Alzheimer’s disease were the groups of participants who received at least one dose of lecanemab or placebo and who underwent a baseline PET or CSF eval‎uation and at least one postdose eval‎uation.

The primary analysis was performed without imputation of missing values. The primary analysis of the change from baseline at 18 months in the CDR-SB score was performed to compare lecanemab and placebo with the use of a mixed model for repeated measures that included the baseline CDR-SB score as a covariate, with trial group, visit, stratification variables (i.e., clinical subgroup, use of medication for symptoms of Alzheimer’s disease at baseline [yes or no], ApoE ε4 carrier status [carriers or noncarriers], and geographic region [North America, Europe, and Asia–Pacific]), baseline CDR-SB score–by–visit interaction, and trial group–by–visit interaction as fixed effects. If the between-group difference in primary end-point results was significant, then key secondary end points were to be tested hierarchically in the following order: change from baseline at 18 months in amyloid burden on PET as measured in centiloids in the subgroup tested and change from baseline at 18 months in the ADAS-cog14 score, change from baseline at 18 months in the ADCOMS, and change from baseline at 18 months in the ADCS-MCI-ADL score, all in the modified intention-to-treat population. Each test was performed at an alpha level of 0.05 (two-sided) and was to be performed only if the preceding test was significant at a two-sided level of 0.05. Additional details on the design and analysis methods are provided in the Supplementary Appendix and protocol.

ResultsParticipants

A total of 5967 persons were screened and 1795 underwent randomization; 898 were assigned to receive lecanemab and 897 to receive placebo at 235 sites in North America, Europe, and Asia from March 2019 through March 2021. Of these participants, 729 (81.2%) in the lecanemab group and 757 (84.4%) in the placebo group completed the trial and had data available on the primary end point (Figure 1). The modified intention-to-treat population included 1734 participants (859 in the lecanemab group and 875 in the placebo group), and the safety population included all 1795 randomly assigned participants. Enrollment in three longitudinal substudies included 698 participants in the substudy of amyloid burden on PET, 257 in the study of tau pathologic features on PET, and 281 in the substudy of CSF biomarkers of Alzheimer’s disease. The baseline characteristics of the substudy groups were generally similar to those in the main analysis. This trial made efforts to enhance global enrollment of a diverse group of participants (20% non-White), including in the United States, where 6.1% and 28.1% of the 3638 screened participants and 4.5% and 22.5% of randomly assigned participants were Black and Hispanic, respectively. The characteristics of the participants at baseline were generally similar in the two trial groups (Table 1). These characteristics were similar to what has been observed in population studies involving persons with early Alzheimer’s disease, although there was an underrepresentation of Black persons in the United States and an overrepresentation of Hispanic persons in the United States. The representativeness of the trial population is shown in Table S1 in the Supplementary Appendix.

Figure 1

Screening, Randomization, and Follow-up.

Table 1

Characteristics of the Participants at Baseline (Modified Intention-to-Treat Population).

End-Point Results

The mean CDR-SB score at baseline was approximately 3.2 in both the lecanemab and placebo groups, findings consistent with early Alzheimer’s disease (score of 0.5 to 6). The adjusted mean change from baseline at 18 months in the CDR-SB score was 1.21 in the lecanemab group and 1.66 in the placebo group (difference, −0.45; 95% confidence interval [CI], −0.67 to −0.23; P<0.001) (Figure 2A and Table 2).

Figure 2

Primary and Key Secondary End Points.

Table 2

Primary and Secondary End Points (Modified Intention-to-Treat Population).

In the substudy of amyloid burden on PET (a key secondary end point) involving 698 participants, the mean amyloid level at baseline was 77.92 centiloids in the lecanemab group and 75.03 centiloids in the placebo group. The adjusted mean change from baseline at 18 months was −55.48 centiloids in the lecanemab group and 3.64 centiloids in the placebo group (difference, −59.12 centiloids; 95% CI, −62.64 to −55.60; P<0.001) (Figure 2B and Table 2). In the modified intention-to-treat population, the mean ADAS-cog14 scores at baseline were 24.45 in the lecanemab group and 24.37 in the placebo group. The adjusted mean change from baseline at 18 months in the ADAS-cog14 score was 4.14 in the lecanemab group and 5.58 in the placebo group (difference, −1.44; 95% CI, −2.27 to −0.61; P<0.001) (Figure 2C and Table 2). The mean ADCOMS in the modified intention-to-treat population at baseline was 0.398 in the lecanemab group and 0.400 in the placebo group. The adjusted mean change from baseline at 18 months in the ADCOMS was 0.164 in the lecanemab group and 0.214 in the placebo group (difference, −0.050; 95% CI, −0.074 to −0.027; P<0.001) (Figure 2D and Table 2). In the modified intention-to-treat population, the mean ADCS-MCI-ADL scores at baseline were 41.2 for lecanemab and 40.9 for placebo. The adjusted mean change from baseline at 18 months in the ADCS-MCI-ADL score was −3.5 in the lecanemab group and −5.5 in the placebo group (difference, 2.0; 95% CI, 1.2 to 2.8; P<0.001) (Figure 2E and Table 2).

For each of these assessments, separation of the trial groups was apparent by visual inspection of graphs at 3 months. However, no conclusions can be drawn because there was no prespecified plan for analysis that included adjustment of confidence intervals for multiple comparisons at any intermediate time point.

Sensitivity analyses of the CDR-SB score that eval‎uated the effect of Covid-19 (missed doses) and potential for bias from functional unblinding due to ARIA were generally consistent with the primary analysis (Table S2). Results were also consistent across key randomization strata, as well as for other factors that affect Alzheimer’s disease (Figs. S1 through S4). The exploratory subgroup analysis involving ApoE ε4 homozygotes (15% of the trial population) numerically favored lecanemab for the ADAS-cog14 and ADCS-MCI-ADL scores but not for the CDR-SB score and the ADCOMS. Results of prespecified analyses of end points involving CSF and plasma biomarkers showed numerical improvements for all assessments comparing lecanemab with placebo, with the exception of CSF NfL (Fig. S5). In a prespecified, multiplicity-unadjusted analysis of the time to worsening of the global CDR score, the hazard ratio for progression to the next stage of dementia (0.69) numerically favored lecanemab over placebo (Fig. S6).

Safety

Deaths occurred in 0.7% of the participants in the lecanemab group and 0.8% of those in the placebo group (Table 3). No deaths were considered by the investigators to be related to lecanemab or occurred with ARIA. Serious adverse events occurred in 14.0% of the participants in the lecanemab group and 11.3% of those in the placebo group. The most commonly reported serious adverse events were infusion-related reactions (in 1.2% of the participants in the lecanemab group and 0 participants in the placebo group), ARIA-E (in 0.8% and 0, respectively), atrial fibrillation (in 0.7% and 0.3%), syncope (in 0.7% and 0.1%), and angina pectoris (in 0.7% and 0). The overall incidence of adverse events was similar in the two groups (Table 3). Adverse events leading to discontinuation of the trial agent occurred in 6.9% of the participants in the lecanemab group and 2.9% of those in the placebo group. The most common adverse events (affecting >10% of the participants) in the lecanemab group were infusion-related reactions (26.4% with lecanemab and 7.4% with placebo); ARIA with cerebral microhemorrhages, cerebral macrohemorrhages, or superficial siderosis (ARIA-H; 17.3% with lecanemab and 9.0% with placebo); ARIA-E (12.6% with lecanemab and 1.7% with placebo); headache (11.1% with lecanemab and 8.1% with placebo); and falls (10.4% with lecanemab and 9.6% with placebo). Infusion-related reactions were largely mild to moderate (grade 1 or 2, 96%) and occurred with the first dose (75%). A total of 56% of the participants did not take preventative medications (i.e., nonsteroidal antiinflammatory drugs, antihistamines, or glucocorticoids) for infusion-related reactions. Of those who took preventative medications for subsequent doses, 63% did not have additional reactions.

Table 3

Adverse Events.

Events of ARIA-E with lecanemab were mostly mild to moderate (91%) on the basis of central reading of imaging with the use of protocol definitions. These events were mostly asymptomatic (78%), occurred during the first 3 months of the treatment period (71%), and resolved within 4 months after detection (81%). A total of 2.8% of the participants in the lecanemab group had symptomatic ARIA-E; commonly reported symptoms were headache, visual disturbance, and confusion. The incidence of isolated ARIA-H (i.e., ARIA-H in participants who did not also have ARIA-E) was 8.9% in the lecanemab group and 7.8% in the placebo group. The incidence of isolated symptomatic ARIA-H was 0.7% in the lecanemab group and 0.2% in the placebo group. The most common symptom associated with isolated symptomatic ARIA-H was dizziness. Macrohemorrhage occurred in 5 of 898 participants (0.6%) in the lecanemab group and 1 of 897 participants (0.1%) in the placebo group. ARIA-H that occurred with ARIA-E tended to occur early (within 6 months). Isolated ARIA-H occurred throughout the trial. ARIA-E and ARIA-H were numerically less common among ApoE ε4 noncarriers than among carriers, with higher frequency among ApoE ε4 homozygotes than among ApoE ε4 heterozygotes (Table 3).

Discussion

In this phase 3 trial, the change from baseline at 18 months in the CDR-SB score (primary end point) was less with lecanemab than with placebo, favoring lecanemab. Results for secondary clinical end points were in the same direction as those for the primary end point. Lecanemab has high selectivity for soluble aggregated species of Aβ as compared with monomeric amyloid, with moderate selectivity for fibrillar amyloid; this profile is considered to target the most toxic pathologic amyloid species.4,7,8,13,14 After 18 months of treatment in the amyloid substudy, the mean amyloid level of 22.99 centiloids in the lecanemab group was below the threshold for amyloid positivity of approximately 30 centiloids, above which participants are considered to have elevated brain amyloid levels.25 In the CSF substudy and in plasma analyses involving the overall population, markers of amyloid, tau, neurodegeneration, and neuroinflammation (plasma GFAP) were reduced to a greater extent with lecanemab than with placebo, with the exception of NfL, which is less sensitive to neurodegeneration than the other markers and has a slower time course for change than the others.

A definition of clinically meaningful effects in the primary end point of the CDR-SB score has not been established; however, this trial exceeded the prospectively defined target, with an estimated treatment difference of 0.373 points on a scale range of 18, a baseline value of 3.2, and early Alzheimer’s disease typically characterized by a score of 0.5 to 6. In a prespecified exploratory and multiplicity-unadjusted analysis of the time to worsening (increase) of the global CDR score of at least 0.5 points on two consecutive visits, the hazard ratio for progression to the next stage of dementia numerically favored lecanemab over placebo. An open-label extension study of Clarity AD is ongoing to provide additional safety and efficacy data beyond 18 months.

In the lecanemab group, the incidence of ARIA-E was 12.6%, and the incidence of ARIA-H was 17.3%. These incidences compare with 9.9% and 10.7%, respectively, in the phase 2b trial of lecanemab, in which ApoE ε4 carriers were underrepresented in the group that received 10 mg per kilogram every 2 weeks.15 The incidence of ARIA, including symptomatic ARIA, was numerically lower than in similar clinical trials, but differences in the drugs used and in trial design do not allow direct comparisons.26,27 ARIA-E generally occurred in the first 3 months, was mild and asymptomatic, did not lead to discontinuation of lecanemab or placebo if mild, and resolved within 4 months. The incidences of both overall and symptomatic ARIA-E were highest among ApoE ε4 homozygotes.

Among the limitations of this trial is that it includes data for only 18 months of treatment; an open-label extension study is ongoing. The Clarity AD trial was conducted during the Covid-19 pandemic and encountered obstacles including missed doses, delayed assessments, and intercurrent illnesses. The dropout rate was 17.2%, and a sensitivity analysis that eval‎uated the effect of missed doses was consistent with the primary end-point analysis. An additional potential limitation was the use of modified intention-to-treat analysis without imputation of missing values. However, a sensitivity analysis that was conducted with the use of a standard intention-to-treat population with imputation yielded similar results. Finally, occurrences of ARIA may have caused participants and investigators to be aware of the trial-group assignments. We attempted to minimize this bias by making clinical raters unaware of the safety assessments and the trial-group assignments, and sensitivity analyses that were performed to examine the effect of ARIA on clinical outcomes showed that ARIA had no effect on the results. Additional trials of lecanemab include a 5-year phase 2 long-term extension trial (ClinicalTrials.gov number, NCT01767311) and a 4-year phase 3 long-term extension trial (NCT03887455) in early Alzheimer’s disease, the 4-year AHEAD 3-45 trial (NCT04468659) in preclinical Alzheimer’s disease, and the 4-year DIAN-TU (Dominantly Inherited Alzheimer Network Trials Unit) Next Generation trial (NCT05269394) in dominantly inherited Alzheimer’s disease.

In persons with early Alzheimer’s disease, lecanemab reduced brain amyloid levels and was associated with moderately less decline on clinical measures of cognition and function than placebo at 18 months but was associated with adverse events. Longer trials are warranted to determine the efficacy and safety of lecanemab in early Alzheimer’s disease.

Notes

This article was published on November 29, 2022, at NEJM.org.

A data sharing statement provided by the authors is available with the full text of this article at NEJM.org.

Supported by Eisai (regulatory sponsor), with partial funding by Biogen.

Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

We thank the trial participants and their families, as well as all the investigators and site staff who made the trial possible (see the Supplementary Appendix for a list of collaborators); the members of the data and safety monitoring board and the raters; Lars Lannfelt and the staff of BioArctic for their early research on lecanemab; the staff of the clinical research organization Worldwide Clinical Trials for their ongoing support in conducting the trial; and J. David Cox (Mayville Medical Communications) and Lisa Yarenis (Eisai) for writing and editing assistance with an earlier version of the manuscript, in accordance with Good Publication Practice 4 ethical guidelines.

Supplementary Material

Research Summary (nejmoa2212948_research-summary.pdf)

• Download
• 260.42 KB

Protocol (nejmoa2212948_protocol.pdf)

• Download
• 4.75 MB

Supplementary Appendix (nejmoa2212948_appendix.pdf)

• Download
• 783.04 KB

Disclosure Forms (nejmoa2212948_disclosures.pdf)

• Download
• 724.97 KB

Data Sharing Statement (nejmoa2212948_data-sharing.pdf)

• Download
• 69.01 KB

References

1.

Marasco RA. Current and evolving treatment strategies for the Alzheimer disease continuum. Am J Manag Care 2020;26:Suppl:S167-S176.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

2.

Long JM, Holtzman DM. Alzheimer disease: an update on pathobiology and treatment strategies. Cell 2019;179:312-339.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

3.

Avgerinos KI, Ferrucci L, Kapogiannis D. Effects of monoclonal antibodies against amyloid-β on clinical and biomarker outcomes and adverse event risks: a systematic review and meta-analysis of phase III RCTs in Alzheimer’s disease. Ageing Res Rev 2021;68:101339-101339.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

4.

Tucker S, Möller C, Tegerstedt K, et al. The murine version of BAN2401 (mAb158) selectively reduces amyloid-β protofibrils in brain and cerebrospinal fluid of tg-ArcSwe mice. J Alzheimers Dis 2015;43:575-588.

Crossref

PubMed

Web of Science

Google Scholar

5.

Sehlin D, Englund H, Simu B, et al. Large aggregates are the major soluble Aβ species in AD brain fractionated with density gradient ultracentrifugation. PLoS One 2012;7(2):e32014-e32014.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

Lecanemab in Early Alzheimer’s Disease

Authors: Christopher H. van Dyck, M.D., Chad J. Swanson, Ph.D., Paul Aisen, M.D., Randall J. Bateman, M.D., Christopher Chen, B.M., B.Ch., Michelle Gee, Ph.D., Michio Kanekiyo, M.S., +11, and Takeshi Iwatsubo, M.D.Author Info & Affiliations

Published November 29, 2022

N Engl J Med 2023;388:9-21

DOI: 10.1056/NEJMoa2212948

VOL. 388 NO. 1

Copyright © 2022

• • Abstract
  • Methods
  • Results
  • Discussion
  • Notes
  • Supplementary Material
  • References
• • Information & Authors
  • Metrics & Citations
  • View Options
  • References
  • Media
  • Tables
  • Share

AbstractBackground

The accumulation of soluble and insoluble aggregated amyloid-beta (Aβ) may initiate or potentiate pathologic processes in Alzheimer’s disease. Lecanemab, a humanized IgG1 monoclonal antibody that binds with high affinity to Aβ soluble protofibrils, is being tested in persons with early Alzheimer’s disease.

Methods

We conducted an 18-month, multicenter, double-blind, phase 3 trial involving persons 50 to 90 years of age with early Alzheimer’s disease (mild cognitive impairment or mild dementia due to Alzheimer’s disease) with evidence of amyloid on positron-emission tomography (PET) or by cerebrospinal fluid testing. Participants were randomly assigned in a 1:1 ratio to receive intravenous lecanemab (10 mg per kilogram of body weight every 2 weeks) or placebo. The primary end point was the change from baseline at 18 months in the score on the Clinical Dementia Rating–Sum of Boxes (CDR-SB; range, 0 to 18, with higher scores indicating greater impairment). Key secondary end points were the change in amyloid burden on PET, the score on the 14-item cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-cog14; range, 0 to 90; higher scores indicate greater impairment), the Alzheimer’s Disease Composite Score (ADCOMS; range, 0 to 1.97; higher scores indicate greater impairment), and the score on the Alzheimer’s Disease Cooperative Study–Activities of Daily Living Scale for Mild Cognitive Impairment (ADCS-MCI-ADL; range, 0 to 53; lower scores indicate greater impairment).

Download a PDF of the Research Summary.

Results

A total of 1795 participants were enrolled, with 898 assigned to receive lecanemab and 897 to receive placebo. The mean CDR-SB score at baseline was approximately 3.2 in both groups. The adjusted least-squares mean change from baseline at 18 months was 1.21 with lecanemab and 1.66 with placebo (difference, −0.45; 95% confidence interval [CI], −0.67 to −0.23; P<0.001). In a substudy involving 698 participants, there were greater reductions in brain amyloid burden with lecanemab than with placebo (difference, −59.1 centiloids; 95% CI, −62.6 to −55.6). Other mean differences between the two groups in the change from baseline favoring lecanemab were as follows: for the ADAS-cog14 score, −1.44 (95% CI, −2.27 to −0.61; P<0.001); for the ADCOMS, −0.050 (95% CI, −0.074 to −0.027; P<0.001); and for the ADCS-MCI-ADL score, 2.0 (95% CI, 1.2 to 2.8; P<0.001). Lecanemab resulted in infusion-related reactions in 26.4% of the participants and amyloid-related imaging abnormalities with edema or effusions in 12.6%.

Conclusions

Lecanemab reduced markers of amyloid in early Alzheimer’s disease and resulted in moderately less decline on measures of cognition and function than placebo at 18 months but was associated with adverse events. Longer trials are warranted to determine the efficacy and safety of lecanemab in early Alzheimer’s disease. (Funded by Eisai and Biogen; Clarity AD ClinicalTrials.gov number, NCT03887455.)

Quick Take

Lecanemab in Alzheimer’s Disease

1m 47s

Current therapeutic agents for Alzheimer’s disease–related dementia temporarily improve symptoms but do not alter the underlying disease course.1,2 Some evidence suggests that amyloid removal slows the progression of disease.3 One anti-amyloid antibody (aducanumab) has received accelerated approval from the Food and Drug Administration.

Lecanemab is a humanized monoclonal antibody that binds with high affinity to soluble amyloid-beta (Aβ) protofibrils, which have been shown to be more toxic to neurons than monomers or insoluble fibrils.4-14 A phase 2b, dose-finding trial involving 854 participants with early Alzheimer’s disease did not show a significant difference between lecanemab and placebo in a Bayesian analysis of 12-month change in a composite score (primary end point). However, analyses at 18 months showed dose- and time-dependent clearance of amyloid with lecanemab, and the drug was associated with less clinical decline on some measures than placebo. In that trial, intravenous administration of 10 mg of lecanemab per kilogram of body weight every 2 weeks was identified as an appropriate dose, with a 9.9% incidence (<3% symptomatic) of amyloid-related imaging abnormalities (ARIA) with edema or effusions (ARIA-E).15 We conducted a phase 3 trial (Clarity AD) to determine the safety and efficacy of lecanemab in participants with early Alzheimer’s disease.

MethodsTrial Design and Oversight

Clarity AD was an 18-month, multicenter, double-blind, placebo-controlled, parallel-group trial involving persons with early Alzheimer’s disease. Eligible participants were randomly assigned in a 1:1 ratio to receive intravenous lecanemab (10 mg per kilogram every 2 weeks) or placebo. The randomization was stratified according to clinical subgroup (mild cognitive impairment due to Alzheimer’s disease or mild Alzheimer’s disease–related dementia on the basis of the criteria noted below), the presence or absence of concomitant approved medication for symptoms of Alzheimer’s disease at baseline (e.g., acetylcholinesterase inhibitors, memantine, or both), apolipoprotein E (ApoE) ε4 carriers or noncarriers, and geographic region. During the trial, participants underwent serial blood testing for plasma biomarkers and could participate in three optional substudies that eval‎uated longitudinal changes in brain amyloid burden as measured by positron-emission tomography (PET), brain tau pathologic features as measured by PET, and cerebrospinal fluid (CSF) biomarkers of Alzheimer’s disease.

The trial was conducted in accordance with International Council for Harmonisation guidelines and the ethical principles of the Declaration of Helsinki. The trial was approved by the institutional review board or independent ethics committee at each center, and all the participants provided written informed consent. The sponsor Eisai designed the trial and analyzed the data in collaboration with the academic authors, provided lecanemab and placebo, provided funding for medical writing, and aided in drafting the manuscript. The sponsor could not delay or interdict publication. The first, second, and sixteenth authors wrote the first draft of the manuscript, with professional medical writing assistance funded by Eisai, and all the authors contributed to subsequent drafts. Confidentiality agreements were in place between the sponsor and the authors and site investigators. Biogen provided partial funding for the trial.

An independent data and safety monitoring board consisting of experts in Alzheimer’s disease and statistics reviewed unblinded safety data during the trial. An independent medical monitoring team, whose members were unaware of the trial-group assignments, reviewed ARIA, infusion-related reactions, and hypersensitivity reactions. Clinical assessment raters were unaware of the safety assessments and the trial-group assignments. All the authors vouch for the completeness and accuracy of the data, the fidelity of the trial to the protocol (available with the full text of this article at NEJM.org), and the full reporting of adverse events.

Eligibility Criteria

The trial included participants 50 to 90 years of age, with either mild cognitive impairment due to Alzheimer’s disease or mild Alzheimer’s disease–related dementia on the basis of National Institute on Aging–Alzheimer’s Association criteria.16,17 Amyloid positivity was determined by PET or CSF measurement of Aβ1–42. All the participants had objective impairment in episodic memory as indicated by at least 1 standard deviation below the age-adjusted mean in the Wechsler Memory Scale IV–Logical Memory II.

End Points

The primary efficacy end point was the change in the score on the Clinical Dementia Rating (CDR)–Sum of Boxes (CDR-SB)18 from baseline at 18 months. The CDR-SB score is a validated outcome measure used in clinical trials of Alzheimer’s disease that is obtained by interviewing patients and their care partners and captures cognition and function. It assesses six domains that patients and caregivers identify as important (Memory, Orientation, Judgment and Problem Solving, Community Affairs, Home and Hobbies, and Personal Care). Scores for each domain range from 0 to 3, with higher scores indicating greater impairment. Total scores range from 0 to 18, with a score of 0.5 to 6 indicating early Alzheimer’s disease.

Key secondary end points were the change from baseline at 18 months in the following: amyloid burden on PET as measured in centiloids (with either florbetaben, florbetapir, or flutemetamol tracers) in a substudy, the score on the 14-item cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-cog14; range, 0 to 90, with higher scores indicating greater impairment),19 the Alzheimer’s Disease Composite Score (ADCOMS; range, 0 to 1.97, with higher scores indicating greater impairment),20 and the score on the Alzheimer’s Disease Cooperative Study–Activities of Daily Living Scale for Mild Cognitive Impairment (ADCS-MCI-ADL; range, 0 to 53, with lower scores indicating greater impairment).21 Biomarker assessments included CSF biomarkers (Aβ1–40, Aβ1–42, total tau, phosphorylated tau 181 [p-tau181], neurogranin, and neurofilament light chain [NfL]) and plasma biomarkers (Aβ42/40 ratio, p-tau181, glial fibrillary acidic protein [GFAP], and NfL). Tau PET and volumetric magnetic resonance imaging (MRI) results have not been fully analyzed.

A prespecified exploratory and multiplicity-unadjusted analysis examined the time to worsening of the global CDR score (range, 0 to 3, with higher scores indicating greater impairment). This end point was defined as the time to the first increase of at least 0.5 points in the global CDR score on two consecutive visits.

Statistical Analysis

The sample size for this trial was estimated on the basis of comparison of lecanemab and placebo with respect to the primary efficacy end point, the change from baseline at 18 months in the CDR-SB score. On the basis of data from the phase 2b trial of lecanemab,15 the estimated standard deviation of the change from baseline at 18 months in the CDR-SB score with placebo was 2.031 points, and the estimated treatment difference between lecanemab and placebo in all the participants was 0.373 points. This estimation corresponds to 25% less decline in cognitive function with lecanemab than with placebo and is consistent with a clinically meaningful difference on the basis of the Alzheimer’s disease literature, statistical principles, and agreements with regulatory authorities.15,22–24 Therefore, under the assumption of an estimated 20% dropout rate at 18 months in this trial, a total sample size of 1566 participants, including 783 participants receiving lecanemab and 783 participants receiving placebo, would provide the trial with 90% power to detect the treatment difference with the use of a two-sample t-test at a two-sided alpha level of 0.05. The sample size was increased by 200 to account for participants who missed three or more consecutive doses during the initial 6-month peak period of coronavirus disease 2019 (Covid-19), in accordance with previous agreement with regulatory authorities. No interim analyses for futility or efficacy were planned or performed.

Efficacy analyses were performed in the modified intention-to-treat population, which was defined as the group of randomly assigned participants who received at least one dose of lecanemab or placebo and who had a baseline assessment and at least one postdose primary efficacy (CDR-SB) measurement. Sensitivity analyses across efficacy end points to assess the robustness of the primary analysis to missing data included rank analysis of covariance with imputation of missing values. Additional sensitivity analyses were performed to eval‎uate potential effects of functional unblinding due to ARIA and effects of missed doses due to Covid-19–related absences (see the Supplementary Appendix, available at NEJM.org). Safety was eval‎uated in the safety population, which was defined as the group of participants who received at least one dose of lecanemab or placebo. Safety eval‎uations included monitoring of adverse events, vital signs, physical examinations, clinical laboratory variables, and 12-lead electrocardiograms. Occurrences of ARIA were monitored throughout the trial by central reading of MRI performed at weeks 9, 13, 27, 53, and 79 as well as at the 3-month follow-up visit (week 91) for safety monitoring. In addition, the populations for the substudies of amyloid burden on PET, tau pathologic features on PET, and CSF biomarkers of Alzheimer’s disease were the groups of participants who received at least one dose of lecanemab or placebo and who underwent a baseline PET or CSF eval‎uation and at least one postdose eval‎uation.

The primary analysis was performed without imputation of missing values. The primary analysis of the change from baseline at 18 months in the CDR-SB score was performed to compare lecanemab and placebo with the use of a mixed model for repeated measures that included the baseline CDR-SB score as a covariate, with trial group, visit, stratification variables (i.e., clinical subgroup, use of medication for symptoms of Alzheimer’s disease at baseline [yes or no], ApoE ε4 carrier status [carriers or noncarriers], and geographic region [North America, Europe, and Asia–Pacific]), baseline CDR-SB score–by–visit interaction, and trial group–by–visit interaction as fixed effects. If the between-group difference in primary end-point results was significant, then key secondary end points were to be tested hierarchically in the following order: change from baseline at 18 months in amyloid burden on PET as measured in centiloids in the subgroup tested and change from baseline at 18 months in the ADAS-cog14 score, change from baseline at 18 months in the ADCOMS, and change from baseline at 18 months in the ADCS-MCI-ADL score, all in the modified intention-to-treat population. Each test was performed at an alpha level of 0.05 (two-sided) and was to be performed only if the preceding test was significant at a two-sided level of 0.05. Additional details on the design and analysis methods are provided in the Supplementary Appendix and protocol.

ResultsParticipants

A total of 5967 persons were screened and 1795 underwent randomization; 898 were assigned to receive lecanemab and 897 to receive placebo at 235 sites in North America, Europe, and Asia from March 2019 through March 2021. Of these participants, 729 (81.2%) in the lecanemab group and 757 (84.4%) in the placebo group completed the trial and had data available on the primary end point (Figure 1). The modified intention-to-treat population included 1734 participants (859 in the lecanemab group and 875 in the placebo group), and the safety population included all 1795 randomly assigned participants. Enrollment in three longitudinal substudies included 698 participants in the substudy of amyloid burden on PET, 257 in the study of tau pathologic features on PET, and 281 in the substudy of CSF biomarkers of Alzheimer’s disease. The baseline characteristics of the substudy groups were generally similar to those in the main analysis. This trial made efforts to enhance global enrollment of a diverse group of participants (20% non-White), including in the United States, where 6.1% and 28.1% of the 3638 screened participants and 4.5% and 22.5% of randomly assigned participants were Black and Hispanic, respectively. The characteristics of the participants at baseline were generally similar in the two trial groups (Table 1). These characteristics were similar to what has been observed in population studies involving persons with early Alzheimer’s disease, although there was an underrepresentation of Black persons in the United States and an overrepresentation of Hispanic persons in the United States. The representativeness of the trial population is shown in Table S1 in the Supplementary Appendix.

Figure 1

Screening, Randomization, and Follow-up.

Table 1

Characteristics of the Participants at Baseline (Modified Intention-to-Treat Population).

End-Point Results

The mean CDR-SB score at baseline was approximately 3.2 in both the lecanemab and placebo groups, findings consistent with early Alzheimer’s disease (score of 0.5 to 6). The adjusted mean change from baseline at 18 months in the CDR-SB score was 1.21 in the lecanemab group and 1.66 in the placebo group (difference, −0.45; 95% confidence interval [CI], −0.67 to −0.23; P<0.001) (Figure 2A and Table 2).

Figure 2

Primary and Key Secondary End Points.

Table 2

Primary and Secondary End Points (Modified Intention-to-Treat Population).

In the substudy of amyloid burden on PET (a key secondary end point) involving 698 participants, the mean amyloid level at baseline was 77.92 centiloids in the lecanemab group and 75.03 centiloids in the placebo group. The adjusted mean change from baseline at 18 months was −55.48 centiloids in the lecanemab group and 3.64 centiloids in the placebo group (difference, −59.12 centiloids; 95% CI, −62.64 to −55.60; P<0.001) (Figure 2B and Table 2). In the modified intention-to-treat population, the mean ADAS-cog14 scores at baseline were 24.45 in the lecanemab group and 24.37 in the placebo group. The adjusted mean change from baseline at 18 months in the ADAS-cog14 score was 4.14 in the lecanemab group and 5.58 in the placebo group (difference, −1.44; 95% CI, −2.27 to −0.61; P<0.001) (Figure 2C and Table 2). The mean ADCOMS in the modified intention-to-treat population at baseline was 0.398 in the lecanemab group and 0.400 in the placebo group. The adjusted mean change from baseline at 18 months in the ADCOMS was 0.164 in the lecanemab group and 0.214 in the placebo group (difference, −0.050; 95% CI, −0.074 to −0.027; P<0.001) (Figure 2D and Table 2). In the modified intention-to-treat population, the mean ADCS-MCI-ADL scores at baseline were 41.2 for lecanemab and 40.9 for placebo. The adjusted mean change from baseline at 18 months in the ADCS-MCI-ADL score was −3.5 in the lecanemab group and −5.5 in the placebo group (difference, 2.0; 95% CI, 1.2 to 2.8; P<0.001) (Figure 2E and Table 2).

For each of these assessments, separation of the trial groups was apparent by visual inspection of graphs at 3 months. However, no conclusions can be drawn because there was no prespecified plan for analysis that included adjustment of confidence intervals for multiple comparisons at any intermediate time point.

Sensitivity analyses of the CDR-SB score that eval‎uated the effect of Covid-19 (missed doses) and potential for bias from functional unblinding due to ARIA were generally consistent with the primary analysis (Table S2). Results were also consistent across key randomization strata, as well as for other factors that affect Alzheimer’s disease (Figs. S1 through S4). The exploratory subgroup analysis involving ApoE ε4 homozygotes (15% of the trial population) numerically favored lecanemab for the ADAS-cog14 and ADCS-MCI-ADL scores but not for the CDR-SB score and the ADCOMS. Results of prespecified analyses of end points involving CSF and plasma biomarkers showed numerical improvements for all assessments comparing lecanemab with placebo, with the exception of CSF NfL (Fig. S5). In a prespecified, multiplicity-unadjusted analysis of the time to worsening of the global CDR score, the hazard ratio for progression to the next stage of dementia (0.69) numerically favored lecanemab over placebo (Fig. S6).

Safety

Deaths occurred in 0.7% of the participants in the lecanemab group and 0.8% of those in the placebo group (Table 3). No deaths were considered by the investigators to be related to lecanemab or occurred with ARIA. Serious adverse events occurred in 14.0% of the participants in the lecanemab group and 11.3% of those in the placebo group. The most commonly reported serious adverse events were infusion-related reactions (in 1.2% of the participants in the lecanemab group and 0 participants in the placebo group), ARIA-E (in 0.8% and 0, respectively), atrial fibrillation (in 0.7% and 0.3%), syncope (in 0.7% and 0.1%), and angina pectoris (in 0.7% and 0). The overall incidence of adverse events was similar in the two groups (Table 3). Adverse events leading to discontinuation of the trial agent occurred in 6.9% of the participants in the lecanemab group and 2.9% of those in the placebo group. The most common adverse events (affecting >10% of the participants) in the lecanemab group were infusion-related reactions (26.4% with lecanemab and 7.4% with placebo); ARIA with cerebral microhemorrhages, cerebral macrohemorrhages, or superficial siderosis (ARIA-H; 17.3% with lecanemab and 9.0% with placebo); ARIA-E (12.6% with lecanemab and 1.7% with placebo); headache (11.1% with lecanemab and 8.1% with placebo); and falls (10.4% with lecanemab and 9.6% with placebo). Infusion-related reactions were largely mild to moderate (grade 1 or 2, 96%) and occurred with the first dose (75%). A total of 56% of the participants did not take preventative medications (i.e., nonsteroidal antiinflammatory drugs, antihistamines, or glucocorticoids) for infusion-related reactions. Of those who took preventative medications for subsequent doses, 63% did not have additional reactions.

Table 3

Adverse Events.

Events of ARIA-E with lecanemab were mostly mild to moderate (91%) on the basis of central reading of imaging with the use of protocol definitions. These events were mostly asymptomatic (78%), occurred during the first 3 months of the treatment period (71%), and resolved within 4 months after detection (81%). A total of 2.8% of the participants in the lecanemab group had symptomatic ARIA-E; commonly reported symptoms were headache, visual disturbance, and confusion. The incidence of isolated ARIA-H (i.e., ARIA-H in participants who did not also have ARIA-E) was 8.9% in the lecanemab group and 7.8% in the placebo group. The incidence of isolated symptomatic ARIA-H was 0.7% in the lecanemab group and 0.2% in the placebo group. The most common symptom associated with isolated symptomatic ARIA-H was dizziness. Macrohemorrhage occurred in 5 of 898 participants (0.6%) in the lecanemab group and 1 of 897 participants (0.1%) in the placebo group. ARIA-H that occurred with ARIA-E tended to occur early (within 6 months). Isolated ARIA-H occurred throughout the trial. ARIA-E and ARIA-H were numerically less common among ApoE ε4 noncarriers than among carriers, with higher frequency among ApoE ε4 homozygotes than among ApoE ε4 heterozygotes (Table 3).

Discussion

In this phase 3 trial, the change from baseline at 18 months in the CDR-SB score (primary end point) was less with lecanemab than with placebo, favoring lecanemab. Results for secondary clinical end points were in the same direction as those for the primary end point. Lecanemab has high selectivity for soluble aggregated species of Aβ as compared with monomeric amyloid, with moderate selectivity for fibrillar amyloid; this profile is considered to target the most toxic pathologic amyloid species.4,7,8,13,14 After 18 months of treatment in the amyloid substudy, the mean amyloid level of 22.99 centiloids in the lecanemab group was below the threshold for amyloid positivity of approximately 30 centiloids, above which participants are considered to have elevated brain amyloid levels.25 In the CSF substudy and in plasma analyses involving the overall population, markers of amyloid, tau, neurodegeneration, and neuroinflammation (plasma GFAP) were reduced to a greater extent with lecanemab than with placebo, with the exception of NfL, which is less sensitive to neurodegeneration than the other markers and has a slower time course for change than the others.

A definition of clinically meaningful effects in the primary end point of the CDR-SB score has not been established; however, this trial exceeded the prospectively defined target, with an estimated treatment difference of 0.373 points on a scale range of 18, a baseline value of 3.2, and early Alzheimer’s disease typically characterized by a score of 0.5 to 6. In a prespecified exploratory and multiplicity-unadjusted analysis of the time to worsening (increase) of the global CDR score of at least 0.5 points on two consecutive visits, the hazard ratio for progression to the next stage of dementia numerically favored lecanemab over placebo. An open-label extension study of Clarity AD is ongoing to provide additional safety and efficacy data beyond 18 months.

In the lecanemab group, the incidence of ARIA-E was 12.6%, and the incidence of ARIA-H was 17.3%. These incidences compare with 9.9% and 10.7%, respectively, in the phase 2b trial of lecanemab, in which ApoE ε4 carriers were underrepresented in the group that received 10 mg per kilogram every 2 weeks.15 The incidence of ARIA, including symptomatic ARIA, was numerically lower than in similar clinical trials, but differences in the drugs used and in trial design do not allow direct comparisons.26,27 ARIA-E generally occurred in the first 3 months, was mild and asymptomatic, did not lead to discontinuation of lecanemab or placebo if mild, and resolved within 4 months. The incidences of both overall and symptomatic ARIA-E were highest among ApoE ε4 homozygotes.

Among the limitations of this trial is that it includes data for only 18 months of treatment; an open-label extension study is ongoing. The Clarity AD trial was conducted during the Covid-19 pandemic and encountered obstacles including missed doses, delayed assessments, and intercurrent illnesses. The dropout rate was 17.2%, and a sensitivity analysis that eval‎uated the effect of missed doses was consistent with the primary end-point analysis. An additional potential limitation was the use of modified intention-to-treat analysis without imputation of missing values. However, a sensitivity analysis that was conducted with the use of a standard intention-to-treat population with imputation yielded similar results. Finally, occurrences of ARIA may have caused participants and investigators to be aware of the trial-group assignments. We attempted to minimize this bias by making clinical raters unaware of the safety assessments and the trial-group assignments, and sensitivity analyses that were performed to examine the effect of ARIA on clinical outcomes showed that ARIA had no effect on the results. Additional trials of lecanemab include a 5-year phase 2 long-term extension trial (ClinicalTrials.gov number, NCT01767311) and a 4-year phase 3 long-term extension trial (NCT03887455) in early Alzheimer’s disease, the 4-year AHEAD 3-45 trial (NCT04468659) in preclinical Alzheimer’s disease, and the 4-year DIAN-TU (Dominantly Inherited Alzheimer Network Trials Unit) Next Generation trial (NCT05269394) in dominantly inherited Alzheimer’s disease.

In persons with early Alzheimer’s disease, lecanemab reduced brain amyloid levels and was associated with moderately less decline on clinical measures of cognition and function than placebo at 18 months but was associated with adverse events. Longer trials are warranted to determine the efficacy and safety of lecanemab in early Alzheimer’s disease.

Notes

This article was published on November 29, 2022, at NEJM.org.

A data sharing statement provided by the authors is available with the full text of this article at NEJM.org.

Supported by Eisai (regulatory sponsor), with partial funding by Biogen.

Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

We thank the trial participants and their families, as well as all the investigators and site staff who made the trial possible (see the Supplementary Appendix for a list of collaborators); the members of the data and safety monitoring board and the raters; Lars Lannfelt and the staff of BioArctic for their early research on lecanemab; the staff of the clinical research organization Worldwide Clinical Trials for their ongoing support in conducting the trial; and J. David Cox (Mayville Medical Communications) and Lisa Yarenis (Eisai) for writing and editing assistance with an earlier version of the manuscript, in accordance with Good Publication Practice 4 ethical guidelines.

Supplementary Material

Research Summary (nejmoa2212948_research-summary.pdf)

• Download
• 260.42 KB

Protocol (nejmoa2212948_protocol.pdf)

• Download
• 4.75 MB

Supplementary Appendix (nejmoa2212948_appendix.pdf)

• Download
• 783.04 KB

Disclosure Forms (nejmoa2212948_disclosures.pdf)

• Download
• 724.97 KB

Data Sharing Statement (nejmoa2212948_data-sharing.pdf)

• Download
• 69.01 KB

References

1.

Marasco RA. Current and evolving treatment strategies for the Alzheimer disease continuum. Am J Manag Care 2020;26:Suppl:S167-S176.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

2.

Long JM, Holtzman DM. Alzheimer disease: an update on pathobiology and treatment strategies. Cell 2019;179:312-339.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

3.

Avgerinos KI, Ferrucci L, Kapogiannis D. Effects of monoclonal antibodies against amyloid-β on clinical and biomarker outcomes and adverse event risks: a systematic review and meta-analysis of phase III RCTs in Alzheimer’s disease. Ageing Res Rev 2021;68:101339-101339.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

4.

Tucker S, Möller C, Tegerstedt K, et al. The murine version of BAN2401 (mAb158) selectively reduces amyloid-β protofibrils in brain and cerebrospinal fluid of tg-ArcSwe mice. J Alzheimers Dis 2015;43:575-588.

Crossref

PubMed

Web of Science

Google Scholar

5.

Sehlin D, Englund H, Simu B, et al. Large aggregates are the major soluble Aβ species in AD brain fractionated with density gradient ultracentrifugation. PLoS One 2012;7(2):e32014-e32014.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

Lecanemab in Early Alzheimer’s Disease

Authors: Christopher H. van Dyck, M.D., Chad J. Swanson, Ph.D., Paul Aisen, M.D., Randall J. Bateman, M.D., Christopher Chen, B.M., B.Ch., Michelle Gee, Ph.D., Michio Kanekiyo, M.S., +11, and Takeshi Iwatsubo, M.D.Author Info & Affiliations

Published November 29, 2022

N Engl J Med 2023;388:9-21

DOI: 10.1056/NEJMoa2212948

VOL. 388 NO. 1

Copyright © 2022

• • Abstract
  • Methods
  • Results
  • Discussion
  • Notes
  • Supplementary Material
  • References
• • Information & Authors
  • Metrics & Citations
  • View Options
  • References
  • Media
  • Tables
  • Share

AbstractBackground

The accumulation of soluble and insoluble aggregated amyloid-beta (Aβ) may initiate or potentiate pathologic processes in Alzheimer’s disease. Lecanemab, a humanized IgG1 monoclonal antibody that binds with high affinity to Aβ soluble protofibrils, is being tested in persons with early Alzheimer’s disease.

Methods

We conducted an 18-month, multicenter, double-blind, phase 3 trial involving persons 50 to 90 years of age with early Alzheimer’s disease (mild cognitive impairment or mild dementia due to Alzheimer’s disease) with evidence of amyloid on positron-emission tomography (PET) or by cerebrospinal fluid testing. Participants were randomly assigned in a 1:1 ratio to receive intravenous lecanemab (10 mg per kilogram of body weight every 2 weeks) or placebo. The primary end point was the change from baseline at 18 months in the score on the Clinical Dementia Rating–Sum of Boxes (CDR-SB; range, 0 to 18, with higher scores indicating greater impairment). Key secondary end points were the change in amyloid burden on PET, the score on the 14-item cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-cog14; range, 0 to 90; higher scores indicate greater impairment), the Alzheimer’s Disease Composite Score (ADCOMS; range, 0 to 1.97; higher scores indicate greater impairment), and the score on the Alzheimer’s Disease Cooperative Study–Activities of Daily Living Scale for Mild Cognitive Impairment (ADCS-MCI-ADL; range, 0 to 53; lower scores indicate greater impairment).

Download a PDF of the Research Summary.

Results

A total of 1795 participants were enrolled, with 898 assigned to receive lecanemab and 897 to receive placebo. The mean CDR-SB score at baseline was approximately 3.2 in both groups. The adjusted least-squares mean change from baseline at 18 months was 1.21 with lecanemab and 1.66 with placebo (difference, −0.45; 95% confidence interval [CI], −0.67 to −0.23; P<0.001). In a substudy involving 698 participants, there were greater reductions in brain amyloid burden with lecanemab than with placebo (difference, −59.1 centiloids; 95% CI, −62.6 to −55.6). Other mean differences between the two groups in the change from baseline favoring lecanemab were as follows: for the ADAS-cog14 score, −1.44 (95% CI, −2.27 to −0.61; P<0.001); for the ADCOMS, −0.050 (95% CI, −0.074 to −0.027; P<0.001); and for the ADCS-MCI-ADL score, 2.0 (95% CI, 1.2 to 2.8; P<0.001). Lecanemab resulted in infusion-related reactions in 26.4% of the participants and amyloid-related imaging abnormalities with edema or effusions in 12.6%.

Conclusions

Lecanemab reduced markers of amyloid in early Alzheimer’s disease and resulted in moderately less decline on measures of cognition and function than placebo at 18 months but was associated with adverse events. Longer trials are warranted to determine the efficacy and safety of lecanemab in early Alzheimer’s disease. (Funded by Eisai and Biogen; Clarity AD ClinicalTrials.gov number, NCT03887455.)

Quick Take

Lecanemab in Alzheimer’s Disease

1m 47s

Current therapeutic agents for Alzheimer’s disease–related dementia temporarily improve symptoms but do not alter the underlying disease course.1,2 Some evidence suggests that amyloid removal slows the progression of disease.3 One anti-amyloid antibody (aducanumab) has received accelerated approval from the Food and Drug Administration.

Lecanemab is a humanized monoclonal antibody that binds with high affinity to soluble amyloid-beta (Aβ) protofibrils, which have been shown to be more toxic to neurons than monomers or insoluble fibrils.4-14 A phase 2b, dose-finding trial involving 854 participants with early Alzheimer’s disease did not show a significant difference between lecanemab and placebo in a Bayesian analysis of 12-month change in a composite score (primary end point). However, analyses at 18 months showed dose- and time-dependent clearance of amyloid with lecanemab, and the drug was associated with less clinical decline on some measures than placebo. In that trial, intravenous administration of 10 mg of lecanemab per kilogram of body weight every 2 weeks was identified as an appropriate dose, with a 9.9% incidence (<3% symptomatic) of amyloid-related imaging abnormalities (ARIA) with edema or effusions (ARIA-E).15 We conducted a phase 3 trial (Clarity AD) to determine the safety and efficacy of lecanemab in participants with early Alzheimer’s disease.

MethodsTrial Design and Oversight

Clarity AD was an 18-month, multicenter, double-blind, placebo-controlled, parallel-group trial involving persons with early Alzheimer’s disease. Eligible participants were randomly assigned in a 1:1 ratio to receive intravenous lecanemab (10 mg per kilogram every 2 weeks) or placebo. The randomization was stratified according to clinical subgroup (mild cognitive impairment due to Alzheimer’s disease or mild Alzheimer’s disease–related dementia on the basis of the criteria noted below), the presence or absence of concomitant approved medication for symptoms of Alzheimer’s disease at baseline (e.g., acetylcholinesterase inhibitors, memantine, or both), apolipoprotein E (ApoE) ε4 carriers or noncarriers, and geographic region. During the trial, participants underwent serial blood testing for plasma biomarkers and could participate in three optional substudies that eval‎uated longitudinal changes in brain amyloid burden as measured by positron-emission tomography (PET), brain tau pathologic features as measured by PET, and cerebrospinal fluid (CSF) biomarkers of Alzheimer’s disease.

The trial was conducted in accordance with International Council for Harmonisation guidelines and the ethical principles of the Declaration of Helsinki. The trial was approved by the institutional review board or independent ethics committee at each center, and all the participants provided written informed consent. The sponsor Eisai designed the trial and analyzed the data in collaboration with the academic authors, provided lecanemab and placebo, provided funding for medical writing, and aided in drafting the manuscript. The sponsor could not delay or interdict publication. The first, second, and sixteenth authors wrote the first draft of the manuscript, with professional medical writing assistance funded by Eisai, and all the authors contributed to subsequent drafts. Confidentiality agreements were in place between the sponsor and the authors and site investigators. Biogen provided partial funding for the trial.

An independent data and safety monitoring board consisting of experts in Alzheimer’s disease and statistics reviewed unblinded safety data during the trial. An independent medical monitoring team, whose members were unaware of the trial-group assignments, reviewed ARIA, infusion-related reactions, and hypersensitivity reactions. Clinical assessment raters were unaware of the safety assessments and the trial-group assignments. All the authors vouch for the completeness and accuracy of the data, the fidelity of the trial to the protocol (available with the full text of this article at NEJM.org), and the full reporting of adverse events.

Eligibility Criteria

The trial included participants 50 to 90 years of age, with either mild cognitive impairment due to Alzheimer’s disease or mild Alzheimer’s disease–related dementia on the basis of National Institute on Aging–Alzheimer’s Association criteria.16,17 Amyloid positivity was determined by PET or CSF measurement of Aβ1–42. All the participants had objective impairment in episodic memory as indicated by at least 1 standard deviation below the age-adjusted mean in the Wechsler Memory Scale IV–Logical Memory II.

End Points

The primary efficacy end point was the change in the score on the Clinical Dementia Rating (CDR)–Sum of Boxes (CDR-SB)18 from baseline at 18 months. The CDR-SB score is a validated outcome measure used in clinical trials of Alzheimer’s disease that is obtained by interviewing patients and their care partners and captures cognition and function. It assesses six domains that patients and caregivers identify as important (Memory, Orientation, Judgment and Problem Solving, Community Affairs, Home and Hobbies, and Personal Care). Scores for each domain range from 0 to 3, with higher scores indicating greater impairment. Total scores range from 0 to 18, with a score of 0.5 to 6 indicating early Alzheimer’s disease.

Key secondary end points were the change from baseline at 18 months in the following: amyloid burden on PET as measured in centiloids (with either florbetaben, florbetapir, or flutemetamol tracers) in a substudy, the score on the 14-item cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-cog14; range, 0 to 90, with higher scores indicating greater impairment),19 the Alzheimer’s Disease Composite Score (ADCOMS; range, 0 to 1.97, with higher scores indicating greater impairment),20 and the score on the Alzheimer’s Disease Cooperative Study–Activities of Daily Living Scale for Mild Cognitive Impairment (ADCS-MCI-ADL; range, 0 to 53, with lower scores indicating greater impairment).21 Biomarker assessments included CSF biomarkers (Aβ1–40, Aβ1–42, total tau, phosphorylated tau 181 [p-tau181], neurogranin, and neurofilament light chain [NfL]) and plasma biomarkers (Aβ42/40 ratio, p-tau181, glial fibrillary acidic protein [GFAP], and NfL). Tau PET and volumetric magnetic resonance imaging (MRI) results have not been fully analyzed.

A prespecified exploratory and multiplicity-unadjusted analysis examined the time to worsening of the global CDR score (range, 0 to 3, with higher scores indicating greater impairment). This end point was defined as the time to the first increase of at least 0.5 points in the global CDR score on two consecutive visits.

Statistical Analysis

The sample size for this trial was estimated on the basis of comparison of lecanemab and placebo with respect to the primary efficacy end point, the change from baseline at 18 months in the CDR-SB score. On the basis of data from the phase 2b trial of lecanemab,15 the estimated standard deviation of the change from baseline at 18 months in the CDR-SB score with placebo was 2.031 points, and the estimated treatment difference between lecanemab and placebo in all the participants was 0.373 points. This estimation corresponds to 25% less decline in cognitive function with lecanemab than with placebo and is consistent with a clinically meaningful difference on the basis of the Alzheimer’s disease literature, statistical principles, and agreements with regulatory authorities.15,22–24 Therefore, under the assumption of an estimated 20% dropout rate at 18 months in this trial, a total sample size of 1566 participants, including 783 participants receiving lecanemab and 783 participants receiving placebo, would provide the trial with 90% power to detect the treatment difference with the use of a two-sample t-test at a two-sided alpha level of 0.05. The sample size was increased by 200 to account for participants who missed three or more consecutive doses during the initial 6-month peak period of coronavirus disease 2019 (Covid-19), in accordance with previous agreement with regulatory authorities. No interim analyses for futility or efficacy were planned or performed.

Efficacy analyses were performed in the modified intention-to-treat population, which was defined as the group of randomly assigned participants who received at least one dose of lecanemab or placebo and who had a baseline assessment and at least one postdose primary efficacy (CDR-SB) measurement. Sensitivity analyses across efficacy end points to assess the robustness of the primary analysis to missing data included rank analysis of covariance with imputation of missing values. Additional sensitivity analyses were performed to eval‎uate potential effects of functional unblinding due to ARIA and effects of missed doses due to Covid-19–related absences (see the Supplementary Appendix, available at NEJM.org). Safety was eval‎uated in the safety population, which was defined as the group of participants who received at least one dose of lecanemab or placebo. Safety eval‎uations included monitoring of adverse events, vital signs, physical examinations, clinical laboratory variables, and 12-lead electrocardiograms. Occurrences of ARIA were monitored throughout the trial by central reading of MRI performed at weeks 9, 13, 27, 53, and 79 as well as at the 3-month follow-up visit (week 91) for safety monitoring. In addition, the populations for the substudies of amyloid burden on PET, tau pathologic features on PET, and CSF biomarkers of Alzheimer’s disease were the groups of participants who received at least one dose of lecanemab or placebo and who underwent a baseline PET or CSF eval‎uation and at least one postdose eval‎uation.

The primary analysis was performed without imputation of missing values. The primary analysis of the change from baseline at 18 months in the CDR-SB score was performed to compare lecanemab and placebo with the use of a mixed model for repeated measures that included the baseline CDR-SB score as a covariate, with trial group, visit, stratification variables (i.e., clinical subgroup, use of medication for symptoms of Alzheimer’s disease at baseline [yes or no], ApoE ε4 carrier status [carriers or noncarriers], and geographic region [North America, Europe, and Asia–Pacific]), baseline CDR-SB score–by–visit interaction, and trial group–by–visit interaction as fixed effects. If the between-group difference in primary end-point results was significant, then key secondary end points were to be tested hierarchically in the following order: change from baseline at 18 months in amyloid burden on PET as measured in centiloids in the subgroup tested and change from baseline at 18 months in the ADAS-cog14 score, change from baseline at 18 months in the ADCOMS, and change from baseline at 18 months in the ADCS-MCI-ADL score, all in the modified intention-to-treat population. Each test was performed at an alpha level of 0.05 (two-sided) and was to be performed only if the preceding test was significant at a two-sided level of 0.05. Additional details on the design and analysis methods are provided in the Supplementary Appendix and protocol.

ResultsParticipants

A total of 5967 persons were screened and 1795 underwent randomization; 898 were assigned to receive lecanemab and 897 to receive placebo at 235 sites in North America, Europe, and Asia from March 2019 through March 2021. Of these participants, 729 (81.2%) in the lecanemab group and 757 (84.4%) in the placebo group completed the trial and had data available on the primary end point (Figure 1). The modified intention-to-treat population included 1734 participants (859 in the lecanemab group and 875 in the placebo group), and the safety population included all 1795 randomly assigned participants. Enrollment in three longitudinal substudies included 698 participants in the substudy of amyloid burden on PET, 257 in the study of tau pathologic features on PET, and 281 in the substudy of CSF biomarkers of Alzheimer’s disease. The baseline characteristics of the substudy groups were generally similar to those in the main analysis. This trial made efforts to enhance global enrollment of a diverse group of participants (20% non-White), including in the United States, where 6.1% and 28.1% of the 3638 screened participants and 4.5% and 22.5% of randomly assigned participants were Black and Hispanic, respectively. The characteristics of the participants at baseline were generally similar in the two trial groups (Table 1). These characteristics were similar to what has been observed in population studies involving persons with early Alzheimer’s disease, although there was an underrepresentation of Black persons in the United States and an overrepresentation of Hispanic persons in the United States. The representativeness of the trial population is shown in Table S1 in the Supplementary Appendix.

Figure 1

Screening, Randomization, and Follow-up.

Table 1

Characteristics of the Participants at Baseline (Modified Intention-to-Treat Population).

End-Point Results

The mean CDR-SB score at baseline was approximately 3.2 in both the lecanemab and placebo groups, findings consistent with early Alzheimer’s disease (score of 0.5 to 6). The adjusted mean change from baseline at 18 months in the CDR-SB score was 1.21 in the lecanemab group and 1.66 in the placebo group (difference, −0.45; 95% confidence interval [CI], −0.67 to −0.23; P<0.001) (Figure 2A and Table 2).

Figure 2

Primary and Key Secondary End Points.

Table 2

Primary and Secondary End Points (Modified Intention-to-Treat Population).

In the substudy of amyloid burden on PET (a key secondary end point) involving 698 participants, the mean amyloid level at baseline was 77.92 centiloids in the lecanemab group and 75.03 centiloids in the placebo group. The adjusted mean change from baseline at 18 months was −55.48 centiloids in the lecanemab group and 3.64 centiloids in the placebo group (difference, −59.12 centiloids; 95% CI, −62.64 to −55.60; P<0.001) (Figure 2B and Table 2). In the modified intention-to-treat population, the mean ADAS-cog14 scores at baseline were 24.45 in the lecanemab group and 24.37 in the placebo group. The adjusted mean change from baseline at 18 months in the ADAS-cog14 score was 4.14 in the lecanemab group and 5.58 in the placebo group (difference, −1.44; 95% CI, −2.27 to −0.61; P<0.001) (Figure 2C and Table 2). The mean ADCOMS in the modified intention-to-treat population at baseline was 0.398 in the lecanemab group and 0.400 in the placebo group. The adjusted mean change from baseline at 18 months in the ADCOMS was 0.164 in the lecanemab group and 0.214 in the placebo group (difference, −0.050; 95% CI, −0.074 to −0.027; P<0.001) (Figure 2D and Table 2). In the modified intention-to-treat population, the mean ADCS-MCI-ADL scores at baseline were 41.2 for lecanemab and 40.9 for placebo. The adjusted mean change from baseline at 18 months in the ADCS-MCI-ADL score was −3.5 in the lecanemab group and −5.5 in the placebo group (difference, 2.0; 95% CI, 1.2 to 2.8; P<0.001) (Figure 2E and Table 2).

For each of these assessments, separation of the trial groups was apparent by visual inspection of graphs at 3 months. However, no conclusions can be drawn because there was no prespecified plan for analysis that included adjustment of confidence intervals for multiple comparisons at any intermediate time point.

Sensitivity analyses of the CDR-SB score that eval‎uated the effect of Covid-19 (missed doses) and potential for bias from functional unblinding due to ARIA were generally consistent with the primary analysis (Table S2). Results were also consistent across key randomization strata, as well as for other factors that affect Alzheimer’s disease (Figs. S1 through S4). The exploratory subgroup analysis involving ApoE ε4 homozygotes (15% of the trial population) numerically favored lecanemab for the ADAS-cog14 and ADCS-MCI-ADL scores but not for the CDR-SB score and the ADCOMS. Results of prespecified analyses of end points involving CSF and plasma biomarkers showed numerical improvements for all assessments comparing lecanemab with placebo, with the exception of CSF NfL (Fig. S5). In a prespecified, multiplicity-unadjusted analysis of the time to worsening of the global CDR score, the hazard ratio for progression to the next stage of dementia (0.69) numerically favored lecanemab over placebo (Fig. S6).

Safety

Deaths occurred in 0.7% of the participants in the lecanemab group and 0.8% of those in the placebo group (Table 3). No deaths were considered by the investigators to be related to lecanemab or occurred with ARIA. Serious adverse events occurred in 14.0% of the participants in the lecanemab group and 11.3% of those in the placebo group. The most commonly reported serious adverse events were infusion-related reactions (in 1.2% of the participants in the lecanemab group and 0 participants in the placebo group), ARIA-E (in 0.8% and 0, respectively), atrial fibrillation (in 0.7% and 0.3%), syncope (in 0.7% and 0.1%), and angina pectoris (in 0.7% and 0). The overall incidence of adverse events was similar in the two groups (Table 3). Adverse events leading to discontinuation of the trial agent occurred in 6.9% of the participants in the lecanemab group and 2.9% of those in the placebo group. The most common adverse events (affecting >10% of the participants) in the lecanemab group were infusion-related reactions (26.4% with lecanemab and 7.4% with placebo); ARIA with cerebral microhemorrhages, cerebral macrohemorrhages, or superficial siderosis (ARIA-H; 17.3% with lecanemab and 9.0% with placebo); ARIA-E (12.6% with lecanemab and 1.7% with placebo); headache (11.1% with lecanemab and 8.1% with placebo); and falls (10.4% with lecanemab and 9.6% with placebo). Infusion-related reactions were largely mild to moderate (grade 1 or 2, 96%) and occurred with the first dose (75%). A total of 56% of the participants did not take preventative medications (i.e., nonsteroidal antiinflammatory drugs, antihistamines, or glucocorticoids) for infusion-related reactions. Of those who took preventative medications for subsequent doses, 63% did not have additional reactions.

Table 3

Adverse Events.

Events of ARIA-E with lecanemab were mostly mild to moderate (91%) on the basis of central reading of imaging with the use of protocol definitions. These events were mostly asymptomatic (78%), occurred during the first 3 months of the treatment period (71%), and resolved within 4 months after detection (81%). A total of 2.8% of the participants in the lecanemab group had symptomatic ARIA-E; commonly reported symptoms were headache, visual disturbance, and confusion. The incidence of isolated ARIA-H (i.e., ARIA-H in participants who did not also have ARIA-E) was 8.9% in the lecanemab group and 7.8% in the placebo group. The incidence of isolated symptomatic ARIA-H was 0.7% in the lecanemab group and 0.2% in the placebo group. The most common symptom associated with isolated symptomatic ARIA-H was dizziness. Macrohemorrhage occurred in 5 of 898 participants (0.6%) in the lecanemab group and 1 of 897 participants (0.1%) in the placebo group. ARIA-H that occurred with ARIA-E tended to occur early (within 6 months). Isolated ARIA-H occurred throughout the trial. ARIA-E and ARIA-H were numerically less common among ApoE ε4 noncarriers than among carriers, with higher frequency among ApoE ε4 homozygotes than among ApoE ε4 heterozygotes (Table 3).

Discussion

In this phase 3 trial, the change from baseline at 18 months in the CDR-SB score (primary end point) was less with lecanemab than with placebo, favoring lecanemab. Results for secondary clinical end points were in the same direction as those for the primary end point. Lecanemab has high selectivity for soluble aggregated species of Aβ as compared with monomeric amyloid, with moderate selectivity for fibrillar amyloid; this profile is considered to target the most toxic pathologic amyloid species.4,7,8,13,14 After 18 months of treatment in the amyloid substudy, the mean amyloid level of 22.99 centiloids in the lecanemab group was below the threshold for amyloid positivity of approximately 30 centiloids, above which participants are considered to have elevated brain amyloid levels.25 In the CSF substudy and in plasma analyses involving the overall population, markers of amyloid, tau, neurodegeneration, and neuroinflammation (plasma GFAP) were reduced to a greater extent with lecanemab than with placebo, with the exception of NfL, which is less sensitive to neurodegeneration than the other markers and has a slower time course for change than the others.

A definition of clinically meaningful effects in the primary end point of the CDR-SB score has not been established; however, this trial exceeded the prospectively defined target, with an estimated treatment difference of 0.373 points on a scale range of 18, a baseline value of 3.2, and early Alzheimer’s disease typically characterized by a score of 0.5 to 6. In a prespecified exploratory and multiplicity-unadjusted analysis of the time to worsening (increase) of the global CDR score of at least 0.5 points on two consecutive visits, the hazard ratio for progression to the next stage of dementia numerically favored lecanemab over placebo. An open-label extension study of Clarity AD is ongoing to provide additional safety and efficacy data beyond 18 months.

In the lecanemab group, the incidence of ARIA-E was 12.6%, and the incidence of ARIA-H was 17.3%. These incidences compare with 9.9% and 10.7%, respectively, in the phase 2b trial of lecanemab, in which ApoE ε4 carriers were underrepresented in the group that received 10 mg per kilogram every 2 weeks.15 The incidence of ARIA, including symptomatic ARIA, was numerically lower than in similar clinical trials, but differences in the drugs used and in trial design do not allow direct comparisons.26,27 ARIA-E generally occurred in the first 3 months, was mild and asymptomatic, did not lead to discontinuation of lecanemab or placebo if mild, and resolved within 4 months. The incidences of both overall and symptomatic ARIA-E were highest among ApoE ε4 homozygotes.

Among the limitations of this trial is that it includes data for only 18 months of treatment; an open-label extension study is ongoing. The Clarity AD trial was conducted during the Covid-19 pandemic and encountered obstacles including missed doses, delayed assessments, and intercurrent illnesses. The dropout rate was 17.2%, and a sensitivity analysis that eval‎uated the effect of missed doses was consistent with the primary end-point analysis. An additional potential limitation was the use of modified intention-to-treat analysis without imputation of missing values. However, a sensitivity analysis that was conducted with the use of a standard intention-to-treat population with imputation yielded similar results. Finally, occurrences of ARIA may have caused participants and investigators to be aware of the trial-group assignments. We attempted to minimize this bias by making clinical raters unaware of the safety assessments and the trial-group assignments, and sensitivity analyses that were performed to examine the effect of ARIA on clinical outcomes showed that ARIA had no effect on the results. Additional trials of lecanemab include a 5-year phase 2 long-term extension trial (ClinicalTrials.gov number, NCT01767311) and a 4-year phase 3 long-term extension trial (NCT03887455) in early Alzheimer’s disease, the 4-year AHEAD 3-45 trial (NCT04468659) in preclinical Alzheimer’s disease, and the 4-year DIAN-TU (Dominantly Inherited Alzheimer Network Trials Unit) Next Generation trial (NCT05269394) in dominantly inherited Alzheimer’s disease.

In persons with early Alzheimer’s disease, lecanemab reduced brain amyloid levels and was associated with moderately less decline on clinical measures of cognition and function than placebo at 18 months but was associated with adverse events. Longer trials are warranted to determine the efficacy and safety of lecanemab in early Alzheimer’s disease.

Notes

This article was published on November 29, 2022, at NEJM.org.

A data sharing statement provided by the authors is available with the full text of this article at NEJM.org.

Supported by Eisai (regulatory sponsor), with partial funding by Biogen.

Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

We thank the trial participants and their families, as well as all the investigators and site staff who made the trial possible (see the Supplementary Appendix for a list of collaborators); the members of the data and safety monitoring board and the raters; Lars Lannfelt and the staff of BioArctic for their early research on lecanemab; the staff of the clinical research organization Worldwide Clinical Trials for their ongoing support in conducting the trial; and J. David Cox (Mayville Medical Communications) and Lisa Yarenis (Eisai) for writing and editing assistance with an earlier version of the manuscript, in accordance with Good Publication Practice 4 ethical guidelines.

Supplementary Material

Research Summary (nejmoa2212948_research-summary.pdf)

• Download
• 260.42 KB

Protocol (nejmoa2212948_protocol.pdf)

• Download
• 4.75 MB

Supplementary Appendix (nejmoa2212948_appendix.pdf)

• Download
• 783.04 KB

Disclosure Forms (nejmoa2212948_disclosures.pdf)

• Download
• 724.97 KB

Data Sharing Statement (nejmoa2212948_data-sharing.pdf)

• Download
• 69.01 KB

References

1.

Marasco RA. Current and evolving treatment strategies for the Alzheimer disease continuum. Am J Manag Care 2020;26:Suppl:S167-S176.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

2.

Long JM, Holtzman DM. Alzheimer disease: an update on pathobiology and treatment strategies. Cell 2019;179:312-339.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

3.

Avgerinos KI, Ferrucci L, Kapogiannis D. Effects of monoclonal antibodies against amyloid-β on clinical and biomarker outcomes and adverse event risks: a systematic review and meta-analysis of phase III RCTs in Alzheimer’s disease. Ageing Res Rev 2021;68:101339-101339.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

4.

Tucker S, Möller C, Tegerstedt K, et al. The murine version of BAN2401 (mAb158) selectively reduces amyloid-β protofibrils in brain and cerebrospinal fluid of tg-ArcSwe mice. J Alzheimers Dis 2015;43:575-588.

Crossref

PubMed

Web of Science

Google Scholar

5.

Sehlin D, Englund H, Simu B, et al. Large aggregates are the major soluble Aβ species in AD brain fractionated with density gradient ultracentrifugation. PLoS One 2012;7(2):e32014-e32014.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

Lecanemab in Early Alzheimer’s Disease

Authors: Christopher H. van Dyck, M.D., Chad J. Swanson, Ph.D., Paul Aisen, M.D., Randall J. Bateman, M.D., Christopher Chen, B.M., B.Ch., Michelle Gee, Ph.D., Michio Kanekiyo, M.S., +11, and Takeshi Iwatsubo, M.D.Author Info & Affiliations

Published November 29, 2022

N Engl J Med 2023;388:9-21

DOI: 10.1056/NEJMoa2212948

VOL. 388 NO. 1

Copyright © 2022

• • Abstract
  • Methods
  • Results
  • Discussion
  • Notes
  • Supplementary Material
  • References
• • Information & Authors
  • Metrics & Citations
  • View Options
  • References
  • Media
  • Tables
  • Share

AbstractBackground

The accumulation of soluble and insoluble aggregated amyloid-beta (Aβ) may initiate or potentiate pathologic processes in Alzheimer’s disease. Lecanemab, a humanized IgG1 monoclonal antibody that binds with high affinity to Aβ soluble protofibrils, is being tested in persons with early Alzheimer’s disease.

Methods

We conducted an 18-month, multicenter, double-blind, phase 3 trial involving persons 50 to 90 years of age with early Alzheimer’s disease (mild cognitive impairment or mild dementia due to Alzheimer’s disease) with evidence of amyloid on positron-emission tomography (PET) or by cerebrospinal fluid testing. Participants were randomly assigned in a 1:1 ratio to receive intravenous lecanemab (10 mg per kilogram of body weight every 2 weeks) or placebo. The primary end point was the change from baseline at 18 months in the score on the Clinical Dementia Rating–Sum of Boxes (CDR-SB; range, 0 to 18, with higher scores indicating greater impairment). Key secondary end points were the change in amyloid burden on PET, the score on the 14-item cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-cog14; range, 0 to 90; higher scores indicate greater impairment), the Alzheimer’s Disease Composite Score (ADCOMS; range, 0 to 1.97; higher scores indicate greater impairment), and the score on the Alzheimer’s Disease Cooperative Study–Activities of Daily Living Scale for Mild Cognitive Impairment (ADCS-MCI-ADL; range, 0 to 53; lower scores indicate greater impairment).

Download a PDF of the Research Summary.

Results

A total of 1795 participants were enrolled, with 898 assigned to receive lecanemab and 897 to receive placebo. The mean CDR-SB score at baseline was approximately 3.2 in both groups. The adjusted least-squares mean change from baseline at 18 months was 1.21 with lecanemab and 1.66 with placebo (difference, −0.45; 95% confidence interval [CI], −0.67 to −0.23; P<0.001). In a substudy involving 698 participants, there were greater reductions in brain amyloid burden with lecanemab than with placebo (difference, −59.1 centiloids; 95% CI, −62.6 to −55.6). Other mean differences between the two groups in the change from baseline favoring lecanemab were as follows: for the ADAS-cog14 score, −1.44 (95% CI, −2.27 to −0.61; P<0.001); for the ADCOMS, −0.050 (95% CI, −0.074 to −0.027; P<0.001); and for the ADCS-MCI-ADL score, 2.0 (95% CI, 1.2 to 2.8; P<0.001). Lecanemab resulted in infusion-related reactions in 26.4% of the participants and amyloid-related imaging abnormalities with edema or effusions in 12.6%.

Conclusions

Lecanemab reduced markers of amyloid in early Alzheimer’s disease and resulted in moderately less decline on measures of cognition and function than placebo at 18 months but was associated with adverse events. Longer trials are warranted to determine the efficacy and safety of lecanemab in early Alzheimer’s disease. (Funded by Eisai and Biogen; Clarity AD ClinicalTrials.gov number, NCT03887455.)

Quick Take

Lecanemab in Alzheimer’s Disease

1m 47s

Current therapeutic agents for Alzheimer’s disease–related dementia temporarily improve symptoms but do not alter the underlying disease course.1,2 Some evidence suggests that amyloid removal slows the progression of disease.3 One anti-amyloid antibody (aducanumab) has received accelerated approval from the Food and Drug Administration.

Lecanemab is a humanized monoclonal antibody that binds with high affinity to soluble amyloid-beta (Aβ) protofibrils, which have been shown to be more toxic to neurons than monomers or insoluble fibrils.4-14 A phase 2b, dose-finding trial involving 854 participants with early Alzheimer’s disease did not show a significant difference between lecanemab and placebo in a Bayesian analysis of 12-month change in a composite score (primary end point). However, analyses at 18 months showed dose- and time-dependent clearance of amyloid with lecanemab, and the drug was associated with less clinical decline on some measures than placebo. In that trial, intravenous administration of 10 mg of lecanemab per kilogram of body weight every 2 weeks was identified as an appropriate dose, with a 9.9% incidence (<3% symptomatic) of amyloid-related imaging abnormalities (ARIA) with edema or effusions (ARIA-E).15 We conducted a phase 3 trial (Clarity AD) to determine the safety and efficacy of lecanemab in participants with early Alzheimer’s disease.

MethodsTrial Design and Oversight

Clarity AD was an 18-month, multicenter, double-blind, placebo-controlled, parallel-group trial involving persons with early Alzheimer’s disease. Eligible participants were randomly assigned in a 1:1 ratio to receive intravenous lecanemab (10 mg per kilogram every 2 weeks) or placebo. The randomization was stratified according to clinical subgroup (mild cognitive impairment due to Alzheimer’s disease or mild Alzheimer’s disease–related dementia on the basis of the criteria noted below), the presence or absence of concomitant approved medication for symptoms of Alzheimer’s disease at baseline (e.g., acetylcholinesterase inhibitors, memantine, or both), apolipoprotein E (ApoE) ε4 carriers or noncarriers, and geographic region. During the trial, participants underwent serial blood testing for plasma biomarkers and could participate in three optional substudies that eval‎uated longitudinal changes in brain amyloid burden as measured by positron-emission tomography (PET), brain tau pathologic features as measured by PET, and cerebrospinal fluid (CSF) biomarkers of Alzheimer’s disease.

The trial was conducted in accordance with International Council for Harmonisation guidelines and the ethical principles of the Declaration of Helsinki. The trial was approved by the institutional review board or independent ethics committee at each center, and all the participants provided written informed consent. The sponsor Eisai designed the trial and analyzed the data in collaboration with the academic authors, provided lecanemab and placebo, provided funding for medical writing, and aided in drafting the manuscript. The sponsor could not delay or interdict publication. The first, second, and sixteenth authors wrote the first draft of the manuscript, with professional medical writing assistance funded by Eisai, and all the authors contributed to subsequent drafts. Confidentiality agreements were in place between the sponsor and the authors and site investigators. Biogen provided partial funding for the trial.

An independent data and safety monitoring board consisting of experts in Alzheimer’s disease and statistics reviewed unblinded safety data during the trial. An independent medical monitoring team, whose members were unaware of the trial-group assignments, reviewed ARIA, infusion-related reactions, and hypersensitivity reactions. Clinical assessment raters were unaware of the safety assessments and the trial-group assignments. All the authors vouch for the completeness and accuracy of the data, the fidelity of the trial to the protocol (available with the full text of this article at NEJM.org), and the full reporting of adverse events.

Eligibility Criteria

The trial included participants 50 to 90 years of age, with either mild cognitive impairment due to Alzheimer’s disease or mild Alzheimer’s disease–related dementia on the basis of National Institute on Aging–Alzheimer’s Association criteria.16,17 Amyloid positivity was determined by PET or CSF measurement of Aβ1–42. All the participants had objective impairment in episodic memory as indicated by at least 1 standard deviation below the age-adjusted mean in the Wechsler Memory Scale IV–Logical Memory II.

End Points

The primary efficacy end point was the change in the score on the Clinical Dementia Rating (CDR)–Sum of Boxes (CDR-SB)18 from baseline at 18 months. The CDR-SB score is a validated outcome measure used in clinical trials of Alzheimer’s disease that is obtained by interviewing patients and their care partners and captures cognition and function. It assesses six domains that patients and caregivers identify as important (Memory, Orientation, Judgment and Problem Solving, Community Affairs, Home and Hobbies, and Personal Care). Scores for each domain range from 0 to 3, with higher scores indicating greater impairment. Total scores range from 0 to 18, with a score of 0.5 to 6 indicating early Alzheimer’s disease.

Key secondary end points were the change from baseline at 18 months in the following: amyloid burden on PET as measured in centiloids (with either florbetaben, florbetapir, or flutemetamol tracers) in a substudy, the score on the 14-item cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-cog14; range, 0 to 90, with higher scores indicating greater impairment),19 the Alzheimer’s Disease Composite Score (ADCOMS; range, 0 to 1.97, with higher scores indicating greater impairment),20 and the score on the Alzheimer’s Disease Cooperative Study–Activities of Daily Living Scale for Mild Cognitive Impairment (ADCS-MCI-ADL; range, 0 to 53, with lower scores indicating greater impairment).21 Biomarker assessments included CSF biomarkers (Aβ1–40, Aβ1–42, total tau, phosphorylated tau 181 [p-tau181], neurogranin, and neurofilament light chain [NfL]) and plasma biomarkers (Aβ42/40 ratio, p-tau181, glial fibrillary acidic protein [GFAP], and NfL). Tau PET and volumetric magnetic resonance imaging (MRI) results have not been fully analyzed.

A prespecified exploratory and multiplicity-unadjusted analysis examined the time to worsening of the global CDR score (range, 0 to 3, with higher scores indicating greater impairment). This end point was defined as the time to the first increase of at least 0.5 points in the global CDR score on two consecutive visits.

Statistical Analysis

The sample size for this trial was estimated on the basis of comparison of lecanemab and placebo with respect to the primary efficacy end point, the change from baseline at 18 months in the CDR-SB score. On the basis of data from the phase 2b trial of lecanemab,15 the estimated standard deviation of the change from baseline at 18 months in the CDR-SB score with placebo was 2.031 points, and the estimated treatment difference between lecanemab and placebo in all the participants was 0.373 points. This estimation corresponds to 25% less decline in cognitive function with lecanemab than with placebo and is consistent with a clinically meaningful difference on the basis of the Alzheimer’s disease literature, statistical principles, and agreements with regulatory authorities.15,22–24 Therefore, under the assumption of an estimated 20% dropout rate at 18 months in this trial, a total sample size of 1566 participants, including 783 participants receiving lecanemab and 783 participants receiving placebo, would provide the trial with 90% power to detect the treatment difference with the use of a two-sample t-test at a two-sided alpha level of 0.05. The sample size was increased by 200 to account for participants who missed three or more consecutive doses during the initial 6-month peak period of coronavirus disease 2019 (Covid-19), in accordance with previous agreement with regulatory authorities. No interim analyses for futility or efficacy were planned or performed.

Efficacy analyses were performed in the modified intention-to-treat population, which was defined as the group of randomly assigned participants who received at least one dose of lecanemab or placebo and who had a baseline assessment and at least one postdose primary efficacy (CDR-SB) measurement. Sensitivity analyses across efficacy end points to assess the robustness of the primary analysis to missing data included rank analysis of covariance with imputation of missing values. Additional sensitivity analyses were performed to eval‎uate potential effects of functional unblinding due to ARIA and effects of missed doses due to Covid-19–related absences (see the Supplementary Appendix, available at NEJM.org). Safety was eval‎uated in the safety population, which was defined as the group of participants who received at least one dose of lecanemab or placebo. Safety eval‎uations included monitoring of adverse events, vital signs, physical examinations, clinical laboratory variables, and 12-lead electrocardiograms. Occurrences of ARIA were monitored throughout the trial by central reading of MRI performed at weeks 9, 13, 27, 53, and 79 as well as at the 3-month follow-up visit (week 91) for safety monitoring. In addition, the populations for the substudies of amyloid burden on PET, tau pathologic features on PET, and CSF biomarkers of Alzheimer’s disease were the groups of participants who received at least one dose of lecanemab or placebo and who underwent a baseline PET or CSF eval‎uation and at least one postdose eval‎uation.

The primary analysis was performed without imputation of missing values. The primary analysis of the change from baseline at 18 months in the CDR-SB score was performed to compare lecanemab and placebo with the use of a mixed model for repeated measures that included the baseline CDR-SB score as a covariate, with trial group, visit, stratification variables (i.e., clinical subgroup, use of medication for symptoms of Alzheimer’s disease at baseline [yes or no], ApoE ε4 carrier status [carriers or noncarriers], and geographic region [North America, Europe, and Asia–Pacific]), baseline CDR-SB score–by–visit interaction, and trial group–by–visit interaction as fixed effects. If the between-group difference in primary end-point results was significant, then key secondary end points were to be tested hierarchically in the following order: change from baseline at 18 months in amyloid burden on PET as measured in centiloids in the subgroup tested and change from baseline at 18 months in the ADAS-cog14 score, change from baseline at 18 months in the ADCOMS, and change from baseline at 18 months in the ADCS-MCI-ADL score, all in the modified intention-to-treat population. Each test was performed at an alpha level of 0.05 (two-sided) and was to be performed only if the preceding test was significant at a two-sided level of 0.05. Additional details on the design and analysis methods are provided in the Supplementary Appendix and protocol.

ResultsParticipants

A total of 5967 persons were screened and 1795 underwent randomization; 898 were assigned to receive lecanemab and 897 to receive placebo at 235 sites in North America, Europe, and Asia from March 2019 through March 2021. Of these participants, 729 (81.2%) in the lecanemab group and 757 (84.4%) in the placebo group completed the trial and had data available on the primary end point (Figure 1). The modified intention-to-treat population included 1734 participants (859 in the lecanemab group and 875 in the placebo group), and the safety population included all 1795 randomly assigned participants. Enrollment in three longitudinal substudies included 698 participants in the substudy of amyloid burden on PET, 257 in the study of tau pathologic features on PET, and 281 in the substudy of CSF biomarkers of Alzheimer’s disease. The baseline characteristics of the substudy groups were generally similar to those in the main analysis. This trial made efforts to enhance global enrollment of a diverse group of participants (20% non-White), including in the United States, where 6.1% and 28.1% of the 3638 screened participants and 4.5% and 22.5% of randomly assigned participants were Black and Hispanic, respectively. The characteristics of the participants at baseline were generally similar in the two trial groups (Table 1). These characteristics were similar to what has been observed in population studies involving persons with early Alzheimer’s disease, although there was an underrepresentation of Black persons in the United States and an overrepresentation of Hispanic persons in the United States. The representativeness of the trial population is shown in Table S1 in the Supplementary Appendix.

Figure 1

Screening, Randomization, and Follow-up.

Table 1

Characteristics of the Participants at Baseline (Modified Intention-to-Treat Population).

End-Point Results

The mean CDR-SB score at baseline was approximately 3.2 in both the lecanemab and placebo groups, findings consistent with early Alzheimer’s disease (score of 0.5 to 6). The adjusted mean change from baseline at 18 months in the CDR-SB score was 1.21 in the lecanemab group and 1.66 in the placebo group (difference, −0.45; 95% confidence interval [CI], −0.67 to −0.23; P<0.001) (Figure 2A and Table 2).

Figure 2

Primary and Key Secondary End Points.

Table 2

Primary and Secondary End Points (Modified Intention-to-Treat Population).

In the substudy of amyloid burden on PET (a key secondary end point) involving 698 participants, the mean amyloid level at baseline was 77.92 centiloids in the lecanemab group and 75.03 centiloids in the placebo group. The adjusted mean change from baseline at 18 months was −55.48 centiloids in the lecanemab group and 3.64 centiloids in the placebo group (difference, −59.12 centiloids; 95% CI, −62.64 to −55.60; P<0.001) (Figure 2B and Table 2). In the modified intention-to-treat population, the mean ADAS-cog14 scores at baseline were 24.45 in the lecanemab group and 24.37 in the placebo group. The adjusted mean change from baseline at 18 months in the ADAS-cog14 score was 4.14 in the lecanemab group and 5.58 in the placebo group (difference, −1.44; 95% CI, −2.27 to −0.61; P<0.001) (Figure 2C and Table 2). The mean ADCOMS in the modified intention-to-treat population at baseline was 0.398 in the lecanemab group and 0.400 in the placebo group. The adjusted mean change from baseline at 18 months in the ADCOMS was 0.164 in the lecanemab group and 0.214 in the placebo group (difference, −0.050; 95% CI, −0.074 to −0.027; P<0.001) (Figure 2D and Table 2). In the modified intention-to-treat population, the mean ADCS-MCI-ADL scores at baseline were 41.2 for lecanemab and 40.9 for placebo. The adjusted mean change from baseline at 18 months in the ADCS-MCI-ADL score was −3.5 in the lecanemab group and −5.5 in the placebo group (difference, 2.0; 95% CI, 1.2 to 2.8; P<0.001) (Figure 2E and Table 2).

For each of these assessments, separation of the trial groups was apparent by visual inspection of graphs at 3 months. However, no conclusions can be drawn because there was no prespecified plan for analysis that included adjustment of confidence intervals for multiple comparisons at any intermediate time point.

Sensitivity analyses of the CDR-SB score that eval‎uated the effect of Covid-19 (missed doses) and potential for bias from functional unblinding due to ARIA were generally consistent with the primary analysis (Table S2). Results were also consistent across key randomization strata, as well as for other factors that affect Alzheimer’s disease (Figs. S1 through S4). The exploratory subgroup analysis involving ApoE ε4 homozygotes (15% of the trial population) numerically favored lecanemab for the ADAS-cog14 and ADCS-MCI-ADL scores but not for the CDR-SB score and the ADCOMS. Results of prespecified analyses of end points involving CSF and plasma biomarkers showed numerical improvements for all assessments comparing lecanemab with placebo, with the exception of CSF NfL (Fig. S5). In a prespecified, multiplicity-unadjusted analysis of the time to worsening of the global CDR score, the hazard ratio for progression to the next stage of dementia (0.69) numerically favored lecanemab over placebo (Fig. S6).

Safety

Deaths occurred in 0.7% of the participants in the lecanemab group and 0.8% of those in the placebo group (Table 3). No deaths were considered by the investigators to be related to lecanemab or occurred with ARIA. Serious adverse events occurred in 14.0% of the participants in the lecanemab group and 11.3% of those in the placebo group. The most commonly reported serious adverse events were infusion-related reactions (in 1.2% of the participants in the lecanemab group and 0 participants in the placebo group), ARIA-E (in 0.8% and 0, respectively), atrial fibrillation (in 0.7% and 0.3%), syncope (in 0.7% and 0.1%), and angina pectoris (in 0.7% and 0). The overall incidence of adverse events was similar in the two groups (Table 3). Adverse events leading to discontinuation of the trial agent occurred in 6.9% of the participants in the lecanemab group and 2.9% of those in the placebo group. The most common adverse events (affecting >10% of the participants) in the lecanemab group were infusion-related reactions (26.4% with lecanemab and 7.4% with placebo); ARIA with cerebral microhemorrhages, cerebral macrohemorrhages, or superficial siderosis (ARIA-H; 17.3% with lecanemab and 9.0% with placebo); ARIA-E (12.6% with lecanemab and 1.7% with placebo); headache (11.1% with lecanemab and 8.1% with placebo); and falls (10.4% with lecanemab and 9.6% with placebo). Infusion-related reactions were largely mild to moderate (grade 1 or 2, 96%) and occurred with the first dose (75%). A total of 56% of the participants did not take preventative medications (i.e., nonsteroidal antiinflammatory drugs, antihistamines, or glucocorticoids) for infusion-related reactions. Of those who took preventative medications for subsequent doses, 63% did not have additional reactions.

Table 3

Adverse Events.

Events of ARIA-E with lecanemab were mostly mild to moderate (91%) on the basis of central reading of imaging with the use of protocol definitions. These events were mostly asymptomatic (78%), occurred during the first 3 months of the treatment period (71%), and resolved within 4 months after detection (81%). A total of 2.8% of the participants in the lecanemab group had symptomatic ARIA-E; commonly reported symptoms were headache, visual disturbance, and confusion. The incidence of isolated ARIA-H (i.e., ARIA-H in participants who did not also have ARIA-E) was 8.9% in the lecanemab group and 7.8% in the placebo group. The incidence of isolated symptomatic ARIA-H was 0.7% in the lecanemab group and 0.2% in the placebo group. The most common symptom associated with isolated symptomatic ARIA-H was dizziness. Macrohemorrhage occurred in 5 of 898 participants (0.6%) in the lecanemab group and 1 of 897 participants (0.1%) in the placebo group. ARIA-H that occurred with ARIA-E tended to occur early (within 6 months). Isolated ARIA-H occurred throughout the trial. ARIA-E and ARIA-H were numerically less common among ApoE ε4 noncarriers than among carriers, with higher frequency among ApoE ε4 homozygotes than among ApoE ε4 heterozygotes (Table 3).

Discussion

In this phase 3 trial, the change from baseline at 18 months in the CDR-SB score (primary end point) was less with lecanemab than with placebo, favoring lecanemab. Results for secondary clinical end points were in the same direction as those for the primary end point. Lecanemab has high selectivity for soluble aggregated species of Aβ as compared with monomeric amyloid, with moderate selectivity for fibrillar amyloid; this profile is considered to target the most toxic pathologic amyloid species.4,7,8,13,14 After 18 months of treatment in the amyloid substudy, the mean amyloid level of 22.99 centiloids in the lecanemab group was below the threshold for amyloid positivity of approximately 30 centiloids, above which participants are considered to have elevated brain amyloid levels.25 In the CSF substudy and in plasma analyses involving the overall population, markers of amyloid, tau, neurodegeneration, and neuroinflammation (plasma GFAP) were reduced to a greater extent with lecanemab than with placebo, with the exception of NfL, which is less sensitive to neurodegeneration than the other markers and has a slower time course for change than the others.

A definition of clinically meaningful effects in the primary end point of the CDR-SB score has not been established; however, this trial exceeded the prospectively defined target, with an estimated treatment difference of 0.373 points on a scale range of 18, a baseline value of 3.2, and early Alzheimer’s disease typically characterized by a score of 0.5 to 6. In a prespecified exploratory and multiplicity-unadjusted analysis of the time to worsening (increase) of the global CDR score of at least 0.5 points on two consecutive visits, the hazard ratio for progression to the next stage of dementia numerically favored lecanemab over placebo. An open-label extension study of Clarity AD is ongoing to provide additional safety and efficacy data beyond 18 months.

In the lecanemab group, the incidence of ARIA-E was 12.6%, and the incidence of ARIA-H was 17.3%. These incidences compare with 9.9% and 10.7%, respectively, in the phase 2b trial of lecanemab, in which ApoE ε4 carriers were underrepresented in the group that received 10 mg per kilogram every 2 weeks.15 The incidence of ARIA, including symptomatic ARIA, was numerically lower than in similar clinical trials, but differences in the drugs used and in trial design do not allow direct comparisons.26,27 ARIA-E generally occurred in the first 3 months, was mild and asymptomatic, did not lead to discontinuation of lecanemab or placebo if mild, and resolved within 4 months. The incidences of both overall and symptomatic ARIA-E were highest among ApoE ε4 homozygotes.

Among the limitations of this trial is that it includes data for only 18 months of treatment; an open-label extension study is ongoing. The Clarity AD trial was conducted during the Covid-19 pandemic and encountered obstacles including missed doses, delayed assessments, and intercurrent illnesses. The dropout rate was 17.2%, and a sensitivity analysis that eval‎uated the effect of missed doses was consistent with the primary end-point analysis. An additional potential limitation was the use of modified intention-to-treat analysis without imputation of missing values. However, a sensitivity analysis that was conducted with the use of a standard intention-to-treat population with imputation yielded similar results. Finally, occurrences of ARIA may have caused participants and investigators to be aware of the trial-group assignments. We attempted to minimize this bias by making clinical raters unaware of the safety assessments and the trial-group assignments, and sensitivity analyses that were performed to examine the effect of ARIA on clinical outcomes showed that ARIA had no effect on the results. Additional trials of lecanemab include a 5-year phase 2 long-term extension trial (ClinicalTrials.gov number, NCT01767311) and a 4-year phase 3 long-term extension trial (NCT03887455) in early Alzheimer’s disease, the 4-year AHEAD 3-45 trial (NCT04468659) in preclinical Alzheimer’s disease, and the 4-year DIAN-TU (Dominantly Inherited Alzheimer Network Trials Unit) Next Generation trial (NCT05269394) in dominantly inherited Alzheimer’s disease.

In persons with early Alzheimer’s disease, lecanemab reduced brain amyloid levels and was associated with moderately less decline on clinical measures of cognition and function than placebo at 18 months but was associated with adverse events. Longer trials are warranted to determine the efficacy and safety of lecanemab in early Alzheimer’s disease.

Notes

This article was published on November 29, 2022, at NEJM.org.

A data sharing statement provided by the authors is available with the full text of this article at NEJM.org.

Supported by Eisai (regulatory sponsor), with partial funding by Biogen.

Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

We thank the trial participants and their families, as well as all the investigators and site staff who made the trial possible (see the Supplementary Appendix for a list of collaborators); the members of the data and safety monitoring board and the raters; Lars Lannfelt and the staff of BioArctic for their early research on lecanemab; the staff of the clinical research organization Worldwide Clinical Trials for their ongoing support in conducting the trial; and J. David Cox (Mayville Medical Communications) and Lisa Yarenis (Eisai) for writing and editing assistance with an earlier version of the manuscript, in accordance with Good Publication Practice 4 ethical guidelines.

Supplementary Material

Research Summary (nejmoa2212948_research-summary.pdf)

• Download
• 260.42 KB

Protocol (nejmoa2212948_protocol.pdf)

• Download
• 4.75 MB

Supplementary Appendix (nejmoa2212948_appendix.pdf)

• Download
• 783.04 KB

Disclosure Forms (nejmoa2212948_disclosures.pdf)

• Download
• 724.97 KB

Data Sharing Statement (nejmoa2212948_data-sharing.pdf)

• Download
• 69.01 KB

References

1.

Marasco RA. Current and evolving treatment strategies for the Alzheimer disease continuum. Am J Manag Care 2020;26:Suppl:S167-S176.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

2.

Long JM, Holtzman DM. Alzheimer disease: an update on pathobiology and treatment strategies. Cell 2019;179:312-339.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

3.

Avgerinos KI, Ferrucci L, Kapogiannis D. Effects of monoclonal antibodies against amyloid-β on clinical and biomarker outcomes and adverse event risks: a systematic review and meta-analysis of phase III RCTs in Alzheimer’s disease. Ageing Res Rev 2021;68:101339-101339.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

4.

Tucker S, Möller C, Tegerstedt K, et al. The murine version of BAN2401 (mAb158) selectively reduces amyloid-β protofibrils in brain and cerebrospinal fluid of tg-ArcSwe mice. J Alzheimers Dis 2015;43:575-588.

Crossref

PubMed

Web of Science

Google Scholar

5.

Sehlin D, Englund H, Simu B, et al. Large aggregates are the major soluble Aβ species in AD brain fractionated with density gradient ultracentrifugation. PLoS One 2012;7(2):e32014-e32014.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

Lecanemab in Early Alzheimer’s Disease

Authors: Christopher H. van Dyck, M.D., Chad J. Swanson, Ph.D., Paul Aisen, M.D., Randall J. Bateman, M.D., Christopher Chen, B.M., B.Ch., Michelle Gee, Ph.D., Michio Kanekiyo, M.S., +11, and Takeshi Iwatsubo, M.D.Author Info & Affiliations

Published November 29, 2022

N Engl J Med 2023;388:9-21

DOI: 10.1056/NEJMoa2212948

VOL. 388 NO. 1

Copyright © 2022

• • Abstract
  • Methods
  • Results
  • Discussion
  • Notes
  • Supplementary Material
  • References
• • Information & Authors
  • Metrics & Citations
  • View Options
  • References
  • Media
  • Tables
  • Share

AbstractBackground

The accumulation of soluble and insoluble aggregated amyloid-beta (Aβ) may initiate or potentiate pathologic processes in Alzheimer’s disease. Lecanemab, a humanized IgG1 monoclonal antibody that binds with high affinity to Aβ soluble protofibrils, is being tested in persons with early Alzheimer’s disease.

Methods

We conducted an 18-month, multicenter, double-blind, phase 3 trial involving persons 50 to 90 years of age with early Alzheimer’s disease (mild cognitive impairment or mild dementia due to Alzheimer’s disease) with evidence of amyloid on positron-emission tomography (PET) or by cerebrospinal fluid testing. Participants were randomly assigned in a 1:1 ratio to receive intravenous lecanemab (10 mg per kilogram of body weight every 2 weeks) or placebo. The primary end point was the change from baseline at 18 months in the score on the Clinical Dementia Rating–Sum of Boxes (CDR-SB; range, 0 to 18, with higher scores indicating greater impairment). Key secondary end points were the change in amyloid burden on PET, the score on the 14-item cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-cog14; range, 0 to 90; higher scores indicate greater impairment), the Alzheimer’s Disease Composite Score (ADCOMS; range, 0 to 1.97; higher scores indicate greater impairment), and the score on the Alzheimer’s Disease Cooperative Study–Activities of Daily Living Scale for Mild Cognitive Impairment (ADCS-MCI-ADL; range, 0 to 53; lower scores indicate greater impairment).

Download a PDF of the Research Summary.

Results

A total of 1795 participants were enrolled, with 898 assigned to receive lecanemab and 897 to receive placebo. The mean CDR-SB score at baseline was approximately 3.2 in both groups. The adjusted least-squares mean change from baseline at 18 months was 1.21 with lecanemab and 1.66 with placebo (difference, −0.45; 95% confidence interval [CI], −0.67 to −0.23; P<0.001). In a substudy involving 698 participants, there were greater reductions in brain amyloid burden with lecanemab than with placebo (difference, −59.1 centiloids; 95% CI, −62.6 to −55.6). Other mean differences between the two groups in the change from baseline favoring lecanemab were as follows: for the ADAS-cog14 score, −1.44 (95% CI, −2.27 to −0.61; P<0.001); for the ADCOMS, −0.050 (95% CI, −0.074 to −0.027; P<0.001); and for the ADCS-MCI-ADL score, 2.0 (95% CI, 1.2 to 2.8; P<0.001). Lecanemab resulted in infusion-related reactions in 26.4% of the participants and amyloid-related imaging abnormalities with edema or effusions in 12.6%.

Conclusions

Lecanemab reduced markers of amyloid in early Alzheimer’s disease and resulted in moderately less decline on measures of cognition and function than placebo at 18 months but was associated with adverse events. Longer trials are warranted to determine the efficacy and safety of lecanemab in early Alzheimer’s disease. (Funded by Eisai and Biogen; Clarity AD ClinicalTrials.gov number, NCT03887455.)

Quick Take

Lecanemab in Alzheimer’s Disease

1m 47s

Current therapeutic agents for Alzheimer’s disease–related dementia temporarily improve symptoms but do not alter the underlying disease course.1,2 Some evidence suggests that amyloid removal slows the progression of disease.3 One anti-amyloid antibody (aducanumab) has received accelerated approval from the Food and Drug Administration.

Lecanemab is a humanized monoclonal antibody that binds with high affinity to soluble amyloid-beta (Aβ) protofibrils, which have been shown to be more toxic to neurons than monomers or insoluble fibrils.4-14 A phase 2b, dose-finding trial involving 854 participants with early Alzheimer’s disease did not show a significant difference between lecanemab and placebo in a Bayesian analysis of 12-month change in a composite score (primary end point). However, analyses at 18 months showed dose- and time-dependent clearance of amyloid with lecanemab, and the drug was associated with less clinical decline on some measures than placebo. In that trial, intravenous administration of 10 mg of lecanemab per kilogram of body weight every 2 weeks was identified as an appropriate dose, with a 9.9% incidence (<3% symptomatic) of amyloid-related imaging abnormalities (ARIA) with edema or effusions (ARIA-E).15 We conducted a phase 3 trial (Clarity AD) to determine the safety and efficacy of lecanemab in participants with early Alzheimer’s disease.

MethodsTrial Design and Oversight

Clarity AD was an 18-month, multicenter, double-blind, placebo-controlled, parallel-group trial involving persons with early Alzheimer’s disease. Eligible participants were randomly assigned in a 1:1 ratio to receive intravenous lecanemab (10 mg per kilogram every 2 weeks) or placebo. The randomization was stratified according to clinical subgroup (mild cognitive impairment due to Alzheimer’s disease or mild Alzheimer’s disease–related dementia on the basis of the criteria noted below), the presence or absence of concomitant approved medication for symptoms of Alzheimer’s disease at baseline (e.g., acetylcholinesterase inhibitors, memantine, or both), apolipoprotein E (ApoE) ε4 carriers or noncarriers, and geographic region. During the trial, participants underwent serial blood testing for plasma biomarkers and could participate in three optional substudies that eval‎uated longitudinal changes in brain amyloid burden as measured by positron-emission tomography (PET), brain tau pathologic features as measured by PET, and cerebrospinal fluid (CSF) biomarkers of Alzheimer’s disease.

The trial was conducted in accordance with International Council for Harmonisation guidelines and the ethical principles of the Declaration of Helsinki. The trial was approved by the institutional review board or independent ethics committee at each center, and all the participants provided written informed consent. The sponsor Eisai designed the trial and analyzed the data in collaboration with the academic authors, provided lecanemab and placebo, provided funding for medical writing, and aided in drafting the manuscript. The sponsor could not delay or interdict publication. The first, second, and sixteenth authors wrote the first draft of the manuscript, with professional medical writing assistance funded by Eisai, and all the authors contributed to subsequent drafts. Confidentiality agreements were in place between the sponsor and the authors and site investigators. Biogen provided partial funding for the trial.

An independent data and safety monitoring board consisting of experts in Alzheimer’s disease and statistics reviewed unblinded safety data during the trial. An independent medical monitoring team, whose members were unaware of the trial-group assignments, reviewed ARIA, infusion-related reactions, and hypersensitivity reactions. Clinical assessment raters were unaware of the safety assessments and the trial-group assignments. All the authors vouch for the completeness and accuracy of the data, the fidelity of the trial to the protocol (available with the full text of this article at NEJM.org), and the full reporting of adverse events.

Eligibility Criteria

The trial included participants 50 to 90 years of age, with either mild cognitive impairment due to Alzheimer’s disease or mild Alzheimer’s disease–related dementia on the basis of National Institute on Aging–Alzheimer’s Association criteria.16,17 Amyloid positivity was determined by PET or CSF measurement of Aβ1–42. All the participants had objective impairment in episodic memory as indicated by at least 1 standard deviation below the age-adjusted mean in the Wechsler Memory Scale IV–Logical Memory II.

End Points

The primary efficacy end point was the change in the score on the Clinical Dementia Rating (CDR)–Sum of Boxes (CDR-SB)18 from baseline at 18 months. The CDR-SB score is a validated outcome measure used in clinical trials of Alzheimer’s disease that is obtained by interviewing patients and their care partners and captures cognition and function. It assesses six domains that patients and caregivers identify as important (Memory, Orientation, Judgment and Problem Solving, Community Affairs, Home and Hobbies, and Personal Care). Scores for each domain range from 0 to 3, with higher scores indicating greater impairment. Total scores range from 0 to 18, with a score of 0.5 to 6 indicating early Alzheimer’s disease.

Key secondary end points were the change from baseline at 18 months in the following: amyloid burden on PET as measured in centiloids (with either florbetaben, florbetapir, or flutemetamol tracers) in a substudy, the score on the 14-item cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-cog14; range, 0 to 90, with higher scores indicating greater impairment),19 the Alzheimer’s Disease Composite Score (ADCOMS; range, 0 to 1.97, with higher scores indicating greater impairment),20 and the score on the Alzheimer’s Disease Cooperative Study–Activities of Daily Living Scale for Mild Cognitive Impairment (ADCS-MCI-ADL; range, 0 to 53, with lower scores indicating greater impairment).21 Biomarker assessments included CSF biomarkers (Aβ1–40, Aβ1–42, total tau, phosphorylated tau 181 [p-tau181], neurogranin, and neurofilament light chain [NfL]) and plasma biomarkers (Aβ42/40 ratio, p-tau181, glial fibrillary acidic protein [GFAP], and NfL). Tau PET and volumetric magnetic resonance imaging (MRI) results have not been fully analyzed.

A prespecified exploratory and multiplicity-unadjusted analysis examined the time to worsening of the global CDR score (range, 0 to 3, with higher scores indicating greater impairment). This end point was defined as the time to the first increase of at least 0.5 points in the global CDR score on two consecutive visits.

Statistical Analysis

The sample size for this trial was estimated on the basis of comparison of lecanemab and placebo with respect to the primary efficacy end point, the change from baseline at 18 months in the CDR-SB score. On the basis of data from the phase 2b trial of lecanemab,15 the estimated standard deviation of the change from baseline at 18 months in the CDR-SB score with placebo was 2.031 points, and the estimated treatment difference between lecanemab and placebo in all the participants was 0.373 points. This estimation corresponds to 25% less decline in cognitive function with lecanemab than with placebo and is consistent with a clinically meaningful difference on the basis of the Alzheimer’s disease literature, statistical principles, and agreements with regulatory authorities.15,22–24 Therefore, under the assumption of an estimated 20% dropout rate at 18 months in this trial, a total sample size of 1566 participants, including 783 participants receiving lecanemab and 783 participants receiving placebo, would provide the trial with 90% power to detect the treatment difference with the use of a two-sample t-test at a two-sided alpha level of 0.05. The sample size was increased by 200 to account for participants who missed three or more consecutive doses during the initial 6-month peak period of coronavirus disease 2019 (Covid-19), in accordance with previous agreement with regulatory authorities. No interim analyses for futility or efficacy were planned or performed.

Efficacy analyses were performed in the modified intention-to-treat population, which was defined as the group of randomly assigned participants who received at least one dose of lecanemab or placebo and who had a baseline assessment and at least one postdose primary efficacy (CDR-SB) measurement. Sensitivity analyses across efficacy end points to assess the robustness of the primary analysis to missing data included rank analysis of covariance with imputation of missing values. Additional sensitivity analyses were performed to eval‎uate potential effects of functional unblinding due to ARIA and effects of missed doses due to Covid-19–related absences (see the Supplementary Appendix, available at NEJM.org). Safety was eval‎uated in the safety population, which was defined as the group of participants who received at least one dose of lecanemab or placebo. Safety eval‎uations included monitoring of adverse events, vital signs, physical examinations, clinical laboratory variables, and 12-lead electrocardiograms. Occurrences of ARIA were monitored throughout the trial by central reading of MRI performed at weeks 9, 13, 27, 53, and 79 as well as at the 3-month follow-up visit (week 91) for safety monitoring. In addition, the populations for the substudies of amyloid burden on PET, tau pathologic features on PET, and CSF biomarkers of Alzheimer’s disease were the groups of participants who received at least one dose of lecanemab or placebo and who underwent a baseline PET or CSF eval‎uation and at least one postdose eval‎uation.

The primary analysis was performed without imputation of missing values. The primary analysis of the change from baseline at 18 months in the CDR-SB score was performed to compare lecanemab and placebo with the use of a mixed model for repeated measures that included the baseline CDR-SB score as a covariate, with trial group, visit, stratification variables (i.e., clinical subgroup, use of medication for symptoms of Alzheimer’s disease at baseline [yes or no], ApoE ε4 carrier status [carriers or noncarriers], and geographic region [North America, Europe, and Asia–Pacific]), baseline CDR-SB score–by–visit interaction, and trial group–by–visit interaction as fixed effects. If the between-group difference in primary end-point results was significant, then key secondary end points were to be tested hierarchically in the following order: change from baseline at 18 months in amyloid burden on PET as measured in centiloids in the subgroup tested and change from baseline at 18 months in the ADAS-cog14 score, change from baseline at 18 months in the ADCOMS, and change from baseline at 18 months in the ADCS-MCI-ADL score, all in the modified intention-to-treat population. Each test was performed at an alpha level of 0.05 (two-sided) and was to be performed only if the preceding test was significant at a two-sided level of 0.05. Additional details on the design and analysis methods are provided in the Supplementary Appendix and protocol.

ResultsParticipants

A total of 5967 persons were screened and 1795 underwent randomization; 898 were assigned to receive lecanemab and 897 to receive placebo at 235 sites in North America, Europe, and Asia from March 2019 through March 2021. Of these participants, 729 (81.2%) in the lecanemab group and 757 (84.4%) in the placebo group completed the trial and had data available on the primary end point (Figure 1). The modified intention-to-treat population included 1734 participants (859 in the lecanemab group and 875 in the placebo group), and the safety population included all 1795 randomly assigned participants. Enrollment in three longitudinal substudies included 698 participants in the substudy of amyloid burden on PET, 257 in the study of tau pathologic features on PET, and 281 in the substudy of CSF biomarkers of Alzheimer’s disease. The baseline characteristics of the substudy groups were generally similar to those in the main analysis. This trial made efforts to enhance global enrollment of a diverse group of participants (20% non-White), including in the United States, where 6.1% and 28.1% of the 3638 screened participants and 4.5% and 22.5% of randomly assigned participants were Black and Hispanic, respectively. The characteristics of the participants at baseline were generally similar in the two trial groups (Table 1). These characteristics were similar to what has been observed in population studies involving persons with early Alzheimer’s disease, although there was an underrepresentation of Black persons in the United States and an overrepresentation of Hispanic persons in the United States. The representativeness of the trial population is shown in Table S1 in the Supplementary Appendix.

Figure 1

Screening, Randomization, and Follow-up.

Table 1

Characteristics of the Participants at Baseline (Modified Intention-to-Treat Population).

End-Point Results

The mean CDR-SB score at baseline was approximately 3.2 in both the lecanemab and placebo groups, findings consistent with early Alzheimer’s disease (score of 0.5 to 6). The adjusted mean change from baseline at 18 months in the CDR-SB score was 1.21 in the lecanemab group and 1.66 in the placebo group (difference, −0.45; 95% confidence interval [CI], −0.67 to −0.23; P<0.001) (Figure 2A and Table 2).

Figure 2

Primary and Key Secondary End Points.

Table 2

Primary and Secondary End Points (Modified Intention-to-Treat Population).

In the substudy of amyloid burden on PET (a key secondary end point) involving 698 participants, the mean amyloid level at baseline was 77.92 centiloids in the lecanemab group and 75.03 centiloids in the placebo group. The adjusted mean change from baseline at 18 months was −55.48 centiloids in the lecanemab group and 3.64 centiloids in the placebo group (difference, −59.12 centiloids; 95% CI, −62.64 to −55.60; P<0.001) (Figure 2B and Table 2). In the modified intention-to-treat population, the mean ADAS-cog14 scores at baseline were 24.45 in the lecanemab group and 24.37 in the placebo group. The adjusted mean change from baseline at 18 months in the ADAS-cog14 score was 4.14 in the lecanemab group and 5.58 in the placebo group (difference, −1.44; 95% CI, −2.27 to −0.61; P<0.001) (Figure 2C and Table 2). The mean ADCOMS in the modified intention-to-treat population at baseline was 0.398 in the lecanemab group and 0.400 in the placebo group. The adjusted mean change from baseline at 18 months in the ADCOMS was 0.164 in the lecanemab group and 0.214 in the placebo group (difference, −0.050; 95% CI, −0.074 to −0.027; P<0.001) (Figure 2D and Table 2). In the modified intention-to-treat population, the mean ADCS-MCI-ADL scores at baseline were 41.2 for lecanemab and 40.9 for placebo. The adjusted mean change from baseline at 18 months in the ADCS-MCI-ADL score was −3.5 in the lecanemab group and −5.5 in the placebo group (difference, 2.0; 95% CI, 1.2 to 2.8; P<0.001) (Figure 2E and Table 2).

For each of these assessments, separation of the trial groups was apparent by visual inspection of graphs at 3 months. However, no conclusions can be drawn because there was no prespecified plan for analysis that included adjustment of confidence intervals for multiple comparisons at any intermediate time point.

Sensitivity analyses of the CDR-SB score that eval‎uated the effect of Covid-19 (missed doses) and potential for bias from functional unblinding due to ARIA were generally consistent with the primary analysis (Table S2). Results were also consistent across key randomization strata, as well as for other factors that affect Alzheimer’s disease (Figs. S1 through S4). The exploratory subgroup analysis involving ApoE ε4 homozygotes (15% of the trial population) numerically favored lecanemab for the ADAS-cog14 and ADCS-MCI-ADL scores but not for the CDR-SB score and the ADCOMS. Results of prespecified analyses of end points involving CSF and plasma biomarkers showed numerical improvements for all assessments comparing lecanemab with placebo, with the exception of CSF NfL (Fig. S5). In a prespecified, multiplicity-unadjusted analysis of the time to worsening of the global CDR score, the hazard ratio for progression to the next stage of dementia (0.69) numerically favored lecanemab over placebo (Fig. S6).

Safety

Deaths occurred in 0.7% of the participants in the lecanemab group and 0.8% of those in the placebo group (Table 3). No deaths were considered by the investigators to be related to lecanemab or occurred with ARIA. Serious adverse events occurred in 14.0% of the participants in the lecanemab group and 11.3% of those in the placebo group. The most commonly reported serious adverse events were infusion-related reactions (in 1.2% of the participants in the lecanemab group and 0 participants in the placebo group), ARIA-E (in 0.8% and 0, respectively), atrial fibrillation (in 0.7% and 0.3%), syncope (in 0.7% and 0.1%), and angina pectoris (in 0.7% and 0). The overall incidence of adverse events was similar in the two groups (Table 3). Adverse events leading to discontinuation of the trial agent occurred in 6.9% of the participants in the lecanemab group and 2.9% of those in the placebo group. The most common adverse events (affecting >10% of the participants) in the lecanemab group were infusion-related reactions (26.4% with lecanemab and 7.4% with placebo); ARIA with cerebral microhemorrhages, cerebral macrohemorrhages, or superficial siderosis (ARIA-H; 17.3% with lecanemab and 9.0% with placebo); ARIA-E (12.6% with lecanemab and 1.7% with placebo); headache (11.1% with lecanemab and 8.1% with placebo); and falls (10.4% with lecanemab and 9.6% with placebo). Infusion-related reactions were largely mild to moderate (grade 1 or 2, 96%) and occurred with the first dose (75%). A total of 56% of the participants did not take preventative medications (i.e., nonsteroidal antiinflammatory drugs, antihistamines, or glucocorticoids) for infusion-related reactions. Of those who took preventative medications for subsequent doses, 63% did not have additional reactions.

Table 3

Adverse Events.

Events of ARIA-E with lecanemab were mostly mild to moderate (91%) on the basis of central reading of imaging with the use of protocol definitions. These events were mostly asymptomatic (78%), occurred during the first 3 months of the treatment period (71%), and resolved within 4 months after detection (81%). A total of 2.8% of the participants in the lecanemab group had symptomatic ARIA-E; commonly reported symptoms were headache, visual disturbance, and confusion. The incidence of isolated ARIA-H (i.e., ARIA-H in participants who did not also have ARIA-E) was 8.9% in the lecanemab group and 7.8% in the placebo group. The incidence of isolated symptomatic ARIA-H was 0.7% in the lecanemab group and 0.2% in the placebo group. The most common symptom associated with isolated symptomatic ARIA-H was dizziness. Macrohemorrhage occurred in 5 of 898 participants (0.6%) in the lecanemab group and 1 of 897 participants (0.1%) in the placebo group. ARIA-H that occurred with ARIA-E tended to occur early (within 6 months). Isolated ARIA-H occurred throughout the trial. ARIA-E and ARIA-H were numerically less common among ApoE ε4 noncarriers than among carriers, with higher frequency among ApoE ε4 homozygotes than among ApoE ε4 heterozygotes (Table 3).

Discussion

In this phase 3 trial, the change from baseline at 18 months in the CDR-SB score (primary end point) was less with lecanemab than with placebo, favoring lecanemab. Results for secondary clinical end points were in the same direction as those for the primary end point. Lecanemab has high selectivity for soluble aggregated species of Aβ as compared with monomeric amyloid, with moderate selectivity for fibrillar amyloid; this profile is considered to target the most toxic pathologic amyloid species.4,7,8,13,14 After 18 months of treatment in the amyloid substudy, the mean amyloid level of 22.99 centiloids in the lecanemab group was below the threshold for amyloid positivity of approximately 30 centiloids, above which participants are considered to have elevated brain amyloid levels.25 In the CSF substudy and in plasma analyses involving the overall population, markers of amyloid, tau, neurodegeneration, and neuroinflammation (plasma GFAP) were reduced to a greater extent with lecanemab than with placebo, with the exception of NfL, which is less sensitive to neurodegeneration than the other markers and has a slower time course for change than the others.

A definition of clinically meaningful effects in the primary end point of the CDR-SB score has not been established; however, this trial exceeded the prospectively defined target, with an estimated treatment difference of 0.373 points on a scale range of 18, a baseline value of 3.2, and early Alzheimer’s disease typically characterized by a score of 0.5 to 6. In a prespecified exploratory and multiplicity-unadjusted analysis of the time to worsening (increase) of the global CDR score of at least 0.5 points on two consecutive visits, the hazard ratio for progression to the next stage of dementia numerically favored lecanemab over placebo. An open-label extension study of Clarity AD is ongoing to provide additional safety and efficacy data beyond 18 months.

In the lecanemab group, the incidence of ARIA-E was 12.6%, and the incidence of ARIA-H was 17.3%. These incidences compare with 9.9% and 10.7%, respectively, in the phase 2b trial of lecanemab, in which ApoE ε4 carriers were underrepresented in the group that received 10 mg per kilogram every 2 weeks.15 The incidence of ARIA, including symptomatic ARIA, was numerically lower than in similar clinical trials, but differences in the drugs used and in trial design do not allow direct comparisons.26,27 ARIA-E generally occurred in the first 3 months, was mild and asymptomatic, did not lead to discontinuation of lecanemab or placebo if mild, and resolved within 4 months. The incidences of both overall and symptomatic ARIA-E were highest among ApoE ε4 homozygotes.

Among the limitations of this trial is that it includes data for only 18 months of treatment; an open-label extension study is ongoing. The Clarity AD trial was conducted during the Covid-19 pandemic and encountered obstacles including missed doses, delayed assessments, and intercurrent illnesses. The dropout rate was 17.2%, and a sensitivity analysis that eval‎uated the effect of missed doses was consistent with the primary end-point analysis. An additional potential limitation was the use of modified intention-to-treat analysis without imputation of missing values. However, a sensitivity analysis that was conducted with the use of a standard intention-to-treat population with imputation yielded similar results. Finally, occurrences of ARIA may have caused participants and investigators to be aware of the trial-group assignments. We attempted to minimize this bias by making clinical raters unaware of the safety assessments and the trial-group assignments, and sensitivity analyses that were performed to examine the effect of ARIA on clinical outcomes showed that ARIA had no effect on the results. Additional trials of lecanemab include a 5-year phase 2 long-term extension trial (ClinicalTrials.gov number, NCT01767311) and a 4-year phase 3 long-term extension trial (NCT03887455) in early Alzheimer’s disease, the 4-year AHEAD 3-45 trial (NCT04468659) in preclinical Alzheimer’s disease, and the 4-year DIAN-TU (Dominantly Inherited Alzheimer Network Trials Unit) Next Generation trial (NCT05269394) in dominantly inherited Alzheimer’s disease.

In persons with early Alzheimer’s disease, lecanemab reduced brain amyloid levels and was associated with moderately less decline on clinical measures of cognition and function than placebo at 18 months but was associated with adverse events. Longer trials are warranted to determine the efficacy and safety of lecanemab in early Alzheimer’s disease.

Notes

This article was published on November 29, 2022, at NEJM.org.

A data sharing statement provided by the authors is available with the full text of this article at NEJM.org.

Supported by Eisai (regulatory sponsor), with partial funding by Biogen.

Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

We thank the trial participants and their families, as well as all the investigators and site staff who made the trial possible (see the Supplementary Appendix for a list of collaborators); the members of the data and safety monitoring board and the raters; Lars Lannfelt and the staff of BioArctic for their early research on lecanemab; the staff of the clinical research organization Worldwide Clinical Trials for their ongoing support in conducting the trial; and J. David Cox (Mayville Medical Communications) and Lisa Yarenis (Eisai) for writing and editing assistance with an earlier version of the manuscript, in accordance with Good Publication Practice 4 ethical guidelines.

Supplementary Material

Research Summary (nejmoa2212948_research-summary.pdf)

• Download
• 260.42 KB

Protocol (nejmoa2212948_protocol.pdf)

• Download
• 4.75 MB

Supplementary Appendix (nejmoa2212948_appendix.pdf)

• Download
• 783.04 KB

Disclosure Forms (nejmoa2212948_disclosures.pdf)

• Download
• 724.97 KB

Data Sharing Statement (nejmoa2212948_data-sharing.pdf)

• Download
• 69.01 KB

References

1.

Marasco RA. Current and evolving treatment strategies for the Alzheimer disease continuum. Am J Manag Care 2020;26:Suppl:S167-S176.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

2.

Long JM, Holtzman DM. Alzheimer disease: an update on pathobiology and treatment strategies. Cell 2019;179:312-339.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

3.

Avgerinos KI, Ferrucci L, Kapogiannis D. Effects of monoclonal antibodies against amyloid-β on clinical and biomarker outcomes and adverse event risks: a systematic review and meta-analysis of phase III RCTs in Alzheimer’s disease. Ageing Res Rev 2021;68:101339-101339.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar

4.

Tucker S, Möller C, Tegerstedt K, et al. The murine version of BAN2401 (mAb158) selectively reduces amyloid-β protofibrils in brain and cerebrospinal fluid of tg-ArcSwe mice. J Alzheimers Dis 2015;43:575-588.

Crossref

PubMed

Web of Science

Google Scholar

5.

Sehlin D, Englund H, Simu B, et al. Large aggregates are the major soluble Aβ species in AD brain fractionated with density gradient ultracentrifugation. PLoS One 2012;7(2):e32014-e32014.

Go to Citation

Crossref

PubMed

Web of Science

Google Scholar