Eligibility for Anti-Amyloid Treatment in a Population-Based Study of Cognitive Aging

Abstract

Background and Objectives

Treatment options for Alzheimer disease (AD) are limited and have focused mainly on symptomatic therapy and improving quality of life. Recently, lecanemab, an anti–β-amyloid monoclonal antibody (mAb), received accelerated approval by the US Food and Drug Administration for treatment in the early stages of biomarker-confirmed symptomatic AD. An additional anti–β-amyloid mAb, aducanumab, was approved in 2021, and more will potentially become available in the near future. Research on the applicability and generalizability of the anti–β-amyloid mAb eligibility criteria on adults with biomarkers available in the general population has been lacking. The study's primary aim was to apply the clinical trial eligibility criteria for lecanemab treatment to participants with early AD of the population-based Mayo Clinic Study of Aging (MCSA) and assess the generalizability of anti-amyloid treatment. The secondary aim of this study was to apply the clinical trial eligibility criteria for aducanumab treatment in MCSA participants.

Methods

This cross-sectional study aimed to apply the clinical trial eligibility criteria for lecanemab and aducanumab treatment to participants with early AD of the population-based MCSA and assess the generalizability of anti-amyloid treatment.

Results

Two hundred thirty-seven MCSA participants (mean age [SD] 80.9 [6.3] years, 54.9% male, and 97.5% White) with mild cognitive impairment (MCI) or mild dementia and increased brain amyloid burden by PiB PET comprised the study sample. Lecanemab trial's inclusion criteria reduced the study sample to 112 (47.3% of 237) participants. The trial's exclusion criteria further narrowed the number of potentially eligible participants to 19 (overall 8% of 237). Modifying the eligibility criteria to include all participants with MCI (instead of applying additional cognitive criteria) resulted in 17.4% of participants with MCI being eligible for lecanemab treatment. One hundred four participants (43.9% of 237) fulfilled the aducanumab clinical trial's inclusion criteria. The aducanumab trial's exclusion criteria further reduced the number of available participants, narrowing those eligible to 12 (5.1% of 237). Common exclusions were related to other chronic conditions and neuroimaging findings.

Discussion

Findings estimate the limited eligibility in typical older adults with cognitive impairment for anti–β-amyloid mAbs.

Introduction

Treatments that are disease-modifying have now been approved for Alzheimer disease (AD), but it is unclear how many patients might actually be eligible for such medications. Aducanumab, an anti–β-amyloid monoclonal antibody (mAb), is the first disease-modifying therapy (DMT) approved in 2021 for those with mild cognitive impairment (MCI) due to AD and mild AD dementia (i.e., early symptomatic AD).¹ In January 2023, lecanemab, an anti–β-amyloid mAb, was given accelerated approval by the Food and Drug Administration (FDA) to treat early AD.^2–4 Additional mAbs are potentially going to become available.¹ The inclusion and exclusion criteria of the clinical trials that led to FDA approval of these agents form the basis of how patients in clinical practice should be invited, discouraged, or rejected from receiving one of the agents.

Recent studies in specialty clinics (i.e., multidisciplinary cognitive unit [Ireland], outpatient geriatrics, “Centres for Dementia and Cognitive Decline” units [Italy]),^5–8 or using medical claims from the Centers for Medicare & Medicaid Services (United States)⁸ applied the aducanumab clinical trial eligibility criteria to their study populations (e.g., patients with MCI, with cognitive issues and positive AD biomarkers, with AD dementia, or with AD and related disorders [ADRD]) and concluded that a considerable proportion of the patients with early AD would not be eligible to receive aducanumab treatment; similar studies are not yet available for the lecanemab clinical trial eligibility criteria.

In addition, research on the applicability and generalizability of the anti–β-amyloid mAb eligibility criteria on community-dwelling adults with biomarkers available in the general population has been lacking. The study's primary aim was to apply the clinical trial eligibility criteria for lecanemab treatment to participants with early AD of the population-based Mayo Clinic Study of Aging (MCSA) and assess the generalizability of anti–β-amyloid treatment. The secondary aim of this study was to apply the clinical trial eligibility criteria for aducanumab treatment in MCSA participants.

Methods

The Mayo Clinic Study of Aging

The MCSA is a population-based cohort study established in 2004 in Olmsted County, MN, to investigate MCI and dementia risk factors. Using the Rochester Epidemiology Project (REP) resources,⁹ an age-stratified and sex-stratified random sample of Olmsted County residents was invited to participate in the MCSA. Initially, persons age 70–89 years were recruited; in 2012, recruitment expanded to age 50–69 years. Participants undergo baseline and follow-up visits approximately every 15 months, using the same evaluation protocol, and are invited to undergo neuroimaging studies (e.g., MRI; 2006 onward, Pittsburgh Compound B PET scan [Pib PET; 2009 onward]).

At each MCSA visit, a study coordinator conducted a risk factor assessment. Participants answered questions on their personal and familial medical history, demographics, neuropsychiatric symptoms, and medications, and the Clinical Dementia Rating (CDR) scale¹⁰ was administered to an informant. A physician reviewed the medical history, did a neurologic examination, and administered the Short Test of Mental Status, from which the Mini-Mental State Examination (MMSE) score was derived.^11,12 Nine neuropsychological tests administered by a psychometrist were used to assess cognitive performance in 4 cognitive domains (i.e., memory, language, attention/executive, visuospatial).¹³

At each MCSA visit, the final diagnosis (cognitively unimpaired [CU], MCI, dementia) was decided by a consensus agreement between the study coordinator, the physician, and a neuropsychologist after reviewing all the information for each participant. Individuals who performed in the normal range and did not meet the criteria for MCI¹⁴ or dementia^15–17 were classified as CU. Apolipoprotein E (APOE) ε4 status was determined from a blood draw at MCSA baseline assessment.¹⁸ In addition, every 2–3 years, expert RN abstractors review the participants' medical records to assess comorbidities.

Study Sample

There were 5,255 community-dwelling MCSA participants aged 50–90 years (mean [SD] age was 73.0 [10.7] years, and 50.3% were male) with a visit on or after January 1, 2009 (year of PiB PET introduction). Seventeen participants were excluded due to MRI or PET scan contraindications, and 3,086 people did not have amyloid PET. Eight hundred sixty-nine participants had an amyloid-positive PET scan (A+; positivity criteria in the “Neuroimaging biomarkers” section). Of the 869 participants with A + PiB PET, 237 were classified as having MCI or mild dementia (possible or probable AD) and had available CDR global data; thus, in this cross-sectional study, 237 participants comprised the study sample (Figure 1).

Figure 1. Selection of the Study Sample.

Standard Protocol Approvals, Registrations, and Patient Consents

Study approval was obtained from the Institutional Review Boards of the Mayo Clinic and Olmsted Medical Center in Rochester, MN. Participants provided written informed consent before participation. In the case of participants with cognitive impairment sufficient to interfere with capacity, assent was obtained from a legally authorized representative.

Neuroimaging Biomarkers

Neuroimaging biomarker acquisition details have been presented previously.^19,20 MRI imaging was acquired on 3 T scanners (GE from 2009 to 2017 and Siemens from 2017 to now), and details have been described previously.^19,21 In brief, for each participant, both the structural T1-weighted magnetization prepared rapid gradient echo (MPRAGE) image and fluid-attenuated inversion recovery (FLAIR) MRI image were used for WMH segmentation.²² Methods were explained in detail previously.²³ The absolute burden of WMH (cm³) was normalized to total intracranial volume (cm³). In addition, 3D MPRAGE and FLAIR images were reviewed, and lacunar and cortical infarctions were identified and classified as previously described.²⁴ As previously published in detail,²⁵ cerebral microbleeds were quantified on 3-tesla MRI scans with T2* gradient recalled echo sequences based on the published consensus criteria.²⁶

Amyloid PET imaging was performed with ¹¹C-Pittsburgh Compound B (PIB).^27,28 The standard uptake value ratio (SUVR) for amyloid was determined by normalizing prefrontal, orbitofrontal, parietal, temporal, anterior and posterior cingulate, and the precuneus to the cerebellar crus gray matter, and amyloid positivity was defined as an SUVR value greater than or equal to 1.48.^19,29,30 We used the visit closest to the PiB PET to determine the eligibility for lecanemab or aducanumab treatment (i.e., study baseline).

Eligibility Criteria for Lecanemab and Aducanumab

To identify whether the MCSA participants are eligible to receive lecanemab, we applied the clinical trials' criteria published in van Dyck et al.³ In addition, we applied the aducanumab clinical trials' criteria as published in Budd Haeberlein et al.³¹

The MCSA study takes advantage of all the resources of the REP,⁹ which links and archives the medical records of virtually all Olmsted County residents. Using the REP medical records linkage system, we were able to electronically extract the ICD-9/ICD-9 codes associated with any health care visit for specific conditions, as demanded by the lecanemab and aducanumab clinical trial criteria. Exact details of the codes extracted are provided in the detailed eAppendix 1, links.lww.com/WNL/D46. The REP resources also allowed us to extract outpatient medication prescriptions, and MCSA also has available self-reported medications at each study visit that were also used (details included in eAppendix 1).

In addition, the electronic health records of participants who met the inclusion criteria for lecanemab treatment (n = 112) were reviewed (by J.G.R.) for cardiopulmonary conditions the past 1 year which were not “stably and adequately controlled, or which in the opinion of the investigator(s) could affect the subject's safety or interfere with the study assessments” (per lecanemab exclusion criteria).³ This review included active congestive heart failure, coronary artery disease (symptomatic within 12 months of study baseline), angina, significant rhythm abnormalities (detected on EKG or prolonged cardiac monitoring), and other potential cardiopulmonary contraindications present (i.e., amyloid cardiomyopathy, vasculitis, pulmonary embolism, severe pulmonary hypertension, symptomatic aortic stenosis, syncope) (i.e., people who presented to the emergency department within 12 months of study baseline) (Table 1).

Table 1.

Characteristics of Participants Who Met the Lecanemab Inclusion Criteria of the Study

Characteristics	Total (N = 112)
Age at study baseline, mean (SD)	80.3 (6.5)
Sex, female	52 (46.4%)
APOE ε4 positive (yes)a	59 (53.2%)
Education (y), mean (SD)	14.3 (2.6)
CDR memory score
0.5	83 (74.1%)
1	25 (22.3%)
2	4 (3.6%)
Cognitive impairment
Mild cognitive impairment	103 (92.0%)
Dementia	9 (8.0%)
Body mass index (>17 and <35), mean (SD)	26.1 (3.8)
MMSE from short test 22–30, mean (SD)b	25.6 (1.8)
Chronic conditionsa,c
Two or more	88 (79.3%)
Three or more	53 (47.7%)
Total no. of exclusions (max 12)
Mean (SD)	1.5 (1.1)
Median (Q1–Q3)	1.0 (1.0–2.0)
Range	0.0–7.0
Having one or more exclusionsd	93 (83.0%)
CNS-related exclusionse	37 (33.0%)
Cardiopulmonary contraindications, past 1 yf	41 (36.6%)
Active congestive heart failure	2 (1.8%)
Coronary artery disease	12 (10.7%)
Angina	7 (6.3%)
Significant rhythm abnormalities	15 (13.4%)
Other cardiopulmonary conditions	7 (6.3%)
Amyloid cardiomyopathy	1
Pulmonary embolism and vasculitis	1
Pulmonary embolism	1
Severe pulmonary hypertension	1
Symptomatic aortic stenosis	1
Syncope	2
End-stage renal disease in past 2 y	1 (0.9%)
Liver failure or cirrhosis in past 2 y	3 (2.7%)
Immunologic disease (ever)g	12 (10.7%)
Transient ischemic attack or stroke (past 1 y)	7 (6.3%)
Neuroimaging findings at the time of PET scan	42 (37.5%)
Greater than 4 microhemorrhages	9
Definite superficial siderosis	8
Two or more lacunar infarcts	10
Subcortical infarct with diam >15 mm	2
Cortical infarcts (any size)	12
Severe WMH (90th% of CU/MCI cohort)	13
Serious risk of suicide, past 5 y	1 (0.9%)
Psychiatric hospitalization (past 6 mo)	2 (1.8%)
Malignant neoplasms, past 3 yh	20 (17.9%)
Alcohol or substance abuse (past 2 y)	3 (2.7%)

Abbreviations: APOE = apolipoprotein E; CDR = Clinical Dementia Rating; CU = cognitively unimpaired; MCI = mild cognitive impairment; MMSE = Mini-Mental State Examination; WMH = white matter hyperintensities.

N (%) unless otherwise specified.

^a1 missing.

^bDerived from the Short Test of Mental Status.

^cDiabetes, hypertension, dyslipidemia, congestive heart failure, coronary artery disease, based on medical record review by expert nurse abstractors; obesity (BMI ≥30) as measured at baseline.

^dBased on ICD-9/ICD-9 codes from the participants' electronic health records (EHR) using the REP medical records linkage system resources, unless otherwise specified. Neuroimaging findings were assessed by the Mayo Clinic Study of Aging neuroimaging studies.

^eBrain cancer, Parkinson disease, epilepsy, coma/brain damage, or intracranial injury.

^fNot stably and adequately controlled, or which in the opinion of the investigator(s) could affect the participant's safety or interfere with the study assessments (past 1 y), as assessed by medical record review, including active congestive heart failure, coronary artery disease, angina, significant rhythm anomalies, other cardiopulmonary conditions (amyloid cardiomyopathy, vasculitis, pulmonary embolism, severe pulmonary hypertension, symptomatic aortic stenosis, syncope).

^gLupus, Sjogren syndrome, rheumatoid arthritis, and other inflammatory arthritis.

^hDid not consider codes for benign tumors, basal cell carcinoma, squamous cell carcinoma, other nonmelanoma skin cancers, or localized prostate cancer.

Finally, we also used information regarding comorbidities that MCSA expert RN abstractors assess by reviewing the participants' medical records every 2–3 years.

Inclusion criteria for lecanemab and aducanumab treatment are presented in Figure 2, and detailed exclusion criteria used in this study are also described the eAppendix 1, links.lww.com/WNL/D46. For participants missing BMI, MMSE, or Wechsler Memory Scale-Revised Logical Memory II data at the study baseline, we carried forward data from the previous visit if available.

Figure 2. Inclusion Criteria Used in the Study for Lecanemab (A) and Aducanumab (B) Treatment.

Statistical Analysis

Patient characteristics were summarized using descriptive statistics (mean, SD, median, interquartile range, count, percentage). We applied the eligibility criteria of the clinical trials and also ran a sensitivity analysis considering the clinical diagnosis of MCI without restricting CDR global, CDR memory, MMSE, or Wechsler Memory Scale-Revised (WMS-R) Logical Memory II scores into specific boundaries, as defined by the clinical trial criteria. In addition, we also compared the characteristics of participants with and without amyloid PET.

Data Availability

The MCSA makes data available to qualified researchers on reasonable request.

Results

Comparison of MCSA Participants With and Without PET

Comparing the participants with (n = 2,152) and without (n = 3,086) PiB PET (Figure 1), we observed that those with PET were, on average, younger (71.1 [10.3] vs 74.0 [10.9] years, p < 0.001), more frequently male (52.9 vs 48.3%, p = 0.001), more frequently CU (87.2% vs 85.1%, p = 0.016) with higher mean of education years (14.8 [2.6] vs 14.4 [2.7], p ≤ 0.001), higher MMSE score (28.2 [1.8] vs 27.6 [2.3], p < 0.001), and lower frequency of 3+ comorbidities (39.8% vs 47.7%, p < 0.001).

Study Sample

As aforementioned, 237 participants (mean age [SD] 80.9 [6.3] years, 54.9% male, 97.5% White, and 99.6% not Hispanic or Latino) with (MCI; n = 222) or dementia (possible or probable AD; n = 15) and positive brain amyloid PiB PET comprised the study population.

Lecanemab Eligibility Criteria

Lecanemab inclusion criteria reduced the study population (N = 237) to 112 participants. Twenty-one participants were excluded due to their BMI (less than or equal to 17 or 35 or greater) and 48 due to a CDR global score other than 0.5 or 1.0; 46 did not meet the WMS-R Logical Memory II scores [score at least 1 SD below age-adjusted mean (≤15 for age 50–64, ≤12 for age 65–69, ≤11 for age 70–74, ≤9 for age 75–79, and ≤7 for age 80–90 years). In addition, 8 people were excluded due to an MMSE score outside the bounds of 22–30, and an additional 2 were excluded with a CDR memory score <0.5 (Figure 2).

Thus, 112 of the 237 participants (47.3%) could fulfill the lecanemab clinical trial's inclusion criteria (Figure 2). Their mean age (SD) was 80.3 (SD 6.5) years, and the mean education was 14.3 (SD 2.6) years, 52 participants were female (46.4%), and 59 (53.2%) were APOE ε4 positive. All but one participant were White, and all 112 participants were not Hispanic or Latino. By study design, MCSA participants have a designated study partner, fulfilling this inclusion criterion, as well. Sixteen participants (14.3% of the 112) were taking medication for cognitive impairment, as reported by the patients (i.e., self-report) (donepezil [15], memantine [1], galantamine [1]), and 15 participants had a blood clotting disorder (past 1 year; 3) or used anticoagulants (13).

Applying the lecanemab's exclusion criteria further narrowed the number of eligible participants from 112 to 19 (8% of 237). Notable exclusions (Table 1) include cardiopulmonary contraindications (41 [36.6%, as reviewed in the electronic medical record]), CNS-related exclusions (e.g., brain cancer, Parkinson disease, epilepsy, coma/brain damage, or intracranial injury; 37 [33.0%]), imaging exclusions (42 [37.5%]), and history of malignancy (20 [17.9%]). Cancer types detected in participants in 3 years before baseline were bladder (2), bone (secondary; 1), breast (1), chronic myeloid leukemia in remission less than 3 years (1), colon (2), leukemia without the mention of remission (1), liver (secondary; 1), melanoma (1), penile (1), and prostate (9). We did not include basal cell carcinoma, squamous cell carcinoma, other nonmelanoma skin cancers, or localized prostate cancer. As presented in Figure 3, 44% of participants who met inclusion criteria (n = 112) met 2 or more exclusion criteria.

Figure 3. Pie Charts Represent the Participants Who Met the Inclusion Criteria for Lecanemab (A; n = 112) and Aducanumab (B; n = 104) and the Proportion of Participants Who Met None, One, or More Exclusion Criteria.

Eligible participants for lecanemab treatment (i.e., met inclusion criteria and had no exclusions) were more likely to be APOE ε4 positive and less likely to have diabetes, congestive heart failure, or coronary heart disease (vs participants who met inclusion criteria but were excluded by exclusion criteria). Having 2 or more and 3 or more chronic conditions was less frequent in eligible participants. Nevertheless, there were few eligible participants (n = 19) to draw any conclusions (Table 3).

Table 3.

Characteristics of Participants Meeting Inclusion Criteria for Lecanemab and Aducanumab Treatment, by the Number of Exclusion Criteria

Characteristic	Lecanemab exclusion criteria		Aducanumab exclusion criteria
	≥1 exclusion (N = 93)	No exclusions, eligible (N = 19)	≥1 exclusion (N = 92)	No exclusions, eligible (N = 12)
Age at study baseline, mean (SD)	80.7 (6.5)	78.6 (6.4)	78.7 (5.0)	76.5 (5.3)
Sex, female	41 (44.1%)	11 (57.9%)	49 (53.3%)	8 (66.7%)
Education (y), mean (SD)	14.1 (2.6)	15.4 (2.5)	14.4 (2.6)	14.2 (2.5)
APOE ε4 positive (yes)a	45 (48.9%)	14 (73.7%)	48 (53.3%)	10 (83.3%)
Diabetesb	22 (23.9%)	1 (5.3%)	22 (24.2%)	1 (8.3%)
Hypertension (ever)b	73 (79.3%)	14 (73.7%)	69 (75.8%)	9 (75.0%)
Dyslipidemia (ever)b	81 (88.0%)	16 (84.2%)	76 (83.5%)	10 (83.3%)
Congestive heart failureb	12 (13.0%)	0 (0.0%)	11 (12.1%)	0 (0.0%)
Coronary artery diseaseb	39 (42.4%)	2 (10.5%)	37 (40.7%)	2 (16.7%)
Body mass index ≥30	19 (20.4%)	1 (5.3%)	22 (23.9%)	4 (33.3%)
Chronic conditionsc
Two or more	75 (81.5%)	13 (68.4%)	68 (74.7%)	10 (83.3%)
Three or more	50 (54.3%)	3 (15.8%)	48 (52.7%)	4 (33.3%)
Blood clotting disorder (past 1 y) or anticoagulant use	15 (16.1%)	0 (0.0%)	11 (12.0%)	0 (0.0%)
Taking anticoagulants	13 (14.0%)	0 (0.0%)	9 (9.8%)	0 (0.0%)

Abbreviation: APOE = apolipoprotein E.

N (%) unless otherwise specified.

^aMissing: 1 for the lecanemab and 2 for the aducanumab groups.

^bMissing: 1 for the lecanemab and 1 for the aducanumab groups.

^cDiabetes, hypertension, dyslipidemia, congestive heart failure, coronary artery disease, based on medical record review by expert nurse abstractors; obesity (BMI ≥30) as measured at baseline.

Aducanumab Eligibility Criteria

When applying the aducanumab inclusion criteria to our study population (N = 237), 63 participants were excluded based on their age, 1 person was excluded based on their years of education, 52 participants were excluded due to a CDR global score other than 0.5, and 17 participants were excluded with an MMSE below 24 (Figure 2). The remaining 104 participants (43.9% of 237) fulfilled the aducanumab clinical trial's inclusion criteria. The mean age (SD) of those fulfilling the inclusion criteria was 78.4 (5.1) years, with the average years of education at 14.4 (2.6). Forty-seven participants were female (45.2%), and 58 (56.9%) were APOE ε4 positive (Table 2). All but 2 participants were White, and all 104 participants were not Hispanic or Latino.

Table 2.

Characteristics of Participants Who Met the Aducanumab Inclusion Criteria of the Study

Characteristics	Total (N = 104)
Age at study baseline, mean (SD)	78.4 (5.1)
Sex, female	47 (45.2%)
APOE ε4 positive (yes)a	58 (56.9%)
Education ≥6 y, mean (SD)	14.4 (2.6)
Cognitive Impairment
Mild cognitive impairment	102 (98.1%)
Dementia	2 (1.9%)
MMSE from short test 24–30, mean (SD)b	26.0 (1.5)
Chronic conditionsa,c
Two or more	78 (75.7%)
Three or more	52 (50.5%)
Total number of exclusions (max 16)
Mean (SD)	2.1 (1.5)
Median (Q1–Q3)	2.0 (1.0–3.0)
Range	0.0–7.0
Having one or more exclusionsd	92 (88.5%)
CNS-related exclusions (ever)e	34 (32.7%)
Cardiovascular disease (past 1 y)	50 (48.1%)
Uncontrolled hypertension	20 (19.2%)
End-stage renal disease in past 2 y	1 (1.0%)
Liver failure or cirrhosis in past 2 y	3 (2.9%)
Immunologic disease (ever)f	7 (6.7%)
TIA, stroke, or unexplained loss of consciousness (past 1 y)	12 (11.5%)
Neuroimaging findings at the time of PET scan	33 (31.7%)
Greater than 4 microhemorrhages	7
Definite superficial siderosis	4
Two or more lacunar infarcts	10
Subcortical infarct with diam >15 mm	2
Cortical infarcts with diam >15 mm	3
Severe WMH (90th% of CU/MCI cohort)	14
Psychiatric hospitalization (past 6 mo)	2 (1.9%)
Malignancy (ever)g	29 (27.9%)
Alcohol or substance abuse (past 1 y)	3 (2.9%)
Narcotics (prescribed; past 1 y)	15 (14.4%)
Blood clotting conditions (past 1 y) or anticoagulant use (at visit)	11 (10.6%)
Organized care facility (self-reported)	1 (1.0%)

Abbreviations: APOE = apolipoprotein E; CDR = Clinical Dementia Rating; CU = cognitively unimpaired; MCI = mild cognitive impairment; MMSE = Mini-Mental State Examination; TIA = transient ischemic attack; WMH = white matter hyperintensities.

N (%) unless otherwise specified.

^a1 missing.

^bDerived from the Short Test of Mental Status.

^cDiabetes, hypertension, dyslipidemia, congestive heart failure, coronary artery disease, based on medical record review by expert nurse abstractors; obesity (BMI ≥30) as measured at baseline.

^dBased on ICD-9/ICD-9 codes from the participants' electronic health records (EHR) using the REP medical records linkage system resources, unless otherwise specified. Neuroimaging findings were assessed by the Mayo Clinic Study of Aging neuroimaging studies.

^eBrain cancer, Parkinson disease, epilepsy, coma/brain damage, or intracranial injury.

^fLupus, Sjogren syndrome, rheumatoid arthritis, and other inflammatory arthritis.

^gDid not consider codes for benign tumors, basal cell carcinoma, squamous cell carcinoma, other nonmelanoma skin cancers, or localized prostate cancer.

Aducanumab exclusion criteria further reduced the number of available participants, narrowing those eligible to 12 (5.1% of 237). Notable exclusions (Table 2) were cardiovascular disease (50 [48.1%] based on ICD9/10 codes), CNS-related exclusions (e.g., brain cancer, epilepsy, intracranial injury, prior coma; 34 [32.7%]), imaging abnormalities (33 [31.7%]), and history of malignancy (29 [27.9%]). Cancer types (ever) detected in participants were bladder (3), breast (2), chronic myeloid leukemia in remission less than 3 years (1), colon (3), leukemia without mention of remission (1), liver (secondary; 1), melanoma (4), penile (1), prostate (5), adrenal gland (1), eye (choroid; 1), head/face/neck (1), genital organ (secondary; 1), lymphoma (1), lung or mesothelioma (2), and pancreatic cancer (1). As mentioned previously, we did not include basal cell carcinoma, squamous cell carcinoma, other nonmelanoma skin cancers, or localized prostate cancer. As presented in Figure 3, 58% of participants who met inclusion criteria (n = 104) met 2 or more exclusion criteria.

Among the 104 participants meeting the inclusion criteria, those finally eligible for aducanumab treatment (no exclusion criteria met) were more often APOE ε4 positive and had lower frequency of diabetes, congestive heart disease, or coronary heart disease (vs participants who met inclusion criteria but were excluded by exclusion criteria). Having 2 or more conditions was more frequent in eligible participants, but the frequency of having 3 or more conditions was lower in eligible than excluded participants. However, there were few eligible participants (n = 12) to draw any conclusions (Table 3).

Assessing Inclusion Criteria Aside From Amyloid PET

We also examined the number of participants who would need to be assessed at first for the inclusion criteria, regardless of PET availability or positivity. There were 1,077 participants with MCI and 59 with mild dementia (CDR global 0.5 or 1).

For lecanemab, 368 participants met the inclusion criteria aside from imaging (i.e., age 50–90 years, BMI >17 and <35, MMSE 22–30 inclusive, CDR global of 0.5 or 1, CDR memory score >0.5, logical memory II score within range for age group); 106 had a PiB PET scan (74 had amyloid PET positivity), and 232 did not have PiB PET. Participants with PET were significantly younger (76.8 [8.7] vs 80.9 [7.4] years, p < 0.001) and had a higher mean MMSE (25.8 [1.8] vs 25.1 [2.0] in the 22–30 range, p = 0.004) and lower frequency of 3+ comorbidities than those without PET (42.9% vs 57.5%).

For aducanumab, 403 MCSA participants met inclusion criteria aside from imaging (i.e., age 50–85 years, 6+ years of education, CDR global ≥0.5, MMSE 24–30 inclusive); 116 of these people had a PiB PET scan (74 had amyloid PET positivity), and 287 did not have PiB PET. There was a significantly higher proportion of men with PET vs without PET (66.4% vs 52.6%, p = 0.011), and participants with PET had a lower frequency of 3+ comorbidities than those without PET (49.5% vs 67.8%).

Sensitivity Analysis

Considering participants with the clinical diagnosis of MCI,¹⁴ without restricting CDR global, CDR memory, MMSE, or WMS-R Logical Memory II scores into specific boundaries, the fraction of eligible participants was higher, although most participants were again ineligible to receive treatment in this sensitivity analysis. In summary, there were 224 participants with a clinical diagnosis of MCI, 50–90 years, and amyloid-positive PET scan (eFigure 1, links.lww.com/WNL/D47). The sample was reduced to 203 participants after applying the lecanemab clinical trial inclusion criterion for BMI. The mean age (SD) was 81.0 (6.3) years, and 44.3% were female. Of the 203 participants, 197 were White, 2 Asian, 1 with unknown race, and 1 did not disclose race; 201 participants were not Hispanic or Latino, 1 with unknown race, and 1 did not disclose ethnicity. Applying the lecanemab clinical trial exclusion criteria next resulted in 39 (17.4% of 224) participants being eligible to receive treatment.

The sample (n=224) was reduced to 164 participants after applying the aducanumab clinical trial inclusion criteria for age and education. (eFigure 1, links.lww.com/WNL/D47) with mean age (SD) of 78.3 (5.3) years, and 47% were female. Of these 164 participants, 159 were white, 3 Asian, 1 with unknown race, and 1 did not disclose race; 163 were not Hispanic or Latino, and 1 had unknown ethnicity. The sample was further reduced to 20 (8.9% of 224) potentially eligible participants for aducanumab treatment. A similar pattern of chronic diseases in eligible participants for anti–β-amyloid mAb treatment was observed in the sensitivity analysis.

Discussion

Applying the clinical trial eligibility criteria resulted in a small fraction of participants (8% for lecanemab and 5.1% for aducanumab) being eligible to receive treatment, of participants with positive amyloid PET scan and early AD, mainly excluded due to other chronic conditions and neuroimaging findings. A sensitivity analysis, including participants with MCI, without CDR global, CDR-memory, MMSE, or WMS-R Logical Memory II scores restrictions, resulted in a higher percentage of eligible participants, especially for lecanemab (17.4% for lecanemab and 8.9% for aducanumab treatment). The study highlights the limited suitability of most persons with MCI or mild dementia with elevated brain amyloid for treatment with anti–β-amyloid mAbs because of the high prevalence of conditions or neuroimaging findings in older adult populations.

The study findings agree with recent reports^5–8 suggesting that a substantial proportion of patients (e.g., with AD, with ADRD, with MCI, with cognitive concerns and positive amyloid AD biomarkers, or patients attending a memory clinic) would not be eligible to receive aducanumab treatment. The fraction of patients excluded ranged from 99% (in patients attending a university memory clinic, applying the criteria of the clinical trial)⁷ to 91% (in Medicare beneficiaries with AD, applying the clinical trial criteria)⁸ to 43% (in patients with cognitive issues and positive AD biomarkers, applying the appropriate use recommendations for aducanumab therapy derived by an expert panel^5,32; of note, once Togher et al.⁵ applied the aducanumab clinical trial criteria, the proportion excluded increased from 43% to 73%). In the studies mentioned above, patients, in general, were excluded due to age, comorbid conditions, laboratory, and neuroimaging findings.

The number of eligible participants for any of the treatments was too small to draw any conclusions in the study related to chronic diseases in eligible participants; the frequency of chronic diseases such as diabetes, coronary artery disease, or congestive heart failure decreased dramatically when the exclusion criteria were applied, but not hypertension and dyslipidemia. Many conditions rendering patients ineligible are those we would expect in older adults, mainly for whom DMT is developed. This is not a surprising conundrum. Others have discussed it,³³ suggesting that many clinical trials for patients with chronic conditions (including dementia) exclude patients with other concomitant chronic conditions and medications.

Clinical trials usually use stringent eligibility criteria to achieve a homogeneous population, reduce clinical and biological heterogeneity of the targeted condition, and reduce safety and tolerability concerns.⁷ However, such a stringent selection process leads to an underrepresentation of the common comorbidities in the targeted clinical population, reducing the external validity (i.e., generalizability) of trial findings and patients' eligibility in the community.^7,34 Previous studies have noted that clinical trial participants are usually healthier, with higher education, younger, and less diverse in race and ethnicity than patients in real-world clinics, and treated in expert centers with close monitoring.³⁵ Treating older patients or with worse health might result in different treatment efficacy and safety than clinical trials.³⁵

Randomized controlled trials ultimately inform prescribing guidelines; their strict eligibility criteria exclude those with vulnerable characteristics (i.e., comorbidities, medications, vulnerabilities due to age).³⁶ Such exclusions might result in a gap in treatment decision information (risk or benefit) for potential patients falling outside the narrow clinical trial enrollment criteria. It is important to note that after the accelerated drug approval, a confirmatory trial is required to assess clinical benefit, and if findings are not confirmatory, the drug can be withdrawn from the market.³⁵ In addition, postmarketing surveillance of side effects and enrollment of patients receiving mAbs to registries (e.g., the Alzheimer Network for Treatment and Diagnostics)³⁷ has been recommended to assist in collecting data useful for any necessary drug use adjustment.³⁵ The Alzheimer Clinical Trials Consortium—an NIH-funded national consortium—has been created to accelerate the development of effective AD/ADRD interventions and aims to develop and implement cutting-edge strategies to accelerate efficient participant accrual and randomization, increase representativeness through enhanced diversity, and maximize participant retention.³⁸

In addition, developing appropriate resources (e.g., referral to specialists, undergoing amyloid PET, or measuring CSF AD biomarkers) and securing funds for such expenditures could be a great challenge for health care systems, and planning should be underway,⁵ although screening with plasma biomarkers for eligibility should largely eliminate problems of accessibility. In addition, resources needed to monitor anti–β-amyloid mAb side effects (e.g., MRI availability, staff support) might also stress already strained health care systems.

Liu et al.³⁹ assessed preparedness of the US health care system for a new DMT for AD, assuming it could arrive in 2020 and individuals would become eligible for annual MCI screening at age 55 years. In the United States, using PET scans for amyloid assessment, the simulation projected that individuals could wait 18.6 months for treatment in 2020 and 2.1 million patients would progress to Alzheimer dementia while on waiting lists between 2020 and 2040. Major constraints were the availability of specialists, access to imaging, or availability of infusion centers.³⁹ In our study, although the sample would not represent patients visiting a memory clinic, 1,136 patients (mainly with MCI) needed assessment and 32% met lecanemab inclusion criteria and would continue with further assessment (e.g., exclusion criteria and PET scan, as needed). We need to note that the simulation does not consider the use of plasma biomarkers for screening, which would result in increased throughput and is expected to cut waiting times dramatically. The limitations of the health care infrastructure are not unique to the US health care system but supported by a recent report⁴⁰ examining the readiness of the health care infrastructure in European countries for AD treatment projected, with the assumption that 90% of biomarker tests would be performed using CSF and 10% assessed by PET scans.

Adding to these concerns expressed is the potential of APOE genotyping becoming a requirement for amyloid-related imaging abnormalities risk stratification before initiating an anti-amyloid treatment, with inherent ethical and clinical resource consequences for standard care.⁵ These are not concerns that one stakeholder can solve; they need extensive discussion and planning among multiple stakeholders related to payment policy and regulatory requirements, planning for the adequate workforce and resource capacity as planning is underway at the local and national level, together with awareness campaigns for the communities and patients.³⁹ As such, access and equity issues will be addressed for those patients eligible for DMT.⁴¹

Appropriate use recommendations³⁵ have been created to help clinicians implement lecanemab treatment into real-world clinical practice and adhere closely to clinical trial inclusion and exclusion criteria. The use of lecanemab requires clinician and institutional preparedness, including protocols for the management of serious events, appropriate informational aids, and tools to help decision-making to assist patients and their caregivers in understanding potential benefits and adverse events.³⁵

The study sample was not racially and ethnically diverse, and we do not have information to assess how this lack of diversity in race and ethnicity (aspects of an individual's social identity)⁴² could affect the study's findings. Although MCSA's recruitment is limited by the racial and ethnic composition of the Olmsted County population (e.g., overall for all ages: White alone 82.5%, Black/African American alone 7.5%, American Indian and Alaskan Native alone 0.7%, Asian alone 6.7%, Hispanic or Latino 5.5%, White alone, not Hispanic or Latino 77.8%),⁴³ previous studies have reported underrepresentation both in AD observational studies and clinical trials³⁸ of Black and Hispanic persons, who are also more likely to be diagnosed with clinical AD/ADRD.⁴⁴ However, individuals from historically underrepresented groups could be eligible for anti-amyloid mAb treatments, and in discussions with patients and caregivers, the lack of or limited information on treatment response from randomized clinical trials must be transparently acknowledged.³⁵ Previous studies suggested that differences in ADRD risk between non-Hispanic Black (NHB) and non-Hispanic White (NHW) individuals could be driven by several risk factors (e.g., vascular diseases, structural and social determinants of health, genetics).⁴⁵ Ethnic and racial differences have been observed in amyloid PET positivity (e.g., lower odds of amyloid positivity in Asian, Black, and Hispanic participants with MCI and dementia in the IDEAS study⁴⁶ (vs White participants; not Hispanic or Latino, or not reported or unknown Hispanic origin); lower amyloid burden in NHB vs NHW CU participants who passed initial screening in the A4 study,⁴⁵ while within NHB, higher amyloid burden was observed in those with lower percentage of African ancestry); not all studies agree in such differences, and selection bias due to higher rates of screen failure in NHB⁴⁵ could be implicated. Differences in the APOE ε4 association with dementia between African American and White American is another example suggesting our limitations in generalizing findings in AD/ADRD research across diverse groups,³⁵ underlining that those participants in clinical trials need to represent all persons at risk for cognitive impairment.

This study has several strengths. MCSA is a population-based study with rich phenotypic data, comprehensive cognitive evaluations, and state-of-the-art neuroimaging that provides reliable measures of AD imaging biomarkers. In addition, we used the ability of the REP medical records linkage system to capture medication and comorbidities and reduce potential bias from self-report.

Findings need to be considered in light of the study's limitations. We were not able to define all inclusion/exclusion criteria as the clinical trials defined them. However, we strived for close modifications where needed. There is a potential for disease misclassification because of the use of ICD-9 or ICD-9 codes, resulting in misclassification of the exclusion criteria. Participants with PiB PET (vs those without) had a lower frequency of chronic conditions, which could imply that the fraction of finally eligible participants might be even lower if other noninvasive AD biomarkers would have been applied (e.g., plasma biomarkers). Olmsted County has overall a higher median household income than the overall US population,⁹ and the study sample had high mean years of education (that is an individual-level measure of socioeconomic status, as well); in addition, the study sample was not racially and ethnically diverse. Thus, assessing these eligibility criteria in more diverse populations would be crucial.

In conclusion, the eligibility for anti–β-amyloid mAb treatment of patients with early AD is limited due to exclusionary common conditions in the typical older adult populations. Continued, coordinated efforts are needed by many stakeholders to prevent disease, postpone progression, and develop DMT with good external validity (i.e., generalizability) to diverse populations.

Glossary

AD

Alzheimer disease

ADRD

AD and related disorders

APOE

apolipoprotein E

CDR

Clinical Dementia Rating

CU

cognitively unimpaired

DMT

disease-modifying therapy

FDA

Food and Drug Administration

FLAIR

fluid-attenuated inversion recovery

ICD-9/ICD-9

International Classification of Diseases, Ninth and Tenth Revision

mAbs

monoclonal antibodies

MCI

mild cognitive impairment

MCSA

Mayo Clinic Study of Aging

MMSE

Mini-Mental State Examination

MPRAGE

magnetization prepared rapid gradient echo

NHB

non-Hispanic Black

NHW

non-Hispanic White

PIB

C-Pittsburgh Compound B

Pib PET

Pittsburgh Compound B PET scan

REP

Rochester Epidemiology Project

SUVR

standard uptake value ratio

WMH

white matter hyperintensities

WMS-R

Wechsler Memory Scale-Revised

Appendix. Authors

Name	Location	Contribution
Rioghna R. Pittock	Department of Neurology, Mayo Clinic, Rochester, MN; The College, University of Chicago, Chicago, IL	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data; additional contributions (in addition to one or more of the above criteria)
Jeremiah A. Aakre, MPH	Department of Quantitative Health Sciences, Mayo Clinic, Rochester, MN	Drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data; additional contributions (in addition to one or more of the above criteria)
Anna M. Castillo, MSc	Department of Quantitative Health Sciences, Mayo Clinic, Rochester, MN	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data; additional contributions (in addition to one or more of the above criteria)
Vijay K. Ramanan, MD, PhD	Department of Neurology, Mayo Clinic, Rochester, MN	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data; additional contributions (in addition to one or more of the above criteria)
Walter K. Kremers, PhD	Department of Quantitative Health Sciences, Mayo Clinic, Rochester, MN	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data; additional contributions (in addition to one or more of the above criteria)
Clifford R. Jack, Jr., MD	Department of Radiology, Mayo Clinic, Rochester, MN	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data; additional contributions (in addition to one or more of the above criteria)
Prashanthi Vemuri, PhD	Department of Radiology, Mayo Clinic, Rochester, MN	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data; additional contributions (in addition to one or more of the above criteria)
Val J. Lowe, MD	Department of Radiology, Mayo Clinic, Rochester, MN	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data; additional contributions (in addition to one or more of the above criteria)
David S. Knopman, MD	Department of Neurology, Mayo Clinic, Rochester, MN	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data; additional contributions (in addition to one or more of the above criteria)
Ronald C. Petersen, MD, PhD	Department of Neurology, Mayo Clinic, Rochester, MN; Department of Quantitative Health Sciences, Mayo Clinic, Rochester, MN	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data; additional contributions (in addition to one or more of the above criteria)
Jonathan Graff-Radford, MD	Department of Neurology, Mayo Clinic, Rochester, MN	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data; additional contributions (in addition to one or more of the above criteria)
Maria Vassilaki, MD, PhD	Department of Quantitative Health Sciences, Mayo Clinic, Rochester, MN	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data; additional contributions (in addition to one or more of the above criteria)

Study Funding

This manuscript has been funded by NIH (U01 AG006786), NIH (P30 AG062677), NIH (R37 AG011378), NIH (R01 AG041851), NIH (R01 NS097495), NIH (AG 058738).

Disclosure

R.R. Pittock reports that her father has received personal compensation for serving on scientific advisory boards for F. Hoffmann-La Roche AG and his institution has received grants; his institution has received grants, personal fees, nonfinancial support, research support, and compensation for serving as a consultant for Alexion, AstraZeneca Rare Disease; he has a patent (Patent 9,891,219B2, Application 12–573942) issued and for which he has received royalties and also reports grants, personal fees, nonfinancial support, and other support from MedImmune/Horizon and has received a grant from Novelmed. J.A. Aakre and A.M. Castillo report no disclosures relevant to the manuscript. V.K. Ramanan receives research funding from the NIH and the Mangurian Foundation for Lewy Body disease research, has provided educational content for Medscape, and is an investigator in clinical trials sponsored by the Alzheimer's Association, Eisai, and the Alzheimer's Treatment and Research Institute at USC. W.K. Kremers receives research funding from NIH. C.R. Jack, Jr., has no conflicts to report. D.S. Knopman serves on a Data Safety Monitoring Board for the Dominantly Inherited Alzheimer Network Treatment Unit study. He served on a Data Safety Monitoring Board for a tau therapeutic for Biogen (until 2021) but received no personal compensation. He is an investigator in clinical trials sponsored by Biogen, Lilly Pharmaceuticals, and the University of Southern California. He has served as a consultant for Roche, Samus Therapeutics, Magellan Health, Biovie, and Alzeca Biosciences but receives no personal compensation. He attended an Eisai advisory board meeting for lecanemab on December 2, 2022, but received no compensation. He receives funding from the NIH. P. Vemuri receives research funding from NIH. V.J. Lowe serves as a consultant for Bayer Schering Pharma, Piramal Life Sciences, Life Molecular Imaging, Eisai Inc., AVID Radiopharmaceuticals, and Merck Research and receives research support from GE Healthcare, Siemens Molecular Imaging, AVID Radiopharmaceuticals, and the NIH (NIA, NCI). R.C. Petersen serves is a consultant for Roche, Inc., Eisai, Inc., Genentech, Inc. Eli Lilly, Inc., and Nestle, Inc., served on a DSMB for Genentech, receives royalties from Oxford University Press and UpToDate, and receives NIH funding. J. Graff-Radford receives support from the NIH, is an investigator in clinical trials sponsored by Esai and the Alzheimer's Treatment and Research Institute at USC, and is a member of the Data and Safety Monitoring Board (DSMB) for StrokeNet. M. Vassilaki has consulted for F. Hoffmann-La Roche Ltd; currently she receives research funding from NIH; she has equity ownership in Abbott Laboratories, Johnson and Johnson, Medtronic, Merck, and Amgen; her spouse receives research funding from Avobis Bio, LLC. Go to Neurology.org/N for full disclosures.