← Back

GRADUATE I & II

Two Phase 3 Trials of Gantenerumab in Early Alzheimer's Disease

Year of Publication: 2023

Authors: Randall J. Bateman, Janice Smith, Michael C. Donohue, ..., for the GRADUATE I and II Investigators and the Gantenerumab Study Group

Journal: New England Journal of Medicine

Citation: N Engl J Med 2023;389:1862-76

Link: https://www.nejm.org/doi/full/10.1056/NEJMoa2304430

PDF: https://discovery.ucl.ac.uk/10184309/1/nejmoa2304430.pdf

Clinical Question

Does gantenerumab slow cognitive and functional decline compared with placebo in patients with early symptomatic Alzheimer's disease?

Bottom Line

Gantenerumab did not significantly slow clinical decline over 116 weeks despite partial amyloid plaque removal. Only approximately one quarter of participants achieved amyloid-negative status, suggesting that rapid and complete plaque clearance may be necessary for clinical benefit within the trial timeframe.

        Major Points
        Primary endpoint not met in either trial: CDR-SB difference was −0.31 (P=0.10) in GRADUATE I and −0.19 (P=0.30) in GRADUATE II
Amyloid reduction was lower than predicted: mean amyloid levels remained elevated (40-45 centiloids) after treatment
Only 28.0% (GRADUATE I) and 26.8% (GRADUATE II) of gantenerumab participants achieved amyloid-negative status (≤24 centiloids)
No treatment effect on tau accumulation on PET despite reductions in CSF and plasma phosphorylated tau 181
Greater brain volume loss and ventricular enlargement observed with gantenerumab vs placebo on volumetric MRI
ARIA-E incidence was 24.9% overall, approximately doubling with each APOE ε4 allele (13.1% noncarriers, 24.5% heterozygotes, 47.8% homozygotes)
Symptomatic ARIA-E occurred in 5.0% of gantenerumab participants; serious symptomatic ARIA-E in 1.1%
Results suggest rapid plaque reduction to below detection threshold may be necessary for clinical efficacy

      

Design

Study Type: Two parallel, multicenter, randomized, double-blind, placebo-controlled phase 3 trials

Randomization: 1

Blinding: Double-blind; personnel preparing/administering drug not involved in assessments; clinical raters unaware of assignments; CDR raters had no ARIA data; MRIs assessed by independent blinded neuroradiologists

Enrollment Period: Not specified

Follow-up Duration: 116 weeks (extended from planned 104 weeks due to COVID-19)

Centers: 288

Countries: United States, Canada, United Kingdom, Germany, Spain, France, Japan, Australia, and 22 others

Sample Size: 1965

Analysis: Conditional mean imputation followed by analysis of covariance with fixed hierarchical testing procedure; MMRM sensitivity analyses performed

Inclusion Criteria

Age 50 to 90 years
Mild cognitive impairment or mild dementia due to Alzheimer's disease (NIA-AA criteria)
CDR-Global Score of 0.5 or 1
MMSE score ≥22
Cognitive impairment with FCSRT cueing index ≤0.67 and free recall score ≤27
Amyloid-positive on PET visual reading or CSF p-tau181/Aβ42 ratio >0.024

Exclusion Criteria

Use of anticoagulants or GV-971
Clinically significant MRI findings: >5 microhemorrhages, >2 lacunar infarcts, Fazekas score of 3, or other lesions that could cause cognitive impairment

Baseline Characteristics

GRADUATE I - Gantenerumab (N=499):

Age - years: 71.1 ± 7.9
Female sex - %: 58.1
AD medication use - %: 62.5
MCI due to AD - %: 55.1
Mild dementia due to AD - %: 44.9
CDR-GS 0.5 - %: 68.9
CDR-GS 1 - %: 29.9
CDR-SB score: 3.71 ± 1.67
ADAS-Cog13 score: 28.1 ± 7.1
ADCS-ADL score: 67.9 ± 7.2
FAQ score: 8.0 ± 5.9
MMSE score: 23.5 ± 3.3
APOE ε4 no allele - %: 34.7
APOE ε4 one allele - %: 47.1
APOE ε4 two alleles - %: 18.2
Amyloid burden on PET - centiloids: 94.44 ± 26.48

GRADUATE I - Placebo (N=485):

Age - years: 72.1 ± 7.8
Female sex - %: 52.6
AD medication use - %: 60.8
MCI due to AD - %: 54.2
Mild dementia due to AD - %: 45.8
CDR-GS 0.5 - %: 74.0
CDR-GS 1 - %: 25.4
CDR-SB score: 3.71 ± 1.57
ADAS-Cog13 score: 28.1 ± 6.8
ADCS-ADL score: 68.2 ± 6.8
FAQ score: 7.8 ± 5.7
MMSE score: 23.6 ± 3.0
APOE ε4 no allele - %: 32.4
APOE ε4 one allele - %: 49.7
APOE ε4 two alleles - %: 17.9
Amyloid burden on PET - centiloids: 96.07 ± 31.47

GRADUATE II - Gantenerumab (N=498):

Age - years: 71.6 ± 7.8
Female sex - %: 57.8
AD medication use - %: 66.5
MCI due to AD - %: 54.0
Mild dementia due to AD - %: 46.0
CDR-GS 0.5 - %: 69.2
CDR-GS 1 - %: 30.2
CDR-SB score: 3.67 ± 1.61
ADAS-Cog13 score: 28.1 ± 6.9
ADCS-ADL score: 68.3 ± 7.3
FAQ score: 7.7 ± 5.8
MMSE score: 23.6 ± 3.1
APOE ε4 no allele - %: 33.1
APOE ε4 one allele - %: 48.6
APOE ε4 two alleles - %: 18.3
Amyloid burden on PET - centiloids: 95.62 ± 30.76

GRADUATE II - Placebo (N=477):

Age - years: 71.8 ± 7.4
Female sex - %: 59.7
AD medication use - %: 66.0
MCI due to AD - %: 55.8
Mild dementia due to AD - %: 44.2
CDR-GS 0.5 - %: 75.5
CDR-GS 1 - %: 24.3
CDR-SB score: 3.52 ± 1.54
ADAS-Cog13 score: 28.2 ± 7.0
ADCS-ADL score: 68.9 ± 7.2
FAQ score: 6.8 ± 5.3
MMSE score: 23.8 ± 3.2
APOE ε4 no allele - %: 32.7
APOE ε4 one allele - %: 53.2
APOE ε4 two alleles - %: 14.0
Amyloid burden on PET - centiloids: 90.70 ± 30.80

Arms

Field	Control	Gantenerumab
Intervention	Placebo subcutaneous injection every 2-4 weeks matching gantenerumab titration schedule	Gantenerumab subcutaneous injection, titrated over 36 weeks: 120 mg every 4 weeks (3 doses) → 255 mg every 4 weeks (3 doses) → 510 mg every 4 weeks (3 doses) → 510 mg every 2 weeks (maintenance)
Duration	116 weeks	116 weeks

Outcomes

Outcome	Type	Control	Intervention	P-value
Change from baseline in CDR-SB score at week 116 (range 0-18, higher = worse)	Primary	GRADUATE I - Gantenerumab: 3.35 ± 0.14 · GRADUATE I - Placebo: 3.65 ± 0.16 · GRADUATE I - Difference: −0.31 (95% CI −0.66 to 0.05) · GRADUATE I - P-value: 0.10 · GRADUATE II - Gantenerumab: 2.82 ± 0.14 · GRADUATE II - Placebo: 3.01 ± 0.15 · GRADUATE II - Difference: −0.19 (95% CI −0.55 to 0.17) · GRADUATE II - P-value: 0.30 · Pooled - Difference: −0.26 (95% CI −0.51 to −0.01)
ADAS-Cog13 score change at week 116 (GRADUATE I)	Secondary	9.82 ± 0.57	8.57 ± 0.47	Not significant (hierarchical testing failed)
ADAS-Cog13 score change at week 116 (GRADUATE II)	Secondary	7.94 ± 0.49	6.66 ± 0.42	Not significant (hierarchical testing failed)
ADCS-ADL score change at week 116 (GRADUATE I)	Secondary	−12.32 ± 0.69	−11.21 ± 0.60	Not significant
Amyloid PET change at week 116 (GRADUATE I)	Secondary	+9.06 centiloids	−57.38 centiloids	<0.001
Amyloid PET change at week 116 (GRADUATE II)	Secondary	+8.46 centiloids	−48.00 centiloids	<0.001
Amyloid-negative status at week 116 (GRADUATE I)	Secondary	2.4%	28.0%
Amyloid-negative status at week 116 (GRADUATE II)	Secondary	0%	26.8%
Tau PET medial temporal change at week 116 (pooled)	Secondary	Median SUVR 1.50	Median SUVR 1.47	Not significant
Any adverse event (pooled)	Adverse	832 (87.1%)	905 (90.1%)
Serious adverse event (pooled)	Adverse	158 (16.5%)	137 (13.6%)
Death (pooled)	Adverse	14 (1.5%)	10 (1.0%)
Discontinuation due to AE (pooled)	Adverse	17 (1.8%)	91 (9.1%)
Any ARIA-E (pooled)	Adverse	26 (2.7%)	247 (24.9%)
ARIA-E - APOE ε4 noncarriers (pooled)	Adverse	9 (2.9%)	44 (13.1%)
ARIA-E - APOE ε4 heterozygotes (pooled)	Adverse	10 (2.1%)	117 (24.5%)
ARIA-E - APOE ε4 homozygotes (pooled)	Adverse	7 (4.7%)	86 (47.8%)
Symptomatic ARIA-E (pooled)	Adverse	2 (0.2%)	50 (5.0%)
Serious symptomatic ARIA-E (pooled)	Adverse	0	11 (1.1%)
Recurrent ARIA-E (pooled)	Adverse	3 (0.3%)	95 (9.6%)
Any new ARIA-H (pooled)	Adverse	116 (12.3%)	227 (22.9%)
New isolated ARIA-H (pooled)	Adverse	108 (11.4%)	85 (8.6%)
Injection site reactions (pooled)	Adverse	7.7%	16.8%
Intraparenchymal macrohemorrhage (pooled)	Adverse	1.0%	1.4%

Subgroup Analysis

Post hoc exploratory analysis of clinical response in participants achieving amyloid-negative status was performed but no definitive conclusions could be drawn. The pooled analysis of both trials showed a difference directionally favoring gantenerumab (−0.26; 95% CI −0.51 to −0.01) but was not part of hierarchical testing.

Criticisms

Amyloid plaque removal was lower than predicted despite similar ARIA-E incidence as modeled, suggesting different mechanisms may drive ARIA and plaque clearance
Only ~27% achieved amyloid-negative status, potentially insufficient for clinical benefit
The dose-escalation scheme was previously untested in this trial design
Lack of racial diversity in U.S. trial population limits generalizability
COVID-19 pandemic caused missed visits and delayed assessments, requiring extension to 116 weeks
Greater brain volume loss with gantenerumab (though potentially related to amyloid removal rather than neurodegeneration)
No effect on tau PET accumulation despite CSF/plasma p-tau181 reductions
Subcutaneous administration may result in different pharmacokinetics compared to IV antibodies that showed efficacy
Trial design differences from other anti-amyloid antibody trials make cross-trial comparisons difficult

Funding

F. Hoffmann–La Roche

Based on: GRADUATE I & II (New England Journal of Medicine, 2023)