DIAN-TU
(2025)Objective
To assess whether long-term high-dose gantenerumab treatment removes amyloid plaques and delays clinical symptom onset in asymptomatic carriers of dominantly inherited Alzheimer's disease (DIAD) mutations.
Study Summary
• 30% of participants converted from amyloid-abnormal to normal; 45% in longest-treated subgroup
• Longest-treated group (~8.4 years) showed non-significant ~50% slowing of CDR-SB progression (HR 0.57, 95% CI 0.31–1.07)
• CSF Aβ42/40 and p-tau181 significantly improved toward normal levels
• ARIA-E incidence 30%; ARIA-H (microhemorrhage) 42%; no deaths from ARIA
• Study stopped early by sponsor due to lack of regulatory path for gantenerumab
Intervention
Open-label extension of phase 2/3 DIAN-TU-001 platform trial. 73 DIAD mutation carriers received gantenerumab up to 1500 mg SC every 2 weeks for up to 3 years. Compared to external (DIAN Observational Study) and internal (prior placebo-treated) controls. Primary endpoint: PiB-PET SUVR change. Key secondary: CDR-SB and CDR global progression.
Inclusion Criteria
Prior participation in DIAN-TU-001 double-blind period; Known carrier of PSEN1, PSEN2, or APP mutation causing dominantly inherited Alzheimer's disease; Willing to know mutation status; Original enrollment: −15 to +10 years from estimated symptom onset; CDR global 0 (cognitively normal) to 1 (mild dementia) at original enrollment
Study Design
Arms: Array
Patients per Arm: Gantenerumab OLE: 73 (55 mITT for primary); Any Gant: 53 asymptomatic; Longest Gant: 22; Main controls: 74; Extended controls: 86
Outcome
• CDR global (Any Gant vs controls): HR 0.84 (0.40–1.77, NS)
• CDR-SB recurrent progression (Any Gant): HR 0.83 (0.50–1.36, NS)
• Longest Gant CDR-SB: HR 0.57 (0.31–1.07, p=0.08)
• Longest Gant CDR global: HR 0.41 (0.14–1.20)
• CSF Aβ42/40 improved (−42 CentiMarker, p<0.0001)
• CSF p-tau181 improved (−21 CentiMarker, p<0.0001)
• No significant differences in cognitive composite, MMSE, or FAS
Bottom Line
High-dose gantenerumab significantly removed amyloid plaques over 3 years (PiB-PET SUVR −0.71, p<0.0001), with 45% of the longest-treated subgroup achieving amyloid-normal status. While clinical endpoints were not statistically significant overall, the longest-treated group (~8.4 years) showed a clinically meaningful but statistically uncertain ~50% slowing of CDR-SB progression (HR 0.57). CSF biomarkers of amyloid and tau pathology normalized. These findings provide preliminary evidence that early, sustained amyloid removal may delay symptom onset in DIAD, though the study was underpowered and stopped early.
Major Points
- Primary endpoint met: PiB-PET SUVR decreased by −0.71 (95% CI −0.88 to −0.53, p<0.0001) from OLE baseline to year 3, confirming substantial amyloid plaque removal
- Amyloid removal was dose-dependent: 3-fold higher doses produced 3-fold greater annual SUVR reduction (0.35 vs 0.06 per year)
- At OLE end, 15% converted amyloid-abnormal to normal; in the longest-treated group, 45% achieved amyloid-normal status
- Any Gant group (n=53 asymptomatic): CDR-SB HR 0.83 (0.50–1.36) and CDR global HR 0.84 (0.40–1.77) — not significant
- Longest Gant group (n=22, ~8.4 years treatment): CDR-SB HR 0.57 (0.31–1.07), CDR global HR 0.41 (0.14–1.20) — suggestive of ~50% slowing but not significant
- CSF Aβ42/40 improved significantly (−42 CentiMarker at year 3, p<0.0001), approaching normal levels in asymptomatic participants
- CSF p-tau181/tau181 improved significantly (−21 CentiMarker at year 3, p<0.0001)
- No significant treatment effect on Tau-PET, cognitive composite, MMSE, or FAS
- ARIA incidence: ARIA-E 30% (22/73), ARIA-H microhemorrhage 42% (31/73); higher with APOE ε4 carriers. No ARIA-related deaths. Higher doses produced ~33% more ARIA-E than double-blind period
- Study stopped early by sponsor (Roche) due to lack of regulatory path for gantenerumab, after most participants completed ~2 years of OLE treatment
Study Design
- Study Type
- Open-label extension of randomized, double-blind, placebo-controlled phase 2/3 platform trial
- Randomization
- Yes
- Blinding
- Open-label (OLE); assessors blinded to prior treatment assignment
- Sample Size
- 73
- Follow-up
- Median 2.64 years OLE (IQR 1.99–2.90); longest-treated subgroup ~8.4 years total
- Centers
- 18
- Countries
- Australia, Canada, France, Ireland, Puerto Rico, Spain, UK, USA
Primary Outcome
Definition: PiB-PET SUVR change from OLE baseline to year 3: −0.71 (95% CI −0.88 to −0.53, p<0.0001) | PiB-PET SUVR at year 1: −0.12 (p=0.012); year 2: −0.47 (p<0.0001) | 15% converted abnormal to normal; 15% remained normal; 70% remained abnormal | Longest-treated group: 45% achieved amyloid-normal status
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| - | p<0.0001 |
Limitations & Criticisms
- Study stopped early by sponsor, limiting statistical power and duration of follow-up
- Small sample size (n=73 OLE, n=22 longest-treated) insufficient for definitive clinical conclusions
- Open-label design introduces potential measurement and selection bias despite blinding of assessors
- Use of external controls from observational study introduces confounding
- Only 18% completed the planned 3-year treatment; 64% stopped due to sponsor termination
- Incomplete amyloid removal — 70% of overall group remained amyloid-abnormal at study end
- No significant effect on tau PET, cognitive composite, MMSE, or functional measures in overall group
- ~1 year treatment gap between double-blind and OLE periods may have reduced treatment effect
- Results in dominantly inherited AD may not generalize to sporadic late-onset AD
Citation
medRxiv 2025.01.29; doi: 10.1101/2024.10.29.24316289