← Back

AMBAR Plasma Exchange

Year of Publication: 2020

Journal: Alzheimer's & Dementia

Link: https://doi.org/10.1002/alz.12137

PDF: https://alz-journals.onlinelibrary.wiley....1002/alz.12137

Clinical Question

Does plasma exchange with albumin replacement slow cognitive (ADAS-Cog) and functional (ADCS-ADL) decline in mild-to-moderate Alzheimer disease?

Bottom Line

In 347 patients with mild-to-moderate AD randomized to three plasma-exchange regimens with albumin ± IVIG replacement vs sham-PE, the PE-treated pooled group showed 52% less ADCS-ADL decline (p=0.03) and 66% less ADAS-Cog decline (p=0.06, trend) at 14 months. Benefits were largest in moderate AD (MMSE 18-21): 61% less ADCS-ADL and ADAS-Cog decline. CDR-SB and ADCS-CGIC also favored PE. Suggested PE with albumin replacement as a potential disease-modifying approach for AD; Grifols sponsored further study, but definitive phase 3 has not replicated benefit.

        Major Points
        Phase 2b/3 multicenter randomized sham-controlled parallel-group trial at 41 sites (19 Spain, 22 US), 2012-2016 (AMBAR, Boada 2020)
N=347 adults 55-85 with probable AD, MMSE 18-26, stable AChEI/memantine
1:1:1:1 randomization: sham-PE vs three PE regimens (low-alb, low-alb+IVIG, high-alb+IVIG); three PE arms pooled for main analysis
14-month protocol: 6 weeks weekly TPE + 12 months monthly low-volume PE with albumin ± IVIG
Co-primary outcomes ADCS-ADL and ADAS-Cog at 14 months by MMRM
ADCS-ADL: PE 52% less decline vs sham; p=0.03 (met co-primary)
ADAS-Cog: PE 66% less decline; p=0.06 (trend, did not meet co-primary)
Moderate-AD (MMSE 18-21) subgroup: 61% less ADCS-ADL decline (p=0.002) and 61% less ADAS-Cog decline (p=0.05)
CDR-SB 71% less decline (p=0.002); ADCS-CGIC 100% less decline (p<0.0001) favoring PE
No significant benefit in mild-AD (MMSE 22-26) subgroup — opposite to typical AD disease-modifying trials
Procedure burden substantial (weekly venous access for 6 weeks + monthly for 12 months); 20-35% dropout across PE arms
Grifols-sponsored; FDA has not approved PE for AD; definitive phase 3 confirmatory data pending

      

Design

Study Type: Phase 2b/3 multicenter randomized sham-controlled parallel-group trial

Randomization: 1

Blinding: Double-blind (sham PE comparator)

Follow-up Duration: 14 months

Sample Size: 347

Analyzed: 322

Analysis: Mixed model for repeated measures (MMRM); three PE arms pooled for main analysis

Baseline Characteristics

Characteristic	Control	Active
N	80	242
Age mean	68.4 ± 8.4	69.2 ± 7.4
Female	45%	57%
APOE ε4	44-48%	44-63%
MMSE	21.5	~21.5
Moderate AD	~45%	~45%

Arms

Field	Control	Low-albumin PE	Low-albumin + IVIG PE	High-albumin + IVIG PE
N	80	78	86	78
Intervention	Simulated noninvasive PE procedure mimicking active arms without fluid replacement	20g albumin LVPE monthly after 6 wk TPE	20g albumin alternating with 10g IVIG monthly after 6 wk TPE	40g albumin alternating with 20g IVIG monthly after 6 wk TPE
Duration	14 months (6 wk weekly + 12 mo monthly)	14 months	14 months	14 months

Outcomes

Outcome	Type	Control	Intervention	P-value
Change from baseline in co-primary ADCS-ADL and ADAS-Cog at 14 months (pooled PE vs sham)	Primary	Decline on both scales	Reduced decline on both scales	ADCS-ADL p=0.03; ADAS-Cog p=0.06 (trend)
CDR-SB change at 14 months (PE pooled vs sham)	Secondary	Decline	71% less decline	p=0.002
ADCS-CGIC change at 14 months	Secondary	Decline	100% less decline	p<0.0001
Moderate-AD (MMSE 18-21) ADCS-ADL	Secondary	Decline	61% less decline	p=0.002
Moderate-AD ADAS-Cog	Secondary	Decline	61% less decline	p=0.05
Mild-AD (MMSE 22-26) co-primary outcomes	Secondary	Decline	No significant effect	NS
CSF Aβ40, Aβ42, t-tau, p-tau biomarkers	Secondary	Reference	Modest changes; not primary biomarker outcome	Exploratory
PE procedure-related AE (any)	Adverse	Low (sham procedure only)	Common; mostly vascular-access related	Expected
Hypotension / dizziness during PE	Adverse	Rare	More common with TPE phase	PE-specific
Venous access complications	Adverse	None (sham)	Central line infections, thromboses	PE-specific
Transfusion reactions (albumin/IVIG)	Adverse	Rare	Occasional	Expected
Serious AE	Adverse	Similar baseline rate	Modestly higher (device, access)	Expected for invasive arm
Completion rate	Adverse	80.0%	65.1-78.2% across PE arms	Higher dropout with more invasive protocols
Death	Adverse	Few	Few	No imbalance

Subgroup Analysis

Strongest benefit in moderate-AD subgroup (MMSE 18-21) across ADCS-ADL, ADAS-Cog, CDR-SB, and ADCS-CGIC. No significant benefit in mild-AD (MMSE 22-26) subgroup. This paradoxical pattern (opposite to typical AD trials where earlier-stage patients benefit more) has been questioned — possible explanations include ceiling effects in mild disease, or that Aβ sequestration via PE works better when amyloid burden is greater. No clear dose-response between the three PE regimens.

Criticisms

Sponsor (Grifols) manufactures the albumin and IVIG used in the intervention — conflict of interest concern in a complex invasive protocol
Co-primary ADAS-Cog endpoint did not reach significance (p=0.06) — trial technically failed its cognitive co-primary
Moderate-AD subgroup benefit was larger than mild-AD, opposite to most disease-modifying trials — requires independent confirmation
Sham PE is difficult to fully blind — patients may infer treatment from sensation of venous access / procedure duration
High dropout (20-35%) across PE arms, driven by procedure burden; selection bias in completers possible
Mechanism (plasma Aβ sequestration) not directly demonstrated — CSF Aβ biomarker data not conclusive
Procedure burden (weekly venous access for 6 weeks + monthly for 12 months) and cost limit real-world adoption

Funding

Grifols (manufacturer of therapeutic albumin and IVIG)

Based on: AMBAR Plasma Exchange (Alzheimer's & Dementia, 2020)

Content summarized and formatted by NeuroTrials.ai.

AMBAR Plasma Exchange

(2020)

Objective

Plasma exchange with albumin replacement (with or without intravenous immunoglobulin) — Alzheimer's Management By Albumin Replacement (AMBAR) phase 2b/3 trial tested whether sequestration of plasma Aβ via PE slows cognitive and functional decline in mild-to-moderate AD.

Study Summary

• Plasma exchange with albumin replacement slowed ADCS-ADL decline 52% vs sham (p=0.03) at 14 months in pooled PE-treated vs placebo analysis.
• ADAS-Cog trend: 66% less decline (p=0.06) — did not reach statistical significance.
• Moderate-AD subgroup (MMSE 18-21) showed strongest benefit: 61% less ADCS-ADL decline (p=0.002) and 61% less ADAS-Cog decline (p=0.05).
• CDR-SB 71% less decline (p=0.002); ADCS-CGIC 100% less decline (p<0.0001) vs placebo.
• No effect seen in mild-AD (MMSE 22-26) subgroup.
• PE was feasible but invasive (central venous access, weekly for 6 weeks then monthly ×12 months); AEs related to vascular access.

Intervention

6 weeks of weekly conventional therapeutic plasma exchange (TPE) with albumin 5% replacement, followed by 12 months of monthly low-volume PE (LVPE) with albumin 20% ± IVIG. Three PE regimens: low-albumin, low-albumin+IVIG, high-albumin+IVIG. Sham-PE control.

Inclusion Criteria

Adults 55-85 years with probable AD (NINCDS-ADRDA), MMSE 18-26, stable AChEI and/or memantine for ≥3 months, no cerebrovascular disease, no contraindication to PE.

Study Design

Arms: Sham-PE (control) vs Pooled 3 PE regimens (low-alb, low-alb+IVIG, high-alb+IVIG)

Patients per Arm: Placebo (sham) 80; Pooled PE 242 (low-alb 78, low-alb+IVIG 86, high-alb+IVIG 78); total randomized 347

Outcome

• Co-primary ADCS-ADL (14 mo): PE 52% less decline vs placebo; p=0.03
• Co-primary ADAS-Cog (14 mo): PE 66% less decline; p=0.06 (trend, did not meet significance)
• Moderate-AD subgroup: ADCS-ADL p=0.002 (61% less decline); ADAS-Cog p=0.05 (61% less decline)
• Mild-AD subgroup: no significant difference
• Secondary endpoints: CDR-SB (71% less decline; p=0.002), ADCS-CGIC (100% less decline; p<0.0001) both favored PE

Bottom Line

Major Points

Phase 2b/3 multicenter randomized sham-controlled parallel-group trial at 41 sites (19 Spain, 22 US), 2012-2016 (AMBAR, Boada 2020)
N=347 adults 55-85 with probable AD, MMSE 18-26, stable AChEI/memantine
1:1:1:1 randomization: sham-PE vs three PE regimens (low-alb, low-alb+IVIG, high-alb+IVIG); three PE arms pooled for main analysis
14-month protocol: 6 weeks weekly TPE + 12 months monthly low-volume PE with albumin ± IVIG
Co-primary outcomes ADCS-ADL and ADAS-Cog at 14 months by MMRM
ADCS-ADL: PE 52% less decline vs sham; p=0.03 (met co-primary)
ADAS-Cog: PE 66% less decline; p=0.06 (trend, did not meet co-primary)
Moderate-AD (MMSE 18-21) subgroup: 61% less ADCS-ADL decline (p=0.002) and 61% less ADAS-Cog decline (p=0.05)
CDR-SB 71% less decline (p=0.002); ADCS-CGIC 100% less decline (p<0.0001) favoring PE
No significant benefit in mild-AD (MMSE 22-26) subgroup — opposite to typical AD disease-modifying trials
Procedure burden substantial (weekly venous access for 6 weeks + monthly for 12 months); 20-35% dropout across PE arms
Grifols-sponsored; FDA has not approved PE for AD; definitive phase 3 confirmatory data pending

Study Design

Study Type: Phase 2b/3 multicenter randomized sham-controlled parallel-group trial
Randomization: Yes
Blinding: Double-blind (sham PE comparator)
Sample Size: 347
Follow-up: 14 months

Primary Outcome

Definition: Change from baseline in co-primary ADCS-ADL and ADAS-Cog at 14 months (pooled PE vs sham)

Control	Intervention	HR/OR	P-value
Decline on both scales	Reduced decline on both scales	- (Reported as LS means with SEM; CIs narrow but not tabulated here)	ADCS-ADL p=0.03; ADAS-Cog p=0.06 (trend)

Limitations & Criticisms

Sponsor (Grifols) manufactures the albumin and IVIG used in the intervention — conflict of interest concern in a complex invasive protocol
Co-primary ADAS-Cog endpoint did not reach significance (p=0.06) — trial technically failed its cognitive co-primary
Moderate-AD subgroup benefit was larger than mild-AD, opposite to most disease-modifying trials — requires independent confirmation
Sham PE is difficult to fully blind — patients may infer treatment from sensation of venous access / procedure duration
High dropout (20-35%) across PE arms, driven by procedure burden; selection bias in completers possible
Mechanism (plasma Aβ sequestration) not directly demonstrated — CSF Aβ biomarker data not conclusive
Procedure burden (weekly venous access for 6 weeks + monthly for 12 months) and cost limit real-world adoption

Citation

View Original Publication →

AMBAR Plasma Exchange

Clinical Question

Bottom Line

Major Points

Design

Baseline Characteristics

Arms

Outcomes

Subgroup Analysis

Criticisms

Funding

Ask about AMBAR Plasma Exchange