Dementia & Memory Care

Infusion Referral Workflow and Selection Checklist

A referral-ready checklist for anti-amyloid therapy: diagnosis, biomarker confirmation, imaging, APOE, anticoagulation, and the MRI plan for clinicians.

Prepared with AI assistance and checked against the linked sources. The update date records an editorial revision; it does not certify physician review.

Zaka Ahmed, MD

Vascular Neurologist · 28 October 2026 · 8 min read

TL;DR — Selection checklist

→ MCI or mild dementia stage with an Alzheimer's clinical syndrome — anti-amyloid therapy is not indicated for moderate-to-severe dementia or for non-Alzheimer's dementias.
→ Amyloid confirmed — CSF, amyloid PET, or a validated blood biomarker with confirmatory testing per protocol.
→ Baseline MRI reviewed — microbleed count, siderosis, prior hemorrhage, and other exclusionary findings checked.
→ APOE discussed and, per protocol, tested — ε4 homozygotes carry markedly higher ARIA risk and need explicit counselling.
→ Anticoagulation reviewed, MRI monitoring schedule booked, ARIA management plan and patient contact route documented, and consent recorded — including modest average benefit, ARIA risk, burden, and alternatives.

This page is written for referring clinicians. It gives a referral-ready selection checklist and workflow for anti-amyloid monoclonal antibody therapy in early Alzheimer's disease. The patient-facing companion page — "Can I Get an Alzheimer's Infusion?" — covers the same pathway in plain language; give it to patients at referral.

Who to refer

Refer adults with mild cognitive impairment (MCI) or mild dementia in whom Alzheimer's disease is the leading clinical syndrome, and where amyloid confirmation is either established or plausibly obtainable. The evidence base for anti-amyloid therapy comes from trial populations with early symptomatic disease and biomarker-confirmed amyloid; benefit outside that population is unproven, and trial data do not support use in moderate-to-severe dementia, in asymptomatic individuals, or in non-Alzheimer's dementias (including isolated vascular, Lewy body, or frontotemporal syndromes).

Referral is appropriate even when eligibility is uncertain. The infusion centre can complete biomarker confirmation, MRI review, and counselling — the referring clinician's role is to identify the syndrome, flag safety issues early, and send a complete packet (below). Do not delay referral for exhaustive workup; do flag anticoagulation, prior hemorrhage, seizures, and MRI contraindications in the referral letter.

Referral workflow, step by step

1. Referring clinician → referral sent. Include: syndrome summary and stage (MCI vs mild dementia, with the cognitive test used and score), comorbidity list, full medicine list (especially anticoagulants, antiplatelets, and thrombolysis-relevant agents), prior neuroimaging, and known allergies. Note any history of brain hemorrhage, stroke, seizures, or neurosurgery.

2. Records + imaging review. The infusion centre reviews prior MRI/CT and medicine history, and requests missing records. A baseline MRI within the required window is confirmed or ordered.

3. Biomarker confirmation. Amyloid status is confirmed by CSF (Aβ42/40 ratio, p-tau), amyloid PET, or a validated plasma biomarker with confirmatory testing per local protocol. Intermediate or discordant blood-biomarker results proceed to CSF or PET confirmation. Biomarker testing in asymptomatic individuals is a separate, counselled decision and is not part of this pathway.

4. MRI safety review. Baseline MRI is reviewed for microbleed count, superficial siderosis, prior lobar hemorrhage, vasogenic edema, and other exclusionary findings, per the therapy's labelling and local protocol. Findings above protocol thresholds exclude or defer therapy pending specialist review.

5. APOE discussion and testing (per protocol). APOE genotyping is offered and counselled, not assumed. ε4 homozygotes carry markedly higher ARIA risk (in trial data, roughly one in three developed ARIA-E, versus a small minority of non-carriers) and need explicit discussion of that risk, including the option to decline therapy. Testing also has family implications — counsel accordingly, and offer genetic-counselling referral where appropriate.

6. Consent visit. Benefits, risks, burden, and alternatives are documented (see "Consent documentation" below). The patient — and, with permission, family — should leave with written information and a contact route for symptoms.

7. Infusion scheduling + monitoring. Infusions are scheduled at the therapy's labelled interval, with MRI monitoring at protocol intervals (baseline plus scheduled safety scans, typically early in treatment and periodically thereafter — follow current labelling, which varies by agent). The referring clinician is copied on the plan, on each MRI safety result, and on any ARIA event or treatment pause.

Communication target: the referring clinician receives an update at each step — referral acknowledged, biomarker result, MRI decision, consent outcome, and treatment start — rather than a single letter at the end. If a step stalls (records missing, MRI delayed), the referrer is told promptly so the patient is not left waiting without explanation.

What to include in the referral packet

Syndrome and staging summary, with cognitive scores and dates
Full medicine list, including over-the-counter agents and supplements
Anticoagulation/antiplatelet details: agent, indication, and whether interruption is feasible
Prior neuroimaging (reports and, where possible, images)
Relevant history: hemorrhage, stroke, seizures, neurosurgery, prior ARIA-relevant findings
Allergies, MRI contraindications (implants, claustrophobia, renal function if contrast planned)
Patient contact details, language/interpreter needs, and family contact (with permission)
What the patient has already been told about infusions, and any stated preference

MRI, APOE, and anticoagulation — the three reviews that most often change decisions

MRI. Microbleed burden, siderosis, and prior hemorrhage are the most common imaging reasons for exclusion or deferral. A recent baseline MRI also serves as the comparator for ARIA detection on follow-up scans — without it, new edema or hemorrhage cannot be judged. If the baseline MRI is old or of limited quality, repeat it before consent, not after.

APOE. APOE ε4 dose correlates with ARIA risk, most steeply in homozygotes. Testing is a counselling intervention as much as a lab result: discuss risk in absolute terms, the option to decline testing or therapy, and the implications for relatives. Document the discussion and the decision.

Anticoagulation. Anticoagulant use raises the stakes of ARIA-H and of any hemorrhagic complication, and complicates emergency management. Review the indication (for example, atrial fibrillation stroke prophylaxis versus remote provoked thrombosis) — the decision is rarely "stop or continue" in isolation. Coordinate with the prescriber (cardiology, primary care) and document the agreed plan before the first infusion. Thrombolysis for acute stroke during anti-amyloid therapy is a separate emergency decision governed by stroke protocols and the treating team.

ARIA monitoring and management — what the referrer should know

ARIA (amyloid-related imaging abnormalities) comprises ARIA-E (edema/effusion) and ARIA-H (hemosiderin deposition, including microhemorrhages and siderosis). In trial populations, ARIA-E occurred in roughly 13% (lecanemab) to 24% (donanemab) of treated participants, with symptomatic cases in a minority (roughly 3–6%); most ARIA is detected on scheduled safety MRI before symptoms appear. Management is protocol-driven: asymptomatic radiographically mild ARIA is often monitored with continued dosing per labelling; symptomatic or radiographically moderate-to-severe ARIA usually prompts dose suspension, repeat MRI, and specialist review. Serious ARIA, including macrohemorrhage, is uncommon but possible and is a reason for urgent contact.

Patients should be told to report headache, confusion, dizziness, vision change, nausea, or any sudden neurological symptom promptly — and referrers should route such reports to the infusion centre the same day, not to routine follow-up. A one-page symptom card for the patient, with the centre's contact route, reduces delay.

Consent documentation

Document, at minimum:

Diagnosis and stage, and how amyloid was confirmed
Expected benefit framed honestly: modest slowing of decline on average in trial populations over 18 months — not improvement, not a cure, and no guarantee for an individual
ARIA risk in absolute terms, including APOE-related risk where tested
Infusion reactions, burden (visit frequency, travel, MRI schedule), and duration of therapy as currently understood
Alternatives, including symptomatic medicines, trial participation, and no disease-modifying therapy
The patient's decision, and family involvement with permission

Referral FAQs

Should I complete biomarker testing before referring?

Not necessarily. Referral with a clear syndrome summary is enough — the centre can complete biomarker confirmation. But do send any biomarker results already available, and flag discordant or intermediate results explicitly.

Does a patient on anticoagulation automatically fail selection?

No — anticoagulation triggers review, not automatic exclusion. The decision depends on the agent, the indication, bleeding history, and MRI findings, coordinated with the prescriber and documented before the first infusion.

What if the patient declines APOE testing?

Declining testing does not automatically exclude therapy, but it limits risk stratification. Document the discussion, counsel on the higher uncertainty, and follow local protocol on whether treatment proceeds without it.

How quickly should I expect a decision?

Timelines vary with records, imaging, and biomarker availability. Expect acknowledgement promptly and updates at each workflow step; chase the centre if a step stalls, and tell the patient what is pending so expectations stay realistic.

What happens if ARIA appears?

Management follows the therapy's labelling and local protocol: monitoring, dose suspension, or discontinuation depending on severity and symptoms, with repeat MRI. The referring clinician is copied on the event and the plan.

Can therapy start if MRI shows many microbleeds?

Microbleed burden above protocol thresholds generally excludes or defers therapy pending specialist review. Send the imaging with the referral so this is judged early, not after the patient has travelled for consent.

Takeaways

Refer early symptomatic Alzheimer's syndrome (MCI or mild dementia) with amyloid confirmation — or a clear path to it — and a complete packet.
The three reviews that most often change decisions are MRI, APOE, and anticoagulation: flag them in the referral, not after it.
Consent, monitoring, and communication are part of the therapy, not paperwork around it — document benefit honestly, book the MRI schedule, and keep the referrer copied at every step.

Refer a patient — send the packet above, flag MRI, APOE, and anticoagulation issues early, and the centre will confirm eligibility and keep you copied at each step. For the patient-facing version of this pathway, share "Can I Get an Alzheimer's Infusion?"

References: van Dyck CH, et al. Lecanemab in early Alzheimer's disease (Clarity AD). New England Journal of Medicine. 2023 · Sims JR, et al. Donanemab in early symptomatic Alzheimer disease (TRAILBLAZER-ALZ 2). JAMA. 2023 · Jack CR Jr, et al. Revised criteria for diagnosis and staging of Alzheimer's disease (NIA-AA). Alzheimer's & Dementia. 2024.

This page is for clinician education and referral guidance. Verify current labelling, local protocol thresholds, and coverage rules at the time of referral, as these change.