Vascular Neurology Considerations for Antiamyloid Immunotherapy: A Science Advisory From the American Heart Association

Three anti-Aβ monoclonal antibodies are FDA-approved for treatment of mild cognitive impairment or mild dementia attributable to AD (as confirmed by positive amyloid imaging or cerebrospinal fluid testing). Aducanumab, a monoclonal antibody selected for reactivity with Aβ aggregates, was granted accelerated approval on the basis of 1 of 2 phase 3 trials in which high-dose antibody reduced decline in Clinical Dementia Rating Scale Sum of Boxes score¹; however, its commercialization has been discontinued by its manufacturer. Lecanemab, a monoclonal antibody reactive to Aβ protofibrils, was initially granted accelerated approval and subsequently full approval on the basis of phase 3 data showing slowing Clinical Dementia Rating Scale Sum of Boxes decline.² Donanemab, a third monoclonal antibody directed against a pyroglutamate-containing form of Aβ present in AD plaques, slowed decline in the integrated Alzheimer Disease Rating Scale in a phase 3 trial,³ and recently received full FDA approval. ARIA-E and ARIA-H events were associated with drug treatment in each of these phase 3 studies, indicating a class effect. ARIA-E was detected in 35% of participants who received high-dose aducanumab, 12.6% who received lecanemab, and 24% who received donanemab; the corresponding figures for ARIA-H were 28% for high-dose aducanumab, 17.3% for lecanemab, and 31.4% for donanemab. The large majority of these events were asymptomatic or only mildly symptomatic with transient reversible manifestations such as headache, nausea, or altered mental status; however, a subset of ≈1% to 2% of treated individuals had ARIA-E that was classified as a severe adverse event, and ≈0.5% had ICH >1 cm in diameter (Table 1). ^10–13 Whether variations in ARIA incidence across studies represent true differences in antibody properties versus trial-specific features of participant selection or design is unclear.

The appearance of vasogenic edema and hemorrhagic lesions in ARIA closely resembles that of cerebral amyloid angiopathy (CAA)–related inflammation (CAA-ri),¹⁴ suggesting a close relationship between the 2 syndromes. Patients with CAA-ri have been reported to have elevated cerebrospinal fluid concentrations of Aβ autoantibodies,¹⁵ indicating that it may represent a spontaneous version of the iatrogenic ARIA syndrome. The observation that markers of CAA such as lobar cerebral microbleeds or cortical superficial siderosis on baseline magnetic resonance imaging (MRI) predict increased future risk of developing ARIA-E (Table 2)^12,16 supports a model for ARIA based on binding of administered antibodies to the vascular Aβ deposits that comprise CAA (Figure 3). Aβ immunologically cleared from plaques may also relocate into vessels, exacerbating CAA severity.¹⁷

The occurrence of symptomatic ICH as a small subset of ARIA-H events (Table 1 and Figure 2) raises concerns that concomitant use of antithrombotic or thrombolytic agents may increase the likelihood or severity of ICH. Analysis of individuals receiving concomitant lecanemab and anticoagulation reported symptomatic ICH in 2 of 83 (2.4%) in the phase 3 randomized study and 4 of 139 (2.9%, excluding the additional thrombolytic-related ICH discussed in the following) in the phase 3 plus open-label extension.¹³ If these proportions of ICH on anticoagulation are replicated in future studies, they would represent a substantial counterweight to the benefits of immunotherapy, given the nearly 50% short-term mortality risk of anticoagulant-related ICH.²⁷ Antiplatelet monotherapy appears to be reasonably well-tolerated with immunotherapy; lecanemab clinical trial data suggest no excess risk of ARIA when antiplatelet therapy is combined with antiamyloid therapy.¹³ Immunotherapy trial data are not available regarding the effects of dual antiplatelet therapy; however, in patients with cardiovascular or cerebrovascular disease, dual versus single antiplatelet therapy appears to increase risk for intracranial hemorrhage ≈1.4-fold.²⁸ FDA labels have advised “additional caution” in use of concomitant antithrombotic agents, and Appropriate Use Recommendations have recommended against concomitant use of anticoagulants until more data demonstrating safety are available (Table 4).

Considerations around administration of thrombolytic agents to immunotherapy-treated patients are further complicated by the unpredictability and major clinical effects of the events that trigger their emergent use: stroke, pulmonary embolism, and myocardial infarction. In the absence of extensive data on thrombolysis in the setting of Aβ immunotherapy, much attention has focused on reports of fatal multifocal ICHs in 1 lecanemab-treated patient and 1 donanemab-treated patient after intravenous administration of alteplase or tenecteplase, respectively.^6,29 The former event occurred in a 65-year-old woman who presented with a stroke-like syndrome 4 days after her third dose of open-label lecanemab with postmortem pathology demonstrating severe CAA-ri⁶; the latter was a 72-year-old man with stroke-like symptoms 9 days after a fifth dose of open-label donanemab who did not have postmortem examination.²⁹ The FDA labels for aducanumab and lecanemab advise “additional caution” when considering thrombolysis, whereas the FDA label for donanemab notes that “because ARIA-E can cause focal neurologic deficits that can mimic an ischemic stroke, treating clinicians should consider whether such symptoms could be due to ARIA-E before giving thrombolytic therapy.”²³ The Appropriate Use Recommendation for lecanemab⁵ recommends against use of thrombolytics until further data demonstrate safety (Table 4). An additional practical step in anticipation of unexpected events is advance care planning discussions between the patient and physician regarding the patient’s preference.³⁰ In stroke care practice, it will be important to consider the possibility that acute ARIA-related symptoms might appear as a stroke mimic requiring urgent MRI for correct diagnosis, particularly in the context of ARIA risk factors, such as APOE ε4/ε4 genotype or recent initiation of immunotherapy (see Table 2). An additional important consideration for practice is that mechanical thrombectomy performed without thrombolytic administration for large-vessel occlusion does not appear to increase risk for symptomatic ICH even in the presence of cerebral microbleeds³¹ and is therefore likely a safe option to be considered when evaluating possible acute stroke in patients receiving immunotherapy. The considerations discussed here and methods for identifying individuals being treated with amyloid immunotherapy should be disseminated across the acute stroke team.

FDA recommendations for management of ARIA are to suspend dosing for moderate to severe ARIA-E or ARIA-H MRI findings (defined as appearance of a hyperintense lesion at least 5 cm in greatest dimension or multiple hyperintense foci, at least 5 new cerebral microbleeds, or at least 2 new foci of cortical superficial siderosis), moderate to severe ARIA-E clinical symptoms (defined as sufficient to affect daily activities), or any symptomatic ARIA-H (including all ICHs), and to base the decision to resume dosing on resolution of imaging findings and symptoms and on clinical judgement.^19,20,23 The Appropriate Use Recommendations suggest the additional option of anti-inflammatory treatment, such as corticosteroids.^5,21 Looking to the experience with CAA-ri, a retrospective observational study found that patients with CAA-ri whose initial clinical presentations were treated with anti-inflammatory agents (most commonly intravenous methylprednisolone 1 g daily for 3 to 5 days; then, oral prednisone 60 mg daily tapered to discontinuation over several months) were more likely to have clinical and neuroimaging improvement and demonstrated longer time to CAA-ri recurrence than those who were untreated.³² These observations offer some support for high-dose steroid treatment for severe ARIA-E, with the caveat that results from the endogenously generated immune response in spontaneous CAA-ri may not fully generalize to the exogenously generated syndrome of ARIA.

For care of patients with ARIA-H presenting as symptomatic ICH, there is no reason to deviate from existing ICH guidelines.³³ Evacuation of the types of lobar ICHs that are often related to CAA does not appear to carry increased risk of surgical complication.^34,35

Identifying modifications of treatment protocols that reduce risk or severity of ARIA appears a promising future direction for optimizing Aβ immunotherapy. An intriguing recent observation that both lower mean arterial pressure and use of an antihypertensive medication—potentially an indication of better blood pressure control—at initiation of donanemab treatment independently predicted reduced ARIA-E incidence¹² raises the possibility that blood pressure may interact with the vascular pathways that generate ARIA. Previous studies among patients with probable CAA have suggested association of hypertension with increased risk of ICH recurrence³⁶ and greater white matter atrophy,³⁷ highlighting likely interactions between blood pressure and small vessels with CAA.

To summarize key vascular considerations related to amyloid immunotherapy:

Optimizing the benefit to risk tradeoff of Aβ immunotherapy for AD will require better predictors of severe or life-threatening ARIA, mechanism-based prevention and treatment strategies, and delineation of the decision tipping point for determining when benefit outweighs risk. Development and validation of novel approaches and evidence-based treatment algorithms⁴¹ in each of these areas will be advanced by collection of real-world data on treated individuals as required by the Centers for Medicare & Medicaid Services for Medicare coverage of lecanemab. The Centers for Medicare & Medicaid Services–approved ALZ-NET Alzheimer’s National Registry for Treatment and Diagnostics (www.alznetproviders.org) has committed to collecting and sharing (in de-identified form) the full range of patient-level data, outcomes, biomarkers, and neuroimaging studies required for creating firm guidance for implementing this new approach to AD-modifying therapy.