Spectrum of amyloid-related imaging abnormalities associated with aducanumab treatment: a case series report

Abstract

Background

Amyloid-related imaging abnormalities (ARIA) observed on magnetic resonance imaging (MRI) are known adverse events associated with beta-amyloid plaque-lowering monoclonal antibodies. A spectrum of radiographic and clinical presentations of ARIA have been observed. ARIA-E refer to vasogenic edema in the brain parenchyma and/or leptomeningeal/subpial sulcal effusion; ARIA-H refer to hemosiderin deposits including cerebral microhemorrhage and localized superficial siderosis. With clinical availability of anti-amyloid therapies for treatment of early clinical stage Alzheimer's disease, there is an important need for better awareness of ARIA and education among a diversity of clinicians and practice settings.

Methods

To illustrate the wide spectrum of ARIA, we present the clinical history, course, and MRI findings as well as summarize learnings and implications of 10 ARIA-E cases from the aducanumab phase 3 clinical trials EMERGE and ENGAGE (ClinicalTrials.gov identifiers NCT02484547 and NCT02477800). Monitoring and management of ARIA, including criteria and timing of MRI, and drug dosing actions were defined per protocol and based upon radiographic severity and clinical symptoms associated with ARIA. ARIA were detected by MRI performed at pre-specified intervals or in response to symptoms or for safety follow-up reasons.

Results

The majority of ARIA-E events were asymptomatic. Most ARIA-E events occurred within the first 8 aducanumab doses and generally resolved within 3–4 months whether dosing was continued (i.e. when ARIA-E was asymptomatic and of mild radiographic severity) or whether dosing was suspended or discontinued. Occasionally, ARIA-E were recurrent. ARIA-H (microhemorrhages and superficial siderosis) most commonly occurred concurrently with ARIA-E. However, ARIA-H without ARIA-E, termed isolated ARIA-H, occurred at a similar frequency in aducanumab- and placebo-treated groups. 10% of participants with ARIA in the 10 mg/kg group had accompanying symptoms. When present, symptoms of ARIA most frequently included headache, nausea, fatigue, dizziness, confusional state, and visual disturbance. Serious symptoms associated with ARIA were uncommon (occurring in 0.3% of the participants treated with the 10 mg/kg dose regimen).

Conclusions

These illustrative cases provide learnings into the nature and factors associated with ARIA that can increase knowledge for clinicians and trialists regarding the varied radiographic and clinical presentations and evolution of ARIA.

Trial registration

Clinical Trials NCT02484547, registered June 18, 2015, https://clinicaltrials.gov/study/NCT02484547.

Clinical Trials NCT02477800, registered June 18, 2015, https://clinicaltrials.gov/study/NCT02477800.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12883-026-05097-9.

Background

In clinical trials evaluating beta-amyloid plaque-lowering monoclonal antibodies (anti-amyloid mAbs) for the treatment of Alzheimer’s disease (AD), magnetic resonance imaging (MRI) findings known as amyloid-related imaging abnormalities (ARIA) have been observed and reported as adverse events (AEs) [1, 2]. ARIA are defined radiographically as ARIA-E or ARIA-H. ARIA-E refer to vasogenic edema in the brain parenchyma and/or leptomeningeal/subpial sulcal effusions that are radiographically transient in nature with full resolution; ARIA-H refer to hemosiderin deposits including cerebral microhemorrhages (hemorrhages < 10 mm in diameter on T2* sequence) and/or localized superficial siderosis (located in the subarachnoid space) that tend to radiographically stabilize over time [3, 4] and/or rarely cerebral hemorrhage (“macrohemorrhage”; defined as > 1 cm in diameter on T2* MRI sequence) [1, 3, 4]. At times, ARIA-E can also be recurrent (i.e. >1 distinct episode following resolution of the previous ARIA-E episode, irrespective of location). The underlying pathophysiology of ARIA is incompletely understood [1, 2]. However, a leading hypothesis postulates that ARIA may be driven by antibody-mediated breakdown of amyloid beta (Aβ) aggregates in blood vessels and in the brain parenchyma [5, 6]. Removal of Aβ aggregates in the blood vessel may increase the permeability of the vessel and result in leakage of fluid. Removal of Aβ aggregates from the parenchyma may impair perivascular clearance and trigger perivascular inflammation [5, 6].

ARIA are seen across clinical trials of anti-amyloid mAbs at doses when clear and significant lowering of fibrillar Aβ is observed, including those investigating aducanumab (up-titration from 1 to 10 mg/kg, administered monthly), donanemab (700 mg for first 3 doses and 1400 mg thereafter, administered every 4 weeks), lecanemab (10 mg/kg biweekly), bapineuzumab (0.5 mg/kg for apolipoprotein E [ApoE] ε4 carriers; 1 or 2 mg/kg for noncarriers), and gantenerumab (105, 225, 510, or 1020 mg every 4 weeks). The incidence of ARIA-E and ARIA-H in treated trial participants with these anti-Aβ antibodies ranged from 6.5% to 35.2% and 3.3% to 32.7%, respectively [7–13]. Most ARIA cases were asymptomatic (> 70% of ARIA-E cases were asymptomatic across the studies) [7, 8, 10–17], usually occurred during the initial months of treatment, often during dose escalation, and typically resolved over weeks to months (4–23 weeks). Although ARIA can develop throughout treatment duration, the risk of ARIA (ARIA-E in particular) decreased over time [6, 9–13, 15, 18, 19]. The most well-established risk factors for ARIA include the APOE genotype, presence and severity of baseline microhemorrhages or superficial siderosis (potential manifestations of cerebral amyloid angiopathy [CAA]), and the dosing exposure to anti-amyloid mAbs [1, 2]. Emerging independent risk factors for ARIA-E may include elevated mean arterial blood pressure, greater severity of beta-amyloid AD pathology (i.e. from cerebrospinal fluid or amyloid positron emission tomography [PET]), and cerebrovascular white matter burden (i.e. leukoaraiosis on brain MRI) [10, 15]. The incidence of ARIA-E is higher in ApoE ε4 carriers compared with noncarriers [6, 7, 9, 11, 18, 20–22] and is highest in ApoE ε4 homozygous carriers [7, 9–11, 18, 21]. The incidence of ARIA also increases with dose of anti-amyloid mAbs [7, 18, 23], and may be mitigated with gradual titration (i.e. as for donanemab in TRAILBLAZER-ALZ 6 trial) [24].

Detailed analysis of ARIA incidence from an integrated safety data set has been reported for the large phase 3 clinical trials of aducanumab, EMERGE and ENGAGE [10]. Aducanumab received accelerated approval by the FDA, but development of aducanumab was discontinued and it was subsequently withdrawn from the market in 2024 [25].

Here, we present the clinical history, course, management, and imaging findings of 10 ARIA cases from the EMERGE and ENGAGE clinical studies of aducanumab. These illustrative cases were selected to represent the wide spectrum of radiographic and clinical presentations of ARIA.

Case presentation

Methods

Ethics approval

Written informed consent for participation in the study and release of this information was obtained from all participants or their caregiver. This study was conducted in accordance with the Declaration of Helsinki and the International Conference for Harmonisation and Good Clinical Practice guidelines and was approved by ethics committees or institutional review boards at each study site. To avoid identifying details about each participant, specific age, sex, geographic location, and ethnicity are not reported.

Study design and population

The design of the phase 3 studies of aducanumab has been previously published [20]. Briefly, the EMERGE trial (N = 1643) and ENGAGE trial (N = 1653) were identically designed double-blind, placebo-controlled (PC), global phase 3 randomized clinical trials. Trial participants were aged 50–85 years and met clinical criteria for mild cognitive impairment (MCI) due to AD or mild AD dementia with amyloid pathology confirmed by visual assessment of amyloid PET. Additional methodology can be found in Additional file 1.

Neuroimaging

Study protocols of the phase 3 studies of aducanumab have been previously published as supplemental material to Budd Haeberlein et al. [26]. Images of brain MRI, performed at 1.5T or 3T field strengths, were evaluated by central radiologists with expertise in ARIA; all ARIA events were recorded as AEs. T2 fluid-attenuated inversion recovery was the primary diagnostic imaging sequence used to evaluate the presence of ARIA-E [4]. The presence of any new incident (i.e. treatment emergent) ARIA-H brain microhemorrhages or cerebral macrohemorrhages or localized superficial siderosis events was evaluated using T2* gradient-echo sequences.

Radiographic severity of ARIA-E was classified based on the number and size of any edematous region: a single region < 5 cm was considered mild, a single region of 5 to 10 cm or multiple regions all < 10 cm was considered moderate, and any region > 10 cm was considered radiographically severe. ARIA-H brain microhemorrhage radiographic severity was classified based on the number of new incident brain microhemorrhages: 1 to 4 was mild, 5 to 9 was moderate, and ≥ 10 was severe. Likewise, ARIA-H superficial siderosis radiographic severity was classified based on the number of new incident areas of superficial siderosis: 1 was mild, 2 was moderate, and > 2 was severe.

During the PC period, baseline MRI was designated as the first scan before treatment commenced in the PC period. Postbaseline surveillance MRI was performed per protocol at weeks 14, 22, 30, 42, 54, 66, and 78. When an ARIA episode was detected, follow-up MRI scans were made approximately every 4 weeks to document ARIA evolution and continued until ARIA-E resolution or ARIA-H stabilization.

ARIA risk minimization, symptom evaluation, and management

Strategies to minimize the risk of ARIA included dose titration, routine brain MRI monitoring, ad hoc MRI testing (as clinically indicated), and dose suspension or permanent discontinuation.

The criteria for dose suspension or discontinuation were based on the type and radiographic severity of ARIA, as well as the presence of clinical symptoms. Evaluation of any symptoms during an ARIA event was performed by the investigator. Dose suspension was always required per protocol in the setting of clinical symptoms, regardless of severity, and dosing was required to be withheld until resolution of symptoms; dose suspension was also required for asymptomatic radiographically moderate/severe ARIA-E or radiographically moderate ARIA-H and withheld until ARIA-E resolution or ARIA-H stabilization. Participants who developed radiographically severe ARIA-H (microhemorrhages or superficial siderosis), or potentially any cerebral macrohemorrhage, would be permanently discontinued from the study treatment regardless of their ARIA-related symptomatic status. Participants with ARIA-E or ARIA-H accompanied by serious clinical symptoms (except for “other medically important event”) were also discontinued from the study treatment.

Results

Together, the aducanumab phase 3 studies EMERGE and ENGAGE constitute the largest clinical data set of trial participants treated with a beta-amyloid plaque-lowering monoclonal antibody, and the safety and ARIA imaging data collected in these trials are invaluable to understanding the range of clinical and radiographic characteristics of ARIA that can occur during treatment with aducanumab. Ten trial participants across the eligible age range of the study were selected to illustrate the broad spectrum of ARIA radiographic severity and symptomatic presentations that were observed. Although ARIA can occur in both ApoE ε4 carriers and noncarriers, they occur more frequently in ApoE ε4 carriers. The selected cases in this series are mostly ApoE ε4 carrier cases, consistent with the phase 3 studies in which most participants were also ApoE ε4 carriers (approximately two-thirds ApoE ε4 carriers, one-third noncarriers). Table 1 summarizes ApoE ε4 carrier status, aducanumab dose exposure at the time of ARIA events, ARIA severity, symptom presentation (if any), and dosing action in each of these 10 cases.

Table 1.

Summary of cases presented

Case	ApoE*	Remarkable baseline MRI findings	ARIA-E radiographic severity	ARIA-E symptomatic	ARIA-H radiographic severity	ARIA-H symptomatic	Total no. of doses received at event onset	Dose at event (mg/kg)	Dosing action and notes on treatment
1	ε4+/−	None	Mild	Asymptomatic	N/A	N/A	17	10	Continued
2	ε4−/−	None	Mild	Asymptomatic	N/A	N/A	6	6	Continued
3	ε4+/−	Microhemorrhage (×1) ARWMC = 1	N/A	N/A	Mild, microhemorrhage (×1)	Asymptomatic	4	3	Continued
			Mild	Headache	N/A	N/A	11	10	No action taken. Participant withdrew from the study after the last dose received (Day 282) due to termination of the trial by the sponsor.
4	ε4+/−	None	Mild	Asymptomatic	N/A	N/A	4	3	Continued
			Moderate	Asymptomatic	N/A	N/A	5	3	Interrupted and resumed at a dose of 6 mg/kg.†
			Moderate	Asymptomatic	N/A	N/A	9	6	Interrupted and resumed at a dose of 3 mg/kg.
			N/A	N/A	Mild, microhemorrhage (×1)	Asymptomatic	28	10	Continued
5	ε4+/−	ARWMC = 1	Mild	Asymptomatic	N/A	N/A	11	10	Continued
			N/A	N/A	Mild, superficial siderosis (×1)	Asymptomatic	13	10	Continued
			Moderate	Asthenia, vision blurred	N/A	N/A	15	10	Interrupted
6	ε4+/+	ARWMC = 1	Severe	Asymptomatic	N/A	N/A	8	10	Interrupted
			N/A	N/A	Mild, microhemorrhage (×2)	Asymptomatic	8	10	No action due to microhemorrhages; suspended treatment for severe ARIA-E and had not resumed prior to phase 3 program termination. Severe ARIA-E resolved.
7	ε4+/+	Microhemorrhage (×3) ARWMC = 1	Severe	Asymptomatic	Severe, microhemorrhage (> 10)	Asymptomatic	6	3	Discontinued
8	ε4+/+	ARWMC = 1	Severe	Mild headache, mild neck pain, mild sinus pressure	N/A	N/A	4	3	Interrupted and restarted at a dose of 3 mg/kg.
9	ε4+/−	Microhemorrhage (×2) ARWMC = 1	Severe	Psychomotor slow down, language disturbance, dysmetria	Moderate, microhemorrhage (×9)	Psychomotor slow down, language disturbance, dysmetria	4	3	Discontinued
			Severe	Balance disturbances, asymmetry, pyramidal signs	Severe, microhemorrhage (> 10)	Balance disturbances, asymmetry, pyramidal signs	4	3
10	ε4−/−	Microhemorrhage (×1) ARWMC = 1	Severe	Visual disturbances, headache, nausea, confusion, dizziness, fatigue, intermittent tachycardia, loss of balance	N/A	N/A	3	3	Discontinued
			Severe	Confusion, dizziness, fatigue, intermittent tachycardia, loss of balance, visual disturbances, ataxia	Moderate, microhemorrhage (×9)	Confusion, dizziness, fatigue, intermittent tachycardia, loss of balance, visual disturbances, ataxia
			Moderate	Balance disorder, ataxia	Severe, microhemorrhage (> 10)	Balance disorder and ataxia

ApoE, apolipoprotein E; ARIA, amyloid-related imaging abnormalities; ARWMC, age-related white matter changes

*ε4−/− denotes ApoE ε4 noncarrier, ε4+/− denotes ApoE ε4 heterozygous, and ε4+/+ denotes ApoE ε4 homozygous

^†This constitutes an overdose per protocol as, on Day 225, the participant received the next study treatment dose at 6 mg/kg instead of the protocol-specified 3 mg/kg, which was a protocol deviation. There was no reported associated adverse event noted due to the receipt of the 6 mg/kg dose

Case 1: asymptomatic mild ARIA-E

An heterozygous ApoE ε4 carrier trial participant with MCI due to AD had a medical history that included depression and hypertension (Supplementary Information, Additional Fig. 1). Concomitant medications at study enrollment included esomeprazole, nicergoline, perindopril, and sertraline. A baseline MRI (Day − 44) showed no remarkable findings.

On Day 461, a routine MRI revealed new vasogenic edema (location: left frontal extending to left parietal lobe) consistent with mild ARIA-E; the participant was asymptomatic, and no ARIA-H were observed. At ARIA-E onset, the participant had received a total of 17 aducanumab doses in the PC period (2 doses of 1 mg/kg, 4 doses of 3 mg/kg, 6 doses of 6 mg/kg, and 5 doses at the target dose of 10 mg/kg). The most recent 10 mg/kg dose was received 13 days prior to the ARIA-E onset (Day 448).

Per protocol for an asymptomatic and radiographically mild ARIA-E event, the participant continued to receive study treatment (10 mg/kg) as scheduled during and after the event. Resolution of the event was confirmed via the first scheduled follow-up MRI on Day 520. The participant received the final aducanumab dose on Day 817, for a total of 28 doses, including 18 in the PC period (2 doses of 1 mg/kg, 4 doses of 3 mg/kg, 6 doses of 6 mg/kg, and 6 doses at the target dose of 10 mg/kg) and 10 doses of 10 mg/kg aducanumab in the long-term extension (LTE) period.

Key learning points

While this case represents a typical clinical presentation of ARIA-E, in that no symptoms were present and the event resolved within 8 weeks while treatment continued per protocol, the event occurred after 17 doses, later than most ARIA-E events observed in the aducanumab phase 3 clinical trials. The majority (72.7% in the 10 mg/kg aducanumab group) of initial ARIA-E events occurred within the first 8 doses of aducanumab in the EMERGE and ENGAGE studies. However, ARIA-E can, on occasion, occur later during treatment, even > 1 year after treatment initiation with anti-amyloid mAbs [15–17, 19]. In the 10 mg/kg dose group during the PC period, 27.3% of participants experienced their first ARIA-E event after the eighth aducanumab dose [10].

Case 2: asymptomatic and radiographically mild ARIA-E in an ApoE ε4 noncarrier

An ApoE ε4 non-carrier trial participant with MCI due to AD had a medical history that included anemia, right bundle branch block, hyperlipidemia, and Lyme disease (Supplementary Information, Additional Fig. 2). A baseline MRI (Day − 31) showed no remarkable findings. Concomitant medications included aspirin, donepezil, omeprazole, and simvastatin.

On Day 173, routine MRI showed new vasogenic edema (location: right frontal) consistent with mild ARIA-E; the participant was asymptomatic, and no ARIA-H were observed. At event onset, the participant had received a total of 6 aducanumab doses in the PC period (2 doses of 1 mg/kg, 2 doses of 3 mg/kg, and 2 doses of 6 mg/kg). The most recent dose of study treatment (6 mg/kg) was received on Day 144.

On Day 247, MRI showed that the initial focus of ARIA-E in the right frontal lobe was resolved, but a new focus of vasogenic edema developed in the right occipital cortex, consistent with mild ARIA-E. The participant remained asymptomatic. ARIA-E completely resolved on the next follow-up MRI on Day 277. The participant continued in the LTE period with no further ARIA events. The participant received the final aducanumab dose on Day 898, for a total of 32 doses, including 19 in the PC period (2 doses of 1 mg/kg, 2 doses of 3 mg/kg, 2 doses of 6 mg/kg, and 13 doses at the target dose of 10 mg/kg) and 13 doses of 10 mg/kg aducanumab in the LTE period.

Key learning points

Although more frequent in ApoE ε4 carriers, ARIA can occur in noncarriers; therefore, monitoring is important in all patients treated with aducanumab or anti-amyloid mAbs [2, 10, 16, 17].

Case 3: asymptomatic ARIA-H microhemorrhage followed by mild symptomatic and mild radiographic ARIA-E

An heterozygous ApoE ε4 carrier trial participant with MCI due to AD had a medical history that included headache, hypothyroidism, and benign pituitary tumor (Supplementary Information, Additional Fig. 3). Concomitant medications at study enrollment included levothyroxine sodium, ibuprofen, and paracetamol. A baseline MRI (Day − 27) showed mild white matter disease consistent with age-related white matter changes (ARWMC) [27] scale of 1 and a single microhemorrhage. On Day 100 after 4 total aducanumab doses (2 doses of 1 mg/kg, 2 doses of 3 mg/kg), a routine MRI showed the interval development of 1 asymptomatic ARIA-H microhemorrhage. Per protocol, drug treatment was continued at the same dose. The ARIA-H microhemorrhage was considered stabilized on follow-up MRI on Day 157.

On Day 284, after having received a total of 11 doses in the PC period (2 doses of 1 mg/kg, 2 doses of 3 mg/kg, 2 doses of 6 mg/kg, 5 doses of 10 mg/kg), the participant reported a headache lasting 1 day. On Day 295, an MRI revealed the previously identified solitary microhemorrhage and new mild vasogenic edema (location: right occipital cortex) consistent with mild ARIA-E. The headache was assessed by the site investigator as a mild symptom associated with treatment and concurrent with ARIA-E. The most recent dose of study treatment (10 mg/kg) was received on Day 282, 13 days prior to the onset of the mild ARIA-E. At the time of the event, no action was taken with study treatment, as the trial was terminated by the sponsor. ARIA-E were reported to have resolved on Day 381.

Key learning points

In trial participants who experienced ARIA symptoms (10.0% in the 10 mg/kg group), headache was a common associated symptom (4.7% in the 10 mg/kg group) [10]. Most symptoms (> 85%) associated with ARIA in EMERGE and ENGAGE resolved while in the study with a median (interquartile range) duration of 5 (1–13) weeks [10]. In this case, the headache associated with ARIA-E was short in duration and mild in severity.

Case 4: asymptomatic, mild ARIA-E evolving into moderate ARIA-E, subsequent recurrent ARIA-E

An heterozygous ApoE ε4 carrier trial participant with mild AD dementia had a medical history that included anxiety and depression (Supplementary Information, Additional Fig. 4). Concomitant medications at study enrollment included alprazolam, donepezil, and mirtazapine. A baseline MRI (Day − 33) showed no remarkable findings (including no microhemorrhages).

The Day 99 routine MRI showed new mild vasogenic edema (location: left parietal) consistent with mild ARIA-E; no ARIA-H were noted, and the participant was asymptomatic. At the time of the event, the participant had received a total of 4 aducanumab doses during the PC period (2 doses of 1 mg/kg and 2 doses of 3 mg/kg). The most recent dose of study treatment (3 mg/kg) was received 14 days prior to the onset of the event (Day 85). Per protocol, for an asymptomatic and radiographically mild ARIA-E event, the participant continued with scheduled treatment.

A follow-up MRI on Day 127 showed ARIA-E reported as moderate and multifocal vasogenic edema (locations: right occipital cortex; left parietal; and right temporal [non-hippocampal]) that had increased in size since the previous MRI. No ARIA-H were observed. At the time of this second ARIA-E event, the participant had received a total of 5 doses during the PC period (2 doses of 1 mg/kg and 3 doses of 3 mg/kg). The most recent dose of study treatment (3 mg/kg) was received 14 days earlier (Day 113). Per protocol, given the development of moderate (multifocal) ARIA-E, treatment was temporarily suspended. The Day 185 follow-up MRI showed complete resolution of all ARIA-E foci and no ARIA-H. Given the resolution of ARIA-E, study treatment was restarted at the same dose (3 mg/kg) on Day 196. On Day 225, the participant received the next dose at 6 mg/kg instead of the protocol-specified 3 mg/kg, which was a protocol deviation. No AEs associated with the 6 mg/kg dose were noted.

On Day 290, a routine MRI showed new moderate vasogenic edema (locations: left parietal and right occipital cortex) consistent with recurrent and moderate ARIA-E. The participant remained asymptomatic. At the time of the event, the participant had received a total of 9 doses during the PC period (2 doses of 1 mg/kg, 4 doses of 3 mg/kg, and 3 doses of 6 mg/kg). The most recent dose of study treatment (6 mg/kg) was received 10 days prior to the onset of the event on Day 280. Per protocol, study drug dosing was suspended. Follow-up MRI on Day 323 revealed complete resolution of ARIA-E foci, and study treatment was restarted at a reduced dose of 3 mg/kg on Day 337. The participant continued in the study.

On Day 855 of the LTE period, MRI showed 1 new microhemorrhage (location: right parietal) consistent with mild ARIA-H microhemorrhage. The participant was asymptomatic. At the time of the event, the participant had received a total of 28 doses of study treatment, including 17 doses during the PC period (2 doses of 1 mg/kg, 6 doses of 3 mg/kg, and 9 doses of 6 mg/kg) and 11 doses during the LTE period (11 doses of 10 mg/kg). The most recent dose of study treatment (10 mg/kg) was received on Day 841, 14 days prior to the onset of the event. MRI on Day 903 showed that the microhemorrhage was stable. The investigator assessed that the mild ARIA-H microhemorrhage was not related to study treatment and the participant continued in the study.

The participant received the final aducanumab dose on Day 873 for a total of 29 doses, including 17 doses in the PC period (2 doses of 1 mg/kg, 6 doses of 3 mg/kg, and 9 doses of 6 mg/kg) and 12 doses in the LTE period (12 doses of 10 mg/kg).

Key learning points

ARIA-E may recur, increase in radiographic severity, or become multifocal over the course of treatment. Therefore, follow-up monitoring by MRI of patients who have been diagnosed with ARIA-E is an important component of management [10, 15, 19, 28]. Approximately 10.6% of participants in the 10 mg/kg dose group experienced > 1 ARIA-E event [10]. Although ARIA-H is often seen concurrently with ARIA-E [6], ARIA microhemorrhages may occur independently of ARIA-E (1 new and asymptomatic microhemorrhage occurred 76 weeks after resolution of ARIA-E).

Case 5: moderate symptomatic ARIA-E and mild asymptomatic ARIA-H superficial siderosis

An heterozygous ApoE ε4 carrier trial participant with MCI due to AD had a medical history that included depression and hypertension (Supplementary Information, Additional Fig. 5). A baseline MRI (Day − 44) was notable for ARWMC = 1 and no microhemorrhages. Concomitant medications at study enrollment included acetylsalicylate lysine (also known as lysine ASA/aspirin in the USA), atorvastatin, bisoprolol, escitalopram oxalate, galantamine, and paracetamol.

On Day 293, MRI revealed right occipital cortex vasogenic edema consistent with mild ARIA-E; the participant was asymptomatic. At the time of the event, the participant had received a total of 11 doses in the PC period (2 doses of 1 mg/kg, 2 doses of 3 mg/kg, 2 doses of 6 mg/kg, and 5 doses of 10 mg/kg). The most recent dose of 10 mg/kg aducanumab was received on Day 278, 15 days prior to the onset of the event. Per protocol, no dosing action was taken as a result of the event, and the participant continued to receive study treatment (10 mg/kg) as scheduled. On Day 349, MRI showed mild vasogenic edema (ARIA-E) and mild ARIA-H superficial siderosis, both in the right occipital cortex. MRI on Day 379 showed continuing mild vasogenic edema (ARIA-E) and stable ARIA-H superficial siderosis, both in the right occipital cortex.

Starting Day 404, asthenia (assessed as moderate severity, related to study treatment, and a symptom of ARIA-E) was reported and subsequently resolved on Day 439. MRI on Day 414 revealed an increase in size of the ongoing right parietal vasogenic edema, increasing radiographic severity to moderate ARIA-E. At the time of the event, the participant had received a total of 15 doses (2 doses of 1 mg/kg, 2 doses of 3 mg/kg, 2 doses of 6 mg/kg, and 9 doses of 10 mg/kg). The most recent dose of 10 mg/kg aducanumab was received on Day 393, 21 days prior to the onset of the event. Study treatment was interrupted as a result of the event. On Day 433, blurred vision (assessed as mild severity, related to study treatment, and a symptom of ARIA) was reported and subsequently resolved on Day 435. Follow-up MRI on Day 449 showed ongoing moderate and multiregional ARIA-E (right parietal and right temporal). Resolution of the ARIA-E was confirmed by MRI on Day 505.

Key learning points

Although most cases of ARIA are asymptomatic, generalized and nonspecific symptoms such as fatigue and or specific symptoms such as blurred vision are known potential symptoms of ARIA. Sometimes, symptoms may localize with the region of ARIA.

Case 6: asymptomatic severe ARIA-E, asymptomatic mild ARIA-H microhemorrhage

An ApoE ε4 homozygote trial participant with mild AD dementia had a medical history that included epilepsy and headache (Supplementary Information, Additional Fig. 6). Concomitant medications at study enrollment included valproic acid. A baseline MRI (Day − 44) showed mild white matter disease (ARWMC = 1).

Starting on Day 114, intermittent headaches were reported (mild in severity, assessed by site primary investigator to not be related to study treatment or ARIA). No ARIA was found on MRI at the time headaches were reported. The intermittent headaches did not resolve during the study. On Day 210, routine MRI revealed vasogenic edema consistent with severe and multifocal ARIA-E (locations: left occipital cortex; left parietal; left temporal, [non-hippocampal]); the participant was asymptomatic. At the time of the event, the participant had received a total of 8 aducanumab doses (2 doses of 1 mg/kg, 2 doses of 3 mg/kg, 2 doses of 6 mg/kg, and 2 doses of 10 mg/kg). The most recent 10 mg/kg dose was received on Day 197, 13 days prior to the onset of this event. Per study protocol, treatment was suspended because of severe ARIA-E.

MRI on Day 238, while treatment was suspended, showed the previously noted severe multifocal ARIA-E (locations: left occipital cortex; left parietal; left temporal, nonhippocampal), and, compared with the last MRI, 1 new microhemorrhage (location: left occipital cortex) consistent with mild ARIA-H microhemorrhage; the participant was asymptomatic. At the time of the ARIA-H microhemorrhage, the most recent (and final) dose of 10 mg/kg aducanumab had been on study Day 197, 41 days prior to the event. No dose changes were made because of the event. On Day 267, a follow-up MRI demonstrated 1 additional microhemorrhage (mild) located in the left temporal (non-hippocampal) region. On Day 295 (Week 42) MRI, ARIA-H microhemorrhages were stable (and no additional microhemorrhages noted), and resolution of all foci of ARIA-E were confirmed. The participant did not resume study drug due to phase 3 clinical trial program termination.

Key learning points

Radiographic severity may not correlate with clinical severity of ARIA. More specifically, as was observed in this case, radiographically severe ARIA do not always present with symptoms. In the aducanumab phase 3 trials, 73.1% of radiographically severe ARIA cases were asymptomatic. ARIA-E and ARIA-H may also occur concurrently (e.g. in this case, severe ARIA-E and 2 foci of mild ARIA-H microhemorrhages occurred concurrently).

Case 7: asymptomatic severe ARIA-E, asymptomatic severe ARIA-H microhemorrhages

An ApoE ε4 homozygote trial participant with MCI due to AD had a medical history that included 3 microhemorrhages and hypertension at the time of enrollment (Supplementary Information, Additional Fig. 7). Concomitant medications at study enrollment included acetylsalicylic acid 81 mg daily, amlodipine, and naproxen sodium. A baseline MRI (Day − 40) showed 3 microhemorrhages and an ARWMC = 1.

On Day 155, a routine MRI revealed vasogenic edema (locations: left frontal; left temporal (non-hippocampal); right frontal; right occipital cortex; right parietal; right temporal (non-hippocampal) consistent with severe multifocal ARIA-E and > 10 microhemorrhages compared with the most recent MRI consistent with severe ARIA-H microhemorrhages; the participant was asymptomatic. The ARIA-H microhemorrhages were concurrent with and colocalized with the ARIA-E. At the time of the event, the participant had received a total of 6 doses (2 doses of 1 mg/kg and 4 doses of 3 mg/kg). The most recent dose of 3 mg/kg aducanumab was received on Day 134, 21 days prior to the onset of the event. Per study protocol, treatment was permanently discontinued due to severe ARIA-H microhemorrhages. Resolution of ARIA-E and stabilization of ARIA-H microhemorrhages were confirmed via MRI on study Days 239 and 268, respectively.

Key learning points

This participant had risk factors for ARIA, including ApoE ε4 allele homozygosity and evidence of 3 microhemorrhages on baseline MRI (hazard ratio of 4.5 [95% CI, 3.33–6.13] for ε4 homozygous carriers compared with noncarriers; hazard ratio of 1.7 [95% CI, 1.31–2.27] for trial participants with baseline microhemorrhages compared with those with none [10]). The ARIA event occurred during the time of highest risk [10], within the first 8 doses. As in this case, ARIA may be radiographically severe on initial detection. Even if the patient is asymptomatic, permanent treatment discontinuation may be necessary based on radiographic severity of ARIA-H, including in this case of asymptomatic severe ARIA-H microhemorrhages [29–32].

Case 8: symptomatic severe ARIA-E (nonserious)

An ApoE ε4 homozygote trial participant with MCI due to AD had a medical history that included depression, hyperlipidemia, hypothyroidism, and hypertension (Supplementary Information, Additional Fig. 8). Concomitant medications included acetylsalicylic acid (81 mg daily), atorvastatin, diltiazem, levothyroxine, and escitalopram. A baseline MRI (Day − 23) was notable for ARWMC = 1.

On Day 100 of treatment with aducanumab, a routine MRI revealed vasogenic edema (locations: left frontal; left parietal; left temporal [non-hippocampal]; right occipital cortex; right parietal; right temporal [non-hippocampal]) consistent with severe and multifocal ARIA-E; and no ARIA-H. The participant was assessed as having symptomatic ARIA with report of headache (mild severity, related to study treatment), neck pain (mild severity, not related to study treatment), and paranasal sinus discomfort (mild severity, related to study treatment) on Day 101 and resolving on Day 109; all were assessed as symptoms of ARIA-E. Additional nonrelated AEs that required treatment were reported during the ARIA-E event, including mild urinary tract infection between Days 139 and 149 and mild hives between Days 149 and 153, which also resolved. While the ARIA-E was ongoing, the participant also began treatment with oral diphenhydramine 25 mg twice daily for 2 days (for hives), oral doxycycline 100 mg twice daily for 5 days (for sinusitis), oral ibuprofen 800 mg as needed (no frequency listed; for neck pain), oral nitrofurantoin 100 mg twice daily for 11 days (for urinary tract infection), oral prednisone 200 mg daily for 5 days (for sinusitis), oral pseudoephedrine 2 cap 4 times daily for 9 days (for sinus pressure), and ranitidine hydrochloride (for hives).

At the time of the ARIA-E event, the participant had received a total of 4 doses (2 doses of 1 mg/kg, 2 doses of 3 mg/kg). The most recent dose of 3 mg/kg aducanumab was received on Day 86, 14 days prior to the onset of the event. Study treatment was interrupted as a result of the event. Resolution of the ARIA-E was confirmed by MRI on Day 150.

Per protocol, the participant resumed the assigned aducanumab dose of 3 mg/kg and received the final dose on Day 393, for a total of 13 doses (2 doses of 1 mg/kg and 11 doses of 3 mg/kg). Early termination of the study by the sponsor was the reason for withdrawal from the study.

Key learning points

This case represents radiographically severe ARIA-E with the common ARIA symptom of headache. In this case, ARIA-related headache was mild and short-lived, while the radiographic ARIA-E findings persisted longer than symptoms before also resolving. In the phase 3 aducanumab clinical trials, the duration of ARIA-E–related symptoms was usually shorter than the typical (radiographic) duration of ARIA-E.

Case 9: serious symptomatic and radiographically severe ARIA-E with symptomatic ARIA-H microhemorrhages

An ApoE ε4 heterozygote trial participant with mild AD dementia had a medical history that included headache, dyslipidemia, and hypertension (Supplementary Information, Additional Fig. 9). Concomitant medications included fenofibrate, indapamide/perindopril, oxazepam, and ormeloxifene. A baseline MRI (Day − 75) showed 2 microhemorrhages and ARWMC = 1.

On study Days 91 to 98, the participant experienced a nonserious AE of cognitive disorder (assessed as severe, related to study treatment). At the time of the events, the participant had received a total of 4 doses of study treatment (2 doses of 1 mg/kg and 2 doses of 3 mg/kg). The most recent dose (3 mg/kg) was received on Day 85. The participant was found to be in a confused state and was unable to recognize a previously familiar nurse. The following day, clinical examination showed psychomotor slowing, language dysfunction, and dysmetria on finger-to-nose testing with eyes closed.

On Day 93, an MRI showed vasogenic edema (locations: hippocampus; left frontal; left occipital cortex; left parietal; left temporal, non-hippocampal; right frontal; right occipital cortex; right parietal; right temporal, non-hippocampal) consistent with severe multifocal ARIA-E; moderate ARIA-H microhemorrhage with 9 microhemorrhages (locations: right occipital; right temporal), 7 of which were new since the previous MRI at baseline. The participant was assessed as having symptomatic ARIA-E and ARIA-H microhemorrhages (Day 93). The investigator assessed the severity of the ARIA-H event as moderate and the ARIA-E event as severe; both events were considered related to study treatment. Study treatment was permanently discontinued due to the symptomatic ARIA-E; final dose was on Day 85.

On Day 94, cognitive symptoms persisted together with the onset of a balance disorder. The participant was admitted to the hospital (geriatric medicine) for observation due to a weight loss of 3 kg and a reduction in speech. Neurological examination on the same day was otherwise normal except for mild dysmetria. During the ARIA event, the participant began treatment with hydrochlorothiazide, indapamide/perindopril erbumine, and oral nicardipine for hypertension. The AE of cognitive disorder was reported as resolved on Day 98, with no specific treatment provided, and the participant was discharged on Day 99.

On Day 113, MRI showed vasogenic edema (locations: hippocampus; left frontal; left occipital cortex; left parietal; left temporal, non-hippocampal; right frontal; right occipital cortex; right parietal; right temporal, non-hippocampal) consistent with severe multifocal ARIA-E that had increased in size compared with the previous MRI, > 10 new microhemorrhages (locations: left frontal, left parietal, right frontal, right occipital, right parietal, and right temporal). The investigator assessed the severity of the second occurrence of ARIA-H as severe and considered the event related to study treatment. On the same day, the participant was readmitted to the hospital. On Day 114, the participant received intravenous methylprednisolone 80 mg once daily for 7 days and continued orally at tapered doses: (a) 20 mg 4 times daily for 7 days, (b) 20 mg 3 times daily for 7 days, (c) 20 mg twice daily for 7 days, (d) 20 mg daily for 7 days, and (e) 10 mg once daily for 8 days.

An electroencephalogram on Day 116 showed slow right parietal temporal waves with no paroxysmal signs. The additional AE of agitation was reported (Days 119–130) and was assessed as moderate severity, related to study treatment, and a symptom of ARIA. Starting Day 121, methylprednisolone was continued orally at tapered doses: (a) 20 mg 4 times daily for 7 days, (b) 20 mg 3 times daily for 7 days, (c) 20 mg twice daily for 7 days, (d) 20 mg daily for 7 days, and (e) 10 mg once daily for 8 days. The participant was discharged from the hospital on Day 130.

The event of symptomatic ARIA-H (microhemorrhage) was reported as stabilized on Day 144. A follow-up MRI on this day showed vasogenic edema (locations: left frontal; left occipital cortex; right frontal; right occipital cortex; right parietal; right temporal, non-hippocampal) consistent with moderate ARIA-E, with > 10 stable microhemorrhages (locations: left frontal, right frontal, right occipital cortex, right parietal, right temporal). The ARIA-E event was reported as resolved on Day 269. MRI on this day showed complete resolution of vasogenic edema, with > 10 stable microhemorrhages (locations: left frontal, right frontal, right occipital cortex, right parietal, right temporal) and no superficial siderosis. An additional AE of irritability was reported (occurring Days 210–309) and was assessed as moderate severity, not related to study treatment, and a symptom of ARIA. The final assessment was on Day 549.

Key learning points

Confusional state is a common symptom of ARIA-E and occurred in 1.5% of the participants in the 10 mg/kg group of EMERGE and ENGAGE [10]. Early during treatment, this participant developed severe and then serious symptoms in the context of radiographically severe ARIA-E along with ARIA-H microhemorrhages that evolved in severity from moderate to severe as the number of new microhemorrhages increased. Treatment was permanently discontinued per protocol. During the PC period of EMERGE and ENGAGE, 425 of 1029 trial participants (41.3%) treated in the 10 mg/kg aducanumab group experienced ARIA, with serious cases occurring in 14 of these 1029 participants (1.4%). Serious ARIA symptoms occurred in 3 of 1029 participants (0.3%) treated in the 10 mg/kg aducanumab group [10].

Case 10: serious symptomatic and radiographically severe ARIA-E and severe ARIA-H microhemorrhages in an ApoE ε4 noncarrier

An ApoE ε4 noncarrier trial participant with MCI due to AD had a medical history that included depression, hypertension, and hyperlipidemia (Supplementary Information, Additional Fig. 10). Baseline MRI (Day − 61) was notable for 1 microhemorrhage and ARWMC = 1. Concomitant medications included acetylsalicylic acid, bupropion, estradiol, fosinopril/hydrochlorothiazide, paroxetine hydrochloride, and prochlorperazine.

On Day 59, the participant experienced nonserious events of headache, nausea, confusional state, dizziness, fatigue, and loss of balance (all assessed as moderate, related to study treatment, and symptoms of ARIA) and intermittent tachycardia (assessed as mild severity, related to study treatment, and a symptom of ARIA).

On Day 69, MRI showed new vasogenic edema (locations: left occipital cortex; left parietal; left temporal, non-hippocampal; right occipital cortex; right parietal) consistent with severe multifocal ARIA-E; 1 microhemorrhage (stable). The participant was symptomatic with onset of concurrent nonserious events of visual impairment (assessed as mild severity, related to study treatment, and a symptom of ARIA) and ataxia (assessed as moderate severity, related to study treatment, and a symptom of ARIA) on the same day, assessed as related to study treatment, as well as with ongoing nonserious events noted above (from Day 59). At the time of the event, the participant had received a total of 3 aducanumab doses in the PC period (2 doses of 1 mg/kg and 1 dose of 3 mg/kg). The most recent dose of study treatment (3 mg/kg) was received on Day 56. Study treatment was permanently discontinued as a result of the ARIA-E event.

The participant experienced several brief bouts of palpitation (intermittent tachycardia, considered a related, mild AE) during Days 59 to 141, with a heart rate reaching as high as 150 beats per minute. On Day 72, physical examination identified a grade 1/6 ejection-type murmur. An electrocardiogram showed normal sinus rhythm. An echocardiogram showed normal size and normal contracting left ventricle with an ejection fraction of 74%. Active diagnoses included palpitations and hypertension. On Day 85, a repeat urgent MRI showed no new lesions.

On Day 100, the participant experienced fewer episodes of palpitations. Long-term Holter monitor for tachycardia showed 1 episode of a supraventricular arrhythmia, which lasted only 13 beats. The participant remained in sinus rhythm throughout the 13 days and had no other significant arrhythmias. On Day 118, the nonserious events of headache and nausea were reported as resolved without any medication.

On Day 120, MRI showed vasogenic edema (locations: left frontal; left parietal; left temporal, non-hippocampal; right frontal; right occipital cortex; right temporal, non-hippocampal) consistent with severe multifocal ARIA-E, which had increased since the previous MRI, 9 microhemorrhages, 8 of which were new (location: left occipital cortex) since the previous MRI and no superficial siderosis. The participant was symptomatic with ongoing nonserious events (considered related to treatment) of confusion (moderate severity), dizziness (moderate severity), fatigue (moderate severity), intermittent tachycardia (mild severity), blurred vision (mild severity), and ataxia (moderate severity).

On Day 137, the participant experienced nonserious events of fall and contusion (all of these events were assessed as mild in severity, related to study treatment, and a symptom of ARIA). The nonserious symptoms of confusional state, dizziness, fatigue, tachycardia, and visual impairment were reported on Day 141 as resolved without any medication. The participant experienced worsening symptomatic ARIA-H (microhemorrhage) on Day 141. MRI on Day 141 showed moderate vasogenic edema (locations: left frontal; left parietal; left temporal, non-hippocampal; right frontal; right occipital cortex; right temporal, non-hippocampal), consistent with severe and multifocal radiographic ARIA-E, which had decreased compared with the previous MRI; and severe ARIA-H microhemorrhages (> 10 microhemorrhages; location: left frontal, 3 of which were new since the previous MRI). On Day 149, the contusion was reported as resolved, and on Day 167, ataxia was reported as resolved.

The event of ARIA-E was reported as resolved on Day 310 as the MRI showed complete resolution of vasogenic edema; >10 microhemorrhages (locations: left frontal; right frontal; right occipital cortex; right temporal, non-hippocampal), 2 of which were new since the previous MRI, were observed. The event of severe ARIA-H was reported as stabilized on Day 434 as the MRI showed no vasogenic edema, > 10 microhemorrhages (stable; locations: left frontal; right frontal; right occipital cortex; right temporal, non-hippocampal), and no superficial siderosis. The final assessment was on Day 1127.

Key learning points

Although not an ApoE ε4 carrier, the participant had a baseline microhemorrhage which may be a risk factor for ARIA. Early during treatment, this participant developed symptomatic (categorized as mild to moderate) and severe ARIA-E with severe ARIA-H microhemorrhages. Per protocol, treatment was permanently discontinued. This case illustrates several of the clinical symptoms suggestive of ARIA, including headache, nausea, fatigue, dizziness, confusional state, gait imbalance, and blurred vision. This participant also experienced a minor fall assessed as mild, related to study treatment, and a symptom of ARIA. Again, as typically observed in the phase 3 aducanumab clinical trials and studies of other anti-amyloid mAbs, ARIA-related symptoms resolved much earlier than ARIA-E resolved radiographically [10].

Discussion and conclusions

The EMERGE and ENGAGE phase 3 trials of aducanumab have produced a large and valuable data set that allows for the clinical characterization of ARIA, as well as further understanding of ARIA identification, monitoring, and management [10]. This paper provides a spectrum of cases that represent a wide range of radiographic and clinical presentations of ARIA-E and ARIA-H, as well as key factors (both radiographic and clinical) to consider in risk assessment, and monitoring and management of patients. Table 2 summarizes and provides insights and implications for clinical management from each case.

Table 2.

Summary of key implications for clinical practice

Case	Implications for clinical practice
1. Asymptomatic mild ARIA-E	• Most initial ARIA-E events occur within the first 8 doses of aducanumab. However, on occasion, ARIA-E can occur later in the course of treatment • Dosing continued in cases of mild asymptomatic ARIA-E
2. Asymptomatic and radiographically mild ARIA-E in an ApoE ε4 noncarrier	• ARIA-E may occur in ApoE ε4 noncarriers • ARIA-E may evolve over the course of treatment and be recurrent, at the same or, as in this case, at different locations at different times
3. Asymptomatic ARIA-H microhemorrhage followed by symptomatic mild ARIA-E	• The majority of ARIA-E cases are asymptomatic • When symptoms occur, headache is the most common symptom reported with ARIA-E
4. Asymptomatic, mild ARIA-E evolving into moderate multifocal ARIA-E, and subsequent recurrent ARIA-E	• ARIA-E may recur, increase in size, or become multifocal over the course of treatment • ARIA-H microhemorrhages may occur independently of ARIA-E
5. Asymptomatic moderate ARIA-E evolving to symptomatic moderate ARIA-E and mild asymptomatic ARIA-H superficial siderosis	• ARIA-E may evolve from mild and asymptomatic to moderate (or greater) radiographic severity and become symptomatic • The location of the ARIA-E may correspond to clinical symptoms (when present) • Per aducanumab appropriate use recommendations, dose suspension is recommended with the onset of clinical symptoms associated with an ARIA event [30]
6. Asymptomatic severe ARIA-E, asymptomatic mild ARIA-H microhemorrhages	• Radiographic severity may not correlate with clinical severity of ARIA • ARIA-E and ARIA-H may be concurrent
7. Asymptomatic severe ARIA-E, asymptomatic severe ARIA-H microhemorrhages	• Along with ApoE ε4 carrier status, the presence of baseline microhemorrhages is a risk factor for development of ARIA • Even if asymptomatic, permanent treatment discontinuation may be necessitated per protocol based on radiographic severity of ARIA-H, including in this case of severe ARIA-H microhemorrhages
8. Symptomatic severe ARIA-E (nonserious clinically)	• Although most cases of ARIA-E are transient, in cases when ARIA-E is symptomatic, radiographic findings may typically persist longer than symptomology
9. Serious symptomatic and radiographically severe ARIA-E with evolving symptomatic moderate/severe ARIA-H microhemorrhages	• Confusion is a common ARIA-E symptom • This case represents early onset of serious and radiographically severe ARIA-E (after the fourth dose) as well as the evolution of ARIA-H microhemorrhages that went from moderate to severe • Permanent discontinuation of treatment was required per protocol
10. Serious symptomatic and radiographically severe ARIA-E and severe ARIA-H microhemorrhage in an ApoE ε4 noncarrier	• Although ARIA is more common in ApoE ε4 carriers, serious and severe ARIA can occur in ε4 noncarriers, and those with baseline microhemorrhages appear to be at higher risk of developing ARIA • Clinical symptoms suggestive of ARIA include headache, nausea, fatigue, dizziness, confusional state, gait imbalance, and visual disturbance • Permanent discontinuation of treatment was required by protocol • Radiographic findings typically persist longer than symptomology

ApoE apolipoprotein E, ARIA amyloid-related imaging abnormalities

Understanding common clinical symptoms of ARIA is essential for careful safety monitoring. In participants when symptoms occurred (10.0% of participants treated with 10 mg/kg aducanumab who experienced ARIA [10]), commonly reported symptoms included headache (46.6%), confusion (14.6%), dizziness (10.7%), and nausea (7.8%) [10]. Rarely, cases of ARIA observed with aducanumab treatment were defined as serious (both with and without symptoms), highlighting the need for appropriate ARIA detection and monitoring [10, 15, 28–32]. Although corticosteroid treatment following ARIA was not per protocol, in some cases, the investigator or treating physician administered them to participants. During ENGAGE and EMERGE, the proportion of participants who took concomitant medications for ARIA, including corticosteroids, was approximately 1% overall and similar across dose groups. Further research is needed to better assess and delineate the appropriate pharmaceutical interventions for cases of ARIA, if any [29–32].

Clinical management of ARIA across the cases was largely aligned with the actions defined in the study protocols [10, 26]. Firstly, there was enhanced clinical vigilance for symptomatic ARIA during the first 8 doses of treatment with aducanumab (which coincides with the period of highest ARIA risk [10]). Notably, however, as demonstrated by clinical trial data and illustrated in some cases in this series, ARIA can occur well outside of these windows of greatest risk. Secondly, brain MRIs were performed at baseline prior to treatment initiation. To fully characterize ARIA during the study, brain MRIs were also scheduled at study Weeks 14, 22, 30, and 42, as well as additional time points after titration per the study protocol. Thirdly, for participants in whom ARIA was detected, follow-up MRIs were required per protocol every 4 weeks until stabilization (in case of ARIA-H) or resolution (in case of ARIA-E).

As illustrated in some of these case studies, the radiographic and symptomatic (when present) presentations of ARIA do not necessarily follow the same course. As observed in the ENGAGE and EMERGE clinical trials, typically symptoms resolved before radiographic resolution/stabilization. In the phase 3 studies of aducanumab, the median (interquartile range) duration of symptoms was 5 (1–13) weeks in the 10 mg/kg group [10]. Moreover, ARIA-E radiographic severities may not necessarily correlate with clinical severity (i.e. as in Case 6 and Case 7, for which severe ARIA were not associated with symptoms). ARIA-E may also evolve and recur over the course of treatment. Although the presence of ARIA-E is associated with an increased risk for ARIA-H microhemorrhage and ARIA-H superficial siderosis, ARIA-H may also occur independently of ARIA-E.

Cerebral microhemorrhage is a known risk factor for ARIA and is an imaging feature of CAA [33, 34]. Other imaging features associated with CAA, including white matter hyperintensities [34], have been proposed as potential risk factors for ARIA-H [22]. ARIA and CAA-related inflammation (CAA-ri) share imaging characteristics and clinical syndromes, and may have common pathophysiology [35]. The relationship between imaging features indicative of CAA-ri and the development of ARIA has not been established or validated.

While the cases presented in this article represent only a portion of all the ARIA cases from the large phase 3 trials of aducanumab, the broad spectrum they comprise lends insight into the clinical considerations relevant to detecting, monitoring, and management of ARIA both in future clinical trial settings and the real world. The evaluation and care of the patients treated with anti-amyloid mAbs require knowledge of risk factors, the radiographic and clinical presentations of ARIA, as well as close partnership and communication with radiologists and other treating clinicians. Consensus-based appropriate use recommendations are available [28–31] to broadly guide clinicians regarding appropriate patient selection, patient-centered informed discussions and shared decision-making related to treatment expectations and meaningfulness [31], benefits, burdens, risks, and requirements; and safety considerations and processes regarding detection, monitoring and management of ARIA. Ultimately, however, each case is unique, and the appropriate clinical considerations and the detection, monitoring and management of ARIA depend on the judgment of the treating clinician individualized to the particulars of their patient [29–32]. With the clinical availability of anti-amyloid mAbs for the treatment of early-stage symptomatic AD, increased need for ARIA education among clinicians remains.

Abbreviations

Aβ

Amyloid beta

AD

Alzheimer’s disease

AE

Adverse event

ApoE

Apolipoprotein E

ARIA

Amyloid-related imaging abnormalities

ARIA-E

Amyloid-related imaging abnormalities (vasogenic edema)

ARIA-H

Amyloid-related imaging abnormalities (microhemorrhage or superficial siderosis)

ARWMC

Age-related white matter changes

LTE

Long-term extension

MCI

Mild cognitive impairment

MRI

Magnetic resonance imaging

PC

Placebo controlled

Authors' contributions

All authors read and approved the final manuscript.

Funding

Aducanumab clinical trials were sponsored and funded by Biogen. This study was sponsored and funded by Biogen, Inc. Medical writing and editorial support were provided by Nucleus Global in accordance with Good Publication Practice guidelines (http://www.ismpp.org/gpp-2022) and were funded by Biogen, Inc. Drs Atri, Aisen, Barakos, and Purcell and Mr Bracoud are not Biogen employees and did not receive funding to participate as authors in this manuscript.

Data availability

Data generated and/or analyzed in this study are not available due to privacy and ethical restrictions and cannot be shared publicly or transmitted to a third party.

Declarations

Ethics approval and consent to participate

Written informed consent for participation in the study and for the use and publication of study-related information was obtained from all participants or their caregivers. Participants were informed that study results may be presented at scientific meetings or published in scientific journals; however, no participant would be personally identified in any presentation or publication. The phase 3 studies of aducanumab were conducted in accordance with the Declaration of Helsinki and the International Conference for Harmonization and Good Clinical Practice guidelines and were approved by ethics committees or institutional review boards (IRBs) at each study site. Ethics committees or IRBs for ENGAGE included: St Vincent’s Hospital Human Research Ethics Committee (RGO), Eastern Health Office of Research and Ethics, Bellberry Limited, Ethics Committee of the City of Vienna, Western Institutional Review Board, University of British Columbia (UBC) Office of Research Services Clinical Research Ethics Board, Comite d’Ethique de la recherche du CHU de Quebec-Universite Laval, MUHC Neurosciences Ethics Board, Sunnybrook Health Sciences REB, Conjoint Health Research Ethics Board, De Videnskabsetiske Komitéer for Region Hovedstaden, Comité de Protection de Personnes Ile de France VI Hopital de la Pitie Salpetriere, Landesamt fuer Gesundheit und Soziales Berlin, Comitato Etico Area Vasta Nord Ovest, Comitato Etico Milano Area B, Comitato Etico IRCCS Ospedale S. Raffaele di Milano, Comitato Etico Regionale delle Marche, Comitato Etico Regionale della Liguria Largo Rosanna Benzi 10, Comitato Etico Provinciale della Provincia di Brescia, Comitato Etico dell’Università Cattolica del Sacro Cuore e annesso Policlinico “A. Gemelli”, Seconda Università degli Studi di Napoli, Comitato Etico per le attività biomediche “Carlo Romano”, Comitato Etico Palermo 1, Comitato Etico di Area Vasta Sud Est, Comitato Etico Campania Sud, Comitato Etico della Provincia di Bergamo Azienda Socio Sanitaria Territoriale Papa Giovanni XXIII (Presidio Papa Giovanni XXIII), Tokushukai Group Institutional Review Board, Nihon University Hospitals’ Joint Institutional Review Board, Nippon Medical School Chiba Hokusoh Hospital Institutional Review Board, Center Hospital of the National Center for Global Health and Medicine Institutional Review Board, IHL Shinagawa East One Medical Clinic IRB9-3 Shimoomashishinden, Sutokukai Nagaoka Nishi Hospital Institutional Review Board, NHO Tokyo National Hospital Institutional Review Board, Tokyo Metropolitan Geriatric Hospital Institutional Review Board, Tokushukai Group Institutional Review Board, Takeda General Hospital Institutional Review Board, Miyanomori Memorial Hospital Institutional Review Board, Toho University Medical Center Omori Hospital Institutional Review Board, National Center of Neurology and Psychiatry (NCNP) Institutional Review Board, Adachi Kyosai Hospital Institutional Review Board, Hakuyokai Kashiwado Hospital Institutional Review Board, Keijinkai Teine Keijinkai Hospital Institutional Review Board, IHL Shinagawa East One Medical Clinic Institutional Review Board, Yokohama General Hospital Institutional Review Board, NHO Asahikawa Medical Center Institutional Review Board, Nippon Medical School Musashi Kosugi Hospital Drug Institutional Review Board, Adachi Kyosai Hospital Institutional Review Board, Saitama Medical University Hospital Institutional Review Board, Nihonbashi Sakura Clinic Institutional Review Board, Institutional Review Board of Inha University Hospital, Institutional Review Board of Konkuk University Medical Center, Institutional Review Board of Samsung Medical Center, Institutional Review Board of Seoul National University Bundang Hospital, Institutional Review Board of Asan Medical Center, Institutional Review Board of Gachon University Gil Medical Center, CEIC - Parque de Saüde de Lisboa, CEIC Hospital de la Santa Creu i Sant Pau, CEIC-Hospital Clinic de Barcelona, CEIC Hospital Universitario Dr. Peset, CEIC de Euskadi, Hospital Universitario de La Princesa, CEIC-Hospital Universitari i Politecnic La Fe, National Taiwan University Hospital Research Ethics Committee, Chang Gung Medical Foundation Institutional Review Board, Chang Gung Memorial Hospital, Linkou, Kaohsiung Medical University Hospital Institutional Review Board, National Cheng Kung University Hospital Institutional Review Board, Changhua Christian Hospital Institutional Review Board, Taipei Veterans General Hospital Institutional Review Board, Scotland A Research Ethics Committee, NHS Lothian, Copernicus Group Institutional Review Board, MCW/FH Institutional Review Board, University of Miami Jackson Medical Towers East, KUMC Human Subjects Committee, Medical University of South Carolina Institutional Review Board, Indiana University Institutional Review Board, Quorum Review Institutional Review Board, Mayo Clinic Institutional Review Board, Western Institutional Review Board, Brigham and Women's Hospital Institutional Review Board, Biomedical Research Alliance of New York Institutional Review Board, Butler Hospital Institutional Review Board, Dignity Health Institutional Review Board, Cleveland Clinic Institutional Review Board, Georgetown University Hospital Institutional Review Board, University of Texas Southwestern Investigational Review Board, Beth Israel Deaconess Medical Center, Baylor College of Medicine, Washington University in St. Louis Institutional Review Board. Ethics committees or IRBs for EMERGE included: Commissie Medische Ethiek van de Universitaire Ziekenhuizen K.U. Leuven, Douglas Institute Research Ethics Board, Institutional Review Board Services, Western Institutional Review Board, University Health Network Research Ethics Board, Western University Health Sciences Research Ethics Board, Bruyere Continuing Care Research Ethics Board, Varsinais-Suomen sairaanhoitopiiri, Comité de Protection des Personnes Sud-Ouest et Outre Mer 2 Agence Regionale de Sante (ARS) Midi-Pyrenees, Ethik-Kommission der Universitaet Ulm, Comitato Etico Palermo 1, Comitato Etico Indipendente dell'IRCCS Fondazione, Azienda Ospedaliera Universitaria Policlinico Umberto I - Università di Roma La Sapienza, Comitato Etico della Provincia Monza Brianza, Comitato Etico dell'Azienda Ospedaliera Universitaria Maggiore della Carità di Novara, Comitato Etico della ASL LE di Lecce, Comitato Etico delle Aziende Sanitarie dell'Umbria di Perugia, Comitato Etico della Fondazione IRCCS Istituto Neurologico Carlo Besta di Milano, NHO Utano Hospital Institutional Review Board, Kobe City Hospital Organization Kobe City Medical Center General Hospital Institutional Review Board, Kojinkai Himeji Central Hospital Institutional Review Board, Shizuoka City Shimizu Hospital Institutional Review Board, Hyogo Brain and Heart Center Institutional Review Board, KYOSOKAI AMC Nishiumeda Clinic Institutional Review Board, Hamamatsu Clinical Research Network Institutional Review Board, Oita University Hospital Institutional Review Board, Kobe City Hospital Organization Kobe City Medical Center General Hospital Institutional Review Board, Nagasaki Kita Hospital Institutional Review Board, Ehime University Hospital Institutional Review Board, Mie University Hospital Institutional Review Board, National Center for Geriatrics and Gerontology Institutional Review Board, Osaka University Hospital Institutional Review Board, NHO Hiroshima-Nishi Medical Center Institutional Review Board, Osaka City General Hospital Institutional Review Board, Nihonbashi Sakura Clinic Institutional Review Board, Goshozuka Clinic Institutional Review Board, NHO Shizuoka Institute of Epilepsy and Neurological Disorders Institutional Review Board, JOHAS Kansai Rosai Hospital Institutional Review Board, Fukuoka University Hospital Institutional Review Board, Sofukai Okayama Kyokuto Hospital Institutional Review Board, Kagawa University Hospital Institutional Review Board, NHO Minami-Okayama Medical Center Institutional Review Board, Tokushukai Group Institutional Review Board, Rakuwakai Otowa Hospital Institutional Review Board, Haradoi Hospital Institutional Review Board, Independent Review Board Nijmegen, Komisja Bioetyczna przy Okręgowej Izbie Lekarskiej w Warszawie, CEIC Parc Salut Mar IMIM-Hospital del Mar, CEIm Parc Salut Mar IMIM-Hospital del Mar, Regionala etikprövningsnämnden i Stockholm, Commission Cantonale (VD) d'éthique de la recherche sur l'être humain, Kantonale Ethikkommission Zurich Wissenschaftliches Sekretariat, Commission Cantonale d'Ethique de la Recherche (CCER) HUG, Kantonale Ethikkommission Bern, Comitato etico cantonale c/o Ufficio di sanità, Ethikkommission Nordwest-und Zentralschweiz (EKNZ), Kantonale Ethikkommission Bern Wissenschaftliches Sekretariat, Copernicus Group Institutional Review Board, Cleveland Clinic Institutional Review Board, Western Institutional Review Board, Partners Human Research Committee, UCSD Altman Clinical and Translational Institute, Institutional Review Board, Research Protections Office, Weill Cornell Medical College Institutional Review Board, University of Southern California Health Science Campus Institutional Review Board, Stanford Institutional Review BoardResearch Compliance Office, Greenwich Hospital Institutional Review Board, Biomed Institutional Review Board, Human Investigation Committee Institutional Review Board Yale University, Banner Health Institutional Review Board, University of California-Irvine Institutional Review Board, University of Utah Institutional Review Board, University of Pennsylvania Institutional Review Board, Roper St. Francis Healthcare Institutional Review Board, Mount Sinai Program for the Protections of Human Subjects (PPHS), Wake Forest University Health Sciences Institutional Review Board Wake Forest School of Medicine, and Methodist Hospital Research Institute Institutional Review Board. To avoid identifying patients, the sex and ethnicity of each patient are not reported in this article.

Consent for publication

Not applicable.

Competing interests

AA received, over the last 10 years, honoraria or travel for consulting; participated in independent data safety monitoring boards; provided educational lectures, programs, and materials; or served on advisory boards for AbbVie, Acadia, Allergan, Alzheimer’s Association, Alzheimer’s Disease International, AriBio, Axovant, Axsome, AZ Therapies, Biogen, Eisai, Grifols, Harvard Medical School Graduate Continuing Education, JOMDD, Johnson & Johnson, Life Molecular/Lantheus, Lundbeck, Merck, ONO, Prothena, Roche/Genentech, Novo Nordisk, Qynapse, Sunovion, Suven, and Vaxxinity. AA receives institutional research grant/contract funding from NIA/NIH 1P30AG072980, R01AG070883, R01AG086363, U01AG082350, U24AG057437, AZ DHS CTR040636, Washington University St Louis (as Project Arm Leader for DIAN-TU Gantenerumab OLE and Amyloid Removal Trial, ART), Foundation for NIH (FNIH), Michael J Fox Foundation and Gates Ventures. AA served as a site primary investigator (PI) in the EMERGE study (aducanumab) and serves as a site PI in the AHEAD 3-45 study (lecanemab). GGC and KS are employees and shareholders of Biogen. PA was chair of the steering committee; has received research support from Eisai, Eli Lilly, Janssen, the Alzheimer’s Association, NIH, and FNIH; and has consulted for ImmunoBrain Checkpoint, Merck, and Roche. JB is an employee of Clario, a part of Thermo Fisher Scientific and has received speaker fees from Biogen. DP is an employee of Clario, a part of Thermo Fisher Scientific and has no disclosures to report. LB is a full-time employee of Clario, a part of Thermo Fisher Scientific. SBH, RH, CG are former employees of Biogen and were employed by Biogen at the time this work was conducted.