Patients with Alzheimer’s disease (AD) who are being treated with anti-amyloid antibody therapy are at risk for developing amyloid-related imaging abnormalities (ARIA). While not all patients experience ARIA, careful considerations regarding risk factors and monitoring are critical to ensure safe treatment with anti-amyloid therapy.

ARIA covers 2 categories: ARIA-edema (ARIA-E), which involves vasogenic edema, and ARIA-hemosiderin (ARIA-H), which involves microhemorrhages. ARIA was originally observed when we started using anti-amyloid monoclonal antibodies for AD. We have since found out that monoclonal antibodies are not the only things that can cause ARIA. A patient can have ARIA events without receiving monoclonal antibodies. The conventional wisdom is that when these antibodies bind to the amyloid, they bind to not only parenchymal amyloid but also vascular amyloid, which is what could trigger both the vasogenic edema and the microhemorrhages. This is important because, in most cases, ARIA is asymptomatic.

Additionally, the severity of ARIA does not necessarily correlate with clinical severity. A patient can have severe radiographic symptoms and mild clinical symptoms, and vice versa (ie, mild radiographic symptoms and severe clinical symptoms). In my clinical experience, 6 of my last 7 cases of ARIA have been asymptomatic.

I consider the currently available anti-amyloid therapies lecanemab and donanemab to be second-generation anti-amyloid monoclonal antibodies. There are third-generation anti-amyloid antibodies in development, such as PRX012 and ACU193, that appear to remove brain amyloid much faster than the second-generation monoclonals. Investigators think that they are associated with fewer ARIA, although I am not sure that I believe that yet. Time will tell, and we will see.

As we think about detection, it is important to remember that findings of ARIA can be radiographically subtle. There are specific magnetic resonance imaging (MRI) sequences used for the detection of ARIA, and, importantly, ARIA is really not something that can be reliably detected with computed tomography scans. Some of the key MRI sequences that are needed include a T2-weighted fluid-attenuated inversion recovery (FLAIR) to detect ARIA-E, a T2-weighted gradient recalled echo (GRE) or susceptibility-weighted imaging (SWI) to detect ARIA-H, and diffusion-weighted imaging (DWI). Sometimes clinicians will integrate both GRE and SWI, recognizing that GRE was principally used in the clinical trials testing the anti-amyloid medications that are available now. In the right setting, DWI can help differentiate between findings of ARIA and change caused by acute ischemia. Many of the recommendations from the literature and consensus panels are to use a 3-Tesla MRI scanner to ensure high-quality imaging for ARIA detection, although this may not be feasible in all centers. It is also important to ensure that the scanner types are as consistent as possible, the quality of the scan is optimized, and the reporting of the results is uniform. If you are making treatment decisions based on a patient having 4 microbleeds in the brain vs 5 or 6, for example, knowing precisely where those microbleeds are so that they can be followed over time and compared longitudinally becomes critical.

It is therefore important to have your neuroradiology and neurology teams, as well as the rest of the treating community at a practice, on alert to detect these changes. One danger is that a patient can have early ARIA without it causing any symptoms; if an appropriate management plan is not in place, those ARIA changes can potentially bloom and cause severe consequences for your patient. This really does require collaboration to make sure that the detection expertise, fundamentals of imaging, and consistency of reporting are all there so that we are providing the best care to patients who are using these therapies.

For a person with AD who is being treated with anti-amyloid therapy, there is a pretty standard set of MRI intervals to monitor for ARIA. We follow the standard MRI protocol schedule depending on the monoclonal antibody, with more frequent MRI scans in the beginning, as ARIA is much more likely to happen within the first several infusions of anti-amyloid therapy, and a low threshold for additional MRIs if there is any suspicion of symptomatic ARIA.

There are risk factors for the development of ARIA. For example, if a person has the APOE4 allele, especially 2 copies, they are at a pretty high risk of developing at least asymptomatic ARIA, depending on the drug and other scenarios. We now test for the APOE4 gene to understand a person’s genotype and to better inform them of their risk of this side effect prior to initiating treatment. Patients with 2 copies of the APOE4 allele have the highest risk of developing ARIA, patients with 1 copy have an elevated risk, and patients with 0 copies have a neutral risk. There are some institutions that do not want to treat people who are APOE4 homozygotes, and I do not think that this is an unreasonable decision. There are other institutions that do it with some caution, given that those are also the people at highest risk for disease.

There is work being done to identify other risk factors for the development of ARIA. One potential risk factor is blood pressure, so lowering blood pressure is recommended to potentially lower the risk of ARIA. There are also different dosing titrations for anti-amyloid therapy that can reduce the risk of developing ARIA, such as titrating up slowly with the medication over a longer period.