Dennis J. Selkoe, MD (DS): I am Dr Dennis Selkoe. I am a professor of neurology at Harvard Medical School and Brigham and Women’s Hospital.

Michael G. Erkkinen, MD (ME): I am Mike Erkkinen. I am a cognitive behavioral neurologist at the Brigham and Women’s Hospital in Boston.

Lawren VandeVrede, MD, PhD (LV): I am Dr Ren VandeVrede. I am an assistant professor and clinical trialist cognitive behavior neurologist at University of California, San Francisco (UCSF).

Factors Affecting the Choice Between Anti-Amyloid Treatments for Patients With MCI or Mild Dementia

DS: I am also involved in research, as we all are actually, but especially in the biology of this disease over the years. My laboratory discovered the pyroglutamate form of Abeta in a paper we published back in 1992, a mass spec analysis. While we are very familiar with and have antibodies in our lab that recognize pyroglutamate, I do think that Lilly’s donanemab requires some special consideration in terms of how to approach this with patients and have them well informed.

The pyroglutamate form that is targeted by donanemab is unequivocally toxic. It is very pro-aggregating. It is something that arises probably in the middle or later part of the disease, but still preclinically, clearly before patients come forward and get a clinical diagnosis. So it requires some special thinking because as Lilly has said and has included in the label and the appropriate-use recommendations for donanemab, this form of Abeta is a low-abundance form, maybe 2%, 5% where you do not really know in any 1 patient. But it is low abundance and therefore it is possible that successful donanemab treatment over a year could deplete the form.

Then you have to decide: should you continue? The answer would be if you really get rid of this unusual form of pyroglutamate, there would not be a good biological rationale to continue the treatment. Indeed, that is in the prescribing guidelines.

The patient and the family need to understand at the beginning if we start you on donanemab, we need to consider that at some point we will very likely stop it and we will maybe have to get another amyloid PET scan at that point to see if your amyloid has been cleared successfully. But in any case, it is unlikely to meet your required treatment beyond 12 months really mostly.

My point is that donanemab would be intended for a period of time, but not indefinitely. Lecanemab, we do not have that same sense of restriction and we may indeed to choose to treat patients longer, but at less-frequent times, so not every 2 weeks like you start lecanemab with, but eventually every 4 weeks, which is already FDA approved for maintenance therapy.

Then the last point I would make about donanemab vs lecanemab is that the reported rate of ARIA is higher with donanemab in both ApoE4 heterozygotes and homozygotes. That may be mitigated by a protocol that Lilly has established that looks like it can decrease the rate of ARIA. But there is still probably a greater likelihood that donanemab might induce ARIA, including symptomatic ARIA, which is what we really are concerned about and we will get into that a little bit more. So I think it is a more complex treatment, but certainly 1 that is proven to be effective. Some say that it may be even a bit more effective than lecanemab, but I do not think we have clear-cut data for that. Let me stop there.

ME: Yeah, I can add that in terms of weighing these factors, amyloid burden, tau burden, etc, and trying to decide — of course we will talk about ARIA in a moment — in our center we do not typically think about amyloid E burden as either high or low, pushing one toward donanemab or lecanemab vs if they had a low amyloid burden.

There is better evidence for donanemab, but in both agents there is some evidence that starting folks off who have a low tau burden may do better in the long run. But we do not usually get a tau staging in our clinics in part because there is not easy access to tau scans. So we do not really use that as a way of saying because you have a low or high tau burden, you should be on agent X or Y.

I think for us it really comes down to the convenience factors that are associated with donanemab, but weighed against the relatively uncertain safety risks of donanemab, the ARIA risk. The original trial showed a higher ARIA risk, but then there was this follow-up trial that showed perhaps with a modified dosing regimen, that perhaps the ARIA risks were closer to what was seen for lecanemab, that kind of uncertainty around the ARIA safety we are actually putting our patients in when helping to decide which agent is right for them.

The other thing I will add is there is a little bit of a lower risk of an infusion reaction with donanemab as opposed to lecanemab. While infusion reactions are rarely serious, they can certainly be a nuisance or bothersome. That is another factor that patients have to weigh.

LV: I can just echo that. It would be great to be able to use biologic stage in a more precise way to target treatment, but I think we do not really even have quantified tools for tau PET yet to be able to use that biologically. That is still being operationalized and Centiloids are now just recently available even for us to be able to look at, let alone integrate into our decision making.

The way the trials were conducted, the patient pools were largely similar. I think the goal in every case is to detect it early in the disease before the disease has become too entrenched and to clear the amyloid as quickly as possible because no one’s better with amyloid in their brain. Getting that down as quickly as possible seems to be the key to therapy. I am hopeful that we are going to learn more and more about how to use these tools in ways outside of the trial and build more personalized treatment regimens and learn about what happens after 18 months in patients and what to do after 18 months in the real world.

Until then, I think we are really based on reasoning from fundamentals, as Dr Selkoe described the biology in such great detail, reasoning what we think about. Then trying to find a way of communicating these relatively sophisticated nuances, the patients as they try to select between these 2 treatment regimens.

In the initial group of patients at UCSF, these are very sophisticated often as they are thinking about these and you can communicate that. But being able to communicate kind of broadly within our field and with 1 voice about how to use these medicines I think is really important. But there is a lot to learn I think and we are learning that in real time as they are being used in the real world. We are just about to have our first patient come to 18 months with the first treatment, so we are taking our first step down that road here at our center within the next month or so. We are really building that bridge as we are walking on it.

Personalizing ARIA Monitoring: Risk-Based Protocols and Discontinuation Thresholds

LV: I will say we probably do not have a personalized ARIA monitoring. This is sort of prescribed by the label about when you should collect the MRIs to monitor for it. We do use ApoE4 and personal characteristics like CAA to inform risk of patients and try to inform them about their individualized risk for ARIA for each of the 2 therapies that we were talking about early. We have that written out in a table and we communicate that to patients.

Then we break down the ARIA by asymptomatic, symptomatic, and what we really care about, which is severe symptomatic ARIA. That is actually what we are trying to avoid, not sort of any ARIA, which is amyloid-related imaging abnormality, which is an MRI finding that really is monitoring to prevent those later complications from occurring. That is the intention of the screening protocol.

The screening protocol is written. At least at our center we adapted from the trial through the appropriate-use recommendations, but maybe we have a lower threshold to screen patients that we think are at increased risk. Maybe we have chosen to treat the E4/E4 homozygotes at our center. Maybe we are listening very closely to them and asking if we see something on the MRI, trying to tie that to symptoms, potentially holding off the dose if needed.

If they do have that, following the appropriate-use recommendations of withholding treatment, additional screening to ensure resolution or stabilization of the symptoms prior to treatment resolution, except in those cases where it may be severe or symptomatic or recurrent to such a degree that it may be unsafe to proceed is important. I think at our center we do follow the appropriate-use recommendations, not personalized, but we do inform the risk.

ME: Yeah, and we do a very similar approach. I will say there are a couple of things that we have done just because ARIA is what keeps us up at night. So before each infusion during the 6-month higher-risk window for developing ARIA, before each infusion we will do sort of a screening assessment of new symptoms, which is a way to say: Are there any symptoms that raise an orange flag or yellow flag for ARIA? If there are, we ask them up front to report those to us. If they do not, we offer a screen. If there is something there that is concerning, we might get an extra MRI and hold the infusion a few extra days until we get the MRI to look. We have caught ARIA in a few instances in those ways.

The other thing that we have done, for sure using a first principles approach, but I do not think a lot of necessarily evidence for this. We have taken a more stringent approach for ApoE4 homozygotes up front. In our center we have a little bit of a more stringent MRI-based exclusion. So rather than greater than 4 micro leads on a gradient echo (GRE), we will exclude folks for a higher than 1 on the GRE or if they have 4 or more on susceptibility weighted imaging (SWI).

The other thing is if someone an E4 homozygote and they develop moderate ARIA, they will be permanently discontinued, which is a more stringent recommendation than is recommended in the appropriate use. That was just our center’s approach, emphasizing safety. However, that is to be balanced with the loss of efficacy. These folks are taken off the agent early and perhaps they would have benefited. So it is a very challenging area. I am sure every center is doing their best to weigh these risks appropriately.

DS: I would say for the practitioners that we are speaking to that are coming more and more to consider these disease-modifying therapies, it is terribly important to communicate very clearly with a patient and the family that they need to pay attention to the development of symptoms of ARIA. It is a symptomatic ARIA that we worry about, much less so than radiologic ARIA.

That means that they need to be sensitive to new headaches and it does become more challenging if they often have headaches in any event. This headache might be qualitatively different. It might be somewhat more severe. The patient has to be prepared to call the doctor’s office as soon as they recognize that they have a headache that is unusual for them or new.

That also goes on to other symptoms. If they have new dizziness, new nausea, new visual disturbance, new confusion, if their family members think that they are going the opposite direction from how they should go on therapy, those are things that have to be communicated right away to the doctor. When the doctor hears about those, that is the time to get an MRI.

I think ultimately in the long run, we may not rely as much on survey MRIs where we doctors automatically say before the fifth, before the seventh, before the 14th infusion. For example, lecanemab — you need an MRI. I think we would like to have patients be very sensitive to a change and inform the doctor. Then MRI can be done. Then if there is symptomatic ARIA, you see it on the MRI scan and the patient reports symptoms, we would pause if the symptoms are mild. If they are moderate or if the ARIA is moderate, we would discontinue.

That gives the practitioner a chance to really control the safety risk and I think if we really pay close attention to whether there are new symptoms and then is there ARIA on scan, we can manage the risk and not see these very unusual, very infrequent, but devastating events such as a low-bar hemorrhage or even death that has occurred occasionally.

Obviously, patients have to carry cards that indicate they are on lecanemab or donanemab so that as soon as they go to a doctor or an emergency room, they say I am on this medicine and that means I should not automatically be given anticoagulants for sure and certainly not be given so-called clot busters or something like tissue plasminogen activator (TPA) for a presumed stroke. That is where the real tragic events have occurred when this has been misunderstood and ARIA has been thought to be a new stroke.

LV: If I could just add to that, in some ways, the last 25 years of drug development in harnessing the immune system to clear amyloid is managing this symptom and becoming comfortable with it from the earliest trials actually even to the earlier versions of the passive antibodies that we are using. It is learning how to carefully select patients, how to monitor appropriately, how to guide the treatment to be able to limit this.

I think now that we are in the generation of implementation, if you follow very closely and advise anybody that is listening to this, if you follow very closely the clinical trial, we have seen now from the real-world setting that the ARIA in the real world is very similar to that. So not sort of going beyond those patients or treating patients that are too far along in the disease to benefit because the risk-benefit with these drugs really invert as the disease goes along and the risk of ARIA goes up and the likelihood of clinical benefit goes down. But the earlier and earlier you go, the less ARIA you will have.

Maybe the more clinical benefit that you will have, really targeting those early patients and those patients that are within the appropriate-use recommendations will keep you from seeing a lot of these very severe symptomatic ARIA cases, which are the worst-case scenarios, I think. That is my strong recommendation to people that are stepping into it. That is what we did at our center and not kind of go outside that quickly until you are comfortable with ARIA and managing it in all its different forms from mild to asymptomatic to symptomatic severe builds.

Aspects Influencing Deferred Anti-Amyloid Treatment: Balancing Clinical Judgment, Patient Preference, and Quality-of-Life Factors  

LV: Patient preference is everything. This is a shared decision-making model. We are being provided the information of what we know about these treatments and we talk about that in a clinical context with the patient. We make a decision together about what is right.

I think we know a lot about this disease in part due to the people on this call, Dr Selkoe and others, who have told us about what the biology is. If you see amyloid that is present, we know that we are on the road to a place that we do not want to go to, which is dementia, and that we want to prevent that if possible. However, where the evidence base is — it is for symptoms. So just the presence of amyloid is not sufficient yet, though there are prevention trials that we are looking at, whether these medicines may be effective in even asymptomatic stages.

What defines symptomatic? What is the earliest detectable change? I think this is something that we are asking ourselves quite a bit and how we operationalize this to be able to target those patients that are earliest. As I mentioned earlier, this is where the benefit is. If this worked better the later the disease went, it would it be easier. But it works best the earlier that you start it.

I think a talk a lot about patients, about when to initiate therapy. But I think what I said earlier was also true, is that no one is better with amyloid in their brain. Once we have clear evidence for a benefit, I think that we are usually kind of more excited to start with an MMSE of 30 than an MMSE of 20. But the treatments are not without burden. I tell that to patients up front. There are 5 MRIs in the first year, frequent touch points with the infusion center, calls from our nurses. I worry about people becoming professional patients generally. But these drugs do provide a benefit that is unequivocal and patients are more than willing to go through this to be able to get that benefit. So all of that has to be balanced.

There is a certain treatment fatigue after a year, a year and a half on treatment because there is maybe not a perceptible benefit. These are not drugs that you give and they make patients better. They slow the rate of progression. Sometimes that is hard to see as you go along, so counseling patients for what that looks like, the regimen, even providing them schedules when the MRIs are, the expectations of continued treatment I think is really important up front so that does not come as a surprise to patients as the treatment unfolds over the course of a year or even longer.

ME: Yeah, I think that is right. I do not have too much to add, other than I think that from the trials, it is important to recognize there was a real benefit. I believe it is a real benefit and that is why I offer these agents to our patients. But the benefit was modest at 18 months and it is a labor-intensive procedure with all the MRIs and the infusions and there is risk. As we are learning more and more; the risk is manageable I think in most instances, but there is risk.

I think for folks who are on the border between mild and moderate dementia, they have to weigh those things carefully. We do our best to sort of inform them what their likelihood of benefit is, what the risk is, and they have to weigh in their own lives if it is worth all the effort that it goes through to be on these agents. For sure, that is an essential part of counseling patients about whether or not these drugs are appropriate for them.

This discussion was edited for clarity and length.