. Donanemab in Early Alzheimer's Disease. N Engl J Med. 2021 May 6;384(18):1691-1704. Epub 2021 Mar 13 PubMed.


Please login to recommend the paper.


  1. I am very encouraged. This trial met its prespecified primary endpoint, which I believe is a first for an anti-amyloid drug in our field. In addition, the growth of tau, measured by tau PET in the frontal lobes, was slowed; this provides evidence of disease modification.

    Finally, their protocol did not require monoclonal antibody infusions throughout the study, but dosed participants for six to 12 months and then stopped for almost all participants once their amyloid load was brought down to “normal” levels. In my view, this study, together with data from other trials, strongly confirms the “amyloid hypothesis” and demonstrates that treatments aimed at amyloid can slow cognitive decline and modify the progression of AD.

    View all comments by Michael Weiner
  2. Saturday was a very exciting day! The donanemab Phase 2 study offers some of the most compelling evidence to date that amyloid reduction can slow clinical decline, albeit modestly, in early stages of AD. While we have seen a suggestion of similar results before with aducanumab and lecanemab, these trials had known limitations, whereas the donanemab data are more straightforward to interpret. The safety profile is similar to what has been reported previously with potent anti-Aβ antibodies in terms of ARIA-E and ARIA-H, and in my opinion would be considered acceptable by most patients.

    In addition to the encouraging results, there are two unique features of this trial that move the field forward. First is the decision by design to reduce dose or switch to placebo once patients are in the amyloid-negative range on PET. This approach has tremendous appeal when considering infusion therapies, and I hope it proves effective.

    The second is the selection of patients with intermediate tau on PET. This, too, has strong biological face validity, though one can imagine the significant issues it would raise in clinical implementation. It is interesting that the overall clinical benefit was driven by the patients in the lowest tau PET tertile. This suggests that the primary role of amyloid-lowering therapies may be in patients in whom tau is not yet widespread, most of whom will be in the preclinical or very earliest clinical stage. Progress in blood-based biomarkers should greatly facilitate the detection of earliest-stage AD in an accessible, equitable, and cost-effective manner.

    While these results are encouraging for the overall drug class, the FDA needs to consider the aducanumab EMERGE and ENGAGE data on their own merits. There are significant differences between aducanumab and donanemab in the antibody-targeted epitopes, study design and patient populations, and one cannot generalize results from one trial to the other.

    View all comments by Gil Rabinovici
  3. This study used a straightforward trial design with a clear outcome, which met the prespecified outcomes. The rate of amyloid-related imaging abnormalities was not trivial but the consequences were manageable. The effect size of reduction of decline was small, and it remains to be seen if a) the result can be replicated, b) whether the result is durable, c) whether the broader community views the benefit as worthwhile.

    There is no question that the drug hits its target: The degree of amyloid-lowering is impressive, and the fact that patients could be withdrawn from the treatment is a remarkable prospect for broader use. I am ordinarily skeptical of secondary analyses, but in this case, the regional reductions in tau at least suggest that there was some downstream effect on tau.

    The donanemab story is the most encouraging news on the amyloid front, ever, but whether the effect size is clinically meaningful is questionable. Furthermore, given the dramatic effects on amyloid by donanemab, it seems unrealistic to expect that there is some further way to enhance the benefit of an amyloid-lowering drug.

    By way of disclosure, I am a site investigator for the A4 study, which is funded largely by Lilly.

    View all comments by David Knopman
  4. In any therapeutic trial, of course it is much nicer to detect a potential signal of efficacy than not to detect such a signal. The more limited that signal is, though, the more cautious we have to be in interpreting its basis.

    One possibility here is that donanemab truly slowed cognitive decline in Alzheimer’s disease. However, as is the case with other trials of amyloid-β-targeted antibodies, individuals with APOE4 alleles are much, much more likely to develop ARIA than those without APOE4 alleles. This could lead to a selective elimination of APOE4 carriers from the active treatment arm, which is relevant because AD patients who are APOE4 carriers may decline more rapidly than AD patients who are not APOE4 carriers (see Figure 1C in Petersen et al., 2005). 

    ARIA of APOE4 carriers in the treatment arm could also lead to their unblinding, which the authors acknowledge. Table 2 from the Mintun et al. report shows 33 out of 93 APOE4 carriers experienced ARIA-E. Only one out of 90 APOE4 carriers in the control group experienced ARIA-E. Hopefully the planned Phase 3 studies will implement measures to take into account these potential confounders.

    Regardless of the efficacy question, donanemab certainly did do an impressive job of reducing PET-detectable β-amyloid.


    . Vitamin E and donepezil for the treatment of mild cognitive impairment. N Engl J Med. 2005 Jun 9;352(23):2379-88. Epub 2005 Apr 13 PubMed.

    View all comments by Russell Swerdlow
  5. The Phase 2 donanemab results are exciting and important. This innovative trial of a highly potent amyloid-normalizing antibody met its primary cognitive/functional endpoint. As with other such antibodies, a significant number of participants developed ARIA, but this was manageable.

    It is great news for the field that donanemab has already advanced to a Phase 3 trial. The donanemab results provide powerful support to the amyloid therapeutic hypothesis; this strategy will bring the first disease-modifying drugs for AD into clinical use.

    View all comments by Paul Aisen
  6. The Phase 2 data on donanemab is encouraging and raises expectations for Phase 3 trials. The effect size on the primary endpoint (iADRS score) is considerable and very promising. However, secondary endpoints such as CDR-SB, ADAS-Cog13, and MMSE were less impressive. There was an excellent reduction in amyloid PET signal, though tau PET did not change, and no CSF biomarkers were included.

    It is difficult to directly compare TRAILBLAZER (donanemab) to ENGAGE/EMERGE (aducanumab) due to the smaller sample size of TRAILBLAZER and the different primary endpoint, as ENGAGE/EMERGE topline results did not, to my knowledge, include iADRS. EMERGE did show a significant effect on CDR-SB, though the magnitude of effect appears similar to that seen in TRAILBLAZER , though this was not significant in TRAILBLAZER. ENGAGE/EMERGE did show decreases in CSF p-tau markers and tau PET from a small subset of patients, an effect not seen in TRAILBLAZER.

    Safety concerns related to ARIA seem to be similar to aducanumab. During the double-blinded period, the proportion of treated patients with ARIA-E with donanemab (26.7 percent) was similar to low-dose aducanumab (25.7 percent) and less than high-dose aducanumab (34 percent), as reported for EMERGE (similar in ENGAGE). ARIA-H was less common with donanemab (8.4 percent) than with low-dose (16.2 percent) or high-dose (18.6 percent) aducanumab.

    The current Phase 2 data for donanemab is exciting. The efficacy appears similar to or perhaps greater than aducanumab, though the primary endpoints were different. The safety profile is similar to aducanumab. A potential advantage of donanemab is that antibody dosing may be intermittent, with BACE inhibitor treatment in the interim, which may reduce burden on patients and cost.

    Donanemab also engages a different, plaque-related target. Thus, larger studies are certainly warranted and will complement ongoing studies with other drugs, including aducanumab.

    View all comments by Erik Musiek
  7. I am impressed with the donanemab data for several reasons. The design is informed based on previous studies. The titration up is appropriate. The selection of intermediate tau load is novel.

    The pyro-glu epitope is downstream from oligomeric species, suggesting more than one form of amyloid might serve as a target

    The study met its prespecified endpoint on the iADRS, a composite measure that is not confounded by difficult-to-interpret analysis, e.g., Bayesian for lecanemab, and futility for aducanumab.

    The titration down after amyloid PET is negative is a forward-thinking approach and will be the trend in the future. I look at it as induction therapy. Also, all secondary endpoints trend in the same direction.

    This is now the third or fourth mAB to show directional concordance with amyloid removal and clinical signal.

    View all comments by Marwan Sabbagh
  8. I think the donanemab data are interesting and possibly encouraging. The key features are:

    1. Donanemab lowers amyloid promptly and significantly to essentially normal levels;
    2. Once normalized, the sponsor stopped dosing;
    3. The incidence of ARIA was significant and needs to be monitored, but the symptomatic ARIA rate was only 6 percent;
    4. This was the first anti-amyloid AD drug to meet a clinical endpoint in a Phase 2 trial. 

    The clinical endpoint was important, and the implication of the discontinued dosing once the participant’s amyloid level normalizes may suggest that, if replicated, we will not need to treat patients continuously. We may be able to lower their amyloid levels, monitor them and, if the levels rise, re-dose. This would be akin to giving a booster immunization. 

    All of this gives me cautious optimism about donanemab’s potential. I was not involved in the study and do not consult for Lilly.

    View all comments by Ron Petersen
  9. There are several very interesting aspects of the design of the TRAILBLAZER-ALZ study. The approach to reduce, or even stop, treatment when amyloid PET imaging shows that the drug target has been sufficiently diminished (or maybe even eliminated) will substantially facilitate implementation of this treatment in clinical practice.

    Even more importantly, this approach really opens up treatment of AD during the preclinical phase of the disease. In this phase, continuous treatment with immunotherapy for five to 20 years would not be an option for the huge majority of affected individuals. However, treatment of individuals with preclinical AD with donanemab for six to 12 months, until their amyloid PET has normalized, would probably be tolerable for most and might also be cost-effective.

    Following treatment, plasma levels of p-tau, and/or Aβ42/Aβ40, could be analyzed every three to six months, and when the plasma markers start to revert back to abnormal levels, a new amyloid PET could be done, triggering temporary treatment with donamemab if the amyloid PET scan reveals recurrent accumulation of Aβ fibrils.

    It will be very interesting to see the plasma biomarker data from the TRAILBLAZER-ALZ and the TRAILBLAZER-EXT studies to better understand how plasma p-tau and Aβ42/Aβ40 change in response to treatment, and how they relate to changes in amyloid PET in this setting.

    A potential worry is that the effects of the treatment on cognition and ADL function resembles those achieved by symptomatic treatment with cholinesterase inhibitors. The current results imply that removal of Aβ fibrils might improve neuronal function, but the rate of cognitive decline after six to eight months of treatment seems to be the same as in the placebo group. If the same results are obtained in the follow-up study (TRAILBLAZER-ALZ 2), then it will be important to better understand the reasons behind this finding.

    It could be that the Aβ fibrils themselves, or the oligomers formed on the surface of the Aβ fibrils, have direct toxic effects that are mitigated by the treatment, resulting in a relatively improved cognitive function during the six- to eight-month period when most Aβ fibrils are removed from the brain. However, after that six- to eight-month period the cognitive decline seems to continue at a rate similar to the placebo group, which might imply that Aβ fibril-independent toxic mechanisms are at play at this late stage of the disease.

    It is possible, but not yet proven, that upstream events in AD, like Aβ pathology, might trigger downstream events where the latter eventually becomes independent from the initiating event. If this is true, then treatment during the preclinical phase of the disease should be clearly more effective. The preliminary results of the donenemab trial, indicating that individuals with a higher burden of tau pathology, i.e., those at a later stage of the disease, did not respond as well to treatment, could support this theory.

    In conclusion, I think the results of the TRAILBLAZER-ALZ are very promising. I am really looking forward to a cleverly designed trial in preclinical AD with this drug, e.g., including individuals with abnormal amyloid PET scans, and potentially also increased P-tau levels, but as yet no or limited neocortical tau PET signal. 

    View all comments by Oskar Hansson
  10. I was sadly disappointed by the study data. It is curious that there has been little discussion about the acceleration to brain-volume loss (whole brain and ventricular enlargement) in patients treated with donanemab. To preempt the tortured rationalization that this may reflect reduction of plaque load, the total Aβ load in the human brain with AD is ~6.5mg (Roberts et al., 2017), the difference in whole brain volume between placebo and donanemab was ~5ml (Fig 3C, Mintun et al., 2021). It is inconceivable that a reduction in Aβ could account for this difference.*

    A better explanation is that donanemab induced brain damage, but not to the hippocampus, where no difference was detected. Several other anti-amyloid drugs have also reported accelerated brain volume loss, which could be a worrying class effect.

    These anatomical findings are a caveat to the cognitive/functional results, which, on a generous read, are not strong. Donanemab seemed to merely delay decline by ~six months. Disease progression seems to ensue without altered trajectory. While we welcome any positive impact to cognition, the effect appears modest, and our enthusiasm must be tempered by a marginal p value for iADRS and that all other tests were nonsignificant.

    As others have noted, this was a well-performed study, with reasonable N, and stringent selection criteria. If donanemab had a clinically meaningful effect, we would expect to see this reflected in the statistical analysis of the other clinical measures. If we do not observe significance in a highly selected cohort, the chances are low of this drug meaningfully impacting the heterogenous clinical population in our societies.

    * If we assume that all Aβ in the brain is insoluble, therefore taking up volume, and that donanemab removed all 6.5 mg of Aβ protein in the brain, and that the specific density of protein is 1.35g/cm3, then we can calculate that, at most, Aβ would occupy 0.0048 cm3 = 0.0048 mL. This is 1,000 times less than the change in brain volume reported.


    . Biochemically-defined pools of amyloid-β in sporadic Alzheimer's disease: correlation with amyloid PET. Brain. 2017 May 1;140(5):1486-1498. PubMed.

    . Donanemab in Early Alzheimer's Disease. N Engl J Med. 2021 May 6;384(18):1691-1704. Epub 2021 Mar 13 PubMed.

    View all comments by Scott Ayton
  11. How can anyone claim that this study confirms the amyloid hypothesis? While donanemab is able to clear the brains of patients of amyloid very effectively, the impact on cognition is only marginal, and perhaps of little relevance in the daily life of AD patients. Surely, if the amyloid theory was right, disease progression should have been halted for good.

    Putting ideology aside, the outcome of this trial proves that amyloid plays virtually no role in disease progression of AD. It is time to move on and to focus on other parameters that appear to be of importance in AD, such as chronic inflammation and loss of energy utilization in neurons.

    View all comments by Christian Hoelscher
  12. The donanemab data is both encouraging and a little bit disappointing.

    Encouraging because donanemab does the job of amyloid plaque clearance as never before and shows cognitive/functional benefits, albeit modest and with arguable clinical significance. Disappointing because it seems difficult to do better than that on the anti-amyloid front, and the clinical effects are not up to our collective expectations.

    So what now? I would be thrilled to see a long-term follow-up study with donanemab. The A4 study that was actively debated when it started now seems like the perfect test of this therapeutic venue but, unfortunately, solanezumab does not seem to be the drug we would choose today.

    More and more I wonder about the ability any drug can have to show meaningful efficacy during a three-year follow-up period that is dictated more by financial reasons than by scientific or medical ones. Why not include in these trials at least a subgroup of participants who would agree to undergo a loose, digitally assisted follow-up for up to 10 years, maybe through academic/industrial partnership as in A4? The medical community would have much more confidence in a drug that is able to show a long-term positive impact for AD patients.

    View all comments by Stéphane Epelbaum
  13. Aβ immunotherapy has been considered as a treatment for Alzheimer’s disease for some years; however, the lack of clear cognitive efficiency associated with the presence of side effects in some patients has tempered enthusiasm. Nevertheless, the recent publication of a clinical trial of passive Aβ immunotherapy targeting a specific form of post-translationally modified Aβ has confirmed that clearing Aβ seems to slow the cognitive decline in patients treated at a relatively early stage of the disease.

    Two additional observations were highlighted in this publication, consistent with the postmortem studies conducted on the unique cohort of 22 Alzheimer’s patients actively immunized against Aβ42 (AN1792, Elan Pharmaceuticals). Firstly, the treatment not only targets Aβ pathology but also improves tau pathology. Changes in tau pathology were previously reported in immunized patients with resolution of plaque-associated dystrophic neurites and reduction in phosphorylated tau in neuronal processes (Boche et al., 2010), associated with a reduction in the tau-phosphorylating enzyme GSK3β (Amin et al., 2015). Changes were still present 14 years following Aβ immunization with a focal reduction in tangles in the regions free of Aβ plaques (Nicoll et al., 2019). 

    The second observation relates to the acceleration of brain volume loss in the treated patients concordant with the brain atrophy described on in vivo MRI (Fox et al., 2005), and postmortem (Paquet et al., 2015) after active Aβ immunotherapy. Neuronal loss observed as increased interneuronal distance, reduced neuronal density, and increased cerebral cortical neuropil degeneration were enhanced in the brains of immunized Alzheimer’s patients relative to non-immunized patients (Paquet et al., 2015). However, these studies also reported improved health of the residual neurons with better neuritic curvature and less pro-apoptotic neurons in the immunized brains (Paquet et al., 2015; Serrano-Pozo et al., 2010Paquet et al., 2017). 

    Clinically, in the AN1792-treated AD patients, no association between neuronal loss and cognitive decline was detected, while a positive correlation between the number of neurons and the age at death was reported (Paquet et al., 2015). Therefore, the brain volume loss could, at least partially, be explained by a donanemab-induced reduction of injured neurons with, as in AN1792, a predicted decreased expression of apoptotic proteins in the brains of treated patients.

    To conclude, Aβ immunotherapy has likely beneficial effects, clearing the brain of Aβ plaques, reducing Aβ-driven tau accumulation and possibly removing degenerate neurons, but for these effects to impact on cognition, the treatment should be applied at the earliest stages in order to protect the neurons. All the effects detected in imaging the brains of patients involved in clinical trials have been detailed in postmortem studies, and thus a better knowledge and understanding of these neuro-immunopathological changes should be taken in account by the clinical studies (Boche and Nicoll, 2020). 


    . Reduction of aggregated Tau in neuronal processes but not in the cell bodies after Abeta42 immunisation in Alzheimer's disease. Acta Neuropathol. 2010 Jul;120(1):13-20. PubMed.

    . Effect of amyloid-β (Aβ) immunization on hyperphosphorylated tau: a potential role for glycogen synthase kinase (GSK)-3β. Neuropathol Appl Neurobiol. 2015 Jun;41(4):445-57. PubMed.

    . Persistent neuropathological effects 14 years following amyloid-β immunization in Alzheimer's disease. Brain. 2019 Jul 1;142(7):2113-2126. PubMed.

    . Effects of Abeta immunization (AN1792) on MRI measures of cerebral volume in Alzheimer disease. Neurology. 2005 May 10;64(9):1563-72. PubMed.

    . Effect of active Aβ immunotherapy on neurons in human Alzheimer's disease. J Pathol. 2015 Apr;235(5):721-30. Epub 2015 Jan 7 PubMed.

    . Beneficial effect of human anti-amyloid-beta active immunization on neurite morphology and tau pathology. Brain. 2010 May;133(Pt 5):1312-27. PubMed.

    . Downregulated apoptosis and autophagy after anti-Aβ immunotherapy in Alzheimer's disease. Brain Pathol. 2017 Oct 13; PubMed.

    . Invited Review - Understanding cause and effect in Alzheimer's pathophysiology: Implications for clinical trials. Neuropathol Appl Neurobiol. 2020 Dec;46(7):623-640. Epub 2020 Jul 25 PubMed.

    View all comments by James Nicoll

Make a Comment

To make a comment you must login or register.