On June 7 of this year, the FDA approved aducanumab for the treatment of Alzheimer’s, a monoclonal antibody directed against β-amyloid developed by Biogen and Eisai. This decision has received much criticism from researchers, clinicians, and independent organizations [1] [2], because there is no solid evidence to support that this drug may have a palpable benefit for patients. In addition, recent investigations into the interactions between the FDA and Biogen have heightened suspicions about the conflicts of interest that could underlie the FDA’s resolution.
Below, I summarize the process that led to the approval of this drug. In a first article I will address the evidence about the efficacy of aducanumab and in a second, how it has been approved, what has happened since then and what implications this decision has.
- Evidence on the efficacy of aducanumab
So far there is no treatment that can slow down the progression of Alzheimer’s disease. The efficacy of symptomatic treatments (acetylcholinesterase inhibitors and memantine) leaves much to be desired [3] [4]. The hypothesis of amyloid accumulation in the brain as an initial event in the pathogenesis of Alzheimer’s has led to attempts to develop drugs aimed at reducing production, antagonizing aggregation or increasing brain clearance of this substance in the last twenty years. but none have so far shown clear clinical benefits [5] [6]. This suggests that this may not be the best way to modify the course of this disease, since other hypotheses such as that the accumulation of β-amyloid was more of an epiphenomenon than a cause of Alzheimer’s cannot be ruled out [7].
The first data on the efficacy of aducanumab came from a phase Ib trial in 164 Alzheimer’s patients with mild symptoms, divided into four groups who received aducanumab at doses between 1 mg/kg and 10 mg/kg or placebo. In 2016, Biogen reported that there was a strong dose-response relationship between aducanumab and the reduction of β-amyloid as assessed by PET, as well as a partial dose-response relationship for cognitive assessments. After the preliminary results of this trial were known, two randomized clinical trials (EMERGE [8] and ENGAGE [9]) were started in 2015 to evaluate the efficacy and safety of aducanumab in people between 50 and 85 years of age with mild cognitive impairment due to Alzheimer’s or mild Alzheimer’s. They were similar in design: double-blind, placebo-controlled, with 1638 and 1647 participants respectively, and an expected duration of 78 weeks. The primary outcomes for both studies were changes in the CDR-SB dementia scale and other dementia scales (MMSE, ADAS-Cog 13 and ADCS-ADL-MCI) were used as secondary outcomes.
The FDA allowed Biogen to bypass the phase 2 trials to do these phase 3 trials directly, so that at the time of initiation, there was not enough information to know what doses of aducanumab would be optimal. In the phase Ib trial, it was observed that participants treated with high doses of aducanumab and carriers of the ApoE4 allele developed ARIAs (Amyloid β-Related Imaging Abnormalities, mainly cerebral edema and microbleeds) in a significant proportion, therefore the ENGAGE trials and EMERGE were initially designed to test only low doses (up to 6 mg/kg of aducanumab) in ApoE4 carriers, while non-carriers could receive up to doses of 10 mg/kg.
Subsequently, in other trials with other anti-amyloid monoclonal antibodies, it was observed that ARIAs used to be benign and reversible if recognized early, which led to modifications in the protocol in March 2017 to allow carriers of the ApoE4 allele also received doses up to 10 mg/kg. These changes would affect to the validity and interpretability of the futility analyzes that prespecified the protocols if these interim analyzes were not delayed, but these times were not altered [10].
In June 2018, both studies reached full recruitment and in December of that year the latest data was collected for the futility analyzes. In March 2019, Biogen announced in a press release that the futility analyzes indicated that aducanumab did not achieve improvement on cognitive scales even in the high-dose groups, so both studies were stopped. But in October 2019, seven months after this information was released, Biogen changed its mind and claimed that subsequent analyzes – which included data from another 318 participants – showed that those who received high doses of aducanumab at EMERGE had improvement on the dementia scales versus those who received placebo [11]; the low dose aducanumab group did not see this improvement compared to the placebo group. The data from the ENGAGE futility analysis showed that patients who received aducanumab, at high or low doses, had worse results than the placebo group for the primary variable (the CDR-SB scale) and for the MMSE scale, although improvements were observed dose-dependent on other scales (ADAS-Cog13 and ADAS-ADL-MCI). In both trials, PET images showed dose-dependent amyloid volume reductions with aducanumab.
As we can see, the results of the EMERGE and the ENGAGE were discordant regarding the impact of the monoclonal antibody on the cognitive-functional scales. In both studies, the groups that received aducanumab showed a reduction in brain amyloid, but this is a surrogate variable whose correlation with clinical improvement has not been demonstrated; in fact, the fact that all the groups that received the drug presented reductions in the biomarker but not all had benefits on the dementia scales supports the observation we outlined above: the reduction of amyloid does not have to translate into clinical gains. Therefore, it seems clear that if aducanumab can provide a tangible benefit to patients, at least one other study would be necessary to clarify the uncertainty about its efficacy.
But Biogen did a post hoc analysis of the data to understand the different results of the two trials [12] and argued that the small group of ApoE4 allele carriers who received high doses of the drug at ENGAGE would have behaved like EMERGE had they received the antibody for longer [13]. However, as the FDA statistical reviewers argued, “it is not justifiable to look in trial 301 [ENGAGE] for the most similar patients to those in trial 302 [EMERGE] because this may introduce selection bias and gives the impression that 302 is correct and 301 is wrong, which has no justification beyond post-hoc analysis” [14]. At a meeting in December 2019 Biogen announced that high-dose aducanumab was effective and in July 2020 it applied for FDA approval.
In sum, the evidence for the efficacy of aducanumab in Alzheimer’s patients comes from two trials (not published in a peer-reviewed journal, but simply announced in a company presentation) with uncertain and somewhat contradictory results. Both trials were stopped early after an interim analysis [15], in the trial proclaimed «positive» only a subgroup showed clinical benefit, and the hypothesis that a subgroup of ENGAGE participants would have benefited similarly to EMERGE comes from a post hoc analysis. Even if aducanumab could modify the clinical course of Alzheimer’s, a thesis that has not been proven so far, it remains to be seen to what degree the potential benefit would be for patients and their families, if it compensates for the side effects of this drug and at what cost.
If Biogen really believes that aducanumab slows the progression of Alzheimer’s, it should do a new clinical trial, but this time without cheating: looking at the final results (not in an interim analysis or through a post hoc analysis), analyzing the whole total number of participants (not just a subgroup) and for the time deemed necessary to demonstrate a benefit on clinical variables (ignoring biomarkers whose clinical correlate is unknown). But they haven’t. They have not done it because there are other faster ways (and with a greater probability of success) when your objective is for a drug to overcome the approved barrier, whether or not it shows a clinical benefit, but for that, the person who sets that barrier must be capable. to lower it. In the next post we will see how the FDA ended up approving this drug and the implications of this decision.
David Serantes is a first-year resident of Internal Medicine at the Fuenlabrada Hospital.
[2] https://www.statnews.com/2021/06/08/aducanumab-lackluster-drug-high-price/
[3] Sempere M, Palop V. ¿Sí o no a los fármacos actuales para la enfermedad de Alzheimer? Una reflexión obligada. AMF 2019;15(2):84-90
[4] McShane R, Westby MJ, Roberts E, et al. Memantine for dementia. Cochrane Database Syst Rev. 2019;3(3):CD003154. Published 2019 Mar 20. doi:10.1002/14651858.CD003154.pub6
[5] Dunn B, Stein P, Cavazzoni P. Approval of Aducanumab for Alzheimer Disease-the FDA’s Perspective. JAMA Intern Med. 2021;10.1001/jamainternmed.2021.4607. doi:10.1001/jamainternmed.2021.4607
[6] Mintun MA, Lo AC, Duggan Evans C, et al. Donanemab in Early Alzheimer’s Disease. N Engl J Med. 2021;384(18):1691-1704. doi:10.1056/NEJMoa2100708
[7] Panza F, Lozupone M, Logroscino G, Imbimbo BP. A critical appraisal of amyloid-β-targeting therapies for Alzheimer disease. Nat Rev Neurol. 2019;15(2):73-88. doi:10.1038/s41582-018-0116-6
[8] https://clinicaltrials.gov/ct2/show/NCT02484547
[9] https://clinicaltrials.gov/ct2/show/NCT02477800
[10] Knopman DS, Jones DT, Greicius MD. Failure to demonstrate efficacy of aducanumab: An analysis of the EMERGE and ENGAGE trials as reported by Biogen, December 2019. Alzheimers Dement. 2021;17(4):696-701. doi:10.1002/alz.12213
[11] EMERGE and ENGAGE Topline Results: Two Phase 3 Studies to Evaluate Aducanumab in Patients With Early Alzheimer’s Disease. https://investors.biogen.com/static-files/8e58afa4-ba37-4250-9a78-2ecfb63b1dcb
[12] It must be taken into account that these post hoc analyzes do not have a predictive power equivalent to the pre-specified analyzes and that their conclusions may, if anything, generate new hypotheses, not conclude anything solid about those investigated.
[13] https://blogs.sciencemag.org/pipeline/archives/2019/10/23/the-return-of-aducanumab
[14] https://blogs.sciencemag.org/pipeline/archives/2020/11/06/aducanumab-at-the-fda
[15] Trials that are stopped early for efficacy reasons tend to show a point estimator greater than the true treatment effect; that is, they tend to overestimate the effect of the investigated treatment.