On June 7, 2021, the Food and Drug Administration (FDA) granted accelerated approval to aducanumab (Aduhelm) for treatment of Alzheimer’s disease. In many respects, this move represents a landmark moment: not only is aducanumab the first drug approved for Alzheimer’s treatment in nearly 20 years, but as a monoclonal antibody targeting aggregated beta-amyloid, it is the first approved drug to directly modify a core molecular feature of the disease’s pathophysiology.
Many clinicians, researchers, and patient advocates lauded the approval, asserting that even modest slowing of disease progression by aducanumab would translate into meaningful benefits and provide hope for the estimated 6.2 million Americans with Alzheimer’s disease. There was optimism that aducanumab’s approval may spark new investment and innovation in drug development for Alzheimer’s and related disorders, after years in which leading pharmaceutical companies abandoned their neurodegeneration drug pipelines.
Counterbalancing this optimism, however, were concerns that the approval was premature, given conflicting evidence regarding aducanumab’s clinical efficacy from its two phase 3 randomized clinical trials (RCTs).1 In addition, vasogenic edema and cortical microhemorrhages (collectively called amyloid-related imaging abnormalities [ARIA]) occurred in about 40% of participants in both trials’ high-dose aducanumab groups, as compared with 10% of those receiving placebo.1 This complication was safely managed in the trials but raised worries about real-world safety. Moreover, even advocates of the drug were disappointed by the high initial price set by the manufacturer ($56,000 per year), consequently raising concerns about the drug’s potential impact on health expenditures.
Finally, critics questioned the appropriateness of close interactions between the FDA and the drug’s sponsor, Biogen, during the evaluation process, which the agency defended as necessary for interpreting complex data. After the approval, acting FDA Commissioner Janet Woodcock invited the Office of Inspector General to investigate these interactions, and two congressional committees are also conducting independent evaluations.
Aducanumab’s development followed a tumultuous course. A phase 1b study provided strong evidence of target engagement, with dose-dependent reductions in beta-amyloid plaques as measured on positron-emission tomography (PET),2 prompting the launch of two international phase 3 RCTs, ENGAGE (NCT02477800. opens in new tab) and EMERGE (NCT02484547. opens in new tab). Collectively, these trials randomly assigned 3285 patients with mild cognitive impairment or mild Alzheimer’s dementia (global Clinical Dementia Rating [CDR] score, 0.5; Mini–Mental State Examination score range, 24 to 30) and a baseline PET scan revealing amyloid plaques to receive placebo or low-dose or high-dose aducanumab. Both trials’ primary end point was a change in the CDR Sum-of-Boxes score (CDR-SB), an 18-point scale measuring cognition (memory, orientation, judgment, and problem solving) and function (community affairs, home and hobbies, personal care). A global CDR score is calculated from the sum-of-boxes, with a required score of 0.5 at study entry identifying patients with measurable cognitive deficits who nevertheless function largely independently.
In March 2019, both trials were terminated early on the basis of a prespecified, pooled futility analysis indicating that they were unlikely to meet their primary efficacy end points. Seven months later, in a stunning turnaround, Biogen announced it would file a Biologics License Application after an analysis of a larger data set revealed that EMERGE’s high-dose group had met the primary end point, with a 22% relative slowing of progression on the CDR-SB scale as compared with placebo (absolute difference, 0.39; P<0.01).1 No significant benefit was found in ENGAGE’s high-dose group or in either studies’ low-dose group. Biogen asserted that the EMERGE results met criteria for FDA approval on the basis of a single compelling trial.
In November 2020, the FDA convened its Peripheral and Central Nervous System Drugs Advisory Committee to review the trial data and post hoc analyses evaluating potential reasons for the discrepancy between the trials. The committee voted decisively that the data, in totality, did not provide sufficient evidence of efficacy and recommended against approval.
In another surprising twist, the FDA then pivoted from its normal regulatory pathway, which requires evidence of clinical efficacy, and approved the drug on an accelerated pathway in which evidence of biomarker changes deemed “reasonably likely to predict clinical benefit” is sufficient. This pathway, intended for drugs treating serious diseases with limited treatment options, has been used for drugs to treat HIV and cancer.
Aducanumab’s approval was based on the effect on beta-amyloid plaques on PET, a biomarker that was convincingly reduced by high doses in both EMERGE (−64.2% vs. placebo; P<0.0001) and ENGAGE (−53.5% vs. placebo; P<0.0001).1 What is more controversial is whether amyloid reduction constitutes a suitable surrogate for clinical benefit. Beta-amyloid–targeting therapies have an unimpressive track record, with numerous failed RCTs. However, most failed trials preceded the biomarker-based era of Alzheimer’s drug development, and aducanumab was the first drug in this class with robust PET evidence of amyloid lowering.
Subsequently, phase 2 RCTs demonstrated that two additional anti–beta-amyloid monoclonal antibodies, lecanemab and donanemab, cause similar reductions in amyloid on PET, with accompanying slowing of decline on clinical scales.3,4 Moreover, data from these studies and the aducanumab trials revealed that robust amyloid lowering modified biomarkers of tau pathology, including cerebrospinal fluid (CSF) concentrations of phosphorylated tau (indicative of beta-amyloid–mediated changes in tau)1,3 and tau PET (a measure of neurofibrillary tangles).1,4 These changes are thought to occur downstream of beta-amyloid and have more established correlations with clinical outcomes. Tau biomarkers, however, were collected only in a small, nonrandom subgroup of aducanumab and lecanemab trial participants.
These additional data provide supportive (if not definitive) evidence that reducing beta-amyloid during early-stage disease can modify pathophysiology and provide clinical benefit, which is the premise of the FDA’s acceptance of amyloid lowering as a surrogate biomarker. Under the accelerated approval provisions, Biogen must conduct a new RCT to verify clinical benefit, and approval may be withdrawn if it is unsuccessful. Details of this phase 4 trial are not yet available, but the FDA must insist on a rigorous design and timely completion to enhance confidence in aducanumab’s efficacy.
With the FDA decision behind us, practicing clinicians and payers must now set responsible guidelines for clinical use of aducanumab. The initial FDA label included a broad indication for “Alzheimer’s disease,” but on July 8, the label was narrowed to patients with mild cognitive impairment or mild dementia due to Alzheimer’s disease, mirroring the clinical trial population. Aducanumab should not be offered to patients with more advanced Alzheimer’s dementia or those with clinical features suggestive of other dementia syndromes, since both groups were excluded from the RCTs and seem less likely to benefit. Similarly, there is no current rationale for treating persons with “preclinical” Alzheimer’s (biomarker evidence of beta-amyloid deposition but no symptoms), though ongoing RCTs are testing amyloid-lowering antibodies in this group.
Although not currently required by the label, a positive beta-amyloid test (PET scan or CSF assay) should be a prerequisite for therapy, since a clinically determined Alzheimer’s diagnosis does not reliably establish the presence of amyloid. For example, in a recent study, amyloid PET was positive in only 5577 of 8742 patients (63.8%) previously diagnosed with Alzheimer’s.5
Treating clinicians should be well versed in monitoring and managing ARIA, which, though usually transient and often asymptomatic, can rarely be associated with severe symptoms. Patients who are considered at high risk for ARIA because of anticoagulant treatment or because of multiple microhemorrhages found on baseline magnetic resonance imaging should not receive aducanumab.
Treatment should be initiated only after shared decision making between a patient and their clinician, including discussion of current ambiguity regarding clinical benefit. Treatment costs should be balanced against potential benefits, and equitable access should be ensured by robust patient-assistance programs, lest this expensive, specialized therapy exacerbate racial inequities in access to dementia care. The Centers for Medicare and Medicaid Services is undertaking a national coverage analysis regarding drug reimbursement, and coverage by private insurance remains uncertain.
Aducanumab’s road to the clinic has been rocky and contentious. The hope is that despite its limitations and challenges, this first-in-class drug will open the door to more effective therapies, as has happened with other diseases once considered untreatable such as multiple sclerosis, cancer, and HIV. The modest clinical benefit observed with beta-amyloid–targeting antibodies should encourage development of drugs targeting other elements of Alzheimer’s pathophysiology (e.g., tau or neuroinflammation), which in combination with amyloid-based therapies might affect the disease trajectory more profoundly.
Whether aducanumab proves to be a blockbuster or a bust, its approval heralds a new era in Alzheimer’s disease care that will feature early, biomarker-supported diagnosis and biologically specific therapies. The modern age of Alzheimer’s therapy is upon us, and the field may never look back.