A resurrection of aducanumab for Alzheimer's disease
2019; Elsevier BV; Volume: 19; Issue: 2 Linguagem: Inglês
10.1016/s1474-4422(19)30480-6
ISSN1474-4465
Autores Tópico(s)GDF15 and Related Biomarkers
ResumoOn Mar 21, 2019, Biogen announced that the anti-amyloid antibody aducanumab failed futility analyses in two identically designed phase 3 Alzheimer's disease trials, and discontinued its development. However, on Oct 22, 2019, Biogen made the surprise announcement that they were applying for US Food and Drug Administration (FDA) marketing approval of aducanumab. The company explained that they reanalysed data from the trials to include patients who had continued in the studies from the Dec 26, 2018, cut-off date for the futility analyses to Mar 21, 2019, when futility was announced. Biogen stated that one trial showed significant findings and a subset from the second trial supports these positive findings. At a public event on Nov 21, 2019, Biogen challenged the FDA not to approve aducanumab now, stating that they should not be required to do another trial. Aducanumab is a human monoclonal antibody that selectively binds to amyloid β fibrils and soluble oligomers. In an early phase multiple ascending dose trial, cohorts comprising 165 patients with prodromal or mild Alzheimer's disease received monthly intravenous doses of 1 mg/kg, 3 mg/kg, 6 mg/kg, or 10 mg/kg, and showed substantial reduction of amyloid plaques in a dose-dependent and time-dependent manner, such that after 12 months nearly half the patients who received the 10 mg/kg dose no longer had positive amyloid PET scans. This finding was robust and unquestionable; plaques were markedly decreased in all cortical brain regions examined. However, clinical effects, assessed using four clinical scales after 6 and 12 months of treatment, were far less certain. Only three of 16 scales at 12 months (including the Clinical Dementia Rating–Sum of Boxes [CDR–SB] in the 10 mg/kg group) were nominally significant at a p=0·05 threshold, unadjusted for multiple comparisons, compared with placebo. Amyloid-related imaging abnormalities, mainly brain oedema (ARIA-E), occurred at the higher doses, most notably in about 40% of APOE ε4 allele carriers, and nearly half of these patients discontinued treatment. Longer term, open-label treatment showed continued reduction in plaques, such that most patients no longer had amyloid-positive PET scans. This early study demonstrated aducanumab to be a robust amyloid plaque buster. Biogen subsequently undertook two phase 3 efficacy trials of identical design with a goal toward gaining marketing approval if the primary outcomes were positive. The ENGAGE (NCT 02477800) and EMERGE trials (NCT 02484547) enrolled their first patients on Aug 31, 2015, and Sept 30, 2015, respectively. Participants had mild cognitive impairment due to Alzheimer's disease or mild Alzheimer's disease, a Mini Mental State Examination score ≥24, CDR score of 0·5, and positive amyloid PET scans; about two-thirds were APOE ε4 carriers. Participants were randomly assigned to placebo, a low dose of aducanumab (3 mg/kg if an APOE ε4 carrier or 6 mg/kg if not), or to a high dose (6 mg/kg if an APOE ε4 carrier or 10 mg/kg if not), to be given intravenously every 4 weeks over 78 weeks. About 1·5 years into each trial's enrolment, the APOE ε4 carriers in the high dose group had their doses increased to 10 mg/kg. The primary outcome of the trials was the CDR–SB; secondary outcomes were the MMSE, the 13-item Alzheimer's Disease Assessment Scale-cognitive subscale, and the Alzheimer's Disease Cooperative Study-Activities of Daily Living-MCI scale. The original, planned sample size for each trial was 1350, apparently intended to have 80–90% power to allow the detection of a 0·4–0·5 point difference on the CDR-SB. In November 2017, Biogen increased the sample size of each trial to 1650, based on a provision in the protocols to compensate for a larger than expected standard deviation of the outcome in order to maintain the statistical power. The planned futility analysis included 945 and 803 patients in ENGAGE and EMERGE, respectively, who had had the opportunity to complete a trial as of Dec 26, 2018. According to Biogen, the futility criteria were that the high and low dose arms of both studies each had to have less than 20% conditional power (ie, the probability of statistical significance for the CDR-SB at the end of the trials given the results at the time of the futility analyses). These criteria were met and the trials were curtailed on Mar 21, 2019. During this 3 month period, an additional 139 and 179 participants in each trial (a further 12% and 18%, respectively) had the opportunity to complete the trials. At an investors conference on Oct 22, 2019, Biogen disclosed that the ENGAGE trial still did not show statistical significance on the CDR-SB primary outcome in the high-dose group (there was a 6% worsening compared with placebo, p=0·627), but the EMERGE trial now showed a 23% improvement compared with placebo (p=0·031). Their analysis of all patients randomised (1647 and 1638, respectively), of which 563 and 656 in each had not had an opportunity to complete the trials, was similar (2% worsening, p=0·825; and 23% improvement, p=0·010, respectively). About 35% of participants developed ARIA-E. They would likely have been overwhelmingly APOE ε4 carriers randomly assigned to the higher doses of aducanumab. Biogen's slide presentation at the investors conference contained sparse information on clinical outcomes, mainly percentage differences and lone p values. The clinical data that were presented as absolute data were displayed in a figure showing change over 78 weeks on the CDR-SB for the placebo groups, from which the values could be estimated as about +1·5 in ENGAGE and +1·75 in EMERGE. Using the percentage differences relative to placebo (ie, −2% and +23%), the mean difference between groups in CDR-SB change in the two trials can be similarly estimated as +0·03 and −0·4 respectively; and thus, the change in the high-dose aducanumab groups can be estimated as +1·53 and +1·35. Although Biogen claims that the positive results of the EMERGE trial were driven by greater exposure to a higher dose in the larger dataset, the effect could just as likely been due to greater worsening in the placebo group. Biogen also compared with the placebo groups of ENGAGE and EMERGE (545 and 548, respectively), a selected subset of participants (116 and 147, respectively) assigned to 10 mg/kg of aducanumab who were essentially protocol compliers over the 78 weeks of the trial (except for about 15% who discontinued after 50 weeks). These groups showed differences of about −0·45 and −0·75 points between CDR-SB changes compared with placebo, as estimated from the figure; and it is this −0·45 difference in ENGAGE that the company considers as supportive of a positive outcome of EMERGE. Biogen frame their analyses as showing one positive trial and a second, negative trial in which a subset comparison is offered as supportive of the positive trial. They do this because the FDA might accept such result as meeting its regulatory criteria for “substantial evidence of effectiveness”, which is defined as “one adequate and well-controlled study and additional confirmatory evidence”. A conundrum for Biogen, however, is whether the EMERGE trial is, in fact, positive and well-controlled. Hypothesis tests yielding p values less than 0·05 can be compelling when they are carried out as planned, not post hoc, and not one of multiple tests performed. Because these were post hoc, subset analyses with smaller than planned samples (40% of participants did not complete EMERGE), treatment effects are likely exaggerated. The findings turned favorable with added participants after the futility analyses, but they might have regressed to the mean if the sample had increase to its planned numbers. Interim treatment effects can vary randomly in relation to a true effect, and the effects seen in EMERGE could be random (ENGAGE could be also randomly negative, but the futility analyses argue against such conclusion). There are other caveats that could affect the validity of these results. Potential biases could result from increased dropouts, more missed doses and lower compliance, and a substantial amount of unblinding of the high-dose aducanumab treatment due to ARIA-E predominantly in APOE ε4 carriers who make up two-thirds of those randomised, compared with lower dose aducanumab and placebo groups. The onset of ARIA-E is determined by MRI scanning and was managed by the clinical investigators who were effectively unblinded to treatment as few, if any, participants receiving placebo have ARIA-E. The previous trial demonstrated this issue well. Finally, the 23% relative difference on the CDR-SB is smaller than it might seem, corresponding to an absolute difference of about 0·4 CDR-SB points (based on their figure), which is of uncertain clinical relevance, especially in light of the lack of effect in the ENGAGE trial. The second conundrum is the uncertain validity of using a subset of high-dose patients in each trial who were mostly fully compliant with medication and completed the trial, and comparing this group to the larger randomly assigned placebo group. The subset is only 20% of the high-dose group and most likely includes a majority of APOE ε4 non-carriers who are far less susceptible to ARIA and likely to have less decline than the APOE ε4 carriers in the placebo group, making this comparison both extreme and biased. Biogen will need to clarify their results in their presentation at the 12th Clinical Trials on Alzheimer's Disease (CTAD) conference on Dec 4–7, 2019, in San Diego, USA. Hopefully, they will present descriptive statistics, demographic, and clinical details. They might fully describe dropouts and when they occurred. Sensitivity analyses should address dropouts, potential sources of bias, and potential effects of sites and dates of participants entering the trials. Subgroups should be based on baseline and stable characteristics, and analyses should include interaction tests. All subgroup analyses must be reported. The original futility analyses should be presented in detail and then show how the added participants seemed to reverse the treatment effect. With a full presentation we might better appreciate what aducanumab has to offer and what the next steps might be. A contentious FDA ruling for Alzheimer's diseaseOn June 7, 2021, the US Food and Drug Administration (FDA) finally made its decision on aducanumab, the anti-amyloid drug that has been the subject of a prolonged and controversial approval process . Contrary to the recommendation of their Peripheral and Central Nervous System Drug Advisory Committee (PCNSDAC), the FDA surprisingly approved aducanumab for the treatment of Alzheimer's disease. The ruling has split the neurology community, with some clinicians and researchers welcoming the approval as a needed boost to the field and others believing that the decision does not follow the data available. Full-Text PDF
Referência(s)