- Since releasing initial BAN2401 results in July, excitement among investors has waned as detailed analysis has shown confounding results. However, many investors still see promise in the drug.
- When viewed within the context of Biogen’s entire Alzheimer’s portfolio, Biogen remains well positioned to get an Alzheimer’s drug to market with aducanumab, BAN2401, and elenbecestat
- Amyloid beta hypothesis still has legs as drugs that target subtypes beyond monomers have shown some degree of clinical efficacy
- Possible futility analyses for aducanumab/crenezumab and FDA statement on ADCOMS as a legitimate endpoint are catalysts to watch, while aducanumab P3 readout in 2020 remains a key conclusive event.
My last article looked at Biogen's controversial Alzheimer's Disease (AD) drug, BAN2401, shortly after Biogen first released initial results back in early July. Since that time, a lot has happened that led the stock to give up just about all of its initial pop. In late July, Biogen released more detailed results from its phase 2 trials at the Alzheimer’s Association International Conference (AAIC). The results led to a heated debate about the reliability of the data presented. In response, Biogen then released an additional subgroup analysis at the Clinical Trials on Alzheimer's (CTAD) in late October. This addressed some concerns, but raised new ones as well and muddied the picture even further.
Today, the overall picture is a very mixed bag. On one hand, many still see promise in the compound, and in the amyloid beta hypothesis. With a better trial design and larger sample sizes, the company could get more conclusive data. On the other hand, the inconclusive nature of the existing phase 2 results prevents one from a confident investment in the company on the promise of BAN2401. The company will decide whether to pursue the drug in phase 3 trials, and if they choose to do so, it will be an expensive and risky proposition.
For those who are looking for a promising Alzheimer's drug with clear-cut evidence of a home run drug, I would advise against investing in Biogen. However, for those who are interested in diving deeper into the details and taking a position based on imbalanced risk/reward, Biogen might present a compelling opportunity when looking at the totality of its Alzheimer’s Disease assets.
Detailing The APOE4 and Dose Response Debate for BAN2401
After reporting limited, but exciting topline data from its phase 2 trial with BAN2401 in early July (which I detailed here), Biogen then released more detailed data at AAIC in late July that led to heated debates about the significance of the trial results. In particular, there were two issues that key opinion leaders (KOLs) and investors raised: 1) the ApoE4 imbalance, and 2) the lack of a clear dose response.
The primary concern raised by numerous investors after analyzing the detailed results was the imbalance of ApoE4 carriers in the placebo and highest dose arm (the arm that showed the statistically significant effect).
ApoE4 is a gene that is associated with 1) a higher risk of developing Alzheimer's, 2) a higher risk of having an earlier onset of the disease, and 3) a higher likelihood of amyloid-related imaging abnormalities events (ARIA). Additionally, many KOLs believe that having the ApoE4 gene could lead to a faster progression of the disease as well.
Within the placebo arm, 71% of patients had the ApoE4 gene, while only 30% of the highest dose arm had the gene. The imbalance was caused by a decision by a health authority outside of the US in 2014 to stop enrollment of ApoE4 patients for safety reasons.
As a result, investors feared that the imbalance between the placebo and highest dose cohort may have led to the entirety of the benefit that was seen in the results - with a higher rate of ApoE4 positive patients in the placebo arm, their AD may progress at faster rate due to the ApoE4 gene, thereby making the high-dose cohort appear to have a less-severe decline.
The key question is whether the ApoE4 gene is associated with faster AD progression. Numerous industry experts and sellside analysts believe it is, and it appears to be the working assumption by many. According to Leerink, there's a substantial amount of literature that suggests that microglia, ApoE4, and amyloid beta antibodies are linked together and could interact with each other enough to make the ApoE4 imbalance an important bias in the results. Leerink's interviews with an AD KOL also noted the belief that ApoE4 positive patients experience "greater amyloid, tau, inflammatory, and oxidative stress burden than non-carriers at similar stage of disease." Additionally, according to Stifel's initiation on Biogen, the covering analyst noted that it is well known that ApoE4 positive patients progress at a faster rate than their negative counterparts.
However, just as many industry experts have noted that the evidence is inconclusive. Numerous sellside analysts have conducted a literature review of prior studies that included ApoE4 positive and negative patients, and found that many studies found no statistically significant progression in AD between the two groups. For example, there was no statistically significant progression of AD in the Scarlet Road trial for gantenerumab, nor was there a statistically significant progression in the PRIME trial for aducanumab. Additionally, while many analysts noted that the ApoE4 imbalance likely complicates regulatory discussions further, the effect size was significant enough to not likely be due to the imbalance alone.
The second issue raised by investors was the lack of a clear dose response seen in the results. Investors wanted to see a dose response between the dosage and the amyloid clearance, and between the amyloid clearance and the clinical impact as measured by ADCOMS. Results showed that BAN2401 was successful in reducing amyloid beta levels at all doses, but the reduction in amyloid beta levels did not correlate cleanly with the clinical impact observed on ADCOMS. While the highest dose achieved statistical significance on ADCOMS and ADAS-COG, the lower doses actually performed worse than placebo most of the time throughout the trial. Additionally, the 5mg biweekly and 10mg monthly dose had similar levels of amyloid clearance, but did not see similar levels of clinical effect. While some KOLs speculated on plausible scientific reasons for the lack of a clear dose response, the lack of a strong correlation was not encouraging.
As a result, many analysts and investors were awaiting the subgroup analysis that Biogen was set to present at CTAD in late October. The subgroup analysis would examine differences between ApoE4 carriers and noncarriers, and would hopefully address the issue of whether the ApoE4 imbalance led to the observed effect. Additionally, the analysis could potentially shed more light on the lack of dose response.
CTAD Analysis Address Issues But Raise New Ones As Well; BAN2401 is a Risky Asset
Biogen presented detailed analysis at CTAD that addressed the ApoE4 imbalance, but raised several new questions as well about the quality of its results.
On the positive side, the ApoE4 imbalance was shown to actually have potentially hurt the results, not helped them. The ApoE4 positive patients actually had an enormous 63% reduction on ADCOMS, while the ApoE4 negative patients only saw a 7% reduction on ADCOMS, suggesting that the lack of ApoE4 positive patients in the strongest dose arm may have actually reduced the observed effect (which is the opposite of what skeptical investors had raised heading into the results). Furthermore, within the placebo arm, there was no statistical difference in disease progression between ApoE4 positive and negative patients. Note that the observed effect in ApoE4 positive patients was also seen on ADAS-Cog (84% reduction) and CDR-SB (60% reduction) as well.
Source: Eisai Presentation at CTAD
However, the positives were overshadowed by numerous other concerns. First, there were only 10 ApoE4 positive patients. It's not ideal when just 10 patients appear to be responsible for a significant portion of the observed effect. Secondly, the 7% reduction within the non-positive ApoE4 group was small, and looked even worse on CDR-SB. This was a surprising outcome given that aducanumab showed a somewhat similar effect within both subgroups, and reduces the perceived reliability of BAN2401’s phase 2 data.
Overall, the mixed results do little to change the view of either the bulls or bears. Bears believe the inconsistency of the data suggests that there is high risk in taking this drug into phase 3, and is too noisy to have much confidence in. The bulls argue that the totality of the data was still statistically significant, and that the data is enough to warrant moving onto phase 3 given the indication and what's at stake (the opportunity is significant for AD, as I've detailed before and many others have stated). The drug appears to be a safer but less effective (and less clear cut) aducanumab, and carries all the risks that are associated with developing an Alzheimer's drug. Ultimately, given the small sample sizes involved, we will need to see additional tests with larger sample sizes in order to reach more conclusive results.
The Amyloid Beta Hypothesis Has Warts, but Still Has Believers
The debate on the Amyloid Beta Hypothesis rages on, and it remains fairly divided after numerous mixed results from trials.
On the negative side, there's a loud and vocal group of investors and KOLs that believe the amyloid beta hypothesis has had too many failures and that the industry should move on to targeting other promising candidates. They argue that despite having shown promising results in animal models, and having success in clearing amyloid beta in trials, the clinical effect has not been shown to be efficacious enough. Patients still have Alzheimer's, and their disease progression continues. Many experts are instead pushing to target the tau protein or ApoE4 gene. Note that our understanding of the two targets are still very limited and are multiple years out from seeing initial results.
On the positive side, amyloid beta supporters argue that amyloid beta approaches drastically differ, and cannot be judged as a whole. The issue in many of the failed amyloid beta trials has been in targeting the right subtype of amyloid beta (i.e. monomers, oligomers, and plaque); targeting each subtype has often led to varying degrees of toxic amyloid beta in the brain and clinical effect (see my prior article for an explanation of the different subtypes). The drugs that have effectively targeted a subtype further downstream from monomers and have effectively cleared plaque have shown some degree of clinical efficacy. While the effect size in these drugs has still been admittedly limited, it may be enough to receive approval and drive significant uptake given the challenging Alzheimer's environment and lack of other treatment options.
Source: Eisai Presentation at CTAD
In the end, there is enough evidence to suggest that the amyloid beta hypothesis could still be at a minimum one of the components to stopping Alzheimers. If you are a believer in this hypothesis, Biogen remains one of the best positioned within this camp with three assets that have a shot at approval - aducanumab, BAN2401, and elenbecestat (not detailed in this note). Many investors and KOLs are most optimistic about aducanumab’s prospects given its strong results. As for BAN2401, the results are more mixed than aducanumab, and as a result, probability of success estimates vary from 10-30%. Finally, there is still some hope (albeit faint) that elenbecestat could continue on as well.
Upcoming Catalysts to Watch
We won't hear conclusive results from Biogen until early 2020 when aducanumab data is expected to readout. However, there are several important events that could move the stock significantly and provide further visibility on the amyloid beta hypothesis.
For Biogen, 2019 could see an FDA ruling on the legitimacy of ADCOMS as an endpoint. Additionally, further clarity on whether BAN2401 will move into pivotal trials could come in 2019 as well. Furthermore, aducanumab and Roche's crenezumab (which I detailed in my last article) could have futility analyses read out in 1Q19 that will be interpreted as either supportive of or undermining the amyloid beta hypothesis.