2023-07-07 11:24:23 ET
Summary
- Cassava stock crashed in reaction to top-line data from a Phase 2 trial.
- Both treatment and placebo groups showed little decline on a cognitive test.
- Market doesn’t realize patient and design characteristics differ from past studies.
Previous coverage on Cassava Sciences (SAVA) listed the ongoing main bear theses : 1) management spins clinical trial results as being positive, 2) simufilam doesn't have a credible mechanism of action ((MOA)), and 3) Cassava gets plenty of bad press. Accordingly, market response on June 27 was muted to published research showing a simufilam MOA suppressing the mammalian target of rapamycin (mTOR)/Akt (serine/threonine kinase) pathway, which has been linked to two hallmarks of Alzheimer's disease ('AD'): the buildup of amyloid-? deposits into senile plaques and tau hyperphosphorylation into neurofibrillary tangles. Worse, shares sank 13% on July 5 after announced results of the Cognition Maintenance Study ('CMS'), despite headlines from news outlets being generally positive. A Motley Fool article was an exception, curiously claiming simufilam had a " lower response rate than some other Alzheimer's disease therapies," without specifying which ones. Investors should be reassured that the simufilam MOA profile is getting stronger, and that other therapies have not been demonstrated as better.
The CMS was a randomized withdrawal study open to patients who completed 12 months of simufilam treatment during the open-label study ((OLS)). Bears may be categorized into 3 levels, and their arguments regarding the CMS are similar but have nuances:
- Simufilam is placebo. Given the low bar set by other drugs, in my opinion, even FDA approval won't sway these non-believers, who totally dismiss all research and trial results as 'spin'. They'll only concede that simufilam is as safe as saline.
- There was no significant difference between groups who took simufilam and placebo. They think the CMS failed and may remain unconvinced of the positive OLS results discussed in a previous article in January.
- Simufilam is worse than other therapies. They see some merit in the science, and are waiting to see how good Phase 3 will be.
In the CMS, the simufilam arm (n=78) declined 0.9 points on AD Assessment Scale-Cognitive Subscale (ADAS-Cog) 11-task test (Figure 1), while the placebo arm (n=77) declined 1.5 points, a 38% difference at month 6 (95% Confidence interval [CI], - 2.1 to 1.0; not significant for sample sizes). While caution should be used in comparing non-head-to-head trials, a closer look may be informative. There are some content variations among the ADAS-Cog versions . The approved dose of Biogen's ( BIIB ) Aduhelm 10 mg/kg every four weeks slowed cognitive decline after 6 months in EMERGE (Figure 2), but this result was contradicted in ENGAGE (Figure 3), even though both trials recruited patients using the same exact criteria. At Week 24, Eli Lilly and Co's ( LLY ) donanemab group had the least squares mean improvement of -0.33 points, compared to 0.18 points worsening with placebo (Figure 4). Biogen's follow-up Leqembi 10 mg/kg biweekly also garnered positive p-values at 6 months (Figures 5 and 6).
Figure 1. Cognition Maintenance Study.
Figures 2 and 3. EMERGE and ENGAGE Topline Results : Two Phase 3 Studies of Aduhelm in Patients With "Early AD".
Figure 4. TRAILBLAZER-ALZ Phase 2 Proof of Concept Trial Symposium : Donanemab in "Early AD".
Figure 5. 2018 CTAD Presentations : Leqembi in "Early AD".
Figure 6. 2022 Clarity AD CTAD Presentations: Leqembi in "Early AD".
Remember that CMS was a withdrawal study. Investors should consider that from Figures 2 and 3, EMERGE/ENGAGE patients, after 1 year on Aduhelm, declined 2-2.5 points. From Figures 4 and 5, the donanemab and Leqembi groups declined by at least 1-1.5 points. Figure 6 shows Leqembi worsened by more than 2 points. All these other drugs performed on par or worse than CMS' PLACEBO!
Investors should then also note that the CMS participants had more severe AD than the other trials based on baseline mini-mental state examination ((MMSE)) scores. The vast majority (Tables 1 and 2) of Aduhelm and Leqembi participants had mild cognitive impairment ('MCI') based on Alzheimer's Disease Neuroimaging Initiative ('ADNI') cutoff of MMSE?26. The donanemab population distribution wasn't disclosed, but the mean score was in Mild AD territory. On the other hand, CMS patients' average MMSE put them on, not Mild, but Moderate AD, or two severity stages higher than most!
Table 1. Demographics of treatment arms from AD trials of approved/approvable drugs.
Study | Approved dose | Study duration | ADAS-Cog | n | MCI AD | Mild AD | MMSE score | ADAS-Cog |
version | % | % | mean ± SD | mean ± SD | ||||
CMS | simufilam | 6 months | 11 | 78 | 18.6 | 19.3 | ||
aducanumab | 26 weeks | 13 | 518 | 80.1% | 19.9% | 26.3±1.7 | 22.25±7.07 | |
ENGAGE | aducanumab | 26 weeks | 13 | 533 | 79.6% | 20.4% | 26.4±1.8 | 22.40±6.54 |
donanemab | 24 weeks | 13 | 112 | 23.6±3.1 | 27.6±7.7 | |||
lecanemab | 6 months | 14 | 130 | 59.2% | 40.8% | 25.6±2.4 | 22.1±7.7 | |
lecanemab | 6 months | 14 | 793 | 61.5% | 38.5% | 25.5±2.2 | 24.45±7.08 |
Table 2. Demographics of placebo arms from AD trials of approved/approvable drugs
Study | ADAS-Cog | n | MMSE score | ADAS-Cog | ||
version | MCI AD | Mild AD | mean ± SD | mean ± SD | ||
CMS | 11 | 77 | 18.1 | 21.9 | ||
13 | 528 | 81.4% | 18.6% | 26.4±1.8 | 21.87±6.73 | |
ENGAGE | 13 | 521 | 81.3% | 18.7% | 26.4±1.7 | 22.48±6.56 |
13 | 111 | 23.7±2.9 | 27.5±7.6 | |||
14 | 216 | 65.4% | 34.6% | 26.0±2.3 | 22.6±7.7 | |
14 | 823 | 62.2% | 37.8% | 25.6±2.2 | 24.37±7.56 |
Takeaways
Finances were reviewed before and nothing material has changed. In short Cassava's cash of $187.5 million with no debt provides a runway until 2025, but not sufficient to fund trials for other indications. CMS passes the eyeball test and came within a sliver statistically with the CI touching '1'. Simufilam may have slowed AD enough that its beneficial effects linger after 6 months, such that many patients with moderate AD who went off held up better than groups taking other active drugs. The main risk is that the positive trends in the OLS and CMS were observed in small populations, and may not translate to the larger Phase 3 trial, which would then threaten FDA approval and commercialization, but odds of that catastrophe are shrinking. The market may have preferred that simufilam be like Ozempic, which causes weight rebound after patients stop taking it. After all, it guarantees residual income for Novo Nordisk (NVO) by forcing patients to stay on the drug to keep the pounds off. But Cassava isn't messing around with results this good.
Just like with the OLS, Suzanne Hendrix, PhD, and CEO of Pentara (a biostatistical consulting company specializing in neurodegeneration including AD), was moved to say, "Results for simufilam continue to be noteworthy. The lack of disease progression in cognition, as measured by the ADAS-Cog over 18 months of treatment in mild patients, is well outside the range in historic placebo decline rates from numerous other studies. The placebo group in the CMS has started to decline again but continues to maintain benefit over historical placebo groups."
For further details see:
Why The Market Is Likely Wrong About Cassava Sciences, Again