Celebration Thread! Biogen is going to the FDA with Aducan.

Alzheimer's, cardiovascular, and other chronic diseases; biomarkers, lifestyle, supplements, drugs, and health care.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

The figure a few posts above showing the relationship between placebo dose and SUVR was quite startling. All those patients on placebo who at the end of 18 months were amyloid negative? Almost impossible to imagine.
Amyloid positivity was an inclusion criteria for the aducan trial because those without amyloid are not considered to have Alzheimer's ... so all those placebo patients at 18 months were cured of their Alzheimer's?

The above figure illustrates this once again for the Leca phase 2 trial. Here what we see on the right is that ~22% of the leca placebo were amyloid negative at month 18. Apparently these placebo patients were also cured of their Alzheimer's. 22% placebo cure rate? Is this normal in other diseases? Importantly, amyloid removal is now the biomarker that defines AD ... and 20% + of placebo are amyloid negative at 18 months? This is very surprising to me.

As can be seen in the figure the percentage of these patients increases quite a bit through time (doubles from 12 to 18 months). Presumably if we waited long enough even more patients would be cured by placebo. Unfortunately, it is unclear what proportion of these placebo cures were in MCI and Mild AD. Perhaps at a more advanced stage of cognitive impairment possibly Mild AD, the percent of reversion to amyloid negativity would diminish. This might explain why the MCI patients versus placebo did substantially less well than Mild AD versus placebo.

It is surprising that this issue was not more fully explored in the aducan discussion as it could have greatly distorted the results. If large numbers of placebo are essentially in the treatment arm, then this would make it much more difficult to show a treatment effect.
In the post above the top figure is the aducan trial color coated for placebo and treatment. It was a significant mystery earlier in the thread how so many placebo appeared to perform so well. As seen in the top figure on the bottom, some of the placebo are indicated with purple squares. These placebo did very well. In fact, these placebo had some of the best cognitive responses of any patients in the trials. It would clearly be of interest to know which of these patients were amyloid negative (or with substantial amyloid removal) at the end of the trial. Those who did not believe that amyloid was of importance in AD (which is what the current biomarker argument is all about) should have no objection to removing those placebos with amyloid reduction. It would be of great interest to see how this would change the treatment vs placebo difference.
Last edited by J11 on Mon Jul 05, 2021 7:04 pm, edited 1 time in total.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

One other notable in the bottom figure above on the left is the SE bars. I copied the SE bars down to roughly were amyloid negativity would be reached ~~ -0.40 change in SUVR. There are 27 =13*2 +1 SEs from the placebo at top down to this negativity. One might think that this would mean essentially 0 chance of placebo reaching this low; and yet then 20%+ ??? of the placebo actually reach amyloid negativity. One might easily confuse SE for SD. +/- 3 SD should contain roughly 99% of the points. This is the mental impression when looking at the SE interval. One thinks that +/- 1 SD should contain 65% of the points. This does not appear to be true.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

Leca ARIA.JPG
aducan ARIA.JPG
You do not have the required permissions to view the files attached to this post.
User avatar
Julie G
Mod
Mod
Posts: 9187
Joined: Sat Oct 26, 2013 6:36 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by Julie G »

Sign me up for the placebo! ;)
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

The above post shows a comparison between ARIA and other side effects between the leca phase 2 and the aducan phase 3s. In the recent FDA documents it was noted that 44s had ARIA rates ~60%. What is remarkable in the above figure is how much lower ARIA is with leca. There was 0 ARIA in APOE- s even up to 5 mg/kg? That is impressive. Even with the APOE+ s the ARIA-E was much lower across the dosing range. One wonders what the 34 ARIA rate might have been? In the recent FDA documents posted online 44s had ~50% more ARIA than 34s. With the above one starts to imagine a potential dosing titration that could eliminate ARIA side effects.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

Thank you for commenting JulieG.

Yes, the price would clearly be more affordable on placebo.

I just had not thought of AD as a reversible illness.
In our experience AD is relentlessly progressive over decades into profound disability. I suppose this is what could be called selection bias or first person bias (i.e., my experience must be the experience of others).
For some, AD might be more like a year or two of cognitive impairment and then auto-immunity takes over.
This idea is not clear to me though, as it is possible that some of these converters from positive to negative might at some
point reconvert to positives and then enter more of a progressive impairment stage. Would be interesting to know what might have caused their conversion.

It does though reinforce the idea that there are safety concerns that arise with amyloid burden even when you choose not to do anything. It would not be entirely unexpected that some of those on placebo (or basically just untreated Alzheimer patients in the community) would experience amyloid clearance which could then trigger ARIA events. Indeed the placebo did in fact have ARIA events. Seeking treatment would allow people to proactively manage their ARIA risk. As amyloid treatment migrates to early stages, anti-amyloids will be able to reduce the inherent risk of ARIA in amyloid positives.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

There's lots of exciting aducan research still to discuss!

The FDA has posted more internal aducan documents that help to clarify a number of issues that I had not been clear about. The clinical pharmacology report is especially increasing and section 4.5 page 134 helps to resolve three of uncertainties. Namely the placebo leap that was widely noted as suspicious, lack of correlational and the order of secondaries analysis.

The placebo leap seems to have been just a big typo. Most of the difference can be explained .

The correlation question which has been considered on several occasions on this thread (admittedly this has been an ongoing confusion for me) also has a clear explanation in the FDA document (see extended quote below). This lack of individual level correlation for the aducan trials has developed quite a following online and has been advanced as a weakness in the argument for aducan efficacy.

Yeah! The idea of using group level data has strong support from the below statement. Trying to take individual level data as was attempted ~20 posts ago is clearly convoluted with many other covariates obscuring a clear view of the data. For example, I read that the APOE- had lower cognitive benefit and they also had more amyloid clearance. So the patients to the left of the CDR-sb vs SUVR figures would tend to pull up the regression line (this would have worked against the 302 line). Somewhere else in these documents they actually give the composition of the covariates by SUVR location. It is very enlightening. The SUVR region between 0.20 and 0.30 is particularly favorable for aducan, while the ends >0.40 and <0.20 are loaded with higher risk patients. My parabolic (U curve) above actually might not have been far off. Without correcting for the differences, you would be left with the U curve ( CDR-sb vs. SUVR) which did not seem entirely sensible to me at the time.

One qualm that I have with the group level figure (Figure 44) is that now we have correlation of up to 99%. Problem here is that there might be a few to few points for these group correlation plots. The linear correlation of 2 points is by definition 100%. The figure uses 3 points. I like my plot that tries to include more of the anti-amyloids in one figure.



" Is “lack of correlation between week 78 CDR-SB changes and SUVR week 78 changes for high dose” presented at the PCNS-AC meeting valid? FDA included Figure 43 shown below in slide 20 of the PCNS-AC presentation8. Based on these data, it was concluded that: “lack of correlation between week 78 CDR-SB changes and SUVR week 78 changes for high dose”.

This univariate analysis between CDR-SB and SUVR within one dose level is incorrect because the patients with different levels of SUVR within the same dose group had imbalanced distribution of multiple prespecified covariates for CDR-SB (Table 29). This is the major reason for the conclusion above, which is clearly not supported by the group level analysis (Figure 44). When individual data from one dose group were used to quantify the relationship between two endpoints, it is extremely difficult to correctly adjust
the imbalance of multiple confounders across different individuals based on multivariate analyses due to potential nonlinear relationship and complex interactions. This challenge is well recognized in the field of pharmacometrics and FDA’s 2003 exposure-response guidance (https://www.fda.gov/media/71277/download) clearly explains these challenges and recommends alternative methods."

In order to correctly assess the relationship between two endpoints based on a univariate analysis, the key assumption or condition is that the only difference among the data points is the SUVR value and all other potential confounders or covariates for CDR-SB are
balanced across the data points. Therefore, the analysis should be conducted at the randomized group level (Figure 44). Even though these patients with both CDR-SB and SUVR observations represent a subset of the originally randomized groups, rigorous analyses confirmed a lack of systematic difference between these subgroups and the originally randomized overall groups (Section 4.2.2.4). The Figure 44 shows a clear positive relationship between CDR-SB change from baseline and SUVR change from baseline at week 78 for study 302 at the group least-square mean (LSM) level based on the three randomized arms with a correlation coefficient of 0.99. This strong positive relationship is supported by the results from study 103 with a correlation coefficient of 0.95. The lack of the expected positive relationship in study 301 is driven by the outlier CDR-SB observation for the high dose group because 80% of the patients belong to the Pre-PV4 subgroup (see Section 4.3 for additional evidence to support Pre-PV4 subgroup in high dose of study 301 as an outlier subgroup). The strong group-level relationship between SUVR and CDR-SB is further supported by a meta-analysis combining data from multiple compounds (Section 3.3.1.4.2)."

Another good one is talking about the order of analysis of the secondaries which has been previously noted. Only question here is they mentioned the tertiary (i.e., NPI) yet the NPI is not actually given in the flow chart.
Last edited by J11 on Mon Jul 05, 2021 7:48 pm, edited 5 times in total.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

Corr 1a.jpg
You do not have the required permissions to view the files attached to this post.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

The above figure shows the lack of correlation. Yet, the points highlighted in pink show points that have leverage which might be related to covariates such as APOE- genotype. We saw earlier how only even single points could substantially influence the slope of the regression lines. The red line is especially "tippy" because it is so low in the graph; brought down by many patients in the 0.20-0.40 range that had strong showings on CDR-sb (i.e., negative values on the y coordinate).

It would certainly be interesting if a "true" regression had been provided that showed the best estimate from the models. Perhaps the points could have been color coated for total covariate risk with the best fit risk adjusted regression. I would be particularly interested in seeing what the slope of this best fit line fit be.
User avatar
floramaria
Support Team
Support Team
Posts: 1423
Joined: Tue Jul 04, 2017 11:22 am
Location: Northern New Mexico

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by floramaria »

J11 wrote: The placebo leap seems to have been just a big typo.
DARN! It did sound too good to be true….
Functional Medicine Certified Health Coach
IFM/ Bredesen Training in Reversing Cognitive Decline (March 2017)
ReCODE 2.0 Health Coach with Apollo Health
Post Reply