Wednesday, September 23, 2015

AAT Completes a Phase-I Trial (No Strong Results)

AAT is an anti-inflammatory / immunomodulatory drug, which the body makes naturally, and which is already FDA approved for people who have a rare condition where they don't make enough of it on their own. Using AAT to treat type-1 diabetes is based on the idea that one of AAT's effects (lowering inflammation, immune modulation, or wound healing) can cure/prevent/treat the disease.

Kamada (makers of one form of AAT) recently published the results of a Phase-I trial in honeymoon type-1 diabetics.  There are several AAT clinical trials going on, and I've reported on different results before:

AAT Completes a Phase-I Trial (But No Strong Results)

This study was open label, with no control group.  A total of 24 people were divided into three groups and each group got a different dose [d1].  Primary end points were safety related, but effectiveness was measured in secondary end points.  The study lasted 37 weeks.  The patients got a total of 18 doses of AAT, spread out over the first 28 weeks of the trial.

The results were:
  • No serious adverse events occurred, and non-serious adverse events were not dose dependant.
  • Average hemoglobin A1c decreased from 8.4% to 7.1% [my rounding]. 
  • C-peptide levels dropped during the study, but the researchers felt that they dropped less than seen in untreated people from other studies.  At the end of the study, 18 subjects (75%) had a peak C-peptide ≥0.2 pmol/mL. 
  • At the end of the study 1/3 of the subjects met the definition of "possible responder" meaning their C-peptide numbers had gone down 7.5% or less.
Opinions of Results

First, no serious adverse events is a good safety result, and having non-serious adverse events be non-dose dependent is also a good safety result.  If there were safety issues, they would likely get stronger as the dose gets higher, since that did not happen, it is likely that these adverse events were not related to the treatment.   So it looks like it passed the safety part of the trial.

Because this study was done on youth between 10 and 18 years old, showing a good safety profile is particularly important.  It is likely to open up future trials to youth.  Recruiting youth speeds the clinical trial process (especially in the honeymoon phase) because type-1 is so often first diagnosed in children.  The safety profile might also lower barriers to "off label" use of this drug [d2]. 

Now an optimist would say "phase-I tests safety and safety is OK, so trial is a success". However, they did measure effectiveness also, and so I do think it is important to look at the effectiveness numbers that we have.  I have noticed that if the phase-I is not effective, it makes it less likely that the phase-II will be.  So with that in mind:

The most important thing to remember is that this study did not have a control group, and that makes interpreting the results difficult.  The A1c group, the C-peptide group, and the "responder" group all have the same fundamental difficulty: we know what happened to the treated people, but there is no untreated group to directly compare them to.  Therefore the researchers compare these people with untreated people from other studies; and I'm doing the same.   

I think that the drop in A1c levels was meaningless.  This was a honeymoon trial; people were recruited within 6 months of diagnosis.  That first A1c number covers either the early months of type-1 diabetes self treatment or the weeks just before diagnosis (or some of each).  These are both times of high A1c numbers.  Conversely, the second A1c number covers a time when the patient has between 6 and 12 months of experience with type-1, and is therefore better at treating their type-1. So of course the A1c numbers are better.

The researchers compare these A1c numbers to average A1c numbers in adults, and note that the first is above average while the second is below.  However, for the reasons described above, I don't think this is a case where comparing to average is an appropriate thing to do.  

For the C-peptide data, I think the data is hard to interpret, but disappointing.  The researchers summarized their findings this way:
subjects treated with AAT showed less of a decline in C-peptide levels as compared with historical controls.  However, in the absence of a randomized control group, these findings should not be interpreted as showing a beneficial effect on beta cell preservation
and glycemic control.
For me "less of a decline" is the very smallest sign of success a researcher can talk about.  I'd put these results below results from TOL-3021 and Alefacept  (both which held C-peptides constant), and maybe even Teplizumab (which held C-peptides constant in some people).

Each person has to decide for themselves which results excite them, and which don't.  For my part, in honeymoon trials, results where the C-peptide numbers go up (after a year) excite me.  Results where they stay constant are unexciting.  And, results where they go down are disappointing.  Here the average went down.

People in this study will be offered a chance to participate in a long term (3 year) follow up study. Those who still generate C-peptides [d3] will be kept on their current dose of AAT.  Those who don't (or who don't want to continue, will not get AAT, but will be followed as a comparison group.

Overall Status of AAT As A Cure For Type-1

AAT is unusual, in that there are two different types of information on it.  There are results of previous clinical trials.  But there are also anecdotal reports from the popular press.  The anecdotal reports are far more positive.   Unfortunately, these phase-I reports are clearly in line with the previous clinical trial results.  They are lackluster and will not lead to a cure without a large improvement in effectiveness.

The good news is that there are two phase-II clinical trials of AAT already underway, so there is nothing to do but wait.  Especially since one of those phase-II trials is being done by the same company as did this trial: they have incorporated what was learned from the older trial into the design of the newer trial.  If there are good results from the phase-II trials, then AAT is in good shape.   Mediocre results in phase-I mean nothing if the phase-II results are good.  If those results are as good as the anecdotal results, then AAT is in great shape.  But if those phase-II results are similar to these phase-I results, then that would be bad news for AAT.

Opinionated Discussion of Footnote 30

Because this study did not have a control group of it's own, the researchers used control group data from other studies.  In particular, some control group data from the phase-II trial of DiaPep 277, and this reference was provided in footnote 30.   In my opinion, using this data raises a red flag.  One of DiaPep's phase-III trials was canceled because of serious scientific misconduct [d4].  The article reporting the results of the other phase-III trial was retracted because the same misconduct was found in that study as well. To my knowledge, no misconduct has been found in the phase-II trial, but none was particularly looked for, either.  And there are authors in common between the retracted phase-III paper, and the cited phase-II paper [d5]. Finally, the manipulation that was done in the phase-III paper changed both the treated group and the control group [d6].  So if that same manipulation was done to the phase-II trial, (that is a big "if" of course) then it would effect the exact data that these AAT researchers are using.

For all these reasons, I think using data from the DiaPep 277 phase-II study is a mistake.  The AAT researchers were using control group data.  Lots of studies have control groups.  Why use potentially tainted data, when untainted data is available?


I want to specifically thank the author of this paper who sent me a copy of the full paper, so I could comment on all of the paper, and not just the abstract.

More Discussion

[d1] The researchers listed this trial as "Phase-I/II".  Generally I consider phase-I to be less than 20 people, and phase-II to be larger than that, but I also expect a Phase-II to have a control group, which this did not.  So based on the whole trial, I consider this a phase-I trial.

[d2] In the USA once a drug or device is approved, a doctor can prescribe it in situations different than it was originally approved for.  As an example, a doctor may prescribe it for a different disease, at a different dose, or for a different type of person than it has been approved for.  In the world of type-1 treatments, drugs or devices that have been approved for use on adults are often prescribed for children.  This is a classic "off label" use.  Off label use is based on the professional opinion of a doctor, and consent of the patient.  Therefore safety data (such as from this study) can make doctors more willing to prescribe "off label", and patients more interested in trying it.

[d3] Specifically, patients who have C-peptide levels of ≥ 0.2nmol/L will be able to continue their AAT regimen.  Note also that I refer to the other group as a "comparison" group and not a "control" group.  These two groups start off different, so I don't consider one to be a good control for the other.

[d4] Hyperion used the term "serious misconduct" in describing the situation, and Globe News used the term "fraud".

[d5] I want to stress that there were authors who only worked on the phase-II paper, and there were authors who only worked on the phase-III paper.  And, there were authors who worked on both papers.  To my knowledge, there has never been a public naming of who was involved in the misconduct, so there is no way to know if one of the authors was involved or if it was someone else in the organization, or even how many people were involved.

[d6] A more complete discussion of the misconduct alleged in this case is in my previous blog posting:


Press Release:
Clinical Trial Record:
Follow On Study Clinical Trial Record

Joshua Levy 
publicjoshualevy at gmail dot com 
All the views expressed here are those of Joshua Levy, and nothing here is official JDRF or JDCA news, views, policies or opinions. My daughter has type-1 diabetes and participates in clinical trials, which might be discussed here. My blog contains a more complete non-conflict of interest statement. Thanks to everyone who helps with the blog.


David Ronaldo said...

it looks great information shared about diabetes ..
urgent care centers brooklyn

Mark Lee said...
This comment has been removed by a blog administrator.