Sunday, January 24, 2010

Pagoclone: puzzling report.

A reader has sent me a first-person report on his Pagoclone trial experience. I am very puzzled by his doctor's comments. Here is his report:
I received an invitation to be part of the next Pagoclone Phase IIb trial last March and started on April 15th. As you know the Phase IIb was to be a double blind study whereby they would also be trying several different doses, .15mg, .30mg, and .60mg all twice a day. I believe that I was on Pagoclone for the majority of the study and came off of Pagoclone and went on the Placebo at least at the very end of the study to properly transition to the Open Label Phase that I am currently on. I truly believe that the medication works. I do believe that I experienced a modest improvement in my fluency, somewhere around maybe 40% better fluency, and then after the dosage change, back to the placebo, I experienced a pretty dramatic drop in fluency to almost pre-study levels. I have not had any problems with side effects, at least none that I am aware of, blood work might show something, etc. However, over the course of the study, I have gained five pounds, but I believe that can be attributed to my horrible diet over the last year....

At my visit on Tuesday with the clinic, the Doctor asked me to describe how the medication works. I explained to him that I feel it reduces the stress, tension and pressure in the articulators (lips, mouth and jaw) which allows for a free flow of speech, or a much more free flow of speech. There is still room for improvement, even on the highest dose, which I am on now, .60mg. I am not aware why they are stopping at .60mg. Why not double it? or Triple it?
The doctor that is over the clinic, that I meet with every visit, had some pretty candid information for me. He said that the results that he has seen over the last year indicate to him that the medication works in about 2/3rds of the participants and of those 2/3rds that it works on, it has the most dramatic effect on the highest dose (similar to my experience). He also stated that most people do not have much, if any, reaction to the drug and he called it a fairly benign drug.
I asked him when we would know the results from Phase IIb and he said that he's surprised that they have not reported on it yet. I also asked him to peer into his crystal ball and predict for me what the final outcome will be relative to the future of Pagoclone. He said that he believes that it has a better than 50/50 chance of ultimately being approved. He said that his typical response for the other drugs that they test for is, "it's a 50/50 chance". He stated that because of its benign nature, the efficacy of the drug and the fact that it meets an unmet need, concludes him to think that it will have a greater chance of being approved than not being approved.
I am very puzzled by his doctor's comments, if correctly reported and understood. How can the doctor know that the medication works for 2/3rd of the participants if he doesn't know who gets what? After all, it's a double blind trial. And how does he know that it's most effective on the highest dose? Either the participant has misunderstood, the doctor has mis-thought, or the blind has been broken somehow (which I do not believe.)

12 comments:

Gerald Maguire, M.D. said...

Hi, Tom. This is your friend, avid reader and pagoclone investigator, Jerry Maguire. The blind has definitely not been broken. The physician may just be providing his/her own speculative opinion. In addition, the investigator may have participated in the earlier pagoclone trial and may be commenting on the experiences with his/her patients from the open-label phases of that study which are still ongoing. We must be careful not to draw any definitive conclusions about the Phase IIb double-blind study until the trial data have been fully analyzed (the final subjects just completed just a few weeks ago). As the rest of you, I anxiously await the results!

Anonymous said...

Hi Tom,

I have a couple of questions:

1) the email in your blog says: "I believe that I was on Pagoclone for the majority of the study".
This means that even though he may have been taking the placebo, he *believes* he was taking pagoclone, and that would almost certainly be a factor in his 40% improvement. So, the placebo effect has not been eliminated in this case. Apart from it being double-blind, what steps are taken in this trial to minimize the placebo effect?

2) Minimal side-effects are reported. In the past we've heard of weight-gain in other trialled drugs. But we are dealing with a drug that affects the brain and its chemistry. Would it not be wise to conduct MRI scans of the patients before and after the trial? For instance, I know that the use of SSRI drugs result in increased brain cells in certain parts of the brain; I wonder what effects long-term use of pagoclone would have on brain structure. We're dealing with human subjects, not guinea pigs or rats.

Ora said...

Well, one way to guess the effectiveness would work like this...

Presumably the percentage of the trial participants on placebo and pagaclone is known. Imagine 50% are on pagaclone and the rest are on placebo. If 1/3 of the total participants in the trial are improved during the trial, then we can conclude that those are the ones taking the drug. Ergo, 2/3 of the pagaclone participants are improved by the drug.

Tom - before you jump all over my logic... I know that you have to adjust for the fact that a certain proportion of those in the placebo group improve as well. (And in fact, a certain proportion of any untreated group would improve.} Any this assumes that the percentage of a placebo group or an untreated group are known (or can be reasonably estimated). And you have to do appropriate statistical tests to determine whether the results are significant, etc.

But in theory it should be possible to do this type of analysis and arrive at a figure like the doctor's 2/3 estimate.

Tom, do you agree with this? I'd be interested in your reaction.

Tom Weidig said...

Hi Jerry,

I agree with you. I wrote that the blind is very unlikely to have been broken!

I still wanted to post his report to show the discrepancies in reporting. And to show that one shouldn't believe everything that a doctor says. Or maybe the reader just misunderstood.

I am just reporting realities.

best wishes,
tom

Tom Weidig said...

To Anonym,

>>> 1) the email in your blog says: "I believe that I was on Pagoclone for the majority of the study".
This means that even though he may have been taking the placebo, he *believes* he was taking pagoclone, and that would almost certainly be a factor in his 40% improvement. So, the placebo effect has not been eliminated in this case. Apart from it being double-blind, what steps are taken in this trial to minimize the placebo effect?

I agree with you. That could be an issue. We know that Pagoclone is well-tolerated, so the side effects do not reveal whether one is on P or not. But if Pagoclone makes you feel differently, you know that you are on P. so on the short-term there might be a stronger placebo effect in the Pagoclone group. However, in the long term this might be less relevant.

Jerry once gave a counter-argument, but i can't remember. Many different dosis have to tease this out.


>> 2) Minimal side-effects are reported. In the past we've heard of weight-gain in other trialled drugs. But we are dealing with a drug that affects the brain and its chemistry. Would it not be wise to conduct MRI scans of the patients before and after the trial? For instance, I know that the use of SSRI drugs result in increased brain cells in certain parts of the brain; I wonder what effects long-term use of pagoclone would have on brain structure. We're dealing with human subjects, not guinea pigs or rats.

There might well be long-term side effects that we are not aware off. so really pagoclone needs to give me a good effect in order for me to accept such costs of uncertainty.

Tom Weidig said...

Hi Ora,

>> Presumably the percentage of the trial participants on placebo and pagaclone is known. Imagine 50% are on pagaclone and the rest are on placebo. If 1/3 of the total participants in the trial are improved during the trial, then we can conclude that those are the ones taking the drug. Ergo, 2/3 of the pagaclone participants are improved by the drug.

Why? imagine Pagoclone is also an empty pill. maybe the 1/3 are just more receptive to placebo because a lot of their stuttering is psychological?

Best wishes,
Tom

Ora said...

Tom - You wrote "Why? imagine Pagoclone is also an empty pill. maybe the 1/3 are just more receptive to placebo because a lot of their stuttering is psychological?"

But you can adjust for the placebo effect. That was the point of my paragraph beginning "Tom - before you jump all over my logic..."

Here's the basic idea.

First you'd need to quantify the placebo effect, but it can certainly be done, at least in principle. Imagine that you do a test: you give a bunch of people a pill and say we're testing a pill against stuttering. In reality the pill is a placebo. But because of the placebo effect, some percentage improve. Imagine it's 20%.

Then you do the pagoclone trial, and based on the test design, you know that the test has been designed to give pagoclone to a certain proportion and a placebo to the rest. Say it's 50/50. Now you observe that 50% of the whole population improve.

Imagine that the trial size is 1000 (500 pagoclone, 500 placebo). You observe that 500 people have improved. Of these, we assume that 20% of these, or 200 participants, are people who have taken the placebo, so that the remaining 300 are are people who have taken pagoclone. That means that of the 500 people in the pagoclone group, 300 have improved, which is 60% of the pagoclone group.

Obviously this is a simplified example, which doesn't account for dosing differences, degree of improvement, issues with test design, or various other issues. And there are the normal issues of statistical reliability and level of confidence. But in theory it should be possible, even without unblinding the study.

This relies on knowing, or being able to estimate, the effect of a placebo on stuttering. I don't know if anyone has ever done a test of a placebo for stuttering. But as a proxy for quantifying the placebo effect, one might use an educated guess. Obviously this would be far from perfect, but it's also not totally valueless.

Anonymous said...

Hi, this is the anonymous participant whose report we are discussing, logging in my opinion to the great debate over, it would seem, our favorite topic of the last couple of years.
First of all, I agree with Dr. Maguire that the investigator is probably basing his opinion of 2/3rds by his involvement in the previous phases of this trial and the open label participants. Furthermore, he talks to the participants about their experiences with the drug/placebo at every visit. So based on his medical training and experience in drug trials, he comes up with his 2/3rds assessment.
Furthermore, I agree with Tom, Dr. Maguire et. al., that the blind has not been broken. Of course, I have no way to prove this, but in my experience with this medical office, EVERYTHING has been on the up and up. They seem to be very meticulous with every minute detail, some of which I think is a little over the top but with them it appears to be very serious and all business.
Relative to the drug being most effective on the highest dose, I think this as well speaks to the Doctor's experience with the participants in the open label phase on the highest dose. I remember discussing with him that most participants on the open label phase state that the drug works but there is room for improvement and most, if not all, request the highest dose currently available. In fact, I don't recall him mentioning that ANY of the open label folks were satisfied with the .15, or .30 mg dose. Some may be, but we did not discuss it.
I hope this helps perhaps clear up any areas of confusion from my report. Please feel free to ask questions, as I will be checking in with Tom's blog regularly.
Thanks as always to Tom for his efforts.

Tom Weidig said...

Hi Ora,

>> assume that 20% of these, or 200 participants, are people who have taken the placebo

The placebo effect is typically across all people, and only modulates the outcomes? So placebo does not occur in a subgroup as far as I am aware?

best wishes,
tom

Gustaf said...

I know very little about pharmacology, but I know there are other already approved drugs which also act on the GABA levels. I'm particularily interested in PharmaGABA. And there are natural herbs such as Kava root which also are said to have an effect on GABA. Would I be wrong to try any of these for the same effect?

Unknown said...

hello all
This is amal from Dubai
Ive tried many & many pills in order 2 reduce my stuttering
Many have worked prtty good for moi but I could NOT really live with the horrible side effects
I hope this pagoclone truly works mmm
I wanna know when it's gonna be released to the market
Thanxxxx

Anonymous said...

Stay away from Pagaclone. Now I have Pre-Diabetes/Diabetes. It's boarder line right now between pre-diabetes & diabetes. They denie any connection, but there is NO family history of diabetes.

Also, Pagaclone did little, if anything for my stuttering.

They don't seem to care, except they want their pills back. Nothing more.