Tuesday, June 19, 2007

False negatives might be a big problem

I have just changed my view on trials after another email by Willme:
With [Correction Tom: Assuming] several possible (neurotransmitter) causes of stuttering identified, the test results for any one drug applied to a seemingly homogeneous group of stutterers will be diluted. The majority may not respond at all. Morever, even if heterogenity is recognized and a subgroup of stutterers is identified (say responding to BZ's) any one particular BZ, like ativan, may not work for all of them. So the test results could be even further diluted. What this means is that the likelihood of "false negative" test results may be more of a threat than "false positives." With false positives, you take the drug and if it is not efficacious you stop taking it. False negatives imply that a promising drug for a subpopulation of stutterers may fall by the wayside forever.
I hope that there are active research programs out there involving DNA studies to identify the cause(s) of stuttering. I'm quite convinced that genetic markers will need to be found for any substantial further progress.

He is completely right in that a medication that works for a small subgroup will be rejected in a standard random control trial. And everyone thinks it's not working. Different stutterers also react differently to real drugs. Per Alm wrote about this in his PhD thesis. So we need to be able to discriminate between different subgroups, and that's very very difficult because it means we understand stuttering!!! Our best hope is genetics, but we also know that not all stuttering is genetics. So this is not an easy avenue either.

I could even turn the argument around, and say that because no trial has been very successful so far, stuttering must have subtypes.

And then I could argue that drugs that do work, must be acting on secondary symptons and not on primary causes. And this might be the case with Pagaclone.

Check out this NY times story on "On the Horizon, Personalized Depression Drugs." for inspiration on what could happen for stuttering.

27 comments:

Anonymous said...

Tom Weidig's argument that pagaclone might be acting on secondary symptonms and not on primary causes is well taken. If pagaclone relieves anxiety and it is the anxiety that causes secondary stuttering such as blocking, then the blocking will be reduced but not the primary stuttering of repetitions and vowel elongations. This is why aggregate measures of success based on the fraction of syllables on which disfluency took place are not adequate. We need to know what kind of stuttering was affected and we need to know the incidence and nature of the improvements within the population. Consider for example a sample of two stutters, one of which had 100 disfluencies upon reading a passage while the other experienced 200. A 30% reduction in disfluency could mean that the person with 100 disfluencies went to 0 disfluencies while the 200 stayed at 200; or it could mean that the 100 was reduced to 66 and the 200 reduced to 134. And of course there are an infinite number of other combinations.

Anonymous said...

"The effect of cyclopyrrolones on GABAA receptor function is different from that of benzodiazepines"
A. Concas, M. Serra, G. Santoro, E. Maciocco, T. Cuccheddu and G. Biggio
Department of Experimental Biology, Chair of Pharmacology, University of Cagliari, Via Palabanda, 12, I-09123 Cagliari, Italy, 1994
http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=Retrieve&db=PubMed&list_uids=7824046&dopt=Abstract


GABA-benzodiazepine receptor complex: focus on receptor subtypes and cyclopyrrolone drugs.
Snyder SH.
Isr J Med Sci. 1987 Jan-Feb;23(1-2):145-52.
http://www.ncbi.nlm.nih.gov/sites/entrez?db=PubMed&cmd=retrieve&dopt=abstract&list_uids=2437071


Cyclopyrrolonderivate
http://www.drogen-forum.com/wiki/index.php/Cyclopyrrolone

Ora said...

willme -

You wrote: "If pagaclone relieves anxiety and it is the anxiety that causes secondary stuttering such as blocking"

I have generally assumed that it has been well established that anti-anxiety drugs do not have any particular benefit against stuttering. Otherwise they would be a standard adjunct to speech therapy, but they're not, based in my experience with multiple speech therapists over the years.

The drugs that I have read about that have shown some promise for stuttering treatment have been some of the recent anti-psychotics, such as risperdal and olanzapine. If anti-anxiety drugs or antidepressants had shown promise for stuttering treatment, I would have expected that we'd be reading more about them, but that's not what I've seen in the literature (though I must admit that while I've reviewed articles from time to time, I don't have a broad knowledge of the full scope of pharmacological research that's been done).

Beyond your own experience, do you know of any evidence that anti-anxiety drugs are effective for stuttering, or for a significant sub-population of stutters?

Tom Weidig said...

I want to clarify that I do not talk about anxiety by itself, rather about the effect that a higher state of emotionality has on many (fine motor control) brain functions in any person.

My view is that stutterers have a fundamental weakness, and that a higher state of emotionality (and/or a higher workload) has a DISPROPORTIONATE impact on fluency.

So I could argue that if Pagaclone is somewhat successful, it might well allow people who stutter to speak as if they are NOT in a state of higher emotionality. Pagaclone activates inhibitors.

Anonymous said...

In answer to ORA: You are right--previous studies have not backed up the effectiveness of anti-anxiety drugs. So does that mean that I am one in a million? Will the other two people step forward, please?
Or I am even considering the possibility that I am a hysterical (in the Freudian sense) stutterer. I don't really stutter but I show all of the symptoms of stuttering. Hysteria is now called conversion disorder--basically, one converts a psychological distress into a physical symptom. If this is so, then taking ativan relieves the psychological distress (I would characterize myself as having generalized anxiety disorder) and hence the physical manifestations of stuttering. Thus, I could join the pantheon of people who have experienced hysterical pregnancies, hysterical polio, etc.

Ora said...

This issue concerning false negatives must be an issue in trials of all sorts of drugs and all sorts of conditions. There's nothing special about stuttering. I wonder if there's any customary way to control for it in medical research? I'd guess that that researchers are looking for subgroups all the time.

I'd imagine this issue would also arise in analysis of adverse effects, where the side effects of a drug are unacceptably high - on average across the entire population - and so the researchers do their best to identify a subgroup which can be identified and excluded in order to improve the results for the remaining population.

Presumably this done by analyzing and reanalyzing the data, and trying to identify correlations among variables. Is there any other way, I wonder.

Anonymous said...

willme/tom: isnt atvian very similar to xanax? did you ever try xanax, and if so, what was the result? I recall being prescribed a generic version of atvian years ago before I went to law school and it was very successful, but for some reason or other the connnection did not stick for me - after reading your guest blog I am interested in getting a prescription and seeing the results. I would like to note that it seems that my reaction to the anti-anxiety drugs is similar to Willme so perhaps we have similar gentic/emotional makeup with regard to stuttering.
Thanks. JK

Anonymous said...

To Anonymous: Yes, Xanax to my knowledge is a BZ, and, no I've never tried it. I want to emphasize that I use ativan SPORADICALLY, for "important events" for two reasons: First I don't want to become addicted and, second, one builds a tolerance for the drug. It is possible that you may have been taking the BZ on a regular basis and, hence, built a tolerance for it. I use ativan sporadically because I don't want to kill the goose that lays the golden eggs.
When I undertook this guest blogger role, one of my hopes was that blog readers would respond as "yes I took it and it worked," or "yes I took it and it didn't work." In this way we could get a handle on the percentage of stutterers who might be positively influenced by the use of BZ's. This percentage would represent an upper bound since some people would experience positive results because of the placebo effect.
DNA studies definitely are needed to identify gene markers for stutterers who respond to different treatments. What may be special about stutterers is that they may be divided into many subgroups. If these subgroups are too small (percentage-wise), then it may be very difficult (or impossible) to tease out of the data, the effects of a drug across the ensemble of stutterers treated as a homogeneous group.

Anonymous said...

willme, this is JK again. to clarify, when I said "the connection did not stick for me", what I meant was although I recall now that I had very good success taking atvian, I never personally made the connection that the medication worked (since no one ever prescribes anti-anxiety meds for stuttering) and never got it prescribed again after the one bottle ran out after a few years (I took it maybe once every month when I had a presentation, interview, etc.). After your blog I will try to get another prescription and see what happens.

I do have some xanax that I take very very rarely, but on the few times I have taken it it seems to make me very sleepy and feel like I am "stoned". I do not recall feeling that way on atvian. Do you know if xanax is stronger or more of a sedative thatn atvian?

Thanks for the great blog, I found this recently. I am a 37 year old attorney who has always been a mild stutterer.

Anonymous said...

To Anonymous (JK): Different BZ's may have different effects on you, but they all have sedation in common. Also, the effect of Xanax vs. ativan may depend upon the relative dosages you take. When I take ativan, I usually take it when the anxiety level is quite high, so the high neuronal activity level effectively cancels or mitigates the sedative effect. I may, however, become sleepy after going through the anxiety producing event. The advantage of pagaclone is that the drug company CLAIMS it has no sedation effect, it is not habit forming, and, presumably, you do not build a tolerance for it (requiring higher dosages). I remain skeptical about these claims since I take Lunesta (to improve the quality of sleep) and this drug is in the same family as pagaclone (i.e., cyclopyrrolones). I find that I do build a tolerance for Lunesta if I take it, for example, 3 days in a row. Therefore, I never take it more than twice a week.

Ora said...

I don't understand the Blogger format. I see 10 responses to the original post, all all showing times but no dates.

Time is interesting, but certainly date is more important. What's going on here? Is there a Blogger option I need to set?

Anonymous said...

Thanks for article!

Anonymous said...

Thanks for interesting article.

Anonymous said...

Glad to read articles like this. Thanks to author!

Anonymous said...

Excellent website. Good work. Very useful. I will bookmark!

Anonymous said...

iqSz1P Your blog is great. Articles is interesting!

Anonymous said...

zOkobs Wonderful blog.

Anonymous said...

lfHwQc Magnific!

Anonymous said...

Nice Article.

Anonymous said...

Nice Article.

Anonymous said...

Good job!

Anonymous said...

Please write anything else!

Anonymous said...

Thanks to author.

Anonymous said...

Nice Article.

Anonymous said...

Please write anything else!

Anonymous said...

DwjjD7 actually, that's brilliant. Thank you. I'm going to pass that on to a couple of people.

Anonymous said...

Wonderful blog.