Domperidone, a drug that is known to raise prolactin levels, did not significantly slow normal disability worsening among patients with secondary progressive MS, a new futility study found.
Luanne Metz, MD, professor of clinical neurosciences, University of Calgary, spoke with MD /alert about the unique trial design and the future of domperidone in the treatment of MS.
What was the idea of the study?
Metz: There's two ideas. One is the trial design itself and the other is around the drug being used. So, as far as this drug itself, this was based on work that me and my colleagues did in the basic science lab, Looking at prolactin, a hormone, and its role in regeneration and repair of myelin in inflammatory demyelination conditions.
There was work done by a team of scientists in Calgary and then some work done in other places that showed that the hormone prolactin could enhance repair. We first explored using prolactin that can be made. It's not an available therapy, but it can be made and then would have to be given as an injectable drug. But the cost and the processes around that would be very expensive. Whereas we know that there are drugs that, as a side effect, are known to increase prolactin levels in most people. We looked to get some of those drugs. Tthe one that is well known and is generally very safe is this drug called domperidone.
What is domperidone used for?
Metz: It’s a drug often used to improve the motility of the stomach in people, particularly for example, people that have diabetes, where one of their problems is poor gastric motility. So, they might be given this drug. And it’s also used in women that are having difficulty with breastfeeding, very specifically, to increase their prolactin levels to enhance their ability to create enough milk and breastfeed. And it is generally very safe. It can cause problems with the QT interval, which is one of the conduction measures that we can measure on electrocardiogram. So we know that we have to be sure that we're not using too high up and that we're only including people into a trial where their risk of having a prolonged QT interval is minimized as much as possible. We also were able to find that work had been done to show that the prolactin molecule gets into the brain, as long as you have it in your blood. And so that work had already been done, so we knew that by creating prolactin it would get into the brain and could potentially enhance repair.
So all of that background work had been done and so what then we wanted to do was to use this drug domperidone, which is very inexpensive, to get to increase prolactin levels and therefore to see if using the drug would then increase repair in multiple sclerosis. And so the initial parts of the project would be, to not necessarily see whether we’re getting the repair, because that would require further studies, expensive studies, with a lot more monitoring of repair itself. But can we give this to people with multiple sclerosis. Is it tolerable to them? And does it increase their prolactin levels? So, that was the initial question.
How was the study set up?
Metz: This trial was set up to look both at the tolerance of the treatment in people with MS and also then to use a new proposed design for progressive trials. The usual way to do a clinical trial is a parallel group trial where you randomize people to the drug or placebo and you require generally hundreds of people in the trial and millions of dollars to run the trial. But there are ways to find out with a smaller number of people and a much lower expense whether a particular drug is worth pursuing in larger trials. That design is very commonly used in cancer studies. But this is, to our knowledge, the first time that we’ve used the design in multiple sclerosis studies.
The idea is to find out if it’s futile to do more trials with this treatment, so it’s called a futility trial design. You have an outcome that is reliable and that isn’t all over the map and random as to what you will see, and you can select participants for the trial that have a reasonable chance of reaching that outcome. And it’s pretty predictable what proportion of participants will reach the outcome.
So my colleague, Dr. Koch is the one that proposed this design and he worked with industry who have done previous trials in both secondary progressive and primary progressive multiple sclerosis before to look at the proportion of people in their studies focusing on the placebo group and then looked at our clinic database, which is pretty large, and used the measure that we propose here, which is the 25-foot timed walk, and found that if we select people that have a walking speed of 9 seconds or longer, we could predict based on those previous trials and our clinical database that about 40% of those people are going to worsen on this measure.
So, it was pretty reliable that if you take this population of people that have secondary progressive MS that have a walking speed that’s already reasonably slow, for this measure, that 40% will worsen, and then use statistics to determine what proportion of people is an improvement.
What did you find out?
Metz: The previous literature suggested that it’s pretty reliable that people will get an increase in their prolactin levels, but it’s always important to find out if that really happens in the population in your trial. And the reason being is that this trial had older people than would ever be having babies for example, and more men as well. Just analysis added to this afterward was that not all people did get an increase in their prolactin levels and there’s still some room for arguments that maybe there’s some value in the treatment if the trial would have been done looking at people that got an actual increase in their prolactin level, perhaps the result would have been different because not everyone got that increase.
The problem at the beginning of the trial is that you don’t know what response people are going to get. So you would need a larger sample and only include the people that got the increase. And that wasn’t what was designed. So we can’t be completely definitive that the treatment should be excluded from further studies.
But at the same time this was being done, I was doing a trial in relapsing-remitting MS, and I can tell you that we didn’t see anything very promising there either. We haven’t finished the write up of that, but, putting it together, we’re not going to be pursuing domperidone or any other thing that looks at use of prolactin. We’ll use this as a signal to say hey, this doesn’t look promising. Let’s look at other therapies and focus on ones that do look promising and put further money and patients into trials where we can give a better signal of them looking promising, and we do have other things in the works that way.
Are your findings more disappointing given the need for more treatments for progressive forms of MS?
Metz: Yeah, in my opinion we don't really have anything that is effective to be honest, because the minimal amount of effect and the fact that the treatments that have now been approved for progressive MS are treatments that otherwise work on the inflammatory component which we know is minimal. We already knew that you could beat inflammation with those drugs. I don't really feel that we truly have any treatment that work on the degenerative part. So, I'm pretty unhappy that this one hasn't looked promising. We will have more soon. We have a paper under review of another therapy that does look promising. By promising I mean, we should consider it for broader trial, not that we have a treatment. At this point we still doubt. However, it’s not totally surprising when we look at neurodegeneration. We don’t really have treatments for neurodegenerative diseases in general.
How important is this new trial design for multiple sclerosis?
Metz: I think the really important thing is that this model can be done even in a single site. You don't need to do a multicenter study to bring in this number of participants. Which means that the many treatments out there that could be promising that there’s some signal that, maybe this will be effective and can be tested very inexpensively and efficiently. So, we’d be looking at a trial that costs in Canada, a in the $300,00 to $500,000 range, which is magnitudes less expensive than jumping right into a big parallel group study, only to find it doesn’t work.
The design also gives you a pretty good appreciation of how the drug is going to be tolerated in people that have the disease which are often going to be different than others that may have used the drug for other reasons. So, this type of trial design has a lot of positives.