The First International Conference On Heroin Maintenance: Research and Evaluation, Wim van den Brink

But before I start my presentation, I would like to take away some common misunderstandings. One is that, I think, in the United States, as in many other places, people think that Holland is a very liberal country, especially with regard to drugs policy; and probably it is. But at the same time, many people believe that because it is such a liberal country, there will be a lot of hard drug addicts and there will be a lot of cannabis users. However, this is not true. In The Netherlands we have, in a population of approximately 16 million people, approximately 25,000 heroin addicts, which is actually rather low. The estimated number of cannabis users in the Netherlands, ranges between 400,000 and 600,000 people, which is also a rather low number.

There is something that is actually true. Most people think that if you have a more liberal situation the number of drug deaths is going down. That is definitely true. In the last few years, the number of overdose deaths in heroin went down to a low of under 100 for the whole country. So I think that is true.

Finally, one other misunderstanding that might have occurred in the morning. Ambros Uchtenhagen said that American scientists were [left out] of the new developments. At least we were able to have our plans [reviewed] by Dr. O'Brien and Dr. Kreek, and they did so with great enthusiasm. That is not to say that they would like to do the project themselves, but they were very willing and very helpful in reviewing our plans. So I think there is some very honest interest in the United States in the scientific experiment that we are planning to execute. So what I would like to share with you this morning is actually the plans that we have developed in the Netherlands, and to see whether this is a reasonable and way to pursue on this trajectory.

What you see here is that the title is slightly changed. The title of the Dutch study is no longer "The Prescription of Heroin to Chronic Treatment Refractory Heroin Addicts", but "The Medical Co-prescription of Heroin to Chronic Treatment Refractory Methadone Patients". And it will be a randomized trial.

What is important here is that we like to emphasis that it is not our intention that the group of patients that you actually try to involve in this study, the methadone patients, should actually give up the methadone treatment, but rather that they can add the (co-) prescription of heroin. I will come back to that a little later.

The issues that I would like to cover in this short talk are: what is the current situation in the Netherlands and in that sense what is the rationale for doing this kind of study? I'm not going through the politics of that but just looking what the epidemiology in the Netherlands says and what the need is for additional treatment options. Then I would like to briefly present the analysis of the Dutch Health Council, and finally I will present a proposal for a randomized trial of heroin co-prescription.

What is the current situation in the Netherlands? I think it's been summarized quite well in this overhead. What you see here is that in the Netherlands, we have approximately 24,000 heroin addicts. Of these 24,000, approximately 70 percent is currently in treatment, which means that they have treatment contacts in the last year, and 29 percent is not in treatment, which is still 7,000 people.

From the ones that are in treatment, the majority, almost three quarters, are on methadone maintenance treatment. The remaining quarter is in some form of drug free treatment, either in a therapeutic setting or a residential treatment or in some kind of ambulatory treatment. So, if you multiply 0.7 and 0.74, then you can see that in the Netherlands, approximately 50 percent of all heroin addicts are involved in methadone maintenance treatment.

Now, the question is, how well are these people doing? Almost eight years ago, we had the last assessment of the status of these patients and what we found is that about a third of these patients in methadone maintenance were functioning rather well. We call them socially integrated and they were in good physical and psychological health. About 20 percent of the total, which is about half of this group of a third, was actually working. But there was another group of about 40 percent that was described as not integrated, meaning that they didn't have work, that they sometimes were in actually quite miserable psychological circumstances, but they were not heavily criminal. And then finally we had a group of about a quarter that was described as extremely problematic and the problems with these patients were basically described in terms of high criminality.

So this was starting situation when the Dutch Health Council started to evaluate the treatment situation and the possible needs for improvements. For the Dutch heroin project we decided to start working with the not integrated and extremely problematic methadone maintenance patients who are actually treatment-refractory. We decided not to go for the 7000 people who were not in treatment for two reasons.

One, if you look at our study protocol, you will see that both the treatment and the scientific evaluation requires quite a lot of effort from the participants. It is very unlikely that the group that is not willing to get into treatment is actually able to participate in a randomized clinical trial. So, in the rest of my talk, I will also look at the group of nonintegrated and the group of extremely problematic heroin addicts that are currently in methadone maintenance treatment.

Given the data I just presented, the Dutch Health Council decided that existing programs should be continued, especially, the drug free programs, the methadone reduction programs and the methadone maintenance programs.

Another advice was that the liaison between the legal system and the treatment system should be improved. And of course, that's very well known to most of the other countries. We should put more efforts in diversion programs and drug free prison programs.

Furthermore, the Dutch Health Council advised that it should be tested whether higher dosages of methadone in the existing methadone maintenance programs would improve the situation. And I have to say that currently there is a randomized trial going on in the Netherlands where they prescribe higher dosages of methadone, i.e. 80 mg/day or more. The problem in that project really is that it is very difficult to recruit patients for this program. So far, the recruitment has been only developed to 50 percent of the patients that they were actually looking for. So it seems that many of the patients that are now in methadone maintenance treatment on low dosage of methadone are not willing to participate in programs with dosages of 80 milligrams or more.

Finally, the Dutch Health Council decided that there is a need for a controlled study regarding the medical prescription of heroin. Similar to Switzerland, the Dutch heroin project is not some kind of a solution to the heroin problem, it is only one of the treatment options that will be offered to patients and basically only to patients who didn't benefit from other programs.

It's not completely new in the Netherlands that we prescribe opiates with a euphoric effect, opiates that actually provide a kick. Of course, we have a long history of methadone maintenance. But in the early 1980s, we also had a small program on the medical prescription of morphine to a very selected group of chronic opiate addicts with severe psychiatric problems. The program was feasible. There were moderate positive effects, but one has to be very careful when an intervention was not studied so well. However, it was clear that there were histamine reactions, which were confirmed later on by the Swiss experiment.

In the early 1990s, there was a small project, again not very well researched, on the medical prescription of methadone intravenously for a special group of people with a terminal state of AIDS. And again, there were moderately positive effects reported, but again there was no real scientific evaluation.

And finally, in the last two, three years, there is a medical prescription of oral dextramoramide, (Palfiumâ) to a small subgroup of smoking heroin addicts in Amsterdam. This is a group of old smoking Surinamese heroin addicts. There is, again, no scientific evaluation. Again, of course, they report moderate positive effects. I think the only real hard data from this project is that, because they had urine samples, the prescription of oral dextramoramide didn't lead to an increase of cocaine use in these patients; a fear that is always around when you talk about the prescription of heroin. The people who actually get heroin prescribed might spend the money they have left on cocaine? This (thought) is on an unrealistic idea that heroin users have any money left. Most of the time they're short of money. So if you prescribe them heroin, most of the time they're just a little less short of money, but they don't have lots of money to spend. But anyway, it's an important thing that there was no increase in cocaine use in this experiment.

I will not go into any of the details of the effects of the British experience. There was only one randomized clinical trial there. However, that was 20 years ago and I think the data and results from that study don't say very much about the possibilities of heroin prescription in this time of the heroin epidemic.

I should say a few words about the Swiss experiment. This overhead is my summary of the Swiss experiment. I think what has been very important is that it shows that the prescription of heroin is feasible. It also shows that if you want to make a choice between the different opiates, that heroin ranks high and that methadone intravenously and morphine intravenously are less likely to give positive results.

I think it's also very important that there were no unexpected numbers of adverse events or fatal overdoses. It's very important for the Netherlands that there were no major public order problems. We now have permission to start our experiment, but only for the first two treatment units and only for the first three months, because the Parliament wants to look at the potential public order problems after the first three months, and then they will give their final approval to go on with the experiment or not.

I think there was a big problem in Switzerland with the smokable heroin. Actually, Dr. Uchtenhagen told you that the bio-availability of smokable heroin was only 10 or 15 percent and that it was very difficult to get people using the smokable form. Now, that might be acceptable in Switzerland where the big majority of the heroin addicts are actually using heroin intravenously. It is not acceptable in the Netherlands where about 70 percent of the population of heroin addicts are chasing the dragon and only 30 percent are injecting.

A heroin prescription trial where you only have a injectable form is therefore not acceptable in the Netherlands, because it could invite smokers to start injecting in order to join the experiment. Thus we cannot start an experiment in the Netherlands without having an elective smokable form of heroin.

And finally, I think that Dr. Uchtenhagen showed you very clearly that there are strong indications of medical, social and criminal improvements in the population that's being treated with heroin. However, I think that the Swiss study -- and I think the Swiss are very willing to admit that – [the study] also has important limitations.

So far the Swiss study is a naturalistic follow-up study, and to date, we have no data on control groups. Another important issue is the fact that the Swiss experiment is a combined pharmacological and psychosocial intervention. People in the experiment got an offer of both heroin and additional psychosocial interventions. As a consequence, we do not know exactly what the results are being accounted for. Is it the heroin or is it the additional psychosocial intervention?

Now, in the Netherlands, we think, at least, that we have methadone maintenance programs that actually do provide quite a lot of psychosocial services to the patients. In fact, there is a strong under-utilization of these services. And we think that we can do an experiment in which we can keep the amount of psychosocial interventions comparable between the experimental and the control groups. This would show you what additional prescription of heroin would do to these kinds of methadone maintenance programs, i.e. methadone maintenance with an adequate offer of psychological services.

I think there are also some unclarities in the Swiss project with regard to the target population, and Dr. Zeltner said something about it, because they were the first to do that, they were developing the project on the way, and I think we have to be very clear in that. And then, of course, there was no adequate evaluation possible of the smokable heroin. In the Dutch study we try to overcome some of the limitations of the Swiss experiment.

So what would be the requirements of a Dutch study? I think these are the main requirements. I think we should have separate protocols for smokable and injectable routes of administration to see whether it works for smokable and intravenous drug users. We should also have very clear, explicit inclusion and exclusion criteria. We would like to have, and we will try to do a randomized allocation to the experimental and the control condition. We will use standardized and validated assessment instruments. And we would like to have a prespecified definition of what is effective and a prespecified analysis plan.

We would also like to have adequate statistical power, which means that we run two parallel experiments, smokable and injectable, of both 375 patients with reliable measures. Finally, the whole study will be performed under the regulations of good clinical practice.

Before going into the details of the design of the Dutch trial I would like to say a few words on the development of inhalable heroin. We say inhalable in order to prevent the misunderstanding that smokable heroin is like reefers or cigarettes. It's really evaporating the heroin and inhaling the vapors.

What we did in the last few months is that we developed a tablet that contains a combination of 50 percent of heroin, pure heroin, and 50 percent of caffeine. The caffeine we need because otherwise we cannot make tablets. You need a kind of a filling substance, and caffeine is well known to be used in the drug using population together with heroin, so we know that no dangerous pyrolytic products will be produced. Caffeine increases the speed with which the heroin emerges when you start to evaporate it.

We did some extensive pharmaceutical tests on purity, homogeneity, stability and evaporation speed of these tablets. We also did pilots on the acceptability of different variants of application, basically comparing a kind of a very static heating device, and the use of aluminum foil in the way people actually use it on the street.

We have also done a pilot on the pharmacokinetics and the pharmacodynamics of this tablet. We have some data now on the pharmacodynamics showing a very clear effect and a dose response relationship with both objective and subjective effect parameters. We are still waiting for the pharmacokinetic results. Finally, we are working on the feasibility of a [chemically] labeling [the heroin] in order to make a distinction between the legal heroin that we prescribe and the heroin that people buy on the street. We can then actually see whether people are using additional street heroin when they're in the project.

Finally, the project itself. These are the objectives. The primary objective of the study is to evaluate the effectiveness of oral methadone and co-prescribed heroin compared to the standard Dutch treatment of oral methadone in a chronic treatment resistant population of heroin users who are currently enrolled in a methadone maintenance treatment program.

The major outcome variables will be illicit drug use, medical status and social integration. And we have a few secondary objectives. One is to compare the effects of a short and a long term treatment with heroin. Another is to evaluate the effects of discontinuation of the co-prescription of heroin; what happens when you stop prescribing? This is a very complicated, but I think very interesting issue. To evaluate the differential effects on client satisfaction and to generate information for a hypothesis regarding patient treatment matching. These are the goals of the study.

And these are the inclusion and exclusion criteria. You see, it's quite extensive, it's very explicit. Patients need to have a diagnosis of heroin dependency of at least five years, they should be no younger than 25, they should be registered in a methadone maintenance treatment in the last 12 months, they should have had at least fifty visits in the last six months, (i.e. that they should come about a third to half of the visits in the methadone maintenance clinic.)

They should have had a period of at least four weeks in which they were using at least 60 milligrams of methadone, because we really want that these are patients who are treatment resistant also to methadone. They should be either in poor physical or mental health or have poor social integration. And they should be legally in the Netherlands, registered in the area for at least three years in order not to attract people in other areas. Finally, they should be willing to [give their informed] consent.

These are some of the inclusion criteria but I will not dwell on it too long. Basically they indicate that you shouldn't recruit patients who are about to die or who are about to go to prison for a very long period. Also, people who need more than 150 milligrams of methadone per day are excluded, not because this is by itself so bad but it is such an exception in the Netherlands that you would probably attract some strange [individuals].

These are the control and the experimental treatments. The control treatment, basically, is oral methadone, up to daily doses of 150 milligrams: methadone prescription once a day, take-home methadone for the weekend, and a standard offer of psychosocial interventions.

In the experimental treatment this is the same, but there is an additional offer of heroin, inhalable or intravenous seven days a week, three times a day, with a maximum amount of heroin per administration of 400 milligrams of pure heroin and a maximum of 1000 milligrams of pure heroin per day. No other illicit drugs will be prescribed and there is no take-home of any of the dosages of the heroin.

Lastly, this is the design. I am sorry that it is a little complicated, but what you see here is basically two parallel protocols. One is the inhaling group the other is for the injecting group. I will describe only the one for the inhaling because basically it is the same for the injecting group.

The study starts with a two month qualification period. We basically do all the assessments there. People are asked to participate in the study in the beginning of the qualification period and they get two months time to test their motivation to participate in this study. At the end of the qualification period, we assess whether they still meet the criteria for inclusion and exclusion and only then do we start the randomization.

People are randomized in three groups. The first group is the group that gets heroin and methadone prescribed during 12 months. The second group gets heroin prescribed only for six months. And then we have a third group that doesn't get any heroin prescribed, they actually stay in the methadone maintenance program.

After the 12 months of the randomized trial, we have a six months naturalistic follow-up period. The people who actually did not get any heroin prescribed in the first 12 months of the randomized trial, do get an offer for heroin prescription if they still meet the criteria then. For the patients who had heroin for six or 12 months, the heroin prescription stops. It is both interesting and somewhat scary to realize that we prescribe heroin for 12 months and then we abruptly stop.

After 6 or 12 months of heroin prescription all patients stop whether they did well or didn't. The patients who actually improve dramatically during this period and show deterioration in the first two months of the naturalistic follow-up period, they can be offered the co-prescription of heroin again on an individual basis; the decision is made by the treating physician.

During the naturalistic follow-up (phase III), we might have some information about the effectiveness already. Basically there are two scenarios: (1)treatment with heroin co-prescription is not effective. In that case there will be very few people who, during heroin prescription, improved dramatically, and deteriorated after stopping the heroin. And so, you could say we're stuck with some people but not a lot that although there is maybe no reason to go on with the whole heroin prescription, we still have to give them heroin for some time.

(2)However, if the study is effective, i.e. if it turns out that co-prescription of heroin is effective, then the people who actually improve a lot and deteriorate later will be a pretty big group. I think, then, there is a strong demand to have treatment available with heroin for this relatively large group also in the future.

Finally, a note on the outcome. This is how we prespecified outcome. Basically, I think it's a very difficult problem. I think the addiction field is on the forefront of science in this respect. How do you define effect in a multidimensional disorder?

The central outcome dimensions in the study are substance use, physical and mental health and social integration. These are the measurements that we use: the European version of the Addiction Severity Index, urinalysis, and the Composite International Diagnostic Interview Schedule for the mental health. For social integration, we use both self-reported data and criminal record data.

What is an effective treatment in our view? There should be at least a difference of 20 percent in the number of responders between the control treatment and the experimental treatment. If the difference is smaller than that, we don't think it's a clinical important effect.

The next question is, what is needed to be a responder? And again, they're you're really in the dark and you will have to make some arbitrary decisions. According to our protocol, a responder is somebody whose medical condition improved substantially, i.e. there should be improvement of at least 40 percent, and/or improvement in terms of social integration of at least 40 percent and, finally, there should be no increase of more than 20 percent in cocaine use. So this is our prespecified effect measure. There is also, a prespecified analysis plan. So right from the start, you can see what we regard to be effective.

My last overhead is about the organization of the heroin trial. Again, like in Switzerland, the Ministry of Health is very helpful so far. They provide the funding for the study and they have been very helpful in planning the study. The study is being prepared by the Central Scientific Committee and a national research group. In contrast to the Swiss, we made one national protocol which will be executed in six cities in the Netherlands in eight treatment centers. Thus it will be a multi-center trial with centers all using the same protocol, in terms of, both, research requirements and the treatment requirements.

On the local level, there will be local research groups and local treatment centers and there will be a central research organization collecting the data and monitoring both the treatment quality and the data collection quality. Finally, there will be control from the Inspectorate for Health on the quality of the delivered services. The study and the treatment starts in July 1998. We hope to finish the study somewhere in the year 2001. Thank you very much.