1. Ed Whitney says

    One quick question about the study: it appears that cocaine-free urines were the main endpoint for efficacy. Were the urines also tested for other drugs? I would assume that you would want to test for other stimulants, especially for methamphetamine and similar drugs, since blunting the effect of cocaine could lead the user to try substitutes like amphetamine. If the user is trying to attain certain levels of dopamine in the reward centers, for example, and the antibody reduces the dopaminergic activity in the parts of the brain usually affected by cocaine, then the user could try to attain the same dopamine levels through other means. The article did not mention whether the urines were also tested for these alternative means for attaining the high normally achieved through use of the target drug.

    • Keith Humphreys says

      Hi Ed,

      Tom Kosten’s study was done on patients who were tested for everything. It is however plausible that some patients would move on to other drugs when cocaine didn’t “work”. Of course the comparison by which to judge the seriousness of that problem is not zero use of other drugs, because many cocaine users also use other drugs, sometimes triggered by the cocaine high (e.g., get high on cocaine and then consume a huge quantity of alcohol).

      One concern I didn’t have time to get to in the article is what if someone on the vaccine takes THE SAME DRUG in higher quantities, in an effort to override the vaccine. Just because it doesn’t get you high doesn’t mean a huge dose of cocaine can’t be cardiotoxic.

      • Ed Whitney says

        Ah, that helps. It seemed logical to assume that the urine would be tested for multiple substances, but the study protocol was not available at NIDA (I could not find it at any rate). For a lot of randomized trials, the registry allows you easily to find the original protocol and any modifications that were made to it. NIDA seems not to do this; entering the grant number in their search box leads nowhere.

        One reason to fuss about this is that a major threat to validity of any study is selective reporting of outcomes. There is good reason to suspect that this happens often; in 2009, a group at Hopkins published an article on outcome reporting in industry-sponsored trials of off-label use of gabapentin. They looked at 12 published studies for which the protocol was available, and found that the primary outcome in the published reports differed from that in the protocol for 8 of them. Trial registries were developed for this reason among others.

        Selective reporting of outcomes is one of the major threats to validity of a study that groups like the Cochrane Collaboration worry about and look for. This is a variant of the selective engagement with information that Nate Silver identified in his book as a universal cognitive blind spot, no? That guy is really onto something in his work.

        Use of extra cocaine to override the effects of the blocking antibody ought to have been detected by the fact that they did quantitative assays of cocaine metabolites for much of the duration of the study, but it is hard to tell if there were differences between groups in the levels of these metabolites. A higher level of metabolites in the group with adequate antibody than in the group without antibody would have been a giant red flag, if it were tested for and reported. Hence the desirability of a protocol at the NIDA website.

        One of the Achilles heels of evidence-based medicine relates to the relevance of reported outcomes to the things that really matter in life; an outcome that is easy to measure and has a numerical distribution that lends itself to statistical testing may or may not have anything to do with what is actually bothering the patient. By itself, the level of cocaine metabolite in the urine is less important than daily life functioning: being there for your kids when they need you, keeping a job, maintaining cordial relationships with friends and neighbors, and the like. Stuff like social adjustment with the SAS-SR are helpful, but they rely on self-report, which is difficult enough to validate in normal circumstances, and must be very difficult in work with addicts, whose self-report is often, shall we say, dubious.

        Overall, this means that research with substance abusers has more obstacles than for most fields of research. It’s gotta be hard to do this stuff. Makes me glad I don’t have to do it!

  2. Sebastian H says

    Any thoughts on ibogane? It appears to somehow reset the receptors so you don’t go through crippling withdrawal. Why aren’t we studying this?

    • Keith Humphreys says

      Sebastian H: I have heard of it but don’t know much about it. Are you sure it isn’t being studied — you might search NIDA’s grants list and see.

  3. says

    Lexical note on ¨vaccine¨. The popular use of this is preventive shots that prime the body´s immune response to a future assault with an antigenic lookalike; the paradigm is Jenner´s cowpox scrape. The trade use seems to be anything (like Pasteur´s rabies shots) that uses such a mechanism either for prevention or treatment. So when we read about HIV or cocaine vaccines, it´s not at first clear what is meant. I think we need distinct terms. Leave the great Dr. Jenner his preventive vaccine. What could we call Pasteur´s treatment one?

    Since Jenner´s zero donor cow was called Blossom, and Pasteur´s zero patient was Joseph Meister (the dog´s name is not recorded), we could go for the double coinage ¨florine¨ and ¨magistrine¨. This is not I fear going to work.

    • Ed Whitney says

      But “immunotherapy” wlll do nicely. It covers all treatments which work by eliciting an immune reponse.

    • Ed Whitney says

      The above utilizes the use-mention distinction to differentiate between statements about the word and statements about the referent of the word. “Immunotherapy” will do nicely to designate the intervention; whether immunotherapy will do nicely to treat cocaine addiction remains an open possibility.