The Conditioned-Tolerance Explanation of “Overdose” Death

I recently blogged about Shepard Siegel‘s idea that heroin “overdose” deaths — such as Philip Seymour Hoffman’s — are often due to a failure of conditioned tolerance. In the 1970s and 80s, Siegel proposed that taking a drug in Situation X causes learning of a situation-drug association. Due to this association, Situation X alone (no drug) will cause an internal response opposite to the drug effect. For example, coffee wakes us up. If you repeatedly drink coffee in Situation X, exposure to Situation X without coffee will make you sleepy. As the learned response opposing the drug effect grows, larger amounts of the drug can be tolerated and the user needs larger amounts of the drug to get the same overall (apparent) effect — the same high, for example. Trying to get the same high, users take larger and larger amounts. But if you take a really large amount of the drug and don’t simultaneously evoke the opposing response, you may die. What is called “overdose” death may be due to a failure to evoke the conditioned response in the opposite direction.

Siegel’s Science paper about this — a demonstration with rats — appeared in 1982. Since then, plenty of evidence suggests the idea is important.

First, “overdose” death has become more common. A Washington Post article prompted by Hoffman’s death says that death due to “overdosing” on drugs — usually opiate drugs — has doubled in the last ten years and is now “the leading cause of accidental death in the United States, accounting for more deaths than traffic fatalities or gun homicides and suicides.”

Second, new data has supported Siegel’s explanation. Alex Schull linked to a 2005 case report that says: “K.J. did not return home with the heroin purchased as he did on other occasions but went to the public toilet in the pedestrian underpass at the Népliget Metro station where he injected the same quantity (0.5 gram) that he had taken the previous day in the accustomed place, at home with his wife.” The report cites other supporting evidence.

Third, new data has contradicted other explanations.The Post article includes an interview with an addiction expert named Keith Humphreys, a Stanford professor of psychiatry. Humphreys said, “Toxicology results after a fatal overdose usually indicate that the victim has consumed either their normal dosage level or a dose slightly lower than their normal level.” He also said, “Toxicology studies of overdosed people very rarely find that impurities played an important role.”

Yet Humphreys appears unaware of Siegel’s idea, even as he provides supporting evidence:

Typically overdose occurs because they’ve had a loss of tolerance. This loss of tolerance often arises because they haven’t used for a while. Maybe they had a voluntary period of abstinence. Maybe they were in jail, and their body can no longer handle the same dose.

The other leading cause of loss of tolerance is consumption of other substances. This is particularly true of alcohol, which seems to lower the body’s ability to tolerate opiates (so do benzodiazepines). Most of what we call “opiate overdoses” are really polydrug overdoses: alcohol and heroin, alcohol and oxycontin, benzodiazepine, alcohol and Vicodin, combinations like that. [This is consistent with Siegel's explanation. The second drug makes the situation less familiar, reducing the conditioned opposing process. -- Seth]

Siegel’s idea was recently mentioned in the New York Times:

A change in where a person uses his or her drug of choice can increase the likelihood of an overdose, studies suggest. “If you habitually use in your car, for example, the body prepares itself to receive the drug when it’s in that environment,” Dr. Rieckmann said. “It’s called conditioned tolerance. When people using are in an unfamiliar places, the body is less physically prepared.”

This was the first mainstream mention I’d ever seen. I told Siegel about it and he said it was the first mainstream mention he’d seen, too. He added, however, that he had come across the idea in a crime novel:

A Scottish constable, Hamish Macbeth, appears in a series of books by M. C. Beaton. In one 1999 book in the series, “Death of an Addict,” Macbeth has a conversation with a Dr. Sinclair, a pathologist on the scene of an apparent heroin overdose: “Dr. Sinclair leaned his cadaverous body against his car and settled down to give a lecture. ‘The reason for tolerance to heroin is partially conditioned by the environment where the drug was normally administered. If the drug is administered in a new setting, much of the conditioned tolerance will disappear and the addict will be more likely to overdose’” (Beaton, M. C. Death of an addict. New York: Warner Books, 1999, p. 23). M. C. Beaton is the pen name of Marion Chesney, and I wrote to her asking how she knew this. She couldn’t recall, but thought that it likely was due to a conversation she had with a Scottish police officer.

There are several similarities between Siegel’s idea and the Shangri-La Diet, which I will point out later.

11 Responses to “The Conditioned-Tolerance Explanation of “Overdose” Death”

  1. Rashad Says:

    Just FYI, I’d definitely heard this idea and assumed it was more widely known before you mentioned it. Probably from my mother who is a retired public health doctor who has done a lot of work on addiction.

  2. Nancy Lebovitz Says:

    Does this imply that people who want to get the most for their money from drugs should keep taking them in different circumstances?

    Seth: Yes. I think if you really want to keep the cost down you should wear headphones and listen to different music each time. Maybe clip your nose shut to reduce smells to focus more attention on sound.

  3. LemmusLemmus Says:

    I, too, had read about the idea, I believe in a 2004 ARP paper on smoking.

    But I had never considered it in connection to a data point that one often sees cited in connection with addiction. The data point is that, apparently, there were many U.S. soldiers who got addicted to Heroin in Vietnam, but many of whom quit when they came home.

    This is quite often cited in support of the idea that addiction/drug consumption is a choice (rather than something that is beyond the addict’s control). I think I may also have seen it linked to the sociological idea of framing (different situations trigger different behaviours).

    Be that as it may, it suggests that the best moment to quit is when you know you’ll be in unfamiliar situations a lot. It may also help explain why I drink so much more coffee in the office than at home.

    Seth: Siegel has emphasized that Vietnam heroin data. I’d say the best moment to quit is when you change where you live — when starting grad school, for example. Or moving to a new apartment. Or even to a new office in the same building. You make a good point, I have never heard “move to a new apartment, change your workplace” as part of how to get rid of an addiction.

  4. Christopher Sturdy Says:

    I mentioned your earlier post and talked about this in my learning class, although I used a different study of Siegel’s (morphine tolerance) to make the point.

  5. Sara Says:

    Someone should develop a blood pressure cuff that automatically injects naltrexone when it can no longer detect a heartbeat.

    Of course, there’s then the issue of moral risk. But if Hoffman had had one he’d still be here.

  6. Sara Says:

    Hmm… That wouldn’t work, because if there’s no pulse, the naltrexone isn’t getting to the brain.

    Maybe it could detect shallow breathing, then sound an alarm, then if no response, an injection?

  7. seth Says:

    Just an FYI: Keith regularly blogs at http://www.samefacts.com

  8. JP Says:

    Seth – I wonder if your 2007 blog post on visual and auditory cues possibly raising the body’s set point could function in a similar manner.

    “I’m not surprised that auditory and visual signals for food cause hunger. There are lots of conditioned cravings like that. Tim goes on to wonder if these learned signals for food raise the body-fat setpoint, as the theory behind the Shangri-La Diet says that food-associated flavors do. If you walk by your favorite bakery every day, will you weigh more than if you don’t? ”

    http://blog.sethroberts.net/2007/08/21/do-all-signals-for-food-raise-the-set-point/

  9. Seth Roberts Says:

    I really don’t know if sounds and sights associated with food raise the setpoint. However, I am sure they cause hunger and I am sure hunger increases the probability of eating. For that reason, it is better to avoid them, if possible, if you want to lose weight.

  10. ‘A change in where a person uses his or her drug of choice can increase the likelihood of an overdose…’ Says:

    […] This was the first mainstream mention I’d ever seen. I told Siegel about it and he said it was the first mainstream mention he’d seen, too. He added, however, that he had come across the idea in a crime novel: […]

  11. bebo Says:

    This phenomenon was discussed in lectures in every psychology course I took in college in the 90′s. Friends I had at low-end state schools at the time also learned the same thing in their intro-level psych courses. Very old news, and widely known.

    They always discussed the case of the girl found ODed in Grand Central Station who took half of her regular dose that she usually injected at home. Sounds suspiciously like the “case study” cited from 2006.