Addiction Treatment, Science, and Dead Rats

In my last post I teased that I would write about fake science. I’ll try to make it interesting.

The internet allows everyone to do research about symptoms and treatments for any condition. If not for need for prescriptions, people could act as their own doctors. But a huge dose of caution is necessary before anyone takes that path.

Realize first that doctors don’t treat themselves or even their family members. The saying that ‘a person representing himself in court has a fool for a lawyer’ applies double in healthcare. Treating someone close to one’s self introduces a bias that is hard to explain, but easy to notice. As an example, I see a doctor annually to monitor a progressive condition that threatens my vision. I would like to know the answer to a simple question: how bad is it? If I have a patient with that condition I can look at images of his/her retina and have an immediate, rough sense about what the person is facing. But when I look at my own images and test results I sense nothing beyond fear or relief. The problems with self-assessment are of course greater in the field of psychiatry and addiction. After my relapse in 2001 I was told I needed treatment, and my assessment called for a brief refresher course on the twelve steps. Three months later, still in residential treatment, I recognized how wrong I was.

A larger problem is that research on the internet is nothing like the research used by doctors or scientists. There are a few sites that offer true research, such as Pub Med, where you can search my name and see the articles from my PhD work in the 1980s. Doctors at academic hospitals or institutions often have access to an electronic database including thousands of peer-reviewed journals. In grad school I spent time each morning in the library, reading the Science Citation Index for new stories about vasopressin and then searching the stacks for the article (medical libraries have so many journals that they take up 4 or 5 floors or more of a large building, with narrow halls between floor-to-ceiling shelves). In the stacks I sometimes realized I was standing amidst the results of the hard work of millions of scientists over the past 50 years.

The information on the internet is useful because it helps patients ask the right questions. But it is a mistake to consider it as research, or even to assume it is correct. Doctors and scientists (and any good health practitioners) rely only on peer-reviewed literature. And even then, a good scientist gathers a sense, over time, of the better peer-reviewed journals vs. the ones with less credence. What is peer review? When a scientist submits research for publication, the article is sent to 3 or 4 independent reviewers who work in the same field but have no connection to the author of the study. I am a peer-reviewer for a couple of journals. When I receive an invitation to review a study I have to disclose any bias or connection to the study or authors. If I accept the invitation I have several weeks to carefully review the study, noting if the findings are valuable, whether the groups were sufficiently randomized and blinded, whether the statistics are correct or if a statistician should be involved, and whether the findings support the conclusions. I then tell the journal editor my opinion, including whether the study should be accepted, rejected, or accepted with certain revisions. Peer reviewers are not paid; they provide the service because they recognize that the process is necessary and valuable.

The FDA regulates medications based on the results of research studies. Some of the studies reviewed by the FDA are already published, and some may never end up in a formal publication. But their process for evaluating medications is similar to the work of a peer-reviewer in that they determine whether the science is ‘good’ – double blinded, properly randomized, good statistics, etc. Any claims about a medication MUST be deemed accurate by the FDA.

This post was inspired by an ad for Declinol, a supplement marketed to ‘treat’ alcoholism. Supplements are not medications, and not subject to the same rules. Read the FAQ on the Declinol web site and note the answer about FDA approval. Declinol is not subject to FDA approval because it is a nutrient, not a medication. The FDA allows greater latitude for promotional claims about nutrients, but even makers of supplements are not allowed to lie. The acrobatics of marketers of such products are sometimes funny, at least to us nerds, and Declinol is a classic example. Note that the web page doesn’t say that it treats alcoholism or cravings; it is a ‘SUPPORT for physical cravings, calmness, and overall well-being’. What is a ‘support’? Your guess is as good as mine.

Instead of making claims that can be found to be false, nutrients often show quotes by ‘satisfied customers’. If the FDA believes that the quotes are misleading, that’s on ‘Bob from California’, not on the marketer of the nutrient. Instead of describing how the nutrient works, nutrient marketers provide citations about the nutrient that support whatever the marketers want you to think. So with Declinol we see ingredients like folic acid, with broad generalizations about the value of that substance. Yes, Folic acid is valuable. You can’t live without it. But that’s a far cry from saying that taking extra folic acid has any value, let alone value in reducing alcohol intake. We give folate to alcoholics in detox because they sometimes have dietary deficiencies caused by consuming nothing but alcoholic beverages. If you eat meals a couple of times per day you almost surely have plenty of folic acid in your body, and any extra is metabolized and excreted.

Must nutrient ‘treatments’ or supplements contain a blend of vitamins. It is very easy to write reassuring and positive statements about vitamins because by definition, vitamins (the term comes from ‘vital amines’) are molecules critical to normal function. But many studies have shown that a typical diet provides adequate amounts of vitamins, even if that diet includes fast food.

Many nutrient ‘treatments’ also contain a couple special ingredients we’ll call ‘secret sauce’. One secret sauce in Declinol is Kudzu, and support for Kudzu in reducing alcohol consumption can be found on Pub Med. Like similar products, Declinol’s marketers take a finding about a substance and grossly generalize the findings to create an impression that was never part of the original finding. According to the study about Kudzu, 20 people in a ‘natural settings laboratory’ (is that an oxymoron?) were given water, juice, and up to six beers, and told to drink at will. And (wow) when people were given 2 grams of Kudzu first, they drank beer more slowly, and opened fewer bottles.

A couple of problems, though, in concluding relevance to treating alcoholism. Were the 20 subjects alcoholics? It doesn’t say, but I would guess not because I don’t know if a study giving beer to alcoholics would pass the ethical review board. Beyond that, WHY did they drink less alcohol? If I gave you syrup of ipecac, you would probably drink less alcohol. If I gave you a tablet of oxycodone, you would probably drink less alcohol. That doesn’t mean that the substances are useful in treating alcoholism or alcohol cravings. Why did the Kudzu group drink less alcohol? Did it truly reduce interest in alcohol in a study with very few subjects who may or may not have alcohol problems? Or did it leave a nasty taste in their mouths or destroy their taste buds? Did it cause nausea or dizziness that made alcohol less enticing? Did it reduce vision so they couldn’t find the beer bottles as easily?

As for the title of this post, when I researched vasopressin one hot idea was that vasopressin enhanced learning and memory. We measured that improvement in studies using ‘passive avoidance.’ We placed rats in a cage that had dark cubbies in one corner, and when rats invariably went into a certain cubby they received an electric shock. We repeated the task with or without putting vasopressin into the rats’ brains and some rats ‘learned’ to avoid the electric cubby, supposedly by remembering the shock better than other rats. There is a major flaw in the study that can often be applied to other ‘experiments’, including the one I cited about Kudzu: the best performer in a passive avoidance task is a dead rat.

I have no idea whether Declinol reduces cravings or generates ‘well being’, whatever that is. But nothing on their website pushes me toward that conclusion. I hope readers will keep some of these comments in mind when the next big cure comes along.

Make Sleep Meds Work For You

I’ve been busier than I like, and haven’t had as much time for posting.  But I spend a lot of time answering emails from my patients, and some of my responses may be useful for others.  Below I’ll share my answer to a patient who has been unable to get quality sleep.  Next week I’ll find another answer to share with readers.
This patient asked whether her insurance would cover Lunesta.  She wrote at 2 AM that she is up most of the night tossing and turning. She now takes 10 mg of Ambien, and wrote that it ‘stopped working’.  She doesn’t think 20 mg of Ambien would be covered by insurance (although Ambien is very inexpensive when purchased for cash).  She takes gabapentin for a pain condition and wonders if increasing it would help with sleep.
My response:
Before getting into a discussion about insurance I want to make sure you have a good understanding about the issues you’re facing when you take sleep medications.  Most sleep medications are subject to tolerance, and some share ‘cross tolerance’.  Lunesta (s-zopiclone) and Ambien (zolpidem) act at the same receptor and have the same actions, so if a person is used to one, she is used to the other.  The situation is analogous to opioids, where a person tolerant to a significant dose of oxycodone will find little effect from morphine.
I have taken Lunesta but stopped it for the same reason you are unhappy with Ambien:  it just didn’t do much.  Lunesta also causes a very bitter taste in the mouth, not from the pill touching taste buds, but from the drug circulating in your bloodstream.  The taste is unpleasant but goes away aftef a few weeks in most people.  If you are not responding to Ambien, you won’t likely respond to Lunesta.
Some people change from an ‘ultrashort’ to a benzodiazepine, typically temazepam.  That change will often offer some benefit, but tolerance will eventually match the effects of temazepam just as with the shorter medications.  Patients often ‘chase’ tolerance higher until their doctor refuses to prescribe larger amounts.  The exercise only deepens the plight of the insomniac, making it harder and harder to sleep without medication.
What is your current dose of gabapentin?  Pain docs sometimes taper that drug up to a total of 3 grams per day, and higher doses cause some degree of sedation.  But again, step back and look at the big picture.  With any sedating substance, including gabapentin, you are only  addressing the short-term.  Your body will adjust to any dose of sedative, just as with opioids, and you will eventually need to address the same symptoms on even higher doses of sedatives.
I am willing to look at the gabapentin, but understand what you are up against.   Patients tend to focus on the short-term at 2 AM.  My job is to know that you will be around for years, even decades!  I often see new patients whose doctor prescribed benzodiazepines for years, repeatedly raising the dose.  At some point the patient ran out too early, prompting discharge.  At that point there is little to be done for such patients beyond helping them through several weeks of severe insomnia caused by benzodiasepine withdrawal.  Sometimes questiapine, clonidine, or hydroxyzine will help patients find some restless sleep while lowering their tolerance to benzodiazepines, but that sleep is usually poor in quality and associated with significant daytime sedation.
My best advice:  When taking sleep medications, the most important point is to NEVER ‘double up’.  If you take two tablets instead of one of any medication, you will never get the same benefit from taking one again.  You will only run out early, and the doctor, pharmacist, and state will all prevent early refills  The second point is that newer sleep meds will not just make a person fall asleep.  For the first few nights a new medication might feel more potent, but over time, the best thing a sleep med will do is HELP a person fall asleep.  Patients must do everything else correctly, including sleep hygiene measures such as using a dark, somewhat cool room, using white noise to help the mind avoid focusing on creaking boards or other noises, avoiding significant alcohol use, avoiding caffeine after noon, avoiding eating or smoking right before bed,  calming down at the end of the day (some people can’t  sleep if they exercise late in the day), dimming lights, avoiding watching football right before bedtime, etc.
If you do all those things, a sleep medication will help ease the transition to sleep.   But the shorter meds like Lunesta and Ambien work best when they take effect in people who are in bed already, working on getting to sleep.  People often make the mistake of taking a sleep medication and wating to go to bed until they feel sleepy.  The newer sleep meds don’t cause the sleepiness that came with the older drugs, and they don’t last near as long.  Waiting ‘until they work’ can also cause amnestic behaviors like sleep-eating, sleep-sexting, sleep-driving, and worse.
Give these issues some thought and tell me the amount of gabapentin you take.  Beyond the controlled substances, consider trazodone or clonidine, two medications that work differently than benzodiazepine and Ambien.
The best sleep, of course, is found in the unmedicated condition desiged by evolution.  Evolution is not fast enough to keep up with changes in communication and the social media phenomenon, so it is often useful to think about the sleep environment of our ancestors, 40,000 years ago.  If I knew how to use twitter, I’d hashtag ‘sleep like a caveman.’

Counseling Schmounseling

I just noticed a couple of my recent posts….  these people have it wrong, and that person has it wrong.  One of these days I really need to print something positive and uplifting.  But not today.
Excuse the self-flattery, but I like to think of myself as a physician scientist.  That concept motivated my PhD work, and cost me friend after friend in the years that followed!  A physician scientist isn’t all that difficult to be from an educational standpoint, especially in the age of the internet.  The one thing that is necessary is the willingness, or need, to question every assumption by the media, the government, physicians, laypersons, and other scientists.   Ideally, the questions are guided by a knowledge of p-values, the process by which scientific grants are awarded, an understanding of the peer-review process, and the realization that anyone elected to office knows less about science than most other humans on the planet.
Last night I came across an opinion piece– I think in the Bangor Daily News, but I could be wrong about that– that argued that we will never stem the heroin epidemic without use of medications.  The comment section after the article was filled with the usual angry banter over methadone and buprenorphine that now follows every article about medication assisted treatment.  As an aside, why are the abstinence-based treatment people so angry about medication?  There are people out there who choose to treat cancer using crystals, but they don’t spend time bashing monoclonal antibodies!
Here is the part of this post where I start losing friends…  but let me first say that I know some counselors.  I like counselors.  In fact, some of my best friends are counselors.  But in the comments after that article I read the same thing over and over–   that meds aren’t the important thing, and that counseling is what really makes all the difference.  A couple weeks ago  the person sitting to my right said the same thing during a discussion about  medication-assisted treatments.  And that same phrase is repeated ad nauseum in lecture after lecture in ASAM lectures and policy statements related to addiction.  The phrase has even been codified into some state laws.  And why not?  It is something we all ‘know’, after all.
If we are going so far as writing laws requiring that people have counseling in order to obtain medication, shouldn’t we do one thing first?  Shouldn’t we determine if the comment is really true?
A couple years ago two papers came out– someone help me with the reference if you have them– that looked at abstinence rates after a year on buprenorphine in patients with or without counseling.  Guess what?  The counseling group did not do better!  In fact, the counseled patients did worse; not sigificantly so, but enough to clearly show that there was no ‘trend’ toward better performance in the counseled group (which would have been pointed out, were it true.)
I could hypothesize many reasons why the counseled groups would do worse.  Maybe they were angered by the forced counseling and therefore bonded less effectively with their physician.  Maybe they obtained a false sense of expertise in dealing with addiction, making them more likely to relapse, whereas the non-counseled group learned to just do as they were told.  Or maybe the counselors send out signals, consciously or unconsciously, that interfered with medication treatment.
The thing is, we have no idea which of these things, if any, are going on!  There have been no systematic studies or other attempts to understand what happens during the combination of counseling and medication treatments.  We just have a bunch of people saying ‘do them both!  do them both!–  a comment that apparently feels so good to some people that they just cannot consider things any other way.
For the record, I see ALL my patients for at least 30 minutes for every appointment.  As a Board Certified Psychiatrist, I guess that means I’m counseling them.  And from what I can tell, it seems to be working pretty well.  But even in my own case, I would never draw firm conclusions unless someone does a double-blind study and collects the data.
I encourage all physicians, scientists or not, to question some of what we ‘know’ about addiction treatment.  Is it really all about the counseling?  Maybe— but then again, smart people used to ‘know’ the world was flat, and the Earth was the center of the Universe.

The Opioid Dependence Big Picture

First Posted 1/16/2014
Below, internet colleague Paul Dessauer shares his extensive knowledge of opioids in comments about my naltrexone post. His comments were particularly interesting in that they provide evidence that at least someone is aware of the big picture about addiction to heroin and pain pills.
In my post about naltrexone, I described how some people favored the drug over buprenorphine because of its lack of opioid effects. Unlike buprenorphine, naltrexone is an antagonist that has no abuse potential.  But I wondered… at a time when so many young people are dying, shouldn’t the primary issue be whether naltrexone saves lives?  Sure, it is ‘safe’– but does it work?
Paul provided references to answer my question.  While everyone is focused on the fact that naltrexone can block opioid receptors, Paul’s data shows that when naltrexone is used in the real world, people die.  I’m excited that someone, somewhere, has the courage to investigate the one thing that is never addressed in discussions about addiction, whether related to residential treatment, counseling, or medication: Does it work?
Paul’s comments:
You wrote;
<<< If ‘success’ consists of moving to naltrexone—a medication that many real-world addicts reject– how long is naltrexone continued, and what happens when it is stopped? Do people go back to heroin again? If not, why not? The cycle of ‘use, treat, cease treatment, use, and repeat’ should be a black box warning on naltrexone >>>
The other “black box” issue is dropped tolerance overdose when someone exits naltrexone treatment and relapses to illicit opioid use.
You asked; <<< How many people will die in their quest—or their doctor’s quest—for ‘abstinent recovery’ with or without naltrexone? >>>
This study used Australian coronial records to compare mortality rates amongst patients in methadone treatment, buprenorphine treatment, and naltrexone treatment. The authors make it clear that they believe the mortality rates for naltrexone calculated here are significant underestimates, as coronial data does not record such deaths consistently, and they found the majority of known naltrexone-related deaths did not appear in this data.
<<< When expressed as deaths per number of treatment episodes, it was estimated that naltrexone had a mortality rate of 10.1 per 1000 treatment episodes. If the mean treatment retention in naltrexone treatment was estimated at 3 months (rather than two months, as assumed in the above estimate), the mortality rate for naltrexone treatment increased to 15.2 deaths per 1000 treatment episodes.
Naltrexone was associated with a mortality rate of 22.1 per 100 person years during the period of high risk (2 weeks post-treatment), and 1 per 100 person years during the period of low risk (during treatment)…. >>>
<<< …The estimated mortality rate was 0.02 per 1000 treatment episodes for buprenorphine and 2.7 per 1000 episodes for methadone.
The mortality rate for naltrexone was four times higher than for methadone when calculated as deaths per number of episodes of treatment, and substantially higher than for buprenorphine.
When considering deaths per periods of high and low risk, the mortality related to naltrexone was approximately seven times that of methadone during the period of high risk and three times the rate during the period of low risk. Naltrexone treatment was associated with a mortality rate of 22.1 per 100 person years during the period of high risk (two weeks following treatment cessation) and 1 per 100 person years during the period of low risk (during treatment). Buprenorphine mortality rates were not expressed in terms of periods of high and low risk due to the low number of deaths detected with this search method. >>>
<<< In comparing mortality rates associated with these pharmacotherapies, it is important to draw the reader’s attention to the rates of mortality for active heroin users. It has been estimated that mortality rates for heroin-dependent persons not in treatment are in the vicinity of 0.9 per 100 person years of risk, very similar to the mortality rate of a person in naltrexone treatment (during the period of low risk) calculated in this study. >>>
Actually, this is slightly less than the risk of naltrexone-associated death calculated in the study, which is believed to be an underestimate.
And the risk of naltrexone-related death calculated for the “high risk” period (the two weeks immediately following cessation of naltrexone) is more than TWENTY TWO TIMES higher than the risk of death estimated for someone using street heroin who is not in any form of treatment at all.
<<< While maintained in methadone or buprenorphine treatment after the initial induction stages, opioid-dependent people are at lower risk of dying. Clearly, an important aspect of methadone and buprenorphine treatment for opioid dependence is the improvement of treatment retention rates.
The mortality risks associated with oral naltrexone treatment, particularly following treatment cessation, warrant serious attention. This is especially the case considering that the majority of unselected opioid-dependent persons will return to opioid use soon after leaving naltrexone treatment. It is recommended that future trials of all treatments for opioid dependence include monitoring of post-treatment mortality risk >>>
Gibson, A. and Degenhardt, L. (2005) Mortality related to naltrexone in the treatment of opioid dependence: A comparative analysis, Sydney: National Drug and Alcohol Research Centre

New Bupe News Section

Wanted to take a second or two to point out a new section to the blog, called ‘Bupe News’.  You’ll see the link at the tope of the page, along with an ever-growing list of links.  The point, of course, is to keep y’all reading, since Google knows EVERYTHING, including how many seconds each and every one of you spends on this (and every other) web site.  Understand that I don’t GIVE that information;  that information is simply that is there for the taking on the internet.  Even I can see the average time people spend on the site, the order they went to one page or another, etc.    I do not ‘collect data’ about any reader, meaning that I do not have information about who any individual is or isn’t…   but I DO get reports on my collective audience.
Check the page out, along with the other new section, and of course tell me what you think.  Positive suggestions are ALWAYS welcome!
Thank you for hanging with me,  by the way, for some tough I.T. times.  I’ve got about 3/4 of the ‘lost’ posts back, and then I’ll be back to firing off the NEW things I’m angry about, rather than replaying the things that irritated me a year ago!
J