The Binge Drinking Pill
This newsletter tells the story of a drug that's effective, safe, cheaply produced, yet seldom prescribed. Decades of research attest to the drug's benefits – we even understand the neurochemistry behind its success – but it remains much less popular than treatments that are no more effective. You might say then that this is a story about why data doesn't always prevail.
A new study
On Tuesday, the New York Times described a new study showing that naltrexone pills can reduce binge drinking. (Naltrexone, sold under names such as Revia and Vivitrol, is an FDA-approved treatment for alcohol as well as opioid use disorder.)
As I read the Times article, I struggled to understand why the study was conducted. Naltrexone was described as a binge drinking pill, but we've known for almost three decades that the drug can reduce alcohol consumption among people with drinking problems. What exactly was "new" about this study? I decided to have a closer look.
Social context
Excessive drinking is a major public health problem. You may have heard the statistics: 15 million Americans have an alcohol use disorder. 65 million Americans binge drink at least once per month. 1 in 8 deaths among adults between the ages of 20 and 64 stem from excessive drinking, and over 10,000 driving fatalities per year are attributed to drunk driving.
There's also the social and emotional turmoil that can't be captured by numbers. Most of us know someone with a drinking problem, if we haven't been there ourselves. My father, for instance, was an alcoholic for most of his adult life. When I was 11, I had to drive us home from a restaurant one night after he passed out getting into the car. (I must've driven too slowly, because I recall almost nonstop honking.) We all have a story...
Because alcohol abuse is so pervasive, but solutions that work for everyone remain elusive, research on treatments continues. The study I'll be discussing, led by Dr. Glenn-Milo Santos at University of California, San Fransisco and published in the prominent American Journal of Psychiatry, looked at how effectively naltrexone can reduce binge drinking.
Neurochemical context
As you drink alcohol, one change that occurs is the release of endorphins (your brain's own naturally-produced opioids) into the synapses between brain cells. Naltrexone blocks the receptors that are ordinarily stimulated by endorphins, and so the euphoria that usually comes when taking a drink is experienced much more weakly, or not at all. Thus, drinking becomes less rewarding, and the cravings diminish. The term "pharmacological extinction" is used to describe this process of the brain unlearning the association between drinking and pleasure.
If you're familiar with this topic or have seen the 2014 documentary "One Little Pill", you'll recognize that this study tested the effectiveness of the Sinclair Method. Dr. David Sinclair, working first with rats and later with humans, was the first to demonstrate that drugs like naltrexone reduce the craving for alcohol as well as actual consumption. What makes his method distinctive is the view of what we now call alcohol use disorder as a learned association that can be broken simply by taking naltrexone at least one hour before drinking.
Study methods
In this study, 120 men who reported occasional binge drinking (defined as five or more drinks on one occasion) were instructed to take a pill at least one hour before they anticipated heavy drinking. Half the men were given naltrexone, while the other half received a placebo. None of the men knew which kind of pill they'd been given. Across a 12-week period, the men reported each week when they took their pill and how much they drank at each sitting.
In other newsletters, I've pointed out startling weaknesses in measures of food and alcohol consumption. The approach to measurement in this study was startlingly rigorous. The pill bottles given to participants had caps with wireless monitors that recorded each opening of the bottle. This allowed the researchers to check on the accuracy of self-reported pill use. Alcohol consumption was recorded once per week, and partially corroborated by blood and urine analyses. Thanks to the pill bottle monitors and biomarker analyses, Santos and colleagues confirmed that the self-report data were quite accurate.
Main findings
Roughly three-quarters of men in both groups (naltrexone vs. placebo) took a pill on days that they anticipated heavy drinking. The main finding was that naltrexone reduced binge drinking as well as cravings for alcohol. Compared to the placebo group, men who took naltrexone exhibited significantly fewer binge-drinking days, significantly fewer weeks with at least one binge drinking day, and significantly fewer total drinks per month. The naltrexone group also showed a significantly greater reduction in craving for alcohol, according to their responses to a standardized survey measure.
Santos and colleagues didn't report their outcomes as concrete numbers (e.g., how many days per month each group binge-drank, on average). Rather, they only reported incidence rate ratios (IRRs). For instance, the IRR for number of binge drinking days was .74, which means that if you divide the number of days that the naltrexone group binge-drank by the number of days the placebo group binge-drank, you obtain a value of 0.74. The placebo group thus binge-drank on 26% more days than the naltrexone group did. If the mean for the naltrexone group had been 7.4 days per month, for example, the mean for the placebo group would've been 10 days per month. (Again, we don't know what the actual means were.)
I have mixed feelings about this approach to data analysis. On the one hand, it's hard to get a feel for the data without the concrete numbers. On the one hand, the risk of reporting those numbers is that some people (clinicians, journalists, policymakers, etc.) will surely overemphasize them. The Sinclair Method website, for example, announces that the method has a "78% long-term success rate", but that's misleading because, regardless of how you define success, each time you run a study, you're not going to obtain exactly 78% success among participants. In the end, the expectation is that naltrexone will help some people; it's exceedingly unlikely you could affix a number to how often it will help.
Limitations?
Santos and colleagues showed that naltrexone reduced craving for alcohol as well as binge-drinking behavior among men with a history of binge-drinking. What makes this news especially good is that the study methodology was strong. The researchers noted some limitations in the Discussion section, but they may have done so merely because one is obliged to do that to get a study published. I'm not sure what they described constitute serious limitations, nor do I see others. For instance, they acknowleged that self-reported measures of behavior such as the ones they used may be subject to bias. That's true, but (a) self-report is, logistically speaking, pretty much the only way of obtaining data on alcohol consumption outside the lab for extended periods, and (b) the researchers' approach to measuring consumption is much more rigorous than what's used in most self-report studies, as I suggested earlier.
One other limitation pertains to the sample. I haven't mentioned this yet, but all 120 men were sexual and gender minorities, or SGM – specifically, men who are transgender and/or report having sex with men. Santos and colleagues commented that the findings may not generalize to the rest of the population.
I have mixed feelings about this too. On the one hand, we understand the neurochemistry behind naltrexone's effectiveness, and I'm not aware of any reason to suspect that the brains of SGM function differently than anyone else's. On the other hand, you never know. It's good science to generalize cautiously from samples. A striking example pertaining to gender is that doctors may still miss, or misdiagnose, heart attacks among women, because their medical school training was informed by studies conducted exclusively with men. (As it turns out, men and women differ in some of the symptoms that tend to accompany a heart attack, and so the research conducted with men doesn't generalize well.)
What's distinctive about the study?
The effectiveness of naltrexone at reducing problem drinking is well-established. What distinguished this study is merely its focus on SGM men, and on a level of prior drinking classified as mild to moderate AUD (most of the research has focused on more severe levels). Thus, Santos and colleagues offer us one more demonstration that the Sinclair method can be helpful for some people with a drinking problem.
Why is naltrexone under-prescribed?
In spite of studies such as this, naltrexone (and two other opioid antagonists approved by the FDA for treating AUD) are prescribed less than 9% of the time for people seeking help with a drinking problem. What gives?
Because naltrexone is cheaply produced, and no company has exclusive rights to its production, manufacturers haven't promoted it very aggressively. The financial incentives are limited, in other words. But there's a bigger issue. A treatment plan that involves naltrexone doesn't ask the person to stop drinking. Rather, it assumes that the person will continue to drink, but to a lesser extent over time, until they've either gone abstinent or reached the point of safely tippling.
I think you can see where this is leading. Naltrexone-based treatments involve giving a drug to people who already have a problem with a drug, in hopes that use of the other drug diminishes. This clashes with mainstream approaches to alcohol abuse treatment, which assume that abstinence from alcohol is both a starting point and the ongoing goal.
The oldest and most popular of these approaches is the Twelve Step program developed by Alcoholics Anonymous (AA) founder Bill Wilson. The only requirement for AA membership is the desire to stop drinking (not reduce it, but stop altogether). Complete sobriety is the goal, and members receive chips indicating how long they've been sober.
A third treatment option, along with pharmacological treatments like naltrexone and 12-step programs, are psychotherapies such as cognitive-behavioral therapy (CBT). AIthough some therapists would accept the Sinclair method or other approaches to controlled drinking, many insist that abstinence should be the goal of AUD treatment.
In sum, naltrexone is rarely prescribed because mainstream approaches to AUD treatment don't leave much room for it. AA opposes use of the drug. Almost all rehab centers rely on 12-step programs and/or psychotherapy, and they seldom have medical staff who are both authorized and willing to prescribe naltrexone. (Substance abuse rehab is a multi-billion dollar industry; some experts claim that industry leaders actively oppose naltrexone, because it would put them out of business.) Meanwhile, therapists rely on psychotherapy and may not suggest that patients seek prescriptions. Physicians often refer patients to 12-step programs or therapists, and some doctors aren't even aware that naltrexone is a treatment option for AUD.
Is naltrexone less effective?
Looking at actual clinical practice, you would think that 12-step programs and psychotherapy are much more effective than naltrexone and related drugs, because the latter are so rarely prescribed. Is that what the data tell us?
In a word, no. Here's what we know:
1. Each of the three approaches is more effective than a placebo. Effectiveness includes reductions in overall drinking, reductions in heavy/binge drinking frequency, reductions in drinks per binge, as well as percentages of people who end up completely abstinent.
2. The specific measures of effectiveness vary too much across studies to allow inferences about which approach is most effective. All three approaches are "somewhat" effective, in some respects, but far from universally effective in any respect.
3. Head-to-head comparisons between 12-step programs and psychotherapy, and between naltrexone and psychotherapy, reveal few differences in outcomes. (As far as I can tell, there haven't been rigorous experimental comparisons of naltrexone vs. 12-step programs.)
4. Combining naltrexone with therapy tends to yield better outcomes than naltrexone alone. I consider this an especially meaningful finding. People drink for different reasons. A pill may help break neurochemical reliance on alcohol, but therapy can help people deal with what prompts them to seek refuge in alcohol in the first place.
From a therapeutic perspective, other important details can be found in the literature. For instance, naltrexone may outperform other drugs at reducing binge drinking, and the effectiveness of this drug may be greatest among people with specific genetic profiles. But the key point is that it's generally effective. It's not better than AA or psychotherapy, but it's not worse either. It's a helpful but under-utilized treatment option for a serious problem.
Conclusion
As suggested in "One Little Pill", people have probably died because they were unable to obtain a prescription for naltrexone. Either their health care provider was unaware of the drug or unwilling to prescribe it. Those who do obtain the drug often speak of it favorably. A friend of mine, a licensed chemical dependency counselor in north Texas, tells me of a client who tried naltrexone last year and within a week had gone from drinking multiple shots every night to filling a single shot glass and leaving it on the counter, untouched, before going to bed. He didn't want it anymore, my friend said. Alcohol has lost its appeal, and he mostly stopped thinking about it. Eventually he stopped pouring the glass.
Each approach to the treatment of alcohol use disorder has its strengths and weaknesses. AA is free and provides a sense of community and faith, but not everyone wants to share their woes with strangers, most or all of whom have no expertise in psychology. Therapy is grounded in psychological support from an expert, but it may be expensive and, like AA, it takes time. Naltrexone is by far more convenient, and it's inexpensive (a little less than $2 per pill), but the pill itself doesn't come with a sense of community or therapeutic guidance from an expert.
All of the treatment approaches I've described here help at least some people, and naltrexone is no less effective than the rest. For this reason I think naltrexone should be more widely available. This means more awareness of it as a treatment option, especially in conjunction with therapy, and less resistance to it on the grounds that abstinence-only should always be the goal.
From talking to people who have struggled with alcohol use disorder or who have been close to an alcoholic, I know that some of them would be dismayed by what I'm recommending here. They've experienced or observed how, after a period of sobriety, one drink is all it takes to precipitate a relapse. For them, nothing short of abstinence feels right.
I can't argue with that. Perhaps naltrexone isn't the ideal option for them or their loved ones. But I also can't turn away from the fact that the drug does help some people, and, among other things, it lowers rates of relapse.
Although some people develop a serious drinking problem and then find a permanent solution, it's at least as common to hear about people who seesaw between dependence and sobriety, often over a period of years. For those people, at least, no inexpensive, evidence-based treatment options should be off the table. I hope that if you or someone you know struggles with alcohol and hasn't found a consistent solution, you'll consider naltrexone or one of the other drugs that may reduce cravings and actual consumption.
Thanks for reading!