Precision medicine is an approach that can help tailor treatments to individuals based on unique characteristics of their condition. This study tested whether individuals with alcohol use disorder would respond better to FDA-approved medications depending on whether they drink “to feel good” or “to avoid feeling bad.”
l
When studies compare different medicines or psychosocial treatments for alcohol use disorder to each other, they often produce similar effects where none have a clear advantage. One explanation may be the substantial variability of individuals with alcohol use disorder; they vary in terms of the severity of their disorder, other medical and mental health conditions with which they are also diagnosed, and resources they have at their disposal (e.g., income, stable housing), to name just a few. The next important question becomes: For whom do these medications work best? This is also known as “precision medicine”; certain healthcare interventions may be better for some than for others and treatment should be tailored to these individuals more precisely. Because the FDA-approved alcohol use disorder medication, naltrexone, affects brain functioning associated with experiences of pleasure and reward (e.g., fun, excitement, euphoria) and the FDA-approved medication, acamprosate, is thought to affect brain functioning associated with relief (e.g., reducing stress, anxiety, depression), Mann and colleagues conducted a “precision medicine” type study to test if naltrexone actually is indeed more effective for individuals who drink for reward and whether acamprosate is actually more effective for individuals who drink for relief.
This study was a secondary analysis of a primary, 12-week randomized controlled trial for alcohol use disorder (“alcohol dependence” in this study based on the 4th edition of the diagnostic and statistical manual of mental disorders) that compared 169 individual receiving daily oral naltrexone, 172 individuals receiving daily oral acamprosate, and 85 receiving a placebo medication. In addition to alcohol use disorder, to be included in the study, individuals were drinking in the month before the study at least 14 standard drinks per week for women and 21 for men. The study excluded individuals with history of other substance dependence (e.g., cannabis use disorder) and current psychiatric disorders such as major depressive or anxiety disorders. The primary study outcome was heavy drinking; that is, drinking five or more drinks in a day for men, or four or more for women. The average participant was 45 years old and 77 percent were male.
From a questionnaire assessing how frequently individuals drank in 15 “reward” situations (e.g., parties/celebration, having fun; out with friends and stopped by a bar) and 15 “relief” situations (e.g., feeling upset, feeling hopeless, criticized by others), authors used a sophisticated statistical procedure to create four groups of drinkers: 1) High Reward/High Relief; 2) High Reward/Low Relief; 3) Low Reward/High Relief; and 4) Low Reward/Low Relief. All analyses controlled statistically for gender, age, drinking severity, and smoking status, to isolate the effects of the medication on drinking outcomes – making a more compelling case that any group differences were caused by the medication and not by these other factors.
Individuals high in “reward” drinking and high in “relief” drinking had other characteristics that fit these profiles
For any heavy drinking, naltrexone was better than placebo for the reward drinkers, but acamprosate was no better than placebo for relief drinkers
For reward drinkers (upper graph), naltrexone produced a 50 percent reduction in the likelihood of having a heavy drinking day compared to placebo, but had similar effects for drinkers who were low in both reward and relief drinking, as the authors hypothesized. For relief drinkers (lower graph), though, acamprosate participants had a similar percentage with any heavy drinking days compared to placebo; descriptively, they did a bit worse. In the graphs below, a shorter bar indicates fewer participants with a heavy drinking day, a better outcome (i.e., shorter is better than taller in these graphs.)
Outcomes were similar when examining time to a heavy drinking day as well
Authors also examined the number of days it took until an individual had a heavy drinking day (note: some made it the entire trial without a single heavy drinking day as described above, and they are counted as having made it 84 days, the length of the 12-week trial). Like the results for any heavy drinking day, for reward drinkers only, naltrexone outperformed placebo (72 days vs. 46 days to first heavy drinking day). For relief drinkers though, acamprosate again did no better than placebo.
Efforts at replicating observational findings where matching effects are initially identified in rigorous trials and effects are hypothesized ahead of time (i.e., “a priori”) have been largely unsuccessful to date. For example, in Project MATCH, a large randomized controlled trial comparing 12-step facilitation, motivational enhancement therapy, and cognitive-behavioral therapy, specifically looked at whether one treatment would be better than another for certain kinds of individuals. Although many of these treatment “matches” had been supported in smaller previous trials, very few matches were actually found. In another medication trial, authors did not find their hypothesized matching effect where they expected individuals with a gene variant (in the OPRM1 gene) to respond better to naltrexone, despite this being found after the fact in a previous study. In this initial study, authors found their hypothesized matching effect, that for reward drinkers, naltrexone may be a better medication for them than acamprosate.
Importantly, if this study’s findings are replicated, based on the types of measures used, treatment programs and other health care providers may be able to identify reward drinkers because reward/relief drinking was assessed using an easy-to-administer questionnaire. Although in this study, reward drinkers only did better on naltrexone compared to those who were neither reward or relief drinkers, separate analyses relevant to clinical settings showed individuals did better on naltrexone too when, on the drinker-type questionnaire, reward scores were greater than relief scores and the reward score is between 12-31. Such a distinction is easy to make in health care settings and could be applied in a straightforward manner.
While authors speculate that acamprosate may indeed be better for relief drinkers if they enter treatment with just a few abstinent days or less (rather than several days or more as was the case in this study), this would need to be tested scientifically. Also, the study examined oral daily dosing of naltrexone; it will be important to assess whether there are similar matching effects from extended release/once per month injectable formulations of naltrexone (also known by its brand name Vivitrol). There are lots of caveats to these findings (see below), and replications to confirm these findings are needed in other samples before being clinically recommended. There is much more to be learned about whether, and how, to match individuals to specific treatments before systematic precision medicine approaches can be undertaken. That said, the model by which authors developed and tested theoretically-driven matching hypotheses to inform a precision medicine approach is one that can be applied in other studies of substance use disorder medication and psychosocial interventions.
Mann, K., Roos, C. R., Hoffmann, S., Nakovics, H., Lemenager, T., Heinz, A., & Witkiewitz, K. (2018). Precision Medicine in Alcohol Dependence: A Controlled Trial Testing Pharmacotherapy Response Among Reward and Relief Drinking Phenotypes. Neuropsychopharmacology, 43(4), 891-899. doi:10.1038/npp.2017.282
l
When studies compare different medicines or psychosocial treatments for alcohol use disorder to each other, they often produce similar effects where none have a clear advantage. One explanation may be the substantial variability of individuals with alcohol use disorder; they vary in terms of the severity of their disorder, other medical and mental health conditions with which they are also diagnosed, and resources they have at their disposal (e.g., income, stable housing), to name just a few. The next important question becomes: For whom do these medications work best? This is also known as “precision medicine”; certain healthcare interventions may be better for some than for others and treatment should be tailored to these individuals more precisely. Because the FDA-approved alcohol use disorder medication, naltrexone, affects brain functioning associated with experiences of pleasure and reward (e.g., fun, excitement, euphoria) and the FDA-approved medication, acamprosate, is thought to affect brain functioning associated with relief (e.g., reducing stress, anxiety, depression), Mann and colleagues conducted a “precision medicine” type study to test if naltrexone actually is indeed more effective for individuals who drink for reward and whether acamprosate is actually more effective for individuals who drink for relief.
This study was a secondary analysis of a primary, 12-week randomized controlled trial for alcohol use disorder (“alcohol dependence” in this study based on the 4th edition of the diagnostic and statistical manual of mental disorders) that compared 169 individual receiving daily oral naltrexone, 172 individuals receiving daily oral acamprosate, and 85 receiving a placebo medication. In addition to alcohol use disorder, to be included in the study, individuals were drinking in the month before the study at least 14 standard drinks per week for women and 21 for men. The study excluded individuals with history of other substance dependence (e.g., cannabis use disorder) and current psychiatric disorders such as major depressive or anxiety disorders. The primary study outcome was heavy drinking; that is, drinking five or more drinks in a day for men, or four or more for women. The average participant was 45 years old and 77 percent were male.
From a questionnaire assessing how frequently individuals drank in 15 “reward” situations (e.g., parties/celebration, having fun; out with friends and stopped by a bar) and 15 “relief” situations (e.g., feeling upset, feeling hopeless, criticized by others), authors used a sophisticated statistical procedure to create four groups of drinkers: 1) High Reward/High Relief; 2) High Reward/Low Relief; 3) Low Reward/High Relief; and 4) Low Reward/Low Relief. All analyses controlled statistically for gender, age, drinking severity, and smoking status, to isolate the effects of the medication on drinking outcomes – making a more compelling case that any group differences were caused by the medication and not by these other factors.
Individuals high in “reward” drinking and high in “relief” drinking had other characteristics that fit these profiles
For any heavy drinking, naltrexone was better than placebo for the reward drinkers, but acamprosate was no better than placebo for relief drinkers
For reward drinkers (upper graph), naltrexone produced a 50 percent reduction in the likelihood of having a heavy drinking day compared to placebo, but had similar effects for drinkers who were low in both reward and relief drinking, as the authors hypothesized. For relief drinkers (lower graph), though, acamprosate participants had a similar percentage with any heavy drinking days compared to placebo; descriptively, they did a bit worse. In the graphs below, a shorter bar indicates fewer participants with a heavy drinking day, a better outcome (i.e., shorter is better than taller in these graphs.)
Outcomes were similar when examining time to a heavy drinking day as well
Authors also examined the number of days it took until an individual had a heavy drinking day (note: some made it the entire trial without a single heavy drinking day as described above, and they are counted as having made it 84 days, the length of the 12-week trial). Like the results for any heavy drinking day, for reward drinkers only, naltrexone outperformed placebo (72 days vs. 46 days to first heavy drinking day). For relief drinkers though, acamprosate again did no better than placebo.
Efforts at replicating observational findings where matching effects are initially identified in rigorous trials and effects are hypothesized ahead of time (i.e., “a priori”) have been largely unsuccessful to date. For example, in Project MATCH, a large randomized controlled trial comparing 12-step facilitation, motivational enhancement therapy, and cognitive-behavioral therapy, specifically looked at whether one treatment would be better than another for certain kinds of individuals. Although many of these treatment “matches” had been supported in smaller previous trials, very few matches were actually found. In another medication trial, authors did not find their hypothesized matching effect where they expected individuals with a gene variant (in the OPRM1 gene) to respond better to naltrexone, despite this being found after the fact in a previous study. In this initial study, authors found their hypothesized matching effect, that for reward drinkers, naltrexone may be a better medication for them than acamprosate.
Importantly, if this study’s findings are replicated, based on the types of measures used, treatment programs and other health care providers may be able to identify reward drinkers because reward/relief drinking was assessed using an easy-to-administer questionnaire. Although in this study, reward drinkers only did better on naltrexone compared to those who were neither reward or relief drinkers, separate analyses relevant to clinical settings showed individuals did better on naltrexone too when, on the drinker-type questionnaire, reward scores were greater than relief scores and the reward score is between 12-31. Such a distinction is easy to make in health care settings and could be applied in a straightforward manner.
While authors speculate that acamprosate may indeed be better for relief drinkers if they enter treatment with just a few abstinent days or less (rather than several days or more as was the case in this study), this would need to be tested scientifically. Also, the study examined oral daily dosing of naltrexone; it will be important to assess whether there are similar matching effects from extended release/once per month injectable formulations of naltrexone (also known by its brand name Vivitrol). There are lots of caveats to these findings (see below), and replications to confirm these findings are needed in other samples before being clinically recommended. There is much more to be learned about whether, and how, to match individuals to specific treatments before systematic precision medicine approaches can be undertaken. That said, the model by which authors developed and tested theoretically-driven matching hypotheses to inform a precision medicine approach is one that can be applied in other studies of substance use disorder medication and psychosocial interventions.
Mann, K., Roos, C. R., Hoffmann, S., Nakovics, H., Lemenager, T., Heinz, A., & Witkiewitz, K. (2018). Precision Medicine in Alcohol Dependence: A Controlled Trial Testing Pharmacotherapy Response Among Reward and Relief Drinking Phenotypes. Neuropsychopharmacology, 43(4), 891-899. doi:10.1038/npp.2017.282
l
When studies compare different medicines or psychosocial treatments for alcohol use disorder to each other, they often produce similar effects where none have a clear advantage. One explanation may be the substantial variability of individuals with alcohol use disorder; they vary in terms of the severity of their disorder, other medical and mental health conditions with which they are also diagnosed, and resources they have at their disposal (e.g., income, stable housing), to name just a few. The next important question becomes: For whom do these medications work best? This is also known as “precision medicine”; certain healthcare interventions may be better for some than for others and treatment should be tailored to these individuals more precisely. Because the FDA-approved alcohol use disorder medication, naltrexone, affects brain functioning associated with experiences of pleasure and reward (e.g., fun, excitement, euphoria) and the FDA-approved medication, acamprosate, is thought to affect brain functioning associated with relief (e.g., reducing stress, anxiety, depression), Mann and colleagues conducted a “precision medicine” type study to test if naltrexone actually is indeed more effective for individuals who drink for reward and whether acamprosate is actually more effective for individuals who drink for relief.
This study was a secondary analysis of a primary, 12-week randomized controlled trial for alcohol use disorder (“alcohol dependence” in this study based on the 4th edition of the diagnostic and statistical manual of mental disorders) that compared 169 individual receiving daily oral naltrexone, 172 individuals receiving daily oral acamprosate, and 85 receiving a placebo medication. In addition to alcohol use disorder, to be included in the study, individuals were drinking in the month before the study at least 14 standard drinks per week for women and 21 for men. The study excluded individuals with history of other substance dependence (e.g., cannabis use disorder) and current psychiatric disorders such as major depressive or anxiety disorders. The primary study outcome was heavy drinking; that is, drinking five or more drinks in a day for men, or four or more for women. The average participant was 45 years old and 77 percent were male.
From a questionnaire assessing how frequently individuals drank in 15 “reward” situations (e.g., parties/celebration, having fun; out with friends and stopped by a bar) and 15 “relief” situations (e.g., feeling upset, feeling hopeless, criticized by others), authors used a sophisticated statistical procedure to create four groups of drinkers: 1) High Reward/High Relief; 2) High Reward/Low Relief; 3) Low Reward/High Relief; and 4) Low Reward/Low Relief. All analyses controlled statistically for gender, age, drinking severity, and smoking status, to isolate the effects of the medication on drinking outcomes – making a more compelling case that any group differences were caused by the medication and not by these other factors.
Individuals high in “reward” drinking and high in “relief” drinking had other characteristics that fit these profiles
For any heavy drinking, naltrexone was better than placebo for the reward drinkers, but acamprosate was no better than placebo for relief drinkers
For reward drinkers (upper graph), naltrexone produced a 50 percent reduction in the likelihood of having a heavy drinking day compared to placebo, but had similar effects for drinkers who were low in both reward and relief drinking, as the authors hypothesized. For relief drinkers (lower graph), though, acamprosate participants had a similar percentage with any heavy drinking days compared to placebo; descriptively, they did a bit worse. In the graphs below, a shorter bar indicates fewer participants with a heavy drinking day, a better outcome (i.e., shorter is better than taller in these graphs.)
Outcomes were similar when examining time to a heavy drinking day as well
Authors also examined the number of days it took until an individual had a heavy drinking day (note: some made it the entire trial without a single heavy drinking day as described above, and they are counted as having made it 84 days, the length of the 12-week trial). Like the results for any heavy drinking day, for reward drinkers only, naltrexone outperformed placebo (72 days vs. 46 days to first heavy drinking day). For relief drinkers though, acamprosate again did no better than placebo.
Efforts at replicating observational findings where matching effects are initially identified in rigorous trials and effects are hypothesized ahead of time (i.e., “a priori”) have been largely unsuccessful to date. For example, in Project MATCH, a large randomized controlled trial comparing 12-step facilitation, motivational enhancement therapy, and cognitive-behavioral therapy, specifically looked at whether one treatment would be better than another for certain kinds of individuals. Although many of these treatment “matches” had been supported in smaller previous trials, very few matches were actually found. In another medication trial, authors did not find their hypothesized matching effect where they expected individuals with a gene variant (in the OPRM1 gene) to respond better to naltrexone, despite this being found after the fact in a previous study. In this initial study, authors found their hypothesized matching effect, that for reward drinkers, naltrexone may be a better medication for them than acamprosate.
Importantly, if this study’s findings are replicated, based on the types of measures used, treatment programs and other health care providers may be able to identify reward drinkers because reward/relief drinking was assessed using an easy-to-administer questionnaire. Although in this study, reward drinkers only did better on naltrexone compared to those who were neither reward or relief drinkers, separate analyses relevant to clinical settings showed individuals did better on naltrexone too when, on the drinker-type questionnaire, reward scores were greater than relief scores and the reward score is between 12-31. Such a distinction is easy to make in health care settings and could be applied in a straightforward manner.
While authors speculate that acamprosate may indeed be better for relief drinkers if they enter treatment with just a few abstinent days or less (rather than several days or more as was the case in this study), this would need to be tested scientifically. Also, the study examined oral daily dosing of naltrexone; it will be important to assess whether there are similar matching effects from extended release/once per month injectable formulations of naltrexone (also known by its brand name Vivitrol). There are lots of caveats to these findings (see below), and replications to confirm these findings are needed in other samples before being clinically recommended. There is much more to be learned about whether, and how, to match individuals to specific treatments before systematic precision medicine approaches can be undertaken. That said, the model by which authors developed and tested theoretically-driven matching hypotheses to inform a precision medicine approach is one that can be applied in other studies of substance use disorder medication and psychosocial interventions.
Mann, K., Roos, C. R., Hoffmann, S., Nakovics, H., Lemenager, T., Heinz, A., & Witkiewitz, K. (2018). Precision Medicine in Alcohol Dependence: A Controlled Trial Testing Pharmacotherapy Response Among Reward and Relief Drinking Phenotypes. Neuropsychopharmacology, 43(4), 891-899. doi:10.1038/npp.2017.282
SIGN UP FOR THE BULLETIN
