GLP-1 receptor agonists for alcohol dependence??

Here are 2 pretty convincing articles suggesting that GLP-1 receptor agonists  (GLP-1RAs) might be important medications to treat alcohol use disorder (AUD) 

 

---------------------------------------------------------

 

1. a rat study: alcohol dec with semaglutide rats lancet2023 in dropbox, or doi.org/10.1016/j.ebiom.2023.104642

-- male and female rats were studied, rats that were bred to have a voluntary high and stable alcohol consumption that created pharmacologically relevant blood alcohol concentrations 

-- acute and repeated semaglutide administration led to significant reductions in alcohol intake and prevented relapse-like drinking in both male and female rats (relapses being simulated by withdrawing semaglutide after it had led to decreased alcohol consumption, rats then resuming drinking, then reinitiating semaglutide with good effect) 

    -- there was no decrease in drinking water or total liquid intake (there actually was a little increase in female rats) 

    -- there was a decrease in food intake 

        -- interestingly, there was also a decrease in "rewarding foods" for rats (peanut butter and Nutella), though an increase in rat chow intake (presumably not so rewarding....) 

-- the nucleus accumbens (NAc), the area of the brain associated with the mesolimbic "reward" dopamine-associated system (see below for more information), had reduced alcohol-induced activation when receiving semaglutide (ie less dopamine present)

    -- there was also fluorescently-labeled semaglutide detected in the NAc system in the brain

-- semaglutide enhanced the metabolism of dopamine in the NAc of mice as well as of rats: the researchers found that the attenuated dopamine release by semaglutide was through enhanced dopamine metabolism rather than increased dopamine reuptake 

-- similar results have been found with dulaglutide (a weaker GLP-1RA than semaglutide), but more so in male than female rodents 

 

2. human exenatide study: alcohol dec with GLP1 JCIinsight2023 in dropbox, or doi.org/10.1172/jci.insight.159863

-- 127 participants randomized to the GLP-1RA exenatide 2mg subq vs placebo weekly for 26 weeks, with followup for 6 months more 

    -- all had at least 5 heavy drinking days in the prior 30 days (60 g of alcohol for men and 48g for women) 

    -- all patients were white, mean age 52, 60% men 

    -- exenatide was chosen because it crossed the blood-brain barrier (not so for semaglutide or dulaglutide), and, along with nicotine patches, has been found effective in nicotine dependence (and also decreased post-cessation weight gain) 

-- standard cognitive behavioral therapy was available for both groups 

-- a subgroup of 25 healthy controls were matched for sex, age, and educational status and received function MRI (fMRI) and single-photon emission CT (SPECT) brain scans 

 

-- primary endpoint: reduction of heavy drinking days, total alcohol consumption in past 30 days, days without alcohol consumption, the PACS craving screen, and AUDIT AUD screen 

 

Results: 

-- 58 if the 127 participants completed the study, 55 had the 6-month followup. no difference between the groups in terms of trial discontinuation

-- exenatide was associated with 0.95 decrease in BMI, 1.6 decrease in A1c, and lower (ie better) Drug Use Disorders Identification Test (DUDIT) by 0.96 points (0.7-1.3), p<0.001 ( a score of at least 6 in men and 2 in women suggests drug-related problems) 

-- endpoints (comparing placebo vs exenatide): 

    -- total number of heavy drinking days: 27 days in the placebo group vs 20 in the exenatide group), non-significant 

    -- total alcohol consumption for 30 days: 1313 grams vs 1,304 grams, non-significant 

    -- days without alcohol consumption (percentage points): 21 vs 11, non-significant 

    -- PACS (Penn Alcohol Craving Scale): -7.3 vs -5.4, non-significant 

    -- AUDIT (Alcohol Use Disorders Identification Test): -8.2 vs 7.0, non-significant 

-- further analysis by BMI subgroups: 

    -- BMI >30: exenatide associated with (vs placebo): 

        -- reduced heavy drinking days by 23.6 percentage points (-44.4 to -2.7), p=0.034

        -- reduced total alcohol intake for the prior month by 1,205 grams (-2206 to -204g), p=0.026 

    -- BMI <25: exenatide associated with (vs placebo):

        -- increased heavy drinking days by 27.5 percentage points (4.7to 50.2), p=0.024

        -- no significant difference in total alcohol intake for prior month

-- there was only a very weak correlation between both total alcohol intake and change in heavy drinking with changes in A1c levels 

-- at 6-month follow-up: not much change, though slightly higher AUDIT score (5.1 points) in the exenatide group 

 

 

-----------

-- significant difference in fMRI for alcohol cue reactivity (ie, response to the alcohol cue) in the ventral striatum and septal areas (crucial for drug reward and addiction) after 26 weeks on exenatide (though was the same as in the placebo group at baseline) 

-- decreased dopamine transporter availability in the exenatide vs placebo group 

-- not much change in the SPECT studies 

 

-- safety issues: higher incidence of GI symptoms, weight loss, fatigue and injection site reactions in the exenatide group: 

    -- nausea: 37% vs 15% 

    -- decreased appetite: 24% vs 9% 

    -- vomiting: 23% vs 8% 

    -- wt loss: 68% vs 40% 

    -- injection site reactions: 41% vs 0% (exenatide injections often cause subcutaneous hard nodules, but not found in other GLP-1RA meds) 

    -- no cases of pancreatitis or elevated pancreatic enzymes

  

Commentary: 

-- AUD has significant morbidity/mortality in Denmark (site of the exenatide study): cumulative 10-yr mortality 30% after first hospital encounter for an alcohol problem 

-- for the FDA approved meds for AUD (disulfuram, naltrexone, acamprosate), 40% alcohol relapse within 3 years 

 

-- there are an array of mechanisms by which GLP-1RA's could be effective in decreasing alcohol use: 

    -- there is a complex set of findings suggesting that the gut-brain peptide GLP-1 modulates the development of AUD: 

        -- different animal studies have found that several GLP-1RAs suppress alcohol intake, alcohol-seeking behavior, and alcohol-induced activation of the mesolimbic dopamine system in male rodents as well as female rats. Activation of this dopamine system is associated with "rewarding" behaviors (ie, makes one feel good) in animals, but also in humans, where there is an alcohol-related  increase in dopamine in the nucleus accumbens (NAc). there is also a study in alcohol-preferring vervet monkeys showing benefit of GLP-1RAs in decreasing alcohol intake (they found that liraglutide was somewhat better than exenatide) 

            -- GLP-1 receptors are found in the areas of the brain associated with reward and addiction in many animal studies: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7848227/

        -- it is this same dopamine "reward" system involved in obesity and drugs, including alcohol, opiates, methamphetamine, cocaine and nicotine (for example, see addiction and obesity both stimulate dopamine ObesRev2013 in dropbox, or doi:10.1111/j.1467-789X.2012.01031.x) 

            -- and we know well that GLP-1RAs are really effective in treating obesity

        -- polymorphisms of the GLP-1 receptor (rendering it less effective) are associated with increased alcohol consumption, as well as AUD, in humans

        -- as with patients with diabetes, people with AUD also have an impaired incretin response (reduction in endogenous secretion of GLP-1 with a meal), leading to less insulin secretion when food hits the stomach https://www.sciencedirect.com/science/article/abs/pii/S0026049503004189?via%3Dihub ).And the responses in both cases to GLP-1RAs are changes in perceptions of food (see below), weight loss, dramatic decreases in A1c in diabetic individuals, and the attendant medication adverse effects

    -- so, it seems that the non-specificity of the dopamine system leads to similar “reward”  responses to food, alcohol, and drugs (there are suggestive anecdotal data as well as animal studies that opiates, cocaine, nicotine and methamphetamine use all decrease), as it does with diabetes. And it makes intuitive sense that GLP-1RAs would have similar benefits for these seemingly quite diverse stimuli by attacking the dopamine reward system

 

-- overall, the exenatide study in humans did not find benefit over placebo in reducing the number of heavy drinking days. Several postulated explanations for this counterintuitive finding: 

    -- animal studies have found more benefit in those consuming higher levels of alcohol. 80% of people in the exenatide study did not meet DSM-5 criteria for severe AUD. And prior studies of meds for AUD (topiramate, naltrexone, acamprosate) had participants with more severe AUD 

    -- the placebo effect in AUD studies is quite profound, as was found in the exenatide study (see figure above), and this would dilute the potential benefit of exenatide (and, perhaps a stronger GLP-1RA than the 2mg dose of exenatide might have been more successful in reaching statistical and clinical significance). Also, the placebo group in this study did receive the same cognitive behavioral therapy against AUD as the exenatide group, perhaps adding to the "placebo benefit" 

    -- only 58 if the 127 participants completed the study, 55 had the 6-month followup. no difference between the groups in terms of trial discontinuation, but this high drop-out rate could affect the impact on the results 

    -- people with AUD have been found to have decreased executive function, and this exenatide study did find reduced "cue reactivity" in the dorsal prefrontal cortex in those on exenatide vs the placebo group (ie, the changes in executive function and response to stimuli might be impaired in those with high levels of chronic alcohol intake, limiting the GLP-1RA effect in some individuals)

 

-- it is not entirely clear why the exenatide study found benefit only in those would had high BMI, but there are a few potential reasons:

    -- it may related to the observed significant heterogeneity of AUD:

        -- the effect of GLP-1RAs on obesity vs AUD may be more complex than the points mentioned above. it could be that those with both obesity and AUD have a more profound disturbance of the incretin system than having either one in isolation. And the effect of GLP-1RAs is more profound in those with both conditions. 

        -- Or, perhaps one needs a stronger GLP-1RA than the exenatide used in the human study above to be effective in AUD in people with lower a normal BMI (eg semaglutide is the most potent one we have now, it has to most weight loss, and it is available in an oral form).

            -- The studies on tirzepatide, which has a double action of GLP1-RA and glucose-dependent insulinotropic polypeptide agonist, seems to be more effective than semaglutide in weight loss (see weight loss tirzepatide better than semaglut DiabObesMed2023 in dropbox, or DOI: 10.1111/dom.14940, which found that tirzepatide was actually less expensive than semaglutide on a % change in weight. for the original SURMOUNT-1 study: https://www.nejm.org/doi/pdf/10.1056/nejmoa2206038. perhaps this double effect of tirzepatide would work better for AUD, etc

 

-- one interesting finding here is a contradiction of the generally held assumption that meds that do not cross the blood-brain barrier cannot really affect the brain directly. With the GLP-1RAs, there is clearly a direct effect on the brain's GLP-1 receptors, even with meds that have been shown not to cross the blood-brain barrier (semaglutide, liraglutide, dulaglutide). so, exenatide was chosen in the above study because it does cross that barrier since it is lipophilic (https://www.dovepress.com/the-relationship-between-the-blood-brain-barrier-and-the-central-effec-peer-reviewed-fulltext-article-DMSO ). Unclear why these other GLP-1RAs work. Perhaps a high blood alcohol level affects the permeability of the blood-brain barrier? Perhaps because alcohol leads to chronic inflammation (https://www.wjgnet.com/1007-9327/full/v16/i11/1304.htm ) and inflammation can decrease the blood-brain barrier? Maybe there are changes of these meds when in the blood than allow them to cross the barrier? But, bottom line, the choice of GLP-1RAs does not seem to matter. it is likely the potency and tolerability of the chosen one that does matter. 

 

-- on the anecdotal side, i was stimulated to do this blog because of a recent NPR report on semaglutide's benefits on alcohol (https://www.npr.org/sections/health-shots/2023/08/28/1194526119/ozempic-wegovy-drinking-alcohol-cravings-semaglutide ), as well as a previous article in the NYTimes https://www.nytimes.com/2023/02/24/well/eat/ozempic-side-effects-alcohol.html. Both have lots of anecdotal reports of semaglutide leading to less desire to drink alcohol, some people who had been regular wine drinkers reporting to being "repelled" by alcohol. A small trial did find that semaglutide led to less hunger, fewer food cravings, lower preference for high-fat foods and lower total energy intake. People with chronic compulsive disorders (eg gambling, online shopping) also reported benefit

 

So, interesting stuff…

    -- as I have mentioned in the slew of blogs on GLP1-RAs for diabetes management, they are a game changer:

        -- huge decreases in A1c levels, some patients of mine going from unmovable A1C concentrations in the 13-16% range now in the 6-7% range

        -- significant positive effects on renal and cardiac clinical effects

        -- not many med-stopping adverse effects (and often lower doses work fine if there is a problem), mostly GI effects, and some patients losing too much weight

        -- they basically work by normalizing the dysfunctional incretin effect in people with diabetes by supplementing their low GLP-1 levels

        -- studies on semaglutide have found that the amount of  associated weight loss is roughly equivalent to the surgical sleeve gastrectomy procedure

    -- and, now as found in many animal studies with some backing from the human exenatide one above, it appears that the GLP-1RAs work for an array of substances that augment the dopamine reward system (eg alcohol, drugs, cigarettes)

        -- but, as per the usual critique, there really need to be high-quality studies to document the benefit of GLP-1RAs

    -- however, based on this alcohol study, i will now check out to see if my patients on GLP-1RAs do decrease alcohol, opiates/cocaine, nicotine, and other "rewarding" events (chocolate, coffee....)

 

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

  

to get access to all of the blogs:

 

 go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

  -- click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

  -- or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them

 

if you would like to see the articles in this blog, please email me. 

 

please feel free to circulate this to others. also, if you send me their emails (gmodest@bidmc.harvard.edu), i can add them to the list