Loading summary
Transcript25 lines
- [00:01]
A
This is endocrine feedback loop. I am your host Chase Hendrickson and welcome you to this Journal Club Podcast series brought to you by the Endocrine Society. Thanks for joining us as we explore an important article recently published in one of the Society's clinical journals. Hello and welcome back to the Endocrine Feedback Loop podcast. For our 71st episode today we return to the JCEM to look at a clinical trial that attempts to understand the weight loss mechanisms of GLP1 receptor agonists, specifically semaglutide's effect on taste. It has long been recognized that we don't understand all the ways that GLP1s exert their beneficial effects, so this article stood out as an important one for us to discuss on this podcast. RCTs often get held up as the gold standard for evidence based medicine and for good reason. However, all study designs have limitations and we will be sure to discuss the ones that relate to this investigation. As usual, we will spend time considering the relevant clinical background for this study and how we think the results advance the care we as endocrinologists provide to our patients. I host the Endocrine Feedback Loop podcast working as a general endocrinologist and Medical Director at Vanderbilt Health in Nashville, Tennessee. Back with me today is this episode's regular contributor is NA Shin, one of the podcast's expert in obesity medicine. She comes to us from the University of California, Los Angeles where she founded and runs the program for reducing obesity in their ambulatory clinics. Her research interests include quality improvement and expanding obesity care access. Our guest expert today joins us from the Mayo Clinic in Rochester, Minnesota. Kolpana Muthasamy focuses her clinical work on weight management and diabetes. You all as our listeners will recognize her from her research and talks in these areas. So as usual, the perfect pair of endocrinologists joins me today to help us understand this research. As is also always the case, everything we discuss will be our opinions only and not necessarily those of our respective institutions or of the immigrant society. Today we review semaglutide and taste in women with obesity and Polycystic Ovary Syndrome, a randomized placebo controlled study which the Journal of Clinical Endocrinology and metabolism published in January 2026. Moytse Jensterli served as the first author for this paper, which comes to us from the University Medical Center Ljubljana in Slovenia. I will now turn things over to na. She will highlight the main points that the authors make in their introduction and get Kolpana to give us some insight along the way.
- [02:41]
B
Na thank you chase for that lovely introduction. Let's get started. So the authors start by stating that the relationship between obesity and sense of taste is complex with inconsistent and conflicting findings. Kalpana, can you describe to us briefly what is known between obesity and taste?
- [03:03]
C
Absolutely not. Taste is something that we experience every single day of our life, but we often don't think about the science behind it. I believe during the COVID pandemic years we talked a little bit more about taste or the lack thereof. But with the rise in use of GLP1 receptor agonists, we've had a little more scientific conversations about eating behaviors and the different factors that influence that. Taste is one dimension of that, but it really doesn't get as much traction. As we dive into the nuances of this paper, I would like to remind ourselves of the basics of taste transduction. When a taste tint is placed in the mouth, it activates the taste receptor cells in the taste buds which are located in the papilla in the tongue. And these signals are carried to the brain. They integrate with smell and other aspects of food like temperature, texture and it creates a complex perception. Now if we look at bioevolution, taste has been a big gatekeeper of what we eat and what we shouldn't eat. But fast forwarding, specifically looking at individuals with obesity, there is some association between taste and weight in individuals with obesity. There is data that shows there is decrease in the density of papilla in the tongue. And also taste perception, particularly for sweet and salt taste, are decreased in individuals with a higher bmi. And there is also a higher threshold for sweet sensation in individuals with obesity, meaning that you need more of the stimuli to create that sweet perception. Some human studies also show a decrease in taste associated gene expression in individuals with obesity. But we don't really know what drives this, what the mechanism behind this is. There are some mice studies where obesity related inflammatory cascades have been associated with decrease in taste bud abundance. And also diet seems to play a role. There are animal studies that show that animals fed with a high fat, high sugar diet seems to decrease the central processing of sweet taste. We really don't know a lot about causality in this area. Does obesity actually impair taste function or whether impaired taste drives overconsumption and weight gain?
- [05:19]
B
Thank you Kalpana for that very lovely and thorough overview. So semaglutide is a glucagon like peptide 1 analog that causes weight loss by reducing appetite and lowering the desire for energy dense foods, reducing food intake. It has been reported that patients treated with semaglutide have reduced desires for sweet, savory and salty foods. However, the mechanisms like you mentioned earlier underlying this are not well known. The hypothesis for this article is we hypothesized that semaglutide would change the taste recognition in women with obesity and polycystic ovarian syndrome, or pcos. All right, Chase, I'll give it back to you.
- [06:07]
A
We will move into the methods and as usual, think about the study design. I mentioned in the introduction of the episode that we always hold up RCTs as the gold standard, and that's for good reason. We look at a lot of observational studies on this podcast. We always talk about some of the challenges putting those into two large categories, and that's of bias and confounding that happens. And RCTs tend to be so effective because they can typically deal with both of those issues. Bias in medical research is when one group is treated differently than another. And the reason that RCTs at least can be helpful with that is because there is blinding. You don't know which group somebody is in. So that's the subjects themselves don't know which group they're in. And the study personnel, anybody looking at the data, they don't know that either. And so you can't treat somebody differently if you don't know which group that they're in. Come back to that here in just a second. The other great benefit is that confounding can be dealt with, and that's with randomizing. Confounding doesn't go away. That is still a potential problem. But if you have a large enough group and you've effectively randomized, then you balance the confounders way of speaking. They cancel each other out so you don't have to worry about them driving the results that you're seeing. So in General, that's why RCTs are so effective. A couple of comments with that. This study that we're going to look at is single blinded. So the subjects themselves didn't know, or at least they weren't told, if they were taking an active medication or a placebo. Those study personnel well or were aware of that. And so one thing that we have to keep in mind is that with the data, if it's focused on objective data, that tends to be less of an issue. When it becomes more subjective data, or at least there is some input that will be necessary from the investigators, from the study personnel, then you would get highly worried about them not being blinded in an investigation. Last comment that I would make in general about study design is that that blinding of subjects, the first way to think about it is whether they were told if they were on a medication, yes or no. But you can also think about effective blinding. And the reason that you have to think about is even if a subject wasn't told whether they were taking a placebo or an active agent, sometimes they are going to be able to figure that out. And if so, even if nobody ever told them, then you would have to think that the blinding itself was not actually effective and then you have to worry about whether that might have an effect. We'll come back to that in a second. But is particularly relevant with medications like GLP1s where there are a substantial number of GI side effects and maybe not bad enough to make somebody drop out of a study, but potentially enough to where you would be aware that you were actually on a medication as opposed to a placebo. Okay, we're going to come back to all those things here in just a second for this investigation. First of all, we'll go through the inclusion criteria. They mentioned this in the title already and you could guess some of this stuff, but these were all women between the ages of 18, all the way up to menopause. BMI had to be 30 or above. And also they all had a diagnosis of pcos and specifically that was phenotype A and they couldn't have had any serious chronic illness. We'll go on to some of the other exclusion criteria, but before we do that, Culp and I wanted to get your thoughts on this. The authors don't go into great detail about why this specific population was used. You would at least wonder if maybe it was a population of convenience that they had already that they were able to investigate. They do make some arguments to this is a relatively specific population and so maybe takes out a few other pieces that you might worry about. But I was interested in your thoughts on this. Should we view this as indeed a strength that we've got a fairly narrowly defined population that we're looking at, or should we view this as a weakness of. Well, is this going to be generalizable to many different individuals? Or maybe it doesn't have an effect one way or the other? What's your take on this specific population being who was investigated here?
- [10:03]
C
Well, Chase, I would say that it falls in both categories. It is an interesting thought provoking strategy when we take into account that there is some data available on estrogen influence, on taste perception especially. There are estrogen receptors in the tongue and the taste buds as well. And in the periobulatory phase, it's known that there can be an increased sensitivity to sweet stimuli and women in their luteal phase actually have increased preference for sweet taste compared to the follicular phase, whereas men don't have the cyclical variation. So we can sort of think about the benefits of using a homogenous group with less cyclical variation being beneficial to detect small changes. But clearly it would affect the generalizability of the outcome, particularly given that just very simply females seem to have better taste ability than males. So I guess that would be excluding generalizability to a big segment of our population, but excluding chronic diseases, smoking and some of the known things that could also have an effect on taste, I think adds to their strength.
- [11:15]
A
There are quite a few exclusion criteria. I'll just highlight the relevant ones given that this is investigation of a GLP1. So specifically you couldn't have diabetes, either type 1 or type 2, couldn't have coronary artery disease or heart failure, and then couldn't have had a history of use of weight loss medications or having had bariatric surgery. And again, quite a few other ones. But I think those are the most important ones that were exclusion criteria in this study, the way the medication was given. So the subjects were randomized and it was a one to one randomization. And so you either got put on a placebo or semaglutide. You were eventually going to get worked up to 1 milligram though it was a fairly standard up titration starting with that low dose of just 0.25 milligrams weekly. And the study collected data. So after randomization and at the end of study, which was 16 weeks, participants underwent both anthropometric and clinical analysis. We're going to hear more about that. They also did extensive biochemical testing and an assessment of taste recognition using taste strips test. And we're going to hear a little bit more about that in a second. More data was obtained with tongue biopsies. So these biopsies were done both at baseline and again after the 16 weeks of this study. They did it in both groups, semaglutide and placebo group. And this was primarily done for RNA testing. And finally the other big piece of this is that there was an assessment of brain responses and those were acquired using functional mri. We'll hear a little bit more about that in a second. The outcomes that the authors were interested here, the primary outcome was a change in overall taste recognition and it was A score, and it was based on different concentrations of basic tests as detected by what are called impregnated strips. Several secondary outcomes that you can guess from what they were measuring here. So one was an alteration of a transcriptomic profile, and that was from that tongue tissue biopsy done with RNA sequencing, and then the other one was changes in brain responses on that functional mri. A couple of comments on the statistics. So the authors did a power calculation and they determined that they needed just 15 subjects in each group in order to carry out their investigation. I'll just make a comment on that. In general, is when you're doing power calculations, the way that you build this and you figure out how many subjects you're going to need in your group is largely dependent upon how much of an effect you think a given intervention is going to have. The idea is that if you think an intervention, whatever it is, is only going to have a fairly small effect, you need way more individuals to be able to prove that that small change was statistically significant. However, if you think that there is going to be a very big, a substantial effect from an intervention, generally you don't need nearly as many subjects. So just based on that general perspective, I'd say the fact that they said, well, we only need 15 subjects in each group, I think you could probably infer that the authors anticipated seeing a fairly large impact, a major impact, just with the inclusion of semaglutide here. There are other things that go into a power calculation. I've oversimplified it substantially, but that was one of the things that I took away from the mention of the statistics and specifically the power calculation. We're going to come back to a lot of these things as we go along. So I'm going to turn things over to Nan now and she's going to start going through the results based on all these tests that the authors did.
- [14:44]
B
All right, so if you look at Table 1, it shows a total of 30 participants completing the study, with 15 in the placebo and 15 in the semaglutide treatment group. Baseline characteristics and demographics were all very similar, with a mean age of 33.7 plus or minus 6.1 years and a mean BMI of 36.4 plus or minus 4.4 kilos over meter squared. If you look at Table 2, it does show that the semaglutide treatment group compared to placebo group had a statistically significant reductions in body weight, BMI, waist circumference, visceral adipose tissue mass, hemoglobin, A1C 120 minute glucose, oral glucose tolerance test, HDL and increase in heart rate and if you look at Figure 3 Panel A shows that the semaglutide treatment group improved their taste recognition score from 11.9 plus or minus 1.9 points to 14.4 plus or minus 1 points out of a total of 16 points with an estimated treatment difference of 2.5 points. So Kalpana, what are your initial thoughts on these findings now?
- [16:05]
C
I guess they were power to detect a difference of 1.6 so a 2.5 point difference met significance. If we look at the taste trip test It's a validated 16 point scale for a score less than 9 is consistent with hypogyceia, meaning that you really don't have a good taste sensation. But a score of 9 or above is considered normal. So from 11 which is sort of low normal it shifts the score towards a high normal range primarily driven by sweet and salt sensations. What we don't know is how much this degree of a taste score difference influences satisfaction and eating behavior, but there are studies that show that a higher taste score is associated with a lower BMI and waist circumference.
- [16:54]
B
Thank you. And Moving forward, genes CRLF1, CYP1, B1, PRMT8 and Eya1 all show differential expression on both salmon and star data sets. If you see in Figure 3 Panel B it shows that all these genes were downregulated after the somazotide treatment and in the placebo group no genes met the threshold for significant differential expression and these genes are all involved with taste transection, neuromaturation, plasticity of gussatory nerve fibers, fibers and also taste bud renewal. So it's very interesting that we do see these results. Also, neuroimaging data shows that after six weeks semaglutide decreased activation of dorsolateral putamin in response to visual cues of calorie, dense, sweet and savory foods after standardized liquid meal intake. Increased activity in the angular gyrus of the parietal cortex in response to tasting the sweet solution compared to the distilled water 30 minutes after a standardized meal were also seen in the semaglutide treatment group and this is all seen in Figure 3 panel C. So authors report that 93% of semaglutide treated group or 14 participants experienced gastrointestinal adverse reactions including nausea, vomiting, diarrhea, constipation, headaches, dyspepsia and abdominal pain which is well known for this class of medications. Adverse events including vomiting, constipation, headaches were reported in 20% of the placebo group or about three participants.
- [18:45]
A
I'll jump in and just say this was what I was referring to earlier is even though the study was designed for the participants to be blinded, I'm not sure that it was very effective. That's not a critique on the authors. This is an effect of semaglutide, so I don't know that you can do anything about that. But you have to be worried, particularly if this was much more subjective based information. If they were doing, say, surveys of a patient experience with something, you'd have to be highly worried that the subjects probably knew whether they were on the medicine or not. Not entirely. 20% of the individuals in the placebo group had some of these side effects. So it wasn't a totally perfect distinguisher, but it was enough to where I think that would have to call it into question. But fortunately most of the results here is objective data. And so I think that gives us a little bit of reassurance because probably not very effective blinding.
- [19:32]
B
All right, well, with that I'll let Chase finish up with the discussion points
- [19:38]
A
and we will start with the authors, the summary that they used to describe the results and I'll quote them here where they say, treatment with semaglutide improved recognition of basic tastes, altered gene expression in the tongue, decreased activation of putamen in response to calorie dense, sweet and savory food cues, and increased cortical brain activity in response to the sweet taste stimulus compared to placebo in women with obesity. And just to summarize a couple of key things there to highlight it, I'd say that the authors found that individuals on semaglutide have an enhanced taste and enhanced brain activity from that stimulus, that sweet taste stimulus in particular. So, Culpin, I'm particularly interested in your thoughts on this. You've alluded to some of this before, but does that make sense physiologically? Is this what we would expect? Is that somebody who is on this medication who has an enhanced taste, enhanced brain activity, that that would in fact potentially lead to less intake of food? Does that make sense?
- [20:45]
C
One would believe so. Specifically, when we take into account the improved taste sensation, it could mean that a person has to consume less to appreciate that taste. The functional MRI findings are particularly very interesting. The dorsolateral putamen, that's mainly related to habitual cue driven eating and showing that there is decreased activation of that in response to visual cues of calorie dense foods would indicate that there is less of that habitual cue driven eating. And also the angular gyrus is more like a hub where all the signals interact. An increased activation also would indicate a greater shift from habit driven eating to more deliberate control. So in combination, the fact that a person is more satisfied and appreciates better taste and also has these central changes would indicate that that individual has more control over their eating.
- [21:43]
A
And the authors then go on to make what I think is an important observation. And they point out, and I mentioned in the results, that individuals on semaglutide lost weight, so exactly what would have expected. And so they point out that, well, you really can't tell if this is a direct effect of the drug that's making some of these changes or potentially this is causing weight loss to happen. And it's actually the weight loss that is leading to these changes. So it's an indirect effect of semaglutide. So Kulpena, obviously just from this investigation, it would be speculation as to which of those it is. But if you had to guess or if there's anything else that we could pull in from other data that's out there, which do you think is the more likely explanation for the changes that the authors have found here?
- [22:24]
C
It clearly is a very complex question than it might seem. There is actually some studies that have specifically looked at taste outcomes with weight loss, and these were mainly behavioral modification trials. There was a meta analysis from 2023 that actually showed that these outcomes were mostly neutral, but there was some improvement, particularly with sweet taste, with weight loss. But these studies were very heterogeneous in their methodologies, the population. So we really cannot make any comparisons and our study itself doesn't answer that question for us. Adding to this complexity, we have to also take into account that there is GLP1 receptors in the nerve fibers in the taste bud. There is GLP1 receptors in the angular gyrus. So it would be very intuitive to think that there must be some direct drug effect. It's also very tempting to invoke the inflammatory pathway related to GLP one that could also have an effect on that taste perception. It is clearly a very small sample size, but if we had data that would tell us if there is variability in these taste scores in correlation to the degree of weight loss, I think that would also tell us something more. Those are my thoughts there.
- [23:43]
A
The authors then go on, as is typical, to compare their findings to the medical literature, Just like Colpin has already alluded to. For us, there's not a lot of investigations reporting the effect of GLP1s on taste. And so it's difficult to compare those results in part because previous studies that are out there had different intervention periods, used different types of GLP1s, different ways of assessing taste components. So very hard to make any real comparisons with the information that's out there. The limitations that the authors point out, we've mentioned some of these already. The first one that they mention is that it is a relatively small sample size and it's a small sample size of a very specific cohort, so specifically PCOS patients. They also point out that the study environment doesn't necessarily reflect an everyday experience, so that may limit the generalizability. And they also point out that taste recognition can vary due to genetic and experiential factors and another thing that would potentially limit generalizability. And then finally they had some comments about the use of RNA sequencing and some limitations related to that. We're going to move on to their conclusions though, and the authors give a nice summary here at the end of their paper. I will quote them again where they say treatment with semaglutide improved an overall taste quality recognition of four basic tastes compared to placebo. Additionally, semaglutide treatment altered RNA expression in the tongue, changed brain activity in response to visual cues of calorie, dense, sweet and savory foods, and changed response in parietal cortex to tasting sweet solution. Where we're going to wrap up, as we usually do, is first of all thinking about the quality of this report overall. And then at the very last step, we'll think about whether this should change our practice. But we'll start with thinking through the quality of this investigation. Nal, let's start with you as you work through this. What is your impression of the quality of the study and write up by the authors here?
- [25:41]
B
Well, Chase, I really thought that this study was overall very good. It was answering a very complex question, like Kapana pointed out earlier. And I really loved how they used a multimodality way to look at it. Not only did they have clinical weight loss data, they also looked at gene expression with tongue biopsies and they did functional mri. So I thought it was very well designed in that sense. I would have liked to see them look at maybe a higher dose of semaglutide. I know they've used 1mg, so it would be interesting to see if more weight loss would alter some of the results. So overall I thought it was a pretty good study.
- [26:24]
A
Yeah, we think about more the dose response effect related to observational studies. But I agree with you on that. It would always bolster your confidence in these findings if you can clearly show that the more of a medication you use, the more of a response that you see. Kalpana, your thoughts on the quality of this report?
- [26:41]
C
I would agree with NA that this is a very well designed, mechanistic study on a very tough question. I think all their methods, especially using the taste trip test, which is a validated standardized method with good reproducibility, and again with their RNA testing, they really went quite a bit further to make sure that there was, they used two methods, that there was agreement and they used a very conservative approach to reduce their false positive rate, an approach which is very suitable for a small sample size. And also we talked about before the homogenous population that they used. We could debate about the weakness and how that limits generalizability, but I really thought that was a very thoughtful design choice. I guess as the authors admit, this cannot help distinguish drug effect from weight loss effect, but I think they have clearly set up a nice model and methodology of doing it that could be adapted as a standard and there could be future designs that are weight matched using different strategies that could teach us a whole lot.
- [27:45]
A
And we will wrap up with thinking about the practice, our practice, the care we deliver to our patients. And as we think about this, maybe not just whether we would use a medication, yes or no, but really broaden that to think how we talk about medications with our patients, how we can advise them on the expected anticipated benefits of a medication. So now give us your thoughts. First of all, do you think this will change our practice or will you do anything different in your clinical care based on these results?
- [28:14]
B
Yeah, no. Great question. And as you alluded to earlier, Chase, I don't think this is going to affect how often I prescribe this class of medications, but I do have a lot of patients, I do see a lot of patients with pcos. So I think this will be a really nice article to cite when they have questions about adverse side effects. But I think it's really just counseling, right? When patients ask whether this is going to affect their taste or not. I think this is a good article to cite, although again, very homogeneous population. So I can counsel this to all my patients, but I think especially my patients who are foodies and they care about that aspect, that joy in their life. I can cite this article as an added kind of warning, but just as additional side effect. Not side effect, but potential thing to watch out for besides the usual, you know, spiel that we give GI side effects, medullary thyroid cancer and pancreatitis. So I think it would increase my counseling time just a little bit.
- [29:19]
A
All right, Copeland, I will end things with you. Same question for you. Either in the practice, your practice, the care that you deliver directly, or just as you advise the rest of us as endocrinologists, what are your thoughts on this study? Should it change the care that we deliver?
- [29:33]
C
Well, I for one, it just makes me a little bit more curious specifically about taste. In my conversations with my patient today, I talk a lot about portion the quality of food, the eating behaviors related to stress, boredom and cravings. But taste is something that I might include more in that conversation going forward. As I think about this with this article, I would consider it at this point as preliminary evidence. We really need large, diverse populations to be studied before we could draw any firm conclusions that would have clinical relevance.
- [30:05]
A
And with that, I would like to thank Na Shin and Kolpana and Muthasamy for joining me for this month's edition of Intercom Feedback Loop. I hope that you all learned as much as I did and that you will join us again next month. And now you're in the loop. This has been Endocrine Feedback Loop Endocrine Feedback Loop is brought to you by the Endocrine Society with Production Oversight by Brandy Brown. If you want to like and subscribe, you can find us on Apple, Spotify, or wherever you get your podcast. We'd love to hear your feedback on this episode of the podcast itself. Please email us@podcastron.org Endocrine Feedback Loop is a free service of the Endocrine Society. To learn more or to become a member, visit the society's website at www.endocrine.org.