This is the Weight and Healthcare newsletter! If you like what you are reading, please consider subscribing and/or sharing!
I have received hundreds of requests from readers to write about this topic, and it’s not surprising. Articles are pouring out of the media claiming that GLP-1 use prevents cancers, reduces progression of existing cancer, and/or reduces mortality. This is also one of these series that has required hours upon hours of research and analysis so of course I appreciate all my readers and I want to give a shout out to my paid subscribers who make the time I spend to create this newsletter possible!
Before I start I want to first say that this would certainly be a benefit if it was true. As I’ve said many times, GLP-1s are solid drugs for the management of Type 2 Diabetes and they may have other health benefits. It’s possible that they may have an impact on cancer development/progression/recurrence/mortality. That said, given the ways we have seen the drug manufacturers and the many, many doctors and researchers on their payrolls push these drugs with what I would describe as, at best, less-than-ethical marketing and practices, these are claims that must be examined and that would need to be based in strong evidence. Otherwise we risk creating side effects (some serious and even fatal) with the only benefit being to the bottom line of the drug manufacturers.
As I read the articles that various readers sent and started to make a list of the cited studies to analyze, I began to also generate a list of general thoughts and questions that I wanted to look for as I analyzed the research.
The first issue (which is not made clear in many of the articles I’ve read) is that the existing studies are all correlational. This means that (unless, of course, I’m missing a study) there is no proven mechanism of action by which GLP1 drugs cause these outcomes. Rather, studies are finding that GLP-1 use and the study outcome (cancer prevention, progression, survival etc.) happens at the same time in some amount of those who used GLP-1s versus those who didn’t.
The studies that I have seen referenced are not clinical trials but, rather, retrospective cohort analyses. This means that study authors are analyzing past medical records. They might, for example, look at records of patients with cancer at certain healthcare facilities, pull the records of those who were also on a GLP-1 medication, then find a way to statistically match them with people who were not on GLP-1s and run a statistical analysis to look at the differences in cancer progression, survival rate etc.
Retrospective cohort analyses can give helpful information, but when it comes to treatment of a health issue with a particular medication they have significant limitations, even with the best researchers. The reality of retrospective analysis is that the researchers make choices around virtually every aspect of the study - the cohort, the variables that will be matched, the time of the study, whether drugs will be considered separately or just as a group of “GLP-1” drugs and more. When the study methodology is written out it can seem obvious or like a forgone conclusion but we always have to remember that every methodological choice includes multiple roads NOT taken that may have led to different conclusions. That also means that, in general, this kind of research can be manipulated by authors who want to do so.
A big concern here is the possibility of unexamined confounding variables (that is, it looks like GLP-1 use is creating a difference in outcomes, but it’s really something else that was not included in the study.) To me the most obvious possible confounding variable would be income/socioeconomic status. GLP-1 use may be acting here as a proxy for money/access/other pertinent factors. So, for example, people in the US who could afford GLP-1s or afford insurance that paid for GLP-1s may also have access to other factors that support their health. This could include things like rest, time off work, additional therapies etc. People who had more access to healthcare in general may have also been prescribed GLP-1s (whereas patients with less access to healthcare may have simply continued with refills of the meds they were already taking prior to GLP-1s coming on the scene.) In that case greater access to healthcare may explain the differences in outcomes.
These aren’t the only difference. Since these drugs may have been prescribed for Type 2 Diabetes management, the difference could be A1C. There are also behavioral factors including movement, sleep, stress, food options, social connection etc. These can also combine - those with less financial security may have experienced significantly higher stress, which can also increase blood sugar and inflammation for which they have less access to healthcare to manage etc. Identifying and analyzing potential confounding variables in retrospective cohort analysis can be incredibly difficult, even with excellent researchers who are absolutely doing their best.
Retrospective cohort analysis creates additional hurdles here because the researchers can only analyze the information that already exists in the records. The medical records that they are analyzing may not have ever been intended for research purposes and so the information may not be as carefully gathered as it would (or at least, should be) in, for example, a clinical trial or prospective study. The researchers may, for example, want to consider the subject’s income, but if that wasn’t recorded then they can’t include it in their analysis. Medical charting between an individual provider and patient is not the same as information gathered for, for example, a clinical trial. There is also rarely a method by which to account for those lost to follow up which can impact the conclusions.
For all of these reasons, in our training researchers are repeatedly cautioned to avoid overgeneralizing the findings or claiming a cause-and-effect relationship when using retrospective analysis.
Another issue that I am wary of here is the “but there’s so much evidence” fallacy that I’ve discussed before. Organizations/authors my flock to a subject when there is interest in that topic such that:
Study publication is likely
Attention to the study is likely
There is money available to fund the research
If it is the kind of research that draws views and clicks, it will be amplified by media from traditional outlets to social media and that can create a perfect storm that amplifies attention-grabbing headlines, leaving science far behind.
In this situation, even if all of the research is of the same type with the same limitations and makes the same mistakes, there will be those who insist that those issues can be overlooked because of the sheer volume of research. That’s a bit like if all the baking groups in the home economics class accidentally baked their brownies with salt instead of sugar, and then the Guardian, Medscape, and social media influencers insisted that this creates delicious brownies because look how many brownies were made like this - they can’t possibly be bad! (Did this simile come from the fact that my home ec teacher insisted that we should be able to tell sugar from salt visually and several groups, including mine, …could not? Yes. Am I still..ahem…salty about it? Yes, thanks for asking.)
Another thing I’m noticing is that a large portion of this research seems to only include higher-weight people (typically based on questionable assumptions about higher-weight and cancer.) The issues with those assumptions will have to be a topic for another day, but of course people of all sizes get cancer and people of all sizes use GLP-1s and so if it is believed/hypothesized that GLP-1s have positive effects on cancer development/progression/mortality, I’m curious of the ethics of ignoring those possible effects in thinner people who both get the same cancers and can take GLP-1s.
At any rate, of course GLP-1s are not FDA approved for cancer prevention/treatment and in order to have responsible clinical decision-making around this I would suggest that we would need the answers to, at least, the following question:
What drug(s)
At what dose(s)?
For what duration?
At what time? (Did the person need to be taking the drugs prior to diagnosis, was there a benefit to starting after diagnosis? etc.)
For what benefit (what types of cancer, prevention/slowed progression/lowered mortality etc.)?
At what effect size?
For which people?
With what side effects?
Are GLP-1s superior to other interventions in terms of a risk/benefit analysis (SGLT1 inhibitors/metformin/weight-inclusive interventions etc.)
Starting in Part 2, we’ll take a look at these questions as they pertain to the research currently being published around GLP-1s and cancer.
The weight loss industry is up to shenanigans that impact the healthcare of higher-weight people. Louise Adams and I are teaming up for this month’s online workshop - Exposing the Weight Loss Industry’s Secrets and Tricks. We’ll talk about what they are doing, why they are doing it, and how we can pushback. All registrants get access to a video and there is a pay-what-you-can-afford option so that money isn’t a barrier. Details and Registration here!
If you find the work I do valuable, you can support my ability to do it by becoming a free or paid subscriber!
Liked the piece? Share the piece!
More research
The Research Post
More resources
The Resource Post
*Note on language: I use “fat” as a neutral descriptor as used by the fat activist community, I use “ob*se” and “overw*ight” to acknowledge that these are terms that were created to medicalize and pathologize fat bodies, with roots in racism and specifically anti-Blackness. Please read Sabrina Strings’ Fearing the Black Body – the Racial Origins of Fat Phobia and Da’Shaun Harrison’s Belly of the Beast: The Politics of Anti-Fatness as Anti-Blackness for more on this.
I've said it before and I'll say it again. I'm not being hyperbolic here. I believe GLP-1 drugs are going to be this decade's PhenFen. While they do have the potential to help some people dealing with type 2 diabetes, they are not benevolent medications. They carry a lot of potential side effects. Some of these side effects are quite serious.
I'm a type 2 diabetic who tried Ozempic after my doctor said it helps some people with chronic inflammation, which I have. I developed brain fog right out of the starting gate, but kept using the medication because I hoped this would pass. Then I started feeling somewhat queasy, but kept using the medication because I hoped the feeling would pass.
Then came the diarrhea. Not just slightly loose poops, but full-on diarrhea. That was followed by the feeling that I was being cooked from the inside. That was the last straw. I stopped the Ozempic and went back to using Januvia with Lantus and Humalog. I recently started taking Metformin ER as well. I'd stopped taking Metformin because of digestive issues, but the ER version helps prevent the GI upset.
GLP-1's were never meant to be used as weight loss drugs. Leave it to the weight loss industry to co-opt a medication meant to treat a serious chronic health condition.
Great deconstruction of retrospective studies, and I love your ironic spin on the fact that thin people could end up de-accessed to a lifesaving drug. But what a great market that would open up! Of course weight loss could be a dangerous side effect for them. So my cartoon mind comes up with the next ironic step - an image of a towering mean-looking PCP leaning over and trilling, "but you don't want to get *Caancerr* do you? " (a nod to the idea that profitability goals spare no one).