GLP-1s for Weight Loss: What the Evidence Actually Shows

GLP-1 drugs have moved from diabetes clinics to dinner parties, and the conversation has gotten sloppy. People are microdosing them for focus. Couples are fighting about them. Investors are trying to figure out which snack food companies are doomed. I want to cut through some of that and tell you what the evidence actually supports, where it gets thin, and where we simply don't know yet.

Where these drugs came from, and why they work

GLP-1 is a peptide your body already makes. It slows gastric emptying and increases satiety. Scientists knew this in the early 1980s, but human GLP-1 has a half-life of about one minute, which makes it useless as a drug. You could infuse it into a vein and people would stop feeling hungry. Unhook the IV, hunger returned.

The breakthrough came from Gila monster venom. A researcher at the NIH was extracting venom for basic science reasons, noticed it did strange things to mouse blood sugar, and in 1992 the active peptide, exendin-4, was isolated. It binds the same receptor as human GLP-1 but survives in the blood for about a day. That became exenatide (Byetta), FDA-approved in 2005. This is not new science.

What has changed is potency and half-life. Semaglutide (Ozempic, Wegovy) moved dosing to once weekly. Tirzepatide (Mounjaro, Zepbound) pushed weight loss further. Both deliver roughly 15% to 20% body weight loss in the first year of trials. A newer Lilly compound, retatrutide, posted around 28% weight loss at a year, which is extraordinary. An oral version is coming, and uptake will be high because most people would rather swallow a pill than inject themselves.

The off-target effects worth taking seriously

Benefits are piling up in cancer, kidney disease, cardiovascular disease, and blood pressure. Most of these are biologically plausible as downstream effects of weight loss itself. Obesity drives a huge share of chronic disease. Reduce obesity by any mechanism and you reduce those diseases. Some cardiovascular trials hint at effects slightly larger than weight loss alone would predict, but separating that cleanly in the data is hard.

More interesting are the effects that can't plausibly be explained by weight loss. People on GLP-1s report drinking less alcohol. Some report smoking less. Patients describe it as the "food noise" going quiet, and that framing seems to extend to other compulsions. There's early interest in studying GLP-1s and recidivism after prison release, on the theory that turning down impulse signals could reduce relapse into drug and alcohol use.

GLP-1 receptors are densely expressed in the brain, and these peptides cross the blood-brain barrier. Once a drug reaches neurons and binds receptors there, a lot becomes possible. That doesn't mean every claim is true. It means the claims aren't crazy.

Muscle loss, weight regain, and the "it's not natural" argument

A common criticism online: GLP-1s cause muscle loss, while "real" weight loss through diet and exercise doesn't. That's wrong. When you lose weight by any method, you lose some fat and some muscle. Typically 30% to 50% of weight lost is lean mass, regardless of how you got there. Your muscles exist in part to move your body around. Less body to move, less muscle needed.

One wrinkle rarely discussed: muscle mass in these studies is usually assessed by DEXA, which can't distinguish true muscle from fat that has infiltrated muscle tissue. Think of the marbling in a steak. Some of the apparent "muscle loss" on GLP-1s may actually be intramuscular fat disappearing, which is a good thing. Either way, if you want to preserve lean mass, the answer is strength training and adequate protein. That's true on a GLP-1 and it's true off one.

Weight regain when you stop is real. Trial data show roughly half the lost weight comes back after discontinuation. This gets framed as a damning feature of the drug. It isn't. If I stop your blood pressure medication, your blood pressure goes back up. We don't call that a failure of the drug. And the track record of maintaining significant weight loss through diet alone is abysmal. Most diets work briefly, then fail. The honest framing is that a GLP-1 is a long-term medication, more like a statin than a cleanse.

In the real world, many people are using what amounts to a maintenance dose: titrate up to lose weight, then titrate down to hold. There's no randomized trial data on this because drug companies trial the dose they're selling. Ideally this titration happens under a physician's supervision rather than patients breaking open pens on their own.

Side effects, and what we genuinely don't know

Nausea and other GI effects are common and largely expected given the mechanism. Pancreatitis is a real but rare risk. The Gila monster's venom causes acute pancreatitis in mice, so the signal is biologically consistent. If you develop a deep burning pain in your upper abdomen on one of these drugs, get checked.

On suicide and self-harm, the best current evidence is a meta-analysis of 27 randomized trials covering about 32,000 people. No difference between GLP-1s and placebo. I don't think that risk is real based on current data.

On sexual function, the data are thin and lopsided. There are roughly ten times more studies on male sexual side effects than female, despite women being more likely to take these drugs and to lose more weight on them. One observational study found about a twofold higher rate of anorgasmia diagnoses among women on GLP-1s, but the same rate appeared with metformin, which suggests the underlying condition driving treatment, not the drug itself, may be the issue. We need better data.

The bigger unknown is time. The first drug in this class was approved in 2005, but today's drugs are different molecules at different doses for different indications. If you want to know what 30 years on a modern GLP-1 looks like, you have to wait 30 years. That uncertainty matters most for people using these drugs off-label, particularly those trying to push a normal BMI into the underweight range. Fat tissue is metabolically and hormonally active. Being underweight carries its own risks. The sweet spot for longevity and resilience against illness sits roughly in the BMI 20 to 25 range.

"Ozempic face" and the downstream economy

Ozempic face is real, and it isn't specific to Ozempic. Plastic surgeons call it midface volume loss, and it happens with any substantial weight loss. We saw it with gastric bypass patients for decades. We just didn't have a catchy name. Facial fat grafting procedures jumped about 50% in 2024. Excess skin removal is going to be a growth industry. So is the protein snack aisle. Carbohydrate-heavy snack brands should be nervous.

Kids

Semaglutide has FDA approval down to age 12. I'm genuinely ambivalent. Kids aren't small adults. They're growing, and calories matter for that. On the other hand, fat cells added in childhood largely persist for life, and food habits formed young tend to stick. What I'm confident about is that this shouldn't be a first-line intervention for pediatric obesity. Better school food, family-level support, and other upstream tools should come before a weekly injection for a twelve-year-old.

Bottom line

GLP-1s are transformative drugs. They produce weight loss at a scale we've never had outside of bariatric surgery, they appear to reduce obesity-driven disease, and they may do interesting things to reward and compulsion pathways in the brain. They also have real side effects, require long-term use to maintain benefit, and shouldn't be used to push people below a healthy weight. Treat them like what they are: a serious medication with a serious upside, prescribed and managed by someone who knows what they're doing.

I covered this in depth on Wellness, Actually — listen below.

Wellness, Actually Podcast

"What's the deal with GLP-1s?" — Listen to the full episode, including the week's health news and listener Q&A.