GLP-1 Drugs After Bariatric Surgery
A weekly injection loses to a second operation by a wide margin, though nobody has randomized people between them, so the gap stays partly unexplained.

Here is a small fact about the gastric sleeve: for a lot of people it is the operation, and for a lot of other people it turns out to be the first operation. The sleeve removes most of the stomach, weight comes off, and then over the years some of it comes back, or it never came off as far as anyone hoped. A comprehensive review of very long-term outcomes puts the weighted mean total weight loss 10 years after sleeve gastrectomy at 24.4%, the lowest of the procedures it covered. People who search “ozempic after gastric sleeve” are mostly asking what happens next.
The plain answer is that the drug does something, it does less than a second operation, and the evidence is thinner than you would like. (People say “Ozempic” for the whole GLP-1 class the way they say “Band-Aid,” while the studies say “GLP-1 receptor agonists” and mean several drugs.) The largest comparison available here, a multicenter retrospective study of 4,901 people with a prior sleeve and either insufficient loss or regain, found that at 3 years, converting the sleeve to a gastric bypass produced 26.1% total body weight loss, versus 13.7% for adding a GLP-1 drug. A randomized trial built to measure the drug cleanly after surgery, called BARI-STEP, exists. Its results are not among the facts here. So that part is not settled.
Why not just take the drug and skip the second surgery
The sensible reaction goes something like this. A weekly injection is not an operation. If the sleeve underdelivered, you reach for the thing that doesn’t involve a hospital, and you only go back to the operating room if the injection fails too. There is also a slightly more cynical version: wasn’t the surgery supposed to be the thing you did instead of drugs? Now the surgery is the thing you do before drugs?
Both reactions are reasonable. That’s fine. The strange part is the structure underneath them. A sleeve does not leave a person back where they started, with the same options as before. It leaves them as a new category of patient, with a new set of definitions, a new set of trials they qualify for, and a new menu that has both a drug and a sequel operation on it. The decision is less “surgery or drugs” than “which add-on to the surgery you already have.”
How anyone decides a sleeve didn’t work
You need a definition of failure before you can study a fix for it, and the field has more than one. A review in Frontiers on endoscopic management says weight recurrence is “an increase of 25% excess weight loss from nadir weight,” which is a measure of how much came back from the lowest point. The BARI-STEP trial uses a different door: its “suboptimal clinical response” is “less than 20% weight loss from surgery,” which is a measure of how far you got in the first place, regardless of what happened after.
These are two different patients. One lost a lot and regained some. The other never lost much. I feel like the literature treats them as one bucket more often than it should, though the large retrospective study above at least names both (“insufficient weight loss or weight recurrence”) in its inclusion criteria.
How big is the bucket? Not small. One study of sleeve conversions says up to half of sleeve patients may eventually need conversion surgery for insufficient loss, regain or complications. The BARI-10-POL study in Poland found revisional procedures in 23.9% of patients over 10 years, including 24.3% after sleeve gastrectomy (and 100% after adjustable gastric banding, which is its own story, told elsewhere on this site). The sleeve is like the first film in a franchise where the studio has already pencilled in a part two, for about a quarter of the audience, maybe half.
What the studies say about Ozempic after gastric sleeve
Back to the 4,901 people: 3,004 had the sleeve converted to a Roux-en-Y bypass and 1,897 got a GLP-1 drug on top of the sleeve. After adjusting for sex, baseline BMI, age and non-white race, conversion was associated with 11% greater total body weight loss than the drug.
Now look at 13.7%. That is the drug arm, the supposedly weaker option, and it represents people who already had most of their stomach removed, whose operation had in some sense stopped working, and who then lost another 13.7% of their body weight over three years from a medication. That is not a rounding error. That is a meaningful amount of person, gone, in people whose bodies had already absorbed one large intervention and adapted to it. It loses the comparison by a lot. It is still a strange and impressive number to lose with.
The 11% gap has an obvious reading: the bypass is a more powerful operation than the sleeve, so converting to it does more than adding a drug. That fits the Oseberg randomized trial, which found bypass beat sleeve at 5 years as a first operation, 22.2% versus 17.2% total body weight loss. But this was a retrospective study, so nobody was randomized, and there is a second reading. The people who choose another operation may be different from the people who choose an injection, in ways that sex, BMI, age and race do not capture (motivation, insurance, how badly the sleeve went, whether a surgeon thought they were a good candidate, stuff like that). On that reading, some of the 11% is the operation and some is the selection, and the study cannot tell you the split. Both readings survive the data. Neither wins.
There is also a systematic review finding that adding liraglutide, an older GLP-1 drug, to endoscopic bariatric therapy produced weight loss comparable to the net loss (lowest weight minus regained weight) after surgery alone. Which is a weird benchmark, I suppose, comparing a combination against surgery’s result after regain, but it is the benchmark the review chose. And a focused review on surgery and diabetes summarizes the general state of things with some restraint: GLP-1 drugs may be “alternatives, bridges, or postoperative adjuncts,” though “post-surgical evidence remains limited.” Adjunct is the sexiest word in this literature, in the sense that everyone wants to hook up a drug with an operation and see what the partnership produces.
The drug before the operation, which is stranger
Some people take a GLP-1 before surgery, and one single-center study looked at what that does. Of 216 screened, 193 were analyzed: 111 on a preoperative GLP-1, 82 on nothing. At 12 months, measured from their starting weight, the GLP-1 group was at 34.5% total weight loss versus 29.5%. Seems good for the drug.
Except the GLP-1 group also lost more before the operation (median 9.0 kg versus 5.0 kg). And when you measure from the day of surgery instead, weight, BMI and excess weight loss at 3, 6 and 12 months did not differ between the groups (all p values at least 0.19). The drug’s head start was the whole gap. After the knife, everyone moved at the same speed. The GLP-1 group did have zero major complications out of 111, versus 5 out of 82, which is a real difference in a small retrospective study, or whatever a small retrospective study can establish.
Where this leaves the comparison
Here is a made-up case, clearly invented. A person gets a sleeve, regains, and adds a GLP-1. The drug produces its 13.7%, more or less, and then plateaus, so the person converts to a bypass anyway. Now they are a post-bypass patient with a new nadir, and therefore a new denominator for a new definition of recurrence, and therefore eligible for a new adjunct study. The surgeon says, “Hey, there’s still intestine to work with.” The trial coordinator says, “Great, you’re our ideal enrollee.” At some point the person has been the subject of more studies than the drug has, and each study needs them to have underperformed the previous one to get in the door.
That sounds dumb, but it is basically how the inclusion criteria work, and you and I can both see the incentive: a field that measures interventions against the last intervention will always have a fresh cohort of last interventions to measure against. The point is that the second treatment is now a decision with real numbers on both sides, 26.1% versus 13.7% at three years, with the caveat that nobody has yet randomized people between them.
And since BARI-STEP enrolls only people with less than 20% weight loss from surgery, the sleeve’s most dependable outcome, for a certain slice of patients, is that it makes them eligible for the trial.
Sources
- Does Preoperative GLP-1 Therapy Improve Postoperative Bariatric Surgery Outcomes? A Single-Centre Retrospective Observational Study
- A Comprehensive Review of Sleeve Gastrectomy, Roux-en-Y Gastric Bypass, One-Anastomosis Gastric Bypass, Duodenal Switch, and SADI-S: Very Long-Term Outcomes at 10 Years and Beyond
- Long-Term Effects of Bariatric Surgery on Type 2 Diabetes: Mechanisms, Outcomes, and Clinical Implications: A Focused Review
- Seven-Year Comparative Outcomes of One Anastomosis Gastric Bypass and Roux-en-Y Gastric Bypass for Weight Recurrence Regain Post Sleeve Gastrectomy
- Sleeve-to-bypass conversion vs. sleeve-with-adjuvant GLP-1 receptor agonists: an academic multicenter retrospective study
- Long-term outcomes of metabolic bariatric surgery: a 10-Year multicenter retrospective study in Poland (BARI-10-POL)
- Frontiers | Endoscopic Management of Weight Recurrence Following Bariatric Surgery
- Weight Loss Effect of GLP-1 RAs With Endoscopic Bariatric Therapy and Bariatric Surgeries
- Semaglutide versus placebo in individuals with poor weight loss after bariatric surgery: a double-blinded, randomized, placebo-controlled trial
- Effect of gastric bypass versus sleeve gastrectomy on the remission of type 2 diabetes, weight loss, and cardiovascular risk factors at 5 years (Oseberg): secondary outcomes of a single-centre, triple-blind, randomised controlled trial.
Bariatric Alt explains published research. It is not medical advice, and it never recommends a surgeon, a clinic, a dose or a product. Talk to a clinician about which option fits you.