Adding Exercise to a GLP-1: What Nine Trials Found

On 13 July 2026, the journal Obesity Reviews published a network meta-analysis ranking GLP-1 receptor agonist therapy, structured exercise, and the two combined against one another and against placebo. Pooling nine randomized controlled trials in 1,009 adults with overweight or obesity, the combination came out ahead of either approach alone on body weight, fat mass, waist-to-hip ratio, insulin sensitivity, and HDL cholesterol.

The direction is not surprising. What this analysis adds is a ranking of all three approaches across the randomized evidence, together with a grade for how much confidence each result deserves.

Why the two do different jobs

GLP-1 receptor agonists work largely through appetite. They slow how fast the stomach empties and act on the brain circuits that govern hunger and fullness, so you eat less without fighting yourself to do it. Less food means an energy deficit, and an energy deficit means the body has to make up the difference from its own tissue.

Here is the catch. The body does not draw that deficit purely from fat. Weight lost by eating less comes off as a mix of fat tissue and lean tissue, muscle included.

Structured exercise pushes a different lever. Loading muscle regularly gives the body a reason to hold onto it, and contracting muscle pulls glucose out of the blood through a route that does not require insulin at all. That is why training can move insulin sensitivity even in someone whose scale weight barely budges.

So the two are not the same tool used twice. One opens the deficit. The other influences what the deficit takes and how the body handles fuel while it is happening.

What the nine trials found

The analysis pooled nine randomized controlled trials covering 1,009 adults, 589 women and 420 men, with a mean age of about 44 and a mean BMI of about 32.6. Each trial compared some combination of a GLP-1 receptor agonist, structured exercise, both together, or placebo, and measured body composition and blood markers of glucose and lipid metabolism.

Results are reported as standardized mean differences, or SMDs. An SMD is a unit-free number: it expresses a change in standard deviations rather than in pounds or kilograms. By common convention, 0.2 is small, 0.5 is moderate, and 0.8 is large. Against placebo, the combination arms showed:

• Body weight: SMD -1.04 (95% CI -1.24 to -0.83)
• Fat mass: SMD -1.01 (95% CI -1.27 to -0.75)
• Waist-to-hip ratio: SMD -0.55 (95% CI -0.91 to -0.18)
• HOMA-IR, a fasting estimate of insulin resistance where lower is better: SMD -0.59 (95% CI -0.89 to -0.29)
• HDL cholesterol: SMD 0.32 (95% CI 0.04 to 0.61)

The authors graded their confidence in the evidence as high for body weight and fat mass, moderate for waist-to-hip ratio and most lipid outcomes, and very low for fasting glucose. That grading is worth as much attention as the numbers. The two findings the authors stand behind most firmly are the two about body composition.

Notice that body weight and fat mass are listed separately. They are not the same measurement, and an approach can move one more than the other.

The one-year trial underneath the average

A pooled estimate is easier to read when you can see one of the trials up close.

In 2021, the New England Journal of Medicine published a randomized trial in 195 adults with obesity (BMI 32 to 43) who did not have diabetes. Everyone first completed an eight-week low-calorie diet. They were then randomized for one year into four groups: exercise plus placebo, liraglutide plus usual activity, exercise plus liraglutide, or placebo plus usual activity.

After that year, body fat percentage in the combination group had fallen by roughly 3.9 percentage points, about double the drop observed in the exercise-only group (1.7 points) or the liraglutide-only group (1.9 points). Compared with placebo, weight change was about 4.1 kg in the exercise group, 6.8 kg with liraglutide, and 9.5 kg with both. The combination was the only arm in which glycated hemoglobin, insulin sensitivity, and cardiorespiratory fitness all improved.

Those are averages observed in that trial's participants, not a forecast for any individual. Two caveats belong right next to them. The trial was funded by the Novo Nordisk Foundation, which is tied to the maker of the medication studied. And there were harms: increased heart rate and gallstones occurred more often in the liraglutide-alone group than in the combination group. GLP-1 medications carry real side effects, and exercise carries its own considerations for people with joint, cardiac, or other limitations.

The label already assumes activity

This is not framed as a bonus in the regulatory language. The FDA-approved label for Zepbound (tirzepatide, Eli Lilly), revised April 2026, indicates the medication "in combination with a reduced-calorie diet and increased physical activity" to reduce excess body weight and maintain that reduction long term. Physical activity is written into the approved use, not appended to it.

What this means, and what it doesn't

If you are on a GLP-1 or considering one, the honest read is that the randomized evidence points toward adding structured activity rather than treating the medication as sufficient on its own, and the strongest part of that evidence concerns what happens to fat mass specifically.

The limits are real:

• Nine trials and 1,009 people is a modest base. The pooled population averaged 44 years old with a BMI around 32.6. Results may not extend cleanly to older adults or to people at much higher BMIs.
• SMDs do not convert to pounds. You cannot read the headline numbers as an amount of weight.
• "Structured exercise" is doing a lot of work as a phrase. The published summary does not specify which modes, intensities, or doses of training the pooled trials used, or how long they ran. Nor does it break out which GLP-1 medications were involved, so how well this maps onto the newest agents is an open question.
• Confidence varies sharply by outcome. The HDL cholesterol interval (0.04 to 0.61) only just clears zero, and the fasting glucose evidence was rated very low.
• Neither analysis reported lean mass. Fat mass and body fat percentage are not the same as a direct measurement of muscle preserved.

None of this tells you whether to start, stop, or change a medication. That decision belongs with a licensed provider who knows your history, and so does the question of what kind of activity is safe and realistic for you to actually sustain.