Surgery Versus GLP-1 Drugs: Reading an Indirect Comparison

A 2026 network meta-analysis of 30 randomized trials in 20,015 people compared bariatric surgery against GLP-1 drugs, and reported that surgery produced about 10 percentage points more total weight loss. Two things about that finding matter more than the number: nobody was randomized to surgery versus a drug, and against tirzepatide alone the difference disappeared.^[1]

First, how the comparison was made

This is not a trial of surgery against drugs. No such trial is in the analysis. What exists is trials of surgery against lifestyle intervention, and trials of GLP-1 drugs against lifestyle intervention. A network meta-analysis uses that shared comparator as a bridge and estimates the surgery-versus-drug comparison indirectly.

The authors state it plainly: all comparisons between surgery and GLP-1 drugs were indirect.^[1] That is a legitimate and widely used method, and it carries an assumption that has to be true for the answer to hold — that the people in the surgery trials and the people in the drug trials were similar enough that a shared comparator connects them fairly.

What was reported

Read as a class comparison, surgery wins on every measure, and the gap persists past two years. Among participants with type 2 diabetes, surgery produced greater reductions in BMI, weight, waist and total weight loss, with similar HbA1c improvement.^[1]

The row that changes the story

In tirzepatide-only analyses, differences versus surgery were not significant.

“GLP-1 receptor agonists” as a class includes liraglutide and other older drugs that produce far less weight loss than the current ones. Averaging them together produces a class figure that no modern patient is actually being offered. When the comparison is restricted to tirzepatide, the surgical advantage stops being statistically distinguishable.^[1]

The authors’ own summary reflects this: surgery remains superior, and tirzepatide is a promising nonsurgical option. Both halves of that sentence are load-bearing, and coverage tends to quote only the first.

Cost is the axis this comparison leaves out, and a separate study puts two-year spending at $63,483 for medication against $51,794 for surgery — an operation being paid for once and a prescription every month. Its weight figures are much weaker than its cost figures, for reasons worth knowing: the weight comparison used 257 people.

What weight loss does not settle

• Surgery is permanent; a drug stops when you stop paying. Our maintenance article covers what discontinuation does, and it is the single largest practical difference between these options.
• Surgery carries operative risk and drugs carry chronic side effects. Neither appears in a weight-loss percentage.
• Nutritional consequences differ enormously. Bypass procedures create lifelong absorption issues that a drug does not.
• Access differs. Surgery is a one-time approval fight; a drug is a recurring one, every month, indefinitely.

⚠ One caveat arrived after this was written. A real-world study of 44,025 people found surgery two to three times better than these drugs in ordinary care, because trial adherence and real adherence are different things — see what these drugs do outside a trial. Under trial conditions the comparison below holds; outside them it does not. The genuinely useful framing is not which is better but that they are now close enough to be a real choice for some people — which was not true five years ago. Our article on weight regain after surgery covers the case where someone ends up needing both.