The comparison is arresting and the date range is the part to read first. This pools studies published between January 2000 and September 2024, which means most of the drug side of it is exenatide, liraglutide and dulaglutide rather than what people take now — the same problem as reading an older trial as evidence about a newer molecule.
What the pooled figures say
Bariatric surgery was associated with lower major adverse cardiovascular events, hazard ratio 0.66, 95% CI 0.60 to 0.74; lower heart failure, 0.45, 95% CI 0.38 to 0.53; and lower myocardial infarction, 0.53, 95% CI 0.45 to 0.63. [1] GLP-1 drugs reduced major cardiovascular events, 0.85, 95% CI 0.77 to 0.94, and did not reach significance for heart failure or heart attack.
Heterogeneity ran from 66% to 93%, which the authors flag themselves while noting the direction was consistent across studies — a spread that puts the pooled numbers closer to a summary of disagreement than to a single estimate.
The other half of the comparison
Bariatric surgery has an eligibility process in front of it. Somebody has to be considered suitable, pass an anesthetic assessment, and be well enough to be operated on, and that selection happens before any outcome is counted.
A network meta-analysis links the two arms largely through their shared comparison with controls rather than through head-to-head studies, so what is being compared is partly two different groups of people — which is the standing limit of evidence assembled indirectly.
What the drug arm does establish
A real reduction in major adverse cardiovascular events, at 0.85 with an interval clear of one, in a pooled population of that size.
Not reaching significance on heart failure and heart attack is not the same as showing no effect on them, and with heterogeneity in this range the pooled estimates carry less than their intervals suggest. The authors’ own framing — GLP-1 drugs as an evidence-based option for people who are not surgical candidates — is the reasonable reading, and it is a clinical decision rather than a ranking anyone shops from.