What are the side effects of semaglutide?

Reported rates

The one head-to-head number

SURMOUNT-5 compared semaglutide against tirzepatide directly, and reported gastrointestinal discontinuation at 5.6% with semaglutide against 2.7% with tirzepatide. Most cross-compound safety comparisons in this field are assembled from separate trials with different populations, which makes them close to meaningless. This one is not, because both arms were randomised in the same study.

Why escalation is when effects cluster

The labelled titration exists for tolerability rather than efficacy — the published rationale is that starting at a maintenance dose raises adverse-event rates without improving outcomes. That is also why the reported rates are not evenly distributed across a course: they concentrate in the weeks following each step up, and the trials record most effects as mild to moderate and diminishing with continued administration.

What the labels warn about that the trials did not measure

The approved labels carry a boxed warning regarding thyroid C-cell tumours, drawn from rodent studies rather than from human trial findings. This is an important distinction to state precisely: it is a warning derived from an animal signal that human trials were not designed and not powered to resolve, so it is neither a rate nor a refutation. Pancreatitis and gallbladder events are reported in the trials at low rates, and hypoglycaemia is reported principally where the drug was combined with insulin or a sulfonylurea.

What this data does not cover

These rates come from controlled trials of an approved, inspected product administered at defined doses under supervision, with adverse events collected systematically. A vial bought on the research market is not that product, and the safety profile does not transfer to it: identity, purity and actual content are separate questions from pharmacology, and no rate table can speak to them.