Evidence review · from the 250-topic plan

The side-effect ledger in words: semaglutide vs tirzepatide, the trial tables translated

The shape of both ledgers

Read enough adverse-event tables and the class signature emerges before any number does: gastrointestinal events dominate the top of both lists, dose-dependence runs through everything (higher arms report more), the curves peak around escalations, and the serious-event rows sit far below the nuisance rows in frequency while sitting far above them in importance. That shape — GI-heavy, dose-linked, titration-timed, mostly mild — is the single most useful thing the tables teach, because it converts “side effects” from a lottery into a schedule you can plan around.

Line by line, in honest approximations

Nausea — both ledgers' headline: reported by roughly four-in-ten semaglutide patients in the pivotal program and by roughly a quarter to a third across tirzepatide's arms — overwhelmingly mild-to-moderate, and the entire reason the comfort protocol is this library's most-linked file. Vomiting — a step less common than nausea in both programs, same dose-and-titration pattern; the practical line it adds is hydration discipline and the missed-dose logic for rough weeks. Diarrhea and constipation — the ledger's odd couple, both appearing at meaningful rates in both programs (the gut's motility is being renegotiated in both directions); fiber, fluids, and patience handle most of it, with the constipation file favoring proactive management over reactive misery. Injection-site reactions — low single digits, usually cosmetic, rotation-responsive (the rituals file maps it). Fatigue, headache, dyspepsia, reflux — the mid-table cluster, real but modest, mostly titration-era. Discontinuation due to adverse events — the ledger's bottom line and its most reassuring row: single digits in both programs, meaning the overwhelming majority who started, finished — the number that keeps every other number in proportion. Exact figures by dose live in the prescribing information; anything more precise than these bands belongs to the labels, not a comparison site's memory.

The time-course that changes everything

The tables are snapshots; the experience is a curve. Events cluster in the first weeks of each new dose — which is why the annotated schedule and the holds file exist: adaptation is the treatment for most of this ledger, and the ladder's pace is the dial. The practical translations: plan escalation weeks like mild-jetlag weeks (lighter calendar, blander plate, more fluids); judge a dose at week three, not day three; and treat a rough patch as a holds conversation with your prescriber rather than a quit decision — the single-digit discontinuation rows say most rough patches pass.

The head-to-head tolerability verdict

The direct trial answered the question comparison shoppers actually ask: does the stronger molecule cost more misery? Broadly, no — GI profiles ran similar across arms even as the efficacy gap (20.2% vs 13.7%) held. Individual variation still rules the last mile (some stomachs prefer one molecule for reasons no table predicts, and GIP's gut effects may explain some of tirzepatide's occasionally gentler reports) — but the population-level verdict removes tolerability as a reason to expect a worse ride from the better-performing option. That's a genuinely useful sentence, and it's the trial's, not ours.

The rare-but-serious column

Shared class rows, each with its own full file: pancreatitis (rare; severe persistent abdominal pain is a stop-and-call symptom — the file covers the biliary neighborhood too); gallbladder disease (elevated with rapid loss itself, therapy or not); the boxed thyroid warning (MTC/MEN2 history only — sized honestly); hypoglycemia (mainly with insulin or sulfonylureas aboard — the interactions file); and the aspiration-under-anesthesia conversation (the perioperative file). The ledger rule for this column: frequencies are low, stakes are high, and every row converts into one intake question or one symptom you'd never sit on.

Ledger → protocol

What the whole translation buys you: expect the GI window and pre-build the counter (protocol, plate, fluids); use holds as instruments, not defeats; rotate sites and log reactions; know your two stop-and-call symptoms (severe persistent abdominal pain; any serious-column sign) versus the big nuisance middle; and pick a program whose clinical team answers Tuesday questions — because a ledger this manageable still deserves management. Care that manages the ledger ↗

FAQ

Which has worse side effects, semaglutide or tirzepatide?

The head-to-head trial found broadly similar GI-forward tolerability even with tirzepatide's efficacy lead — nausea led both ledgers, events clustered in titration, and discontinuation stayed single-digit for both.

How common is nausea on these medications?

Roughly four-in-ten semaglutide patients and a quarter-to-a-third across tirzepatide arms reported it in pivotal trials — mostly mild-to-moderate and concentrated around dose escalations; exact rates live in the labels.

Do most people quit because of side effects?

No — adverse-event discontinuation stayed in the single digits in both programs; most rough patches pass with adaptation, protocol, and titration holds.

What side effects are actually dangerous?

The rare column: pancreatitis (severe persistent abdominal pain = stop and call), gallbladder disease, the MTC/MEN2 thyroid contraindication, hypoglycemia with certain diabetes drugs, and the perioperative aspiration conversation.

Sources

• Pivotal-trial adverse-event tables and prescribing information for both molecules (exact rates).
• Head-to-head trial tolerability reporting.
• Companion files: nausea protocol, holds, interactions, thyroid, gallbladder, perioperative.