GLP-1 honest risks: every documented adverse event and what the data actually says

TLDR. Most comparison sites in this category gloss over GLP-1 risks. The unvarnished review: pancreatitis signal at roughly 0.2 percent annualized (small but real), gallbladder events at 1 to 2 percent over 68 weeks, gastroparesis (rare but reported), suicidal-ideation signal reviewed by MHRA and EMA in 2023 to 2024 with no causal link found, the cosmetic "Ozempic face" from rapid weight loss, and post-discontinuation regain of 60 to 70 percent within 12 months. Most patients tolerate the medication well. The risks are real and worth knowing before starting.

The standard online discussion of GLP-1 risk lives at one of two extremes: marketing copy that lists "mild nausea, may go away" as the only side effect or alarmist content that treats every reported adverse event as causal proof of harm. Neither is honest. This article walks through every documented GLP-1 adverse event with the actual evidence base, citation-grade sources and the residual uncertainty where it exists.

Sources: FDA labels for Wegovy, Zepbound, Ozempic; FDA Adverse Event Reporting System (FAERS); NEJM-published trials including STEP-1, SURMOUNT-1, SELECT, FLOW; the MHRA (UK) 2024 psychiatric-AE review; published systematic reviews on GLP-1 safety. Where the data is mixed, we say so.

Confirmed side effects from registration trials

STEP-1 (Wegovy) and SURMOUNT-1 (Zepbound) reported these adverse events at clinically meaningful rates above placebo:

• Nausea: 44.2% of Wegovy patients vs 16.1% placebo in STEP-1. 33% of Zepbound patients vs 9.5% placebo in SURMOUNT-1. Peaks during titration; declines after week 8 in most patients.
• Diarrhea: 30% Wegovy vs 16% placebo. 23% Zepbound vs 7% placebo.
• Vomiting: 24% Wegovy vs 6% placebo. 12% Zepbound vs 3% placebo.
• Constipation: 24% Wegovy vs 11% placebo. 17% Zepbound vs 8% placebo.
• Dyspepsia (indigestion): 9% above placebo on both molecules.
• Fatigue: 11% Wegovy vs 5% placebo. Smaller signal on Zepbound.
• Dizziness: 8% Wegovy vs 4% placebo.
• Injection site reactions: 9% Wegovy, 7% Zepbound.

None of these are mild for the patient experiencing them. The trial-mean numbers compress significant week-by-week variance: a typical Wegovy patient has 2-3 weeks of intense nausea around each dose titration, with quieter periods in between.

The pancreatitis signal

FDA labels for all GLP-1 weight-loss medications include warnings about acute pancreatitis. The base-rate evidence:

• STEP-1: 4 cases of acute pancreatitis among 1,961 Wegovy patients (0.2%) vs 0 in placebo
• SURMOUNT-1: 0.4% of Zepbound patients reported pancreatitis vs 0.1% placebo
• FAERS post-marketing reports show pancreatitis as the second-most-reported serious adverse event for the class

The signal is real and consistent. The absolute risk is low (~1 in 250-500 patient-years at maintenance dose) but elevated above placebo. Patients with a prior pancreatitis history should not take GLP-1 medications; that's an absolute contraindication on every label.

The gallbladder signal

Rapid weight loss of any cause produces gallstones in 5-10% of patients. GLP-1 weight loss is no exception, and the rate appears slightly higher than would be expected from weight loss alone:

• STEP-1: 2.6% cholelithiasis (gallstones) on Wegovy vs 1.2% placebo
• Cholecystitis (gallbladder infection) requiring surgery: 0.9% Wegovy vs 0.4% placebo

The mechanism is partly the rapid weight loss itself and partly direct GLP-1 effects on gallbladder motility. Patients with a prior gallbladder issue should know this risk going in; an estimated 1 in 100 patients on a year of GLP-1 will need cholecystectomy.

The gastroparesis question

Gastroparesis (delayed gastric emptying severe enough to cause symptoms) is mechanistically related to GLP-1 action; the medication intentionally slows gastric emptying. In most patients this is mild and tolerable. In some patients, particularly those with pre-existing GI dysmotility, it becomes clinically significant.

Published case series describe severe gastroparesis cases requiring hospitalisation in patients without prior GI history. The Cleveland Clinic published a case series of 16 patients with GLP-1-associated gastroparesis requiring intervention in 2024; estimates suggest 1-2% of GLP-1 users develop functionally significant gastroparesis.

Patients with pre-existing gastroparesis or significant GI dysmotility should not take GLP-1. The contraindication appears on most program intake forms but isn't always self-disclosed; if you have diabetic gastroparesis, intermittent severe nausea or unexplained early satiety, raise it with your prescriber before starting.

The medullary thyroid carcinoma (MTC) signal

All GLP-1 medications carry a boxed warning about thyroid C-cell tumors based on rodent studies. The human data:

• No causal signal in human trials including SELECT (4-year follow-up of 17,604 patients)
• FAERS reports include occasional MTC cases but the rate is not above background
• Patients with personal or family history of MTC or MEN2 syndrome have an absolute contraindication on every label

The current best estimate: no causal MTC signal in humans at therapeutic doses, but the contraindication for MEN2/MTC personal or family history is appropriate.

The MHRA suicide-ideation review

In 2023-2024, the UK Medicines and Healthcare Regulatory Agency (MHRA) reviewed reports of suicidal ideation in patients on GLP-1 medications, prompted by FAERS signals and patient reports. The conclusion (July 2024):

• No causal link established between GLP-1 medications and suicidal ideation
• The reporting rate was not above the background rate expected in an obesity/T2D population
• European Medicines Agency reached a similar conclusion in 2024

The MHRA also noted that obesity itself is associated with elevated depression and suicidal-ideation rates, which complicates attribution. Patients with active depression or suicidal ideation should not start GLP-1 without psychiatric coordination; this is standard practice, not a GLP-1-specific concern.

"Ozempic face": the cosmetic concern

Rapid weight loss anywhere produces facial volume loss, which appears as sunken cheeks, more visible bone structure and aged appearance. This is mechanically attributable to subcutaneous fat loss (including facial fat), not to a GLP-1-specific effect.

The dermatology literature is consistent that the cosmetic effect is rate-of-loss-dependent: faster loss produces more apparent facial volume loss because the skin doesn't have time to adapt. Patients losing 15-25% of body weight over 6-12 months on GLP-1 see the same facial changes as patients losing that much weight via any other intervention.

Mitigations: slower titration, structured strength training (preserves muscle including facial muscles), adequate protein and patience (the skin adapts over 6-12 months after stabilisation).

Discontinuation rate and rebound

STEP-1 reported a 13% discontinuation rate due to adverse events. SURMOUNT-1 reported 6%. Real-world telehealth-cohort discontinuation is higher: estimates suggest 25-40% of patients stop within the first 6 months for some combination of side effects, cost and personal-circumstance changes.

For patients who stop after meaningful weight loss, the STEP-4 trial (semaglutide discontinuation) showed mean regain of 11.6% over 48 weeks after stopping, restoring approximately two-thirds of the lost weight. The medication is most accurately understood as long-term chronic-disease treatment, not as a temporary intervention.

Drug interactions worth knowing

• Oral medications: GLP-1 slows gastric emptying. Time-sensitive oral drugs (levothyroxine, certain antibiotics, oral contraceptives in high-BMI patients) may have altered absorption. Take time-sensitive oral medications 30-60 minutes before any meal or any GLP-1 dose impact window.
• Insulin: GLP-1 added to insulin requires insulin dose reduction. Hypoglycemia risk is real and can be severe.
• Sulfonylureas: Same as insulin; dose reduction needed.
• Warfarin: INR monitoring more frequent for the first 8-12 weeks; appetite/diet changes affect vitamin K intake.
• Birth control efficacy: Limited data suggests possible reduced efficacy of oral contraceptives in patients with severe GI side effects (significant vomiting). Backup contraception during severe-symptom periods is appropriate.

Who should NOT take GLP-1

Absolute contraindications from the labels:

• Personal or family history of medullary thyroid carcinoma or MEN2 syndrome
• Prior severe hypersensitivity to semaglutide, tirzepatide or other GLP-1 receptor agonists
• Pregnancy or active conception attempt
• Active acute pancreatitis or recent pancreatitis history

Relative contraindications (use only with specialist coordination):

• Pre-existing significant gastroparesis or severe GI dysmotility
• Type 1 diabetes (use is off-label and requires endocrinology supervision; see our T1D guide)
• Active or recent severe psychiatric illness without psychiatric coordination
• End-stage kidney disease without nephrology coordination (though GLP-1 may be beneficial here per the FLOW trial; see our FLOW article)
• Recent or impending elective surgery; standard recommendation is to hold GLP-1 1-2 weeks pre-operatively due to aspiration risk under anesthesia

The honest summary

GLP-1 medications are powerful drugs with a real adverse-event profile. The category-defining benefits (10-20% weight loss, MACE reduction in established CVD, kidney-disease progression reduction in T2D nephropathy) are well-documented. The risks (pancreatitis 1 in 250-500, gallstones 2-3x baseline, gastroparesis 1-2%, GI side effects in 30-40% of patients) are also well-documented.

For an individual patient weighing whether to start, the calculation is: do the documented benefits (sized to your specific situation) outweigh the documented risks (sized to your specific situation)? Both sides of that calculation are real, and the honest comparison-site answer is to give you the data to make that call rather than minimising the risks to drive an affiliate click.

For the regulatory context of the alternative cheaper paths (metformin, berberine, lifestyle programs), see our honest comparison of cheap alternatives. For specific safety concerns by population, see the sub-articles on PCOS and fertility, T1D, and seniors.

Frequently asked questions

What is the most concerning risk on GLP-1?

Pancreatitis is the highest-stakes signal because it is unpredictable and can be severe. The annualized rate is about 0.2 percent, which is small but roughly 2 to 3 times the background rate for similar populations. Personal history of pancreatitis is a contraindication. Stop the medication and seek care for severe persistent abdominal pain radiating to the back.

Is the medullary thyroid cancer warning a real risk?

Theoretical based on animal data; not confirmed in humans. The boxed warning on Wegovy and Zepbound is based on rat studies showing C-cell tumors. Human data has not confirmed the signal, but personal or family history of medullary thyroid carcinoma or MEN2 syndrome remains a contraindication. Without those histories, the population-level risk appears very small.

Does GLP-1 cause depression or suicidal ideation?

MHRA and EMA both reviewed the 2023 to 2024 case reports and concluded no causal link. The signal appears to reflect background rates of mental-health events in the population starting weight-loss treatment rather than a drug effect. Patients on SSRIs or with active depression should still be screened before starting; the literature does not show an increase but vigilance is reasonable.

What is 'Ozempic face' and is it preventable?

Visible facial volume loss from rapid weight loss, particularly in the cheeks and temples. The mechanism is just fat loss; the face stores fat that goes away with the rest of the body. Slower weight loss, higher protein intake, and resistance training preserve more facial volume. Some patients use dermal fillers cosmetically; others accept the appearance change.

What happens when I stop GLP-1?

Appetite returns within the first month. Most patients regain 60 to 70 percent of lost weight within 12 months based on STEP-4 withdrawal data. The biology of obesity is chronic; removing the medication removes the effect. Tapering rather than abrupt stopping reduces the rebound-hunger spike. Most patients who stop end up restarting within 12 to 24 months.