Safety · Pharmacovigilance · Management Frameworks · May 2026

GLP-1 Receptor Agonists —
Side Effects & Management

The complete safety reference for GLP-1 receptor agonists — covering every adverse effect reported in clinical trials, post-marketing surveillance, and emerging pharmacovigilance data through May 2026. Each entry includes incidence data, mechanistic explanation, graded management protocols, agent-specific differences, and stopping criteria. Structured for practical clinical use at the point of prescribing and patient review.

S.K.H. Kazmi
Updated May 2026
SmPC · EMA EPAR · FDA FAERS · MHRA Yellow Card
25+ adverse effects covered
Safety Overview

The GLP-1 RA Safety Landscape

Class profile · Incidence hierarchy · Red flags

GLP-1 receptor agonists have one of the most extensively characterised safety profiles in modern pharmacology — with over 150,000 patient-years of randomised clinical trial exposure across the SUSTAIN, SURPASS, STEP, SURMOUNT, LEADER, SELECT, FLOW, and PIONEER programmes, supplemented by years of post-marketing surveillance data in millions of real-world patients. The class is broadly well-tolerated, with a safety profile dominated by gastrointestinal effects — most of which are predictable, dose-dependent, and self-limiting. Serious adverse events are rare but include pancreatitis, thyroid malignancy (animal data), retinopathy worsening in T2DM, and aspiration under anaesthesia.

20–30%
Nausea incidence at initiation (most common AE)
4–7%
Discontinuation due to GI side effects
<0.3%
Acute pancreatitis — rare but serious
+2–4 bpm
Mean resting heart rate increase (class effect)
Very common · >10%
Nausea
20–44%
Titration-dependent · Weeks 1–8 · Usually self-limiting
Very common · >10%
Diarrhoea
13–30%
Higher with tirzepatide · Usually transient
Common · 5–15%
Constipation
5–15%
More persistent than diarrhoea · Fibre + fluids
Common · 5–10%
Dyspepsia / Reflux
5–13%
Gastric emptying delay · Heartburn · Bloating
Common · Class effect
Heart rate increase
+2–4 bpm
Sustained · Monitor in arrhythmia · Mechanism known
Common · 1–3%
Cholelithiasis
1.5–3%
Rapid weight loss mechanism · Screen at risk patients
Uncommon · <0.5%
Acute Pancreatitis
<0.3%
Serious · Absolute CI after confirmed episode
Uncommon · T2DM
Retinopathy worsening
HR 1.76
SUSTAIN-6 signal · Screen before sema in T2DM
Emerging · Under review
Neuropsychiatric
Signal
Suicidality signal · FDA review resolved Jan 2026: no causal link established
💡
The Most Important Safety Principle for GLP-1 RAs
The vast majority of GLP-1 RA adverse effects are dose-dependent, titration-related, and self-limiting. The single most effective intervention for managing side effects is slower dose titration — not discontinuation. Patients who discontinue GLP-1 RAs due to GI side effects in weeks 1–4 would often have tolerated the drug if titration had been slowed. Pre-treatment counselling on expected side effect timing, severity, and duration is the most evidence-based intervention for improving tolerability and reducing early discontinuation.
Agent Comparison

Side Effect Profile Differences by Agent

Not all GLP-1 RAs have the same safety profile — clinically important differences

Side effect Sema SC (1–2 mg) Sema 2.4 mg Tirzepatide 15 mg Dulaglutide Liraglutide Exenatide XR
Nausea
~20%
~44%
~31%
~15%
~26%
~11%
Vomiting ~8% ~24% ~19% ~6% ~15% ~4%
Diarrhoea ~9% ~30% ~26% ~13% ~17% ~7%
Constipation ~5% ~24% ~15% ~6% ~11% ~3%
GI discontinuation ~5% ~7% ~4.5% ~5% ~9% ~3%
Heart rate ↑ +2–3 bpm +4 bpm +2–3 bpm +2 bpm +3 bpm +2 bpm
Cholelithiasis ~1.5% ~2.6% ~1.2% ~1.3% ~3.0% ~1.0%
Retinopathy signal SUSTAIN-6 HR 1.76 Not seen (SELECT) Limited data Not seen (REWIND) Not seen (LEADER) Not seen (EXSCEL)
Injection site reactions ~2% ~2% ~3.5% ~2% ~3% ~13% (nodules)
Hair loss ~3% ~3–5% ~5% ~2% ~2% ~1%
📊
Key Pattern: Dose Matters Most
The dominant predictor of GI side effect burden is dose magnitude, not the specific agent. Semaglutide 2.4 mg (Wegovy) causes substantially more nausea and vomiting than semaglutide 1 mg (Ozempic) — despite being the same molecule. Tirzepatide at 15 mg produces more diarrhoea than at 5 mg. This dose-dependence means staying at the lowest effective dose is the most important tolerability strategy — and that escalating patients to maximum dose when lower doses achieve clinical targets is poor practice.
GI Side Effects · Most Common

Nausea & Vomiting

🤢
Nausea & Vomiting
Very Common · >10% Self-limiting · Weeks 1–8 Most common reason for discontinuation
20–44%
Nausea incidence (dose-dependent)
Mechanism
GLP-1R activation in the area postrema (vomiting centre, outside the blood-brain barrier) and nucleus tractus solitarius (NTS) is the primary mechanism for GLP-1 RA-induced nausea. Simultaneously, gastric emptying delay causes gastric distension and early satiety that reinforces the nausea signal. At higher doses (semaglutide 2.4 mg, tirzepatide 15 mg), the hypothalamic appetite-suppressing signal itself may contribute via NTS activation, explaining why nausea is more prominent at initiation when dose escalation is rapid. Tachyphylaxis of the area postrema response explains why nausea improves with continued exposure — the receptor downregulates, attenuating the emetogenic signal while preserving glucose-lowering effects.
Incidence by Agent & Dose
Agent / Dose Nausea Vomiting Severe (>grade 2) Leading to discontinuation
Semaglutide SC 0.5 mg 15% 5% 2% 2%
Semaglutide SC 1 mg 20% 8% 3% 3%
Semaglutide SC 2.4 mg (Wegovy) 44% 24% 8% 5%
Tirzepatide 5 mg 17% 7% 2% 2%
Tirzepatide 10 mg 22% 13% 3% 3%
Tirzepatide 15 mg 31% 19% 5% 4.5%
Dulaglutide 1.5 mg 15% 6% 2% 2.5%
Liraglutide 1.8 mg 26% 15% 4% 7%
Oral semaglutide 14 mg 20% 9% 3% 11%
Management Protocol
📋
Before initiating: Counsel all patients that nausea is expected in weeks 1–8, peaks at dose escalation steps, and improves with time. Set realistic expectations — "this is not a signal to stop; it is expected and temporary." Provide written information on dietary modifications before first injection.
1
Dietary modifications (first-line): Eat small, frequent meals (5–6 per day rather than 3 large). Avoid fatty, spicy, or very sweet foods during dose-escalation period. Eat slowly and stop at first sense of fullness — the drug exaggerates satiety signals. Avoid lying down for 1–2 hours after eating. Cold or room-temperature foods are often better tolerated than hot meals.
2
Dose titration adjustment: If nausea is ≥moderate (affecting quality of life or food intake): do not escalate to the next dose step. Remain at the current dose for an additional 4 weeks (total 8 weeks at current dose) before retrying escalation. There is no clinical mandate to reach maximum dose — some patients do well long-term at sub-maximal doses.
3
Antiemetic therapy (if needed): Metoclopramide 10 mg TDS (short-term, maximum 5 days due to extrapyramidal risk) or domperidone 10 mg TDS (short-term, avoid if QTc prolonged) during dose escalation weeks. Ondansetron 4–8 mg as needed is an alternative (avoid in patients on QT-prolonging drugs). Antiemetics are not a substitute for dietary advice and slower titration.
4
Injection timing adjustment: For weekly SC injections — experiment with timing (morning vs evening). Some patients find evening injections reduce daytime nausea. For oral semaglutide — the fasting requirement amplifies nausea in some patients; ensure tablet is swallowed whole (not chewed) and the 30-minute fast is strictly observed.
⚠
Vomiting persisting >48 hours: Do not administer next dose while actively vomiting. Ensure adequate hydration. Check renal function if on ACEi/ARBi/SGLT2i/NSAIDs (AKI risk from volume depletion). If unable to maintain oral fluids after 48 hours — seek medical review; IV fluids may be required. Withhold SGLT2i and NSAIDs during significant GI illness.
✗
Stop drug if: Persistent severe vomiting (>3 days) despite dose hold and antiemetics; clinical dehydration with hypotension; inability to tolerate oral fluids; or acute pancreatitis suspected (upper abdominal pain radiating to back — check serum lipase urgently). Do not resume until cause clarified.
💡
When Nausea is Actually a Good Sign
Counter-intuitively, patients who experience moderate early nausea with GLP-1 RAs often achieve greater long-term weight loss — the nausea reflects robust GLP-1R activation in the area postrema and NTS, the same signalling that drives appetite suppression. The goal is to manage nausea to a tolerable level, not eliminate it entirely. Patients with no nausea whatsoever may occasionally be non-responders or may have been prescribed counterfeit product (a real pharmacovigilance concern with online/unlicensed purchasing).
GI Side Effects

Diarrhoea

🚽
Diarrhoea
Very Common · 10–30% Higher with tirzepatide & high-dose sema
13–30%
Diarrhoea incidence (by agent)
Mechanism
GLP-1R activation in intestinal enteroendocrine cells and myenteric plexus neurons increases intestinal motility — the opposite effect to gastric emptying delay. In the stomach, GLP-1 RAs slow emptying; in the small and large intestine, they accelerate transit. This discordant effect (slow proximal, fast distal) explains why some patients experience concurrent constipation and diarrhoea — alternating bowel habit is common early in treatment. Tirzepatide's higher diarrhoea incidence (26% at 15 mg) may relate to its GIPR agonism altering intestinal secretory function in addition to GLP-1R effects on motility.
Management
1
Dietary adjustment: Reduce high-fibre foods, lactose, and high-fat foods temporarily during dose escalation. BRAT diet (bananas, rice, applesauce, toast) acutely. Avoid sorbitol-containing foods and drinks. Reintroduce normal diet gradually once bowel habit stabilises.
2
Loperamide 2–4 mg PRN: Safe and effective for GLP-1 RA-associated diarrhoea. Can be used prophylactically on dose escalation days if diarrhoea is predictable. Not suitable if any concern about infective cause or inflammatory bowel disease.
3
Hydration: Oral rehydration salts (ORS) if >4 loose stools per day. Monitor for hyponatraemia in elderly patients. Withhold metformin if significant diarrhoea (lactic acidosis risk from dehydration + metformin). Review concurrent SGLT2i dose (dehydration risk compounds).
⚠
Investigate if: Blood in stool; fever; severe abdominal cramps; diarrhoea lasting >7 days without improvement; new diarrhoea after weeks of stable therapy (consider new cause — C. difficile, IBD, new onset coeliac). Rule out infective cause before attributing to GLP-1 RA.
GI Side Effects

Constipation

🔒
Constipation
Common · 5–24% More persistent than nausea
5–24%
Highest at obesity doses
Mechanism & Clinical Pattern
Gastric emptying delay reduces intestinal motility globally, though the magnitude varies by intestinal segment. In the large intestine, reduced colonic transit time leads to increased water absorption from stool, resulting in constipation. Unlike nausea, constipation may not improve with time and can persist throughout treatment — particularly at higher doses. The combination of reduced food intake (smaller appetite) reducing bowel habit frequency, decreased physical activity (energy conservation during rapid weight loss), and direct GLP-1R colonic effects all contribute. Constipation at high-dose sema 2.4 mg (24%) is particularly common and clinically significant, sometimes requiring pharmacological laxative intervention indefinitely while on therapy.
Management
1
Increase fluid and fibre intake: Target 2L water/day; 25–30g dietary fibre daily. Psyllium husk (ispaghula) 3.5–7g OD in water — first-line bulk-forming agent, safe and effective long-term.
2
Osmotic laxatives if dietary measures insufficient: Macrogol (polyethylene glycol) 17g OD-BD — safe, well-tolerated, non-habit forming. Lactulose 15–30ml BD is an alternative. Avoid stimulant laxatives (senna, bisacodyl) as first-line — reserve for acute constipation relief.
3
Physical activity: Recommend 30 minutes daily walking — exercise is independently associated with improved bowel transit and should be part of obesity management anyway.
⚠
Monitor for faecal impaction in elderly, immobile, or opioid-using patients. Alert re: ileus/obstruction: If constipation is accompanied by abdominal distension, absent bowel sounds, or severe pain — suspect ileus (see bowel obstruction section). This is a rare but serious complication, particularly relevant in the context of FDA safety communications in 2023–2024.
GI Side Effects

Dyspepsia, Reflux & Bloating

🔥
Dyspepsia · Heartburn · Abdominal Bloating
Common · 5–13% Gastric emptying delay mechanism
5–13%
Dyspepsia incidence
Gastric emptying delay causes gastric contents to remain in the stomach longer, increasing intragastric pressure and promoting reflux of acidic content into the oesophagus. Patients describe a sensation of prolonged fullness after small meals, upper abdominal bloating, belching, and heartburn — distinct from the nausea of direct emetic centre activation. Abdominal bloating is particularly prominent with tirzepatide and may persist beyond the initial titration phase. Pre-existing GORD is not a contraindication but symptoms will likely worsen during dose escalation.
Management
1
PPI therapy: Omeprazole 20 mg OD or lansoprazole 15–30 mg OD — first-line for symptomatic dyspepsia/heartburn. Review PPI interaction with oral semaglutide (PPIs reduce gastric acidity, potentially improving Rybelsus absorption; take oral sema ≥30 minutes before PPI).
2
Dietary: Small meals; avoid lying down after eating (2-hour rule); elevate head of bed; avoid alcohol and caffeine; reduce meal size further if bloating is dominant symptom.
3
Simethicone (activated dimethicone) for bloating and flatulence — safe, effective, non-systemic. Peppermint oil capsules (IBS-standard preparation) may reduce colonic spasm and bloating.
Serious GI · Black Box Warning

Acute Pancreatitis

⚠️
Acute Pancreatitis
Serious · Black Box Warning (FDA) Uncommon · <0.3% Absolute CI after confirmed episode
<0.3%
Incidence in RCTs
Mechanism & Evidence
GLP-1R is expressed on pancreatic acinar cells — the exocrine pancreatic cells that secrete digestive enzymes. GLP-1R activation may promote acinar cell proliferation and increase zymogen secretory activity, theoretically predisposing to acinar enzyme auto-activation (the initiating event in acute pancreatitis). However, the clinical evidence is conflicting: no dedicated CVOT or obesity RCT has demonstrated a statistically significant increase in pancreatitis incidence compared with placebo. The FDA and EMA mandated black-box warnings based on spontaneous post-marketing reports and biological plausibility, not RCT evidence. The background rate of pancreatitis in T2DM itself is elevated (2–3× vs general population), making causal attribution difficult. The LEADER (liraglutide), SUSTAIN-6 (semaglutide), REWIND (dulaglutide), and SELECT (sema 2.4 mg) trials showed numerically similar or lower pancreatitis rates in the active arm vs placebo — though none were powered to detect this outcome individually.
Clinical Recognition
Classic presentation: Persistent, severe epigastric pain radiating to the back; onset within hours of a meal; associated nausea/vomiting; patient distressed. Critical distinction from GI side effects: Nausea from GLP-1 RA typically occurs at dose escalation, is not associated with severe abdominal pain, and improves with dietary modifications. Severe, persistent upper abdominal pain with back radiation is a red flag requiring urgent investigation — this is not a normal GLP-1 RA side effect and must not be dismissed as such. Investigations: Serum lipase (>3× upper limit of normal = diagnostic); CT abdomen with contrast (if diagnostic uncertainty or assessing severity); CRP, FBC, LFTs, renal function, bone profile.
Management Protocol
1
STOP the GLP-1 RA immediately if acute pancreatitis is suspected. Do not withhold this step pending confirmation — the risk of continuing through an episode outweighs delay. Do not restart the suspected medicine after confirmed acute pancreatitis, in line with the current UK product information and MHRA warning; use the product-specific SmPC for subsequent decisions.
2
Hospital admission: All cases of suspected pancreatitis should be admitted for monitoring, IV fluid resuscitation (guided by fluid balance and renal function), analgesia (paracetamol IV first; opioids if required), and nutritional support (early enteral nutrition if mild-moderate severity; parenteral if severe).
3
Investigate for aetiology: Gallstones (USS), alcohol history, hypertriglyceridaemia (lipids — GLP-1 RAs reduce TG, so hypertriglyceridaemia at presentation suggests pre-existing dyslipidaemia rather than drug-induced). Exclude other causes — ERCP if biliary pancreatitis suspected. Drug causality assessment per the Naranjo algorithm.
⚠
Pre-prescribing screen: Always take a detailed history of prior pancreatitis before initiating any GLP-1 RA. A previous episode is a caution requiring an individual risk–benefit assessment and product-specific SmPC review, not a universal UK class contraindication. Check that patients have not had GI events attributed to a previous GLP-1 RA that could represent undiagnosed pancreatitis.
Serious GI · Post-Marketing Signal 2023

Bowel Obstruction & Ileus

🚨
Intestinal Obstruction / Paralytic Ileus
FDA Safety Communication · September 2023 Rare · Post-marketing signal
Rare
Post-marketing · FAERS reports
Background & FDA Action
In September 2023, the FDA issued a safety communication adding intestinal obstruction (including ileus) to the GLP-1 RA class label as a potential adverse effect, following analysis of FAERS (FDA Adverse Event Reporting System) spontaneous reports. The cases were predominantly in patients on semaglutide (Ozempic/Wegovy) and liraglutide (Victoza/Saxenda), though the signal extended across the class. A 2023 observational study (JAMA Internal Medicine, Sodhi et al., n=4,855) found GLP-1 RA users had a significantly higher risk of ileus (OR 3.95) and bowel obstruction (OR 2.53) compared with bupropion-naltrexone users — a finding that generated significant clinical and regulatory attention. Importantly, this was an observational study with significant confounders; RCT data do not demonstrate a statistically significant signal in primary analyses.
Clinical Recognition & Management
🚨
Red flag symptoms — urgent ED referral: Abdominal distension; absent or tinkling bowel sounds; absolute constipation (no passage of stool or flatus); progressive abdominal pain; vomiting of faeculent material; clinical dehydration with tachycardia/hypotension. These features are not consistent with routine GLP-1 RA GI side effects and require urgent surgical assessment. AXR and CT abdomen are first-line investigations.
⛔
Stop GLP-1 RA if intestinal obstruction or ileus confirmed. Do not re-challenge until underlying cause (adhesions, volvulus, or drug-related motility disorder) has been clarified. Surgical or gastroenterology review required before any re-prescribing decision.
⚠
High-risk patients to monitor: Prior abdominal surgery (adhesion risk); prior bowel obstruction; severe constipation on therapy that does not respond to laxatives; opioid co-prescribing (compounds GI motility reduction). Counsel all patients: if constipation worsens with bloating and absent flatus — do not take the next GLP-1 RA dose and seek medical review.
Serious GI

Gastroparesis — Worsening & Unmasking

🔄
Gastroparesis — Worsening or New Diagnosis
Severe gastroparesis: specialist caution Unmasking of subclinical disease
GLP-1 RAs delay gastric emptying — the therapeutic mechanism that contributes to postprandial glucose control. In patients with pre-existing subclinical or mild diabetic gastroparesis (estimated to affect 30–50% of longstanding T2DM), this additional emptying delay may unmask or dramatically worsen gastroparesis symptoms: prolonged early satiety, nausea hours after eating (not immediately as with typical GLP-1 RA nausea), vomiting of undigested food, significant weight loss beyond expected, bloating, bezoar formation. Confirmed or severe gastroparesis requires product-specific review and specialist risk assessment; it is not a universal UK class contraindication. Current tirzepatide SmPC wording advises caution in severe gastrointestinal disease, including severe gastroparesis. In practice, pre-existing gastroparesis is frequently undiagnosed; clinicians should ask about symptoms before prescribing and refer for scintigraphy if suspected.
1
If gastroparesis is confirmed on therapy: Withhold treatment and review the product-specific SmPC and individual risk with diabetes/gastroenterology specialists; permanent discontinuation is a clinical decision, not a universal class rule. Refer for gastroparesis management (dietary modification, prokinetics — domperidone, metoclopramide; endoscopic or surgical options in refractory cases). Address underlying diabetic autonomic neuropathy.
⚠
Pre-prescribing: Specifically ask about symptoms of early satiety, postprandial bloating lasting hours, vomiting of undigested food, and unexplained weight loss before GLP-1 RA initiation. If present — arrange gastric emptying scintigraphy before prescribing. Do not attribute these symptoms post-hoc to "GLP-1 RA side effects" without investigation.
⚠
Anaesthetic risk: Gastroparesis significantly amplifies the aspiration risk under general anaesthesia. If a patient on a GLP-1 RA is found to have gastroparesis and requires surgery — this is a critical anaesthetic risk factor. Inform the anaesthetic team immediately. See Anaesthesia section for full protocol.
Endocrine · Regulatory Warning

Thyroid — MTC Risk & C-cell Signal

🦋
Thyroid C-cell Effects — Medullary Thyroid Carcinoma Risk
US FDA boxed warning · jurisdiction-specific Human relevance uncertain Animal data only · No confirmed human MTC cases
Mechanistic Basis & Evidence
GLP-1R is expressed on thyroid parafollicular C-cells in rodents. In rat and mouse studies, long-term GLP-1R agonist exposure produced dose-dependent C-cell hyperplasia and medullary thyroid carcinomas (MTC) — leading to boxed warnings in US prescribing information for affected products. The critical question — whether this translates to humans — remains unresolved: human thyroid C-cells express GLP-1R at approximately 5% the density of rodent C-cells, calcitonin responses to GLP-1 RAs in humans are minimal or absent, and no confirmed drug-attributable human MTC cases have been reported in over 15 years of use and several pharmacoepidemiology studies. A 2022 observational study (Bezin et al., BMJ) found no increased MTC incidence in GLP-1 RA users vs metformin users in a French national database (n=2.5 million). Despite this reassuring human data, the contraindication in personal/family MTC history and MEN2 is jurisdiction- and product-specific. UK clinicians must use the current UK SmPC; Wegovy and Mounjaro section 4.3 currently list hypersensitivity rather than a class-wide MTC/MEN2 contraindication.
Practical Prescribing Implications
🚫
US label contraindications for affected products: Personal history of medullary thyroid carcinoma (MTC); family history of MTC in first-degree relative; Multiple Endocrine Neoplasia type 2A or 2B (MEN2A/2B — RET proto-oncogene mutation). These contraindications should not be transposed automatically to UK/EU products. Check the current jurisdiction-specific prescribing information.
⚠
Thyroid nodules (incidental or known): Not an absolute contraindication. Arrange thyroid ultrasound and serum calcitonin measurement before prescribing. If calcitonin elevated or nodule suspicious on USS (TIRADS 4–5) — refer to endocrinology/ENT for further assessment before initiating. Once nodules assessed as benign — GLP-1 RA may proceed with annual thyroid surveillance.
✓
Counsel all patients: Report new neck masses, hoarseness, dysphagia, or progressive difficulty swallowing during treatment. These symptoms are non-specific but warrant urgent thyroid assessment. Most will not be drug-related but should be assessed promptly.
ℹ
Calcitonin monitoring: NICE and EMA do not recommend routine calcitonin monitoring in the absence of symptoms or nodules. If monitoring is desired (e.g. patient anxiety): baseline calcitonin, repeat at 3–6 months; if stable and no nodules — annual monitoring thereafter. Calcitonin >20 pg/mL: investigate further. Calcitonin >100 pg/mL: urgent endocrinology referral.
Ophthalmology · SUSTAIN-6 Signal

Diabetic Retinopathy Worsening

👁️
Early Worsening of Diabetic Retinopathy
Uncommon · T2DM only SUSTAIN-6: HR 1.76 Not seen in SELECT (no T2DM) Semaglutide-specific signal
HR 1.76
Retinopathy complications (SUSTAIN-6)
Mechanism — The Early Worsening Phenomenon
The SUSTAIN-6 trial (semaglutide SC vs placebo in T2DM, n=3,297) found a significant increase in diabetic retinopathy complications (HR 1.76, 3.0% vs 1.8%, p=0.02). This early worsening phenomenon is well-recognised in diabetes management: rapid normoglycaemia after chronic hyperglycaemia paradoxically worsens retinopathy in the short term before long-term improvement occurs. The mechanism involves: sudden reduction in retinal pericyte perfusion pressure (adapted to chronic high glucose); vasoconstriction and ischaemia unmasking pre-existing fragile neovascularisation; Cotton-wool spots from capillary occlusion in pre-proliferative disease; worsening macular oedema from rapid osmotic gradient changes. This phenomenon was first documented in DCCT (intensive insulin therapy) and has been replicated with every treatment that produces rapid, large HbA1c reductions in patients with pre-existing retinopathy. It is not specific to semaglutide — it relates to the speed of HbA1c reduction. It was not seen in SELECT because that population had no T2DM (and thus no diabetic retinopathy baseline).
Risk Stratification
Retinal status Risk level Action
No diabetic retinopathy Low risk Proceed with standard initiation; routine annual retinal screening
Background/mild NPDR Low-moderate Confirm retinal screening up to date; titrate slowly; rescreen at 3–6 months
Moderate-severe NPDR Moderate-high Ophthalmology review before initiating; slow titration mandatory; 3-month rescreen
Pre-proliferative / PDR High Ophthalmology review mandatory before prescribing; consider withholding if active neovascularisation
Clinically significant macular oedema Very high Ophthalmology must clear before initiation; high-risk of worsening; avoid if CSME present without retinal specialist advice
1
Pre-prescribing: Check retinal screening status (within 12 months) for all T2DM patients before semaglutide. If screening overdue — arrange urgent digital retinal photography before first prescription. Document retinal status in clinical records.
2
Slow titration in moderate-high risk: Extend 0.25 mg starting dose for 8 weeks (rather than 4) before advancing to 0.5 mg. The rate of HbA1c reduction correlates with the degree of retinopathy worsening — a slower reduction is safer for the retina even if it delays glycaemic targets.
⚠
On therapy — urgent symptoms to report: Sudden visual loss or blurring; new floaters or flashes of light; any visual change should prompt urgent ophthalmology assessment within 24–48 hours. Do not wait for routine screening if symptomatic.
ℹ
Tirzepatide and other agents: The retinopathy signal is most clearly established for semaglutide SC based on SUSTAIN-6. Tirzepatide's SURPASS programme did not show a statistically significant retinopathy signal, though the trials were not powered for this endpoint. The same biological mechanism (rapid HbA1c reduction) applies — apply equivalent precautions with tirzepatide in patients with pre-existing retinopathy. Liraglutide (LEADER) and dulaglutide (REWIND) showed no significant retinopathy signal, possibly because their HbA1c reductions were more modest and slower.