Gastrointestinal adverse effects of GLP‐1 receptor agonists and dual GIP/GLP‐1 receptor agonists: A narrative expert review of approved therapies, oral formulations, and pipeline agents

Abstract

Background

Glucagon‐like peptide‐1 receptor agonists (GLP‐1 RAs) and dual glucose‐dependent insulinotropic polypeptide (GIP)/GLP‐1 receptor agonists have transformed the management of type 2 diabetes and obesity. Gastrointestinal (GI) adverse events are the most common limitation of these therapies and a leading cause of discontinuation.

Methods

This narrative expert review synthesizes evidence from randomized controlled trials, observational studies, meta‐analyses, and pharmacovigilance data to characterize the incidence, mechanisms, and clinical implications of GI adverse effects across approved agents and late‐stage pipeline therapies.

Results

Nausea, vomiting, diarrhea, and constipation are established class effects, occurring in approximately 30%–50% of patients, typically during initiation and dose escalation. These events are generally mild to moderate and manageable with dose titration and dietary modification. Mechanisms include delayed gastric emptying, central activation of emetic pathways, altered intestinal motility, and physiologic effects of rapid weight loss itself. Randomized trial data are reassuring for acute pancreatitis, with no class‐level excess risk demonstrated; pharmacovigilance signals are most plausibly explained by diagnostic misclassification rather than true causal risk. Cholelithiasis shows a probable class‐level association, substantially mediated by weight loss rather than direct receptor signaling. Peri‐procedural data show increased residual gastric volume without confirmed aspiration events. Preliminary data on late‐stage pipeline agents (retatrutide, survodutide, and cagrilintide‐semaglutide) suggest a similar GI adverse event profile to approved agents, though evidence remains largely limited.

Conclusion

GI adverse events with GLP‐1‐based therapies are predictable and manageable. A proactive, patient‐centered approach is essential to optimize adherence, safety, and clinical outcomes as newer agents expand the therapeutic landscape.

INTRODUCTION

Glucagon‐like peptide‐1 receptor agonists (GLP‐1 RAs) and glucose‐dependent insulinotropic polypeptide (GIP)/GLP‐1 dual receptor agonists, collectively termed incretin‐based therapies or nutrient‐stimulating hormone (NuSH)‐based therapies, are among the most rapidly adopted pharmacological agents in modern metabolic medicine. ¹ , ² , ³ By mimicking endogenous GLP‐1, these agents stimulate glucose‐dependent insulin secretion, suppress glucagon, slow gastric emptying, and activate central satiety circuits via brainstem and hypothalamic GLP‐1 receptor populations. ¹ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ Tirzepatide additionally engages the GIP receptor, conferring complementary metabolic benefits on energy expenditure and insulin sensitivity. ⁹ , ¹⁰ , ¹¹ Their cardiometabolic efficacy has driven broad adoption across endocrinology, obesity medicine, gastroenterology, and primary care. The pharmacological profiles and GI‐relevant characteristics of approved agents are summarized in Table 1.

Table 1.

Pharmacological profiles of FDA‐approved NuSH therapies.

Agent	Class	Route/Freq	Indication(s)	Half‐life	GI‐relevant notes
Semaglutide SC	GLP‐1 RA	Weekly SC	T2D; Obesity	~7 days	Area postrema/NTS activation; ileal brake engagement 1 , 4 , 12
Tirzepatide	Dual GIP/GLP‐1 RA	Weekly SC	T2D; Obesity	~5 days	GIP co‐agonism attenuating nausea: preclinical hypothesis only 9 , 10
Liraglutide	GLP‐1 RA	Daily SC	T2D; Obesity	~1 day	Rapid peak; higher early nausea vs. weekly agents 13 , 14
Exenatide BID	GLP‐1 RA	Twice‐daily SC	T2D	~2.4 h	Short‐acting; highest nausea in pooled T2D analyses (~31.7%) 15 , 16
Lixisenatide	GLP‐1 RA	Daily SC	T2D	~3 h	Proximal GI preference; more nausea than dulaglutide 17
Dulaglutide	GLP‐1 RA	Weekly SC	T2D	~5 days	Lowest nausea vs. semaglutide in NMA 15 , 17
Oral Semaglutide 14/25 mg	GLP‐1 RA	Daily oral	T2D (14 mg); Obesity (25 mg—Phase 3)	~7 days	Requires fasting + 30‐min post‐dose fast; SNAC enhancer 18 , 19 , 20
Orforglipron (investigational)	Non‐peptide GLP‐1 RA	Daily oral	T2D; Obesity (Phase 3)	Not established	No food/water restrictions; dose‐dependent GI profile 21 , 22

Note: All GI risk statements are descriptive; no head‐to‐head comparative safety rankings are implied. Ileal brake = physiologic negative‐feedback reflex triggered by nutrients reaching the distal ileum.

Abbreviations: BID, twice daily; NMA, network meta‐analysis; NTS, nucleus tractus solitarius; NuSH, nutrient‐stimulating hormone; SC, subcutaneous; SNAC, sodium N‐[8‐(2‐hydroxybenzoyl)amino]caprylate; T2D, type 2 diabetes.

Despite their proven efficacy, GI adverse events are the most consistently reported limitation of GLP‐1 RAs and GIP/GLP‐1 receptor agonists, occurring in 30%–50% of patients across pivotal trials and accounting for the majority of treatment‐related discontinuations. ² , ²³ , ²⁴ , ²⁵ Common events such as nausea, vomiting, diarrhea, and constipation are typically mild to moderate and self‐limited, clustering around treatment initiation and dose escalation. ²⁶ The occurrence of acute pancreatitis and biliary disease under GLP‐1 and GIP/GLP‐1 receptor agonists has been reported, but as discussed in respective sections below, these warrant separate interpretation: randomized trial data are reassuring for pancreatitis, while biliary disease appears largely mediated by weight loss rather than direct receptor signaling. ²⁷ , ²⁸ Delayed gastric emptying with peri‐procedural implications and functional bowel motility changes carry distinct evidentiary bases addressed throughout this review.

This article is a narrative expert review covering US Food and Drug Administration (FDA)‐approved GLP‐1 RAs and GIP/GLP‐1 dual receptor agonists, with additional discussion of the investigational oral GLP‐1 RA orforglipron and three late‐stage pipeline compounds: retatrutide, survodutide, and cagrilintide‐semaglutide (CagriSema). Cross‐agent adverse event comparisons from independent trials are presented as descriptive summaries only and should not be interpreted as head‐to‐head safety rankings. Evidence is explicitly tiered throughout using the four‐level framework defined in the methods section.

METHODS

This is a narrative expert review. Protocol registration was not needed, consistent with the narrative review format. No Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA)‐style study flow, formal risk‐of‐bias assessment, or quantitative meta‐analysis was conducted.

A structured literature search was performed across MEDLINE/PubMed (inception to March 28, 2026), Embase (inception to March 25, 2026), and the Cochrane Library (inception to February 2026 issue). Keyword domains included: GLP‐1 receptor agonists; GIP receptor agonists; tirzepatide; semaglutide; liraglutide; exenatide; dulaglutide; lixisenatide; orforglipron; retatrutide; survodutide; cagrilintide; gastrointestinal adverse events; nausea; vomiting; diarrhea; constipation; gastroparesis; pancreatitis; cholecystitis; peri‐procedural; retained gastric contents. Two reviewers independently screened titles and abstracts; disagreements were resolved by discussion and adjudication by the senior author.

Eligible study types included randomized controlled trials (RCTs), cohort studies, meta‐analyses, and large pharmacovigilance analyses in adults aged ≥18 years, published in peer‐reviewed English‐language journals. Pipeline agents are discussed in a dedicated section; non‐peer‐reviewed sources (press releases) are explicitly labeled as provisional at every point of use. Case reports, conference abstracts without full text, non‐systematic narrative reviews, editorials, and letters were excluded.

Evidence hierarchy: (1) Established: consistent findings from multiple adjudicated RCTs or high‐quality meta‐analyses; (2) Probable: consistent observational cohort or pooled trial data with a biologically plausible mechanism; (3) Possible: limited RCT data, pharmacovigilance signals, or single cohort findings; (4) Conflicting or insufficient: discordant findings across study types or a small, potentially biased evidence base. Food and Drug Administration (FDA) Adverse Event Reporting System (FAERS) pharmacovigilance signals represent disproportionate reporting, not confirmed causality. FAERS was used only to identify disproportionality signals that informed hypothesis generation and signal contextualization. FAERS data were not used to estimate incidence or causal risk.

UPPER GASTROINTESTINAL ADVERSE EFFECTS

Upper GI symptoms are the most consistently reported adverse effects of GLP‐1 receptor agonists and tirzepatide, peaking during initiation and dose escalation. ²⁶ , ²⁹ A Bayesian network meta‐analysis of 48 RCTs (n = 27,729 patients with type 2 diabetes (T2D)) found dose‐dependent increases in nausea, vomiting, diarrhea, constipation, and dyspepsia across the class. ¹⁵ A separate systematic review and dose‐response network meta‐analysis specifically in non‐diabetic overweight or obese patients reached similar conclusions. ¹⁷ Note: diarrhea and constipation are lower GI events discussed fully in the Lower Gastrointestinal Adverse Effects section; they are mentioned here only as co‐reported outcomes in upper GI‐focused analyses. Mechanistic drivers include delayed gastric emptying mediated by GLP‐1 receptor activation at enteric myenteric neurons via nitrergic and cyclic adenosine monophosphate (cAMP)‐dependent pathways, ³⁰ and direct receptor activation in brainstem emetic centers (area postrema and nucleus tractus solitarius). ⁵

Nausea and vomiting

Evidence level: Established. Nausea occurs in approximately 10%–44% of recipients across approved agents and doses; vomiting occurs in 8%–25%. ³¹ , ³² , ³³ , ³⁴ Symptoms peak within the 1st week of initiation or dose escalation and decrease with continued use. ²⁶ A Bayesian network meta‐analysis found semaglutide associated with more nausea than dulaglutide (odds ratio [OR] 1.79) and exenatide (OR 1.48); these are indirect estimates subject to between‐trial heterogeneity and should not be read as clinical rankings. ¹⁵ Short‐acting GLP‐1 receptor agonists produce more nausea than long‐acting formulations, consistent with higher peak receptor activation, a finding supported by LEAD‐6 direct comparison data and Cochrane pooled analyses. ¹⁴ , ¹⁶ , ³⁵ Within‐trial placebo‐controlled data: STEP 1 reported vomiting 25% versus 6% placebo and SURMOUNT‐1 reported 13% versus 2% (Table 2; T2D‐trial rates in Table 3). ³⁶ , ³⁸ GI‐related discontinuation affected fewer than 10% of participants in most pivotal obesity RCTs. ⁴⁸ , ⁴⁹ , ⁵⁰

Table 2.

GI adverse events—Approved agents, obesity trials.

Agent (Dose)	Trial [Ref]	N active	Nausea A/Pbo (%)	Vomiting A/Pbo (%)	Diarrhea A/Pbo (%)	Constipation A/Pbo (%)	GI discont. (%)	AE discont. (%)
Semaglutide SC 2.4 mg	STEP 1 36	1306	44/16	25/6	30/16	24/11	4.3	6.8
Liraglutide 3.0 mg	SCALE Obesity 37	2487	39.3/13.8	13.0/3.5	NR/NR	NR/NR	~6.0	~9.9
Tirzepatide 15 mg	SURMOUNT‐1 38	2539	31/11	13/2	23/10	15/5	4.3	6.2
Tirzepatide 15 mg	SURMOUNT‐2a , 39	938	NR/NR	NR/NR	NR/NR	NR/NR	~7	13.8

Note: INTERPRETIVE CAUTION: All data are from independent, non‐head‐to‐head trials with different populations, doses, titration schedules, event definitions, and trial eras. Side‐by‐side presentation is descriptive only. Only within‐trial placebo‐controlled data support causal inference.

Abbreviations: A/Pbo, active arm/placebo arm; NR, not reported; SC, subcutaneous; T2D, type 2 diabetes.

^a SURMOUNT‐2 enrolled adults with obesity and T2D.

Table 3.

GI adverse events—Approved agents, T2D trials.

Agent (Dose)	Trial/Programme [Ref]	Source type	N Active	Nausea (%)	Diarrhea (%)	Any GI AE (%)	AE Discont. (%)
Exenatide BID	Cochrane pooled T2D RCTs 16	Cochrane systematic review	Pooled	31.7	~17	~50	~19.4
Liraglutide 1.8 mg	LEADER 40	Adjudicated CVOT	4668	15–20	~17	41.9	NR
Dulaglutide 1.5–4.5 mg	AWARD‐11 41	RCT (Lancet DE 2021)	~600/dose	~14	~8	NR	3.9
Semaglutide SC 1.0 mg	SUSTAIN‐6 42	Adjudicated CVOT	3297	~21.5	NR	41.9	NR
Oral Semaglutide 14 mg	PIONEER programme 20	Pooled programme review	Pooled	11–16	NR	39.1	7–15
Oral Semaglutide 25/50 mg	PIONEER PLUS 43	RCT (Lancet 2023)	1606	NR	NR	53–54	12–13
Tirzepatide vs. Semaglutide SC	SURPASS‐2 9	Active‐comparator RCT (NEJM 2021)	1133	17–22	NR	NR	6.0–8.5

Note: LEADER enrolled high‐CV‐risk T2D; rates may not generalize. SURPASS‐2 sema comparator arm: 4.1% AE discontinuation. Cross‐trial comparisons are descriptive only.

Abbreviations: CVOT, cardiovascular outcome trial; NR, not reported; SC, subcutaneous; T2D, type 2 diabetes.

Dyspepsia, early satiety, and belching

Evidence level: Established for occurrence; mechanism characterized in human myenteric neuron studies. ³⁰ , ⁵¹ Dyspepsia occurred in approximately 9%–10% versus 3%–4% placebo with semaglutide 2.4 mg and tirzepatide. ⁴⁸ Belching was reported in approximately 7% versus <1% placebo in STEP‐programme trials. ⁴⁸ GLP‐1 receptor activation at myenteric neurons reduces antral contractions and delays gastric emptying via nitrergic and cAMP‐dependent pathways; vagal afferent activation also contributes to early satiety signaling. ³⁰ , ⁵¹ Early satiety is both an intended therapeutic endpoint and an adverse symptom when it restricts adequate nutritional intake. Dietary adjustments, including smaller portions and slower eating, frequently provide relief. ⁵²

Abdominal pain

Evidence level: Probable for occurrence; agent‐specific comparative ranking is not established. Abdominal pain occurred in approximately 10%–20% of active‐arm participants versus 5%–10% placebo in pivotal RCTs. ⁴⁸ The 57.6% rate in the National Institutes of Health (NIH) All of Us cohort reflects broad International Classification of Diseases, 10th Revision (ICD‐10) Electronic Health Record (EHR) coding encompassing bloating, cramping, and distension, and is not equivalent to trial adverse event definitions. ⁵³ A comparison of RCT‐derived incidence ranges with NIH All of Us real‐world data across all major GI adverse events is provided in Table 4. Severe or persistent abdominal pain warrants a thorough medical evaluation to exclude conditions, including but not limited to pancreatitis, biliary disease, bowel obstruction, mesenteric ischemia, diverticulitis, peptic ulcer, and other causes of an acute abdomen. GLP‐1 RA use should not anchor the differential.

Table 4.

GI adverse effect incidence—RCT ranges vs. NIH All of Us real‐world data.

Adverse event	RCT active‐arm range	NIH All of Us (%) 53	Notes
Nausea	10%–44%	23.4 (combined w/vomiting)	RCT range across different trials/populations; not directly comparable. FAERS ROR highest with exenatide BID 15 , 16
Vomiting	8%–25%	—	Combined with nausea in All of Us report. STEP 1: 25% vs. 6% pbo 36 ; SURMOUNT‐1: 13% vs. 2% pbo 38
Abdominal Pain	10%–20%	57.6%	57.6% = broad ICD‐10 EHR coding; NOT equivalent to RCT adverse event definitions 53
Diarrhea	17%–30%	32.7%	Dose‐dependent. FAERS ROR 3.55 semaglutide 54 : reporting signal, not incidence. First‐month predominance
Constipation	11%–31%	30.4%	FAERS ROR 6.17 (95% CI 5.72–6.66) semaglutide 54 : reporting signal only. CagriSema REDEFINE 1: 30.7% vs. 11.6% pbo 44

Note: NIH All of Us Cohort: Aldhaleei et al. Pharmaceuticals 2024;17:199. ⁵³ EHR‐based cross‐sectional analysis; values represent coded diagnoses, not trial‐defined adverse event reporting. The 57.6% abdominal pain rate reflects broad ICD‐10 coding and is not directly comparable to RCT event rates.

Abbreviations: BID, twice daily; FAERS ROR, reporting odds ratio; reflects disproportionate reporting, not incidence or causation.

Gastroesophageal reflux disease (GERD)

Evidence level: Possible (observational signals with high confounding burden; outcome heterogeneity precludes firm conclusions). In pivotal obesity RCTs, GERD was reported in approximately 5% versus 3% placebo with semaglutide 2.4 mg, and approximately 2% with tirzepatide. ⁴⁸ A meta‐analysis of 55 trials found a pooled relative risk of 2.19 (95% confidence interval [CI] 1.48–3.25) for GERD. ²⁸ A large population‐based cohort study found that the increased GERD risk was confined to short‐acting GLP‐1 receptor agonists (hazard ratio [HR] 1.215; 95% CI 1.111–1.328), while long‐acting agents showed no significant association (HR 0.994; 95% CI 0.924–1.069). ⁵⁵ A separate T2D cohort study comparing GLP‐1 RAs with sodium‐glucose cotransporter‐2 (SGLT‐2) inhibitors found a relative risk (RR) of 1.27 (95% CI 1.14–1.42) for GERD. ⁵⁶ Important caveats: obesity independently causes GERD; dietary changes, reduced meal size, delayed gastric emptying, and surveillance bias are confounders that are not fully adjusted in observational studies. The class‐level GERD risk signal appears primarily driven by short‐acting formulations. ⁵⁵

LOWER GASTROINTESTINAL ADVERSE EFFECTS

Diarrhea, constipation, abdominal distension, and flatulence constitute the second major adverse event cluster. Diarrhea is typically early and intermittent, while constipation is more persistent and most prominent in obesity‐focused trials. ¹⁷ , ⁵³

Diarrhea

Evidence level: Established. Diarrhea occurs in approximately one‐quarter to one‐third of participants in pivotal obesity RCTs: 30% versus 16% placebo with semaglutide 2.4 mg in STEP 1, and 23% versus 10% placebo with tirzepatide 15 mg in SURMOUNT‐1. ³⁶ , ³⁸ The mechanism of GLP‑1 RA‐associated diarrhea remains incompletely characterized and may involve the ileal brake, a concept not yet well described in the GLP‐1 RA literature specifically. The ileal brake is a physiological negative‑feedback reflex by which delivery of nutrients, particularly fat and protein, to the distal ileum triggers the release of peptide YY and GLP‑1, reducing upper GI motility and gastric emptying rate, thereby altering small‑bowel transit and fluid handling. The strongest FAERS disproportionality signal is for semaglutide (reporting odds ratio [ROR] 3.55), representing reporting frequency rather than incidence or causality. ⁵⁴

Constipation

Evidence level: Established. Constipation affects approximately 11%–24% in approved‐agent pivotal trials, and up to 30.7% with CagriSema in REDEFINE 1. ⁴⁴ The strongest disproportionality signal is semaglutide (ROR 6.17, 95% CI 5.72–6.66): a reporting signal, not a confirmed incidence estimate. ⁵⁴ The mechanism involves GLP‐1 receptor‐mediated inhibition of myenteric neuron firing through nitrergic and cAMP‐dependent pathways, supported by direct human myenteric neuron studies. ³⁰ Constipation is dose‐dependent and should be anticipated and managed proactively. Patients should be counseled at initiation, and dietary fiber and hydration should be optimized before pharmacological measures are initiated. Refer to general constipation management guidelines (e.g., American College of Gastroenterology Guideline) for stepwise pharmacological therapy when dietary measures fail. ⁵⁷

Abdominal distension and flatulence

Evidence level: Probable. Distension (4%–7%) and flatulence (4%–6%) are reported significantly more often than with placebo with semaglutide and tirzepatide. ⁴⁸ Network meta‐analysis found increased distension with semaglutide (RR 1.42, 95% CI 1.13–1.79) and increased flatulence with semaglutide (RR 1.40) and tirzepatide (RR 2.42). ¹⁷ The physiologic mechanism—slowed intestinal transit promoting increased bacterial fermentation of undigested carbohydrates—is plausible but inferred from general GI physiology principles rather than directly demonstrated in GLP‐1 RA‐specific human mechanistic studies. ⁵⁸

UNCOMMON SERIOUS GI AND PANCREATOBILIARY ADVERSE EVENTS

Serious, uncommon adverse events, including pancreatitis, biliary disease, gastroparesis, bowel obstruction, and GERD, have received disproportionate regulatory and public attention. All observational associations are susceptible to residual confounding from T2D severity, autonomic neuropathy, obesity, concurrent medications, and healthcare utilization. The evidence hierarchy and key supporting studies for these events are summarized in Table 5.

Table 5.

Serious uncommon GI and pancreatobiliary adverse events—Evidence hierarchy.

Adverse event	Best available evidence	Key findings	Confidence rating	Principal confounders/caveats
Acute Pancreatitis	Adjudicated CVOTs; meta‐analysis 28 , 40 , 59 , 60	LEADER: 18/4,668 vs. 23/4,672 pbo. 40 Class meta‐analysis: not increased. 59 Real‐world: lower lifetime risk 60	No increased risk in RCT data; pharmacovigilance signals best explained by diagnostic misclassification	Drug‐induced abdominal pain and enzyme elevation can fulfill biochemical criteria without true pancreatitis 61
Cholelithiasis	Meta‐analysis of 55 RCTs 28	RR 1.46 (95% CI 1.09–1.97) at class level; no significant increase in cholecystitis or cholangitis	Probable (class‐level); weight‐loss mediation is a major unresolved pathway	Rapid weight loss independently promotes gallstones. Drug vs. weight‐loss effect cannot be disentangled
Gastroparesis	Observational cohort 27 ; Scandinavian cohort 62	HR 1.591 (US cohort 27 ); not reproduced in Scandinavian cohort. 62 No RCT signal	Conflicting/insufficient	T2D autonomic neuropathy; healthcare utilization bias. Pharmacological slowing ≠ clinical gastroparesis
Bowel Obstruction	Single observational cohort 27 ; Scandinavian cohort 62	HR 1.383 in US cohort 27 ; not reproduced. 62 No RCT signal	Conflicting/insufficient	Confounding by obesity, prior surgery, analgesic use, and constipation. Functional slowing ≠ mechanical obstruction
RGV (peri‐procedural)	OCULUS RCT 63 ; multi‐society guidance 64	25% continue vs. 3.1% hold (p = 0.003). Liquid diet eliminated RGV. No aspiration events	Established for physiologic RGV effect. Possible for clinical aspiration harm (undemonstrated in RCT)	RGV is a surrogate. OCULUS n = 60; early termination. Two‐center; generalizability limited
GERD	Meta‐analysis 28 ; cohorts 55 , 56 ; NMA 17	Pooled RR 2.19. 28 Short‐acting only: HR 1.215. 55 Long‐acting: no significant association. 55 T2D cohort RR 1.27 56	Possible (high confounding; risk primarily in short‐acting formulations)	Obesity independently causes GERD. Dietary change, meal size, delayed emptying, and surveillance confound attribution

Abbreviations: CVOT, cardiovascular outcome trial; FAERS, FDA Adverse Event Reporting System; HR, hazard ratio; NMA, network meta‐analysis; RGV, residual gastric volume; RR, relative risk; SGLT‐2i, sodium‐glucose cotransporter‐2 inhibitor; T2D, type 2 diabetes.

Acute pancreatitis

Evidence level: No increased risk in RCT data; pharmacovigilance signals best explained by diagnostic misclassification. There is no high‐quality evidence from adjudicated RCTs to suggest an increased risk of acute pancreatitis with GLP‐1 receptor agonists or GIP/GLP‐1 receptor agonists. The LEADER trial reported pancreatitis in 18 of 4668 liraglutide recipients versus 23 of 4672 placebo recipients, a non‐significant difference. ⁴⁰ A meta‐analysis of cardiovascular outcome trials confirmed no excess risk of pancreatitis at the class level. ⁵⁹ A real‐world United States (US) federated cohort analysis found GLP‐1 RA use associated with a significantly lower lifetime pancreatitis risk compared with non‐use. ⁶⁰ Earlier and more recent comparative cohort analyses similarly found no increased risk of acute pancreatitis with GLP‐1 receptor agonist use relative to other glucose‐lowering therapies. ⁶⁵ , ⁶⁶ The most plausible explanation for earlier pharmacovigilance signals is diagnostic misclassification: drug‐associated abdominal pain and transient pancreatic enzyme elevations can fulfill biochemical criteria for pancreatitis in the absence of true clinical pancreatitis. ⁶¹ , ⁶⁷ The preponderance of evidence from adjudicated RCTs is reassuring. Mild asymptomatic elevations in serum lipase and amylase may occur during therapy and should not, by themselves, be interpreted as acute pancreatitis without compatible symptoms and, when appropriate, imaging.

Gastroparesis and bowel obstruction

Evidence level: Conflicting/insufficient. One large U.S. observational study reported hazard ratios of 1.591 for gastroparesis and 1.383 for bowel obstruction with GLP‐1 RA use. ²⁷ These findings were not reproduced in a Scandinavian cohort comparing GLP‐1 receptor agonists with SGLT‐2 inhibitors. ⁶² No Randomized Controlled Trial (RCT) or Cardiovascular Outcome Trial (CVOT) has yielded a significant signal for either outcome. Critical caveats: T2D‐related autonomic neuropathy is an independent and major cause of gastroparesis and cannot be fully adjusted for in observational analyses. Pharmacological gastric motility slowing during active drug exposure must not be conflated with clinically diagnosed gastroparesis or mechanical bowel obstruction; objective motility testing is required before attributing either diagnosis to GLP‐1 RA exposure.

Importantly, the expected pharmacological slowing of gastric emptying, which underpins peri‐procedural caution (see Peri‐Procedural Considerations), is reversible upon drug discontinuation based on limited case data, whereas true gastroparesis implies persistent dysmotility independent of drug exposure. The distinction is clinically essential, as patients with drug‐induced motility slowing typically improve after discontinuation, while those with true diabetic or idiopathic gastroparesis may not. Evidence on the natural history of drug‐induced motility changes after withdrawal is limited and prospective discontinuation studies are lacking.

Biliary disease

Evidence level: Probable class‐level cholelithiasis risk; weight‐loss mediation is a major, unresolved confounder. A meta‐analysis of 55 placebo‐controlled trials reported a class‐level increase in the risk of cholelithiasis (RR 1.46, 95% CI 1.09–1.97), with no significant increase in cholecystitis or cholangitis. ²⁸ Rapid weight loss independently promotes gallstone formation through biliary cholesterol supersaturation and impaired gallbladder contractility. Disentangling a direct drug receptor effect from weight‐loss mediation requires a weight‐loss‐matched comparator, which does not currently exist. Clinicians should monitor for biliary symptoms, particularly in patients experiencing rapid weight loss or with pre‐existing gallbladder disease.

Peri‐procedural considerations

GLP‐1 receptor agonists and tirzepatide slow gastric emptying via GLP‐1 receptor activation at myenteric neurons, an expected class effect that may result in elevated residual gastric volume (RGV) at the time of upper endoscopy or procedures requiring general anesthesia. ³⁰ , ⁶³ Three related but distinct concepts must be clearly separated: (1) elevated RGV, a pharmacological surrogate endpoint; (2) aspiration risk, a biologically plausible downstream consequence; and (3) clinical aspiration, a serious harm outcome that has not been demonstrated in any published RCT.

The OCULUS RCT (n = 60, two centers) randomized patients undergoing elective upper endoscopy to continue or hold their GLP‐1 RA for one dose (weekly agents) or 1 day (daily agents) prior to the procedure. ⁶³ Clinically significant RGV occurred in 25.0% of the continuation arm versus 3.1% of the hold arm (absolute difference 21.9%; 90% CI 7.0%–36.7%; p = 0.003). A protocol‐mandated 24‐h clear liquid diet eliminated clinically significant RGV in all subgroups regardless of drug continuation. No aspiration events occurred in either arm. Generalizability across agents, doses, procedure types, and patient risk profiles requires further study.

Multi‐society guidance from the American Society of Anesthesiologists (ASA) and collaborating GI and surgical societies (October 2024) recommends individualized shared decision‐making rather than blanket drug‐holding rules, reserving specific modifications for higher‐risk patients. ⁶⁴ Specific drug‐holding durations in Table 6 are expert consensus based on pharmacokinetic reasoning. Current evidence does not support extending upper GI peri‐procedural precautions to lower GI procedures (e.g., colonoscopy).

Table 6.

Evidence‐based management of GI adverse effects and peri‐procedural safety.

Setting/symptom	Intervention	Evidence basis	Key reference(s)
Upper GI: Nausea (initiation)	Slow titration (≤monthly); smaller portions; low‐fat meals; upright posture post‐meals	Expert consensus; dose‐escalation rationale from RCT subgroup analyses	[36, 37, 52]
Upper GI: Nausea (persistent)	Dietary optimization first; ginger supplementation (250–500 mg QID); antiemetics if oral intake impaired. Prochlorperazine preferred in constipated patients. Ondansetron acceptable where constipation absent	Expert consensus; no comparative antiemetic RCT in GLP‐1 RA recipients	[24, 68]
Upper GI: Dyspepsia/Early Satiety	Smaller, more frequent meals; avoid high‐fat/carbonated beverages; slow eating; upright 30 min post‐meal	Expert consensus	[52, 68]
Lower GI: Constipation	Fluid ≥1.5–2 L/day; soluble fiber titration (psyllium 3–15 g/day as tolerated); osmotic laxative (PEG 3350, 17 g/day) if dietary measures insufficient; avoid insoluble fiber if nausea/bloating present	Expert consensus; extrapolated from ACG constipation guidelines. No GLP‐1 RA‐specific laxative RCT.	[57]
Lower GI: Diarrhea	Maintain hydration; low‐fat, low‐insoluble‐fiber diet; loperamide for persistent events (after excluding infection)	Expert consensus; no GLP‐1 RA‐specific antidiarrheal RCT	[68]
Peri‐procedural (upper GI): Higher‐risk	24‐h clear liquid diet prior to upper endoscopy (eliminates RGV per OCULUS). Individualize drug‐holding collaboratively with proceduralist/anesthesiologist.	RCT evidence for RGV reduction 63 ; multi‐society guidance 64 ; drug‐holding durations are expert consensus based on PK reasoning	[63, 64]
Peri‐procedural (upper GI): Lower‐risk	Continue GLP‐1 RA unless individualized assessment identifies additional risk factors; standard pre‐procedure fasting	Multi‐society guidance 64 ; OCULUS continuation arm	[63, 64]
Peri‐procedural (lower GI/colonoscopy)	No specific GLP‐1 RA modification currently supported; standard bowel prep and fasting apply	ASGE/ACG/AGA multisociety statement 2023 69 ; no RGV mechanism for lower GI tract	[69]

Note: All recommendations, except where specifically labeled trial‐supported, are expert consensus based on physiologic reasoning, extrapolation from general GI management principles, or multi‐society guidance. No comparative antiemetic or laxative RCTs have been conducted in GLP‐1 RA populations. Prochlorperazine preference over ondansetron in constipated patients reflects mechanistic reasoning only.

Abbreviations: ACG, American College of Gastroenterology; AGA, American Gastroenterological Association; ASGE, American Society for Gastrointestinal Endoscopy; PEG, polyethylene glycol; RGV, residual gastric volume.

ORAL GLP‐1 RECEPTOR AGONISTS: ORAL SEMAGLUTIDE AND ORFORGLIPRON

Two daily oral GLP‐1 receptor agonist formulations have been evaluated in late‐stage pivotal trials: oral semaglutide, which is a peptide formulation using sodium N‐[8‐(2‐hydroxybenzoyl)amino]caprylate (SNAC) absorption technology, and orforglipron, a non‐peptide small‐molecule GLP‐1 receptor agonist requiring no absorption enhancer and no food or water restrictions. ¹⁸ , ²¹ , ²²

Oral semaglutide 14 mg is approved for T2D; the 25 mg dose for obesity was evaluated in OASIS 4, which reported nausea in approximately 47% versus 11% placebo with a 7% adverse event (AE)‐related discontinuation rate. ¹⁹ Dose‐finding and PIONEER programme trials, including cardiovascular outcome and renal‐impairment studies, consistently identified gastrointestinal events as the most common adverse events with oral semaglutide. ⁷⁰ , ⁷¹ , ⁷² The higher GI burden compared with subcutaneous semaglutide 2.4 mg likely reflects faster systemic absorption kinetics via the SNAC mechanism and daily dosing.

Orforglipron is a non‐peptide small‐molecule GLP‐1 receptor agonist that does not require peptide absorption technology, food restrictions, or water restrictions before dosing. In ATTAIN‐1 (obesity Phase 3 RCT), nausea occurred in 33.7% at the 36 mg dose with a 10.3% AE‐related discontinuation rate. ²¹ In ACHIEVE (T2D Phase 3 RCT), diarrhea was the most prominent lower GI event (26%) with an 8.0% AE‐related discontinuation rate. ²² A published head‐to‐head trial (ACHIEVE‐3) comparing once‐daily oral orforglipron with once‐daily oral semaglutide in T2D found higher overall GI event rates and higher AE‐related discontinuation with orforglipron, though between‐arm differences were modest. ⁷³

PIPELINE AGENTS: RETATRUTIDE, SURVODUTIDE, AND CAGRISEMA

Three investigational agents with late‐stage Phase 3 programs are approaching regulatory review. Their GI adverse event profiles are reviewed below based on the evidence available at the time of the literature search (March 2026).

Retatrutide

Evidence level, Phase 2: Probable; Phase 3: Provisional only (press release; not peer‐reviewed). Retatrutide is a triple GLP‐1, GIP, and glucagon receptor agonist. ⁴⁵ In the Phase 2 RCT (n = 338), nausea occurred in approximately 43% and diarrhea in approximately 33% at the 12 mg dose, with a GI adverse event profile broadly similar to semaglutide at equivalent titration stages. ⁴⁵ Phase 3 TRIUMPH‐4 data are available only from an Eli Lilly press release (December 11, 2025) reporting AE‐related discontinuation of 12.1% in the body mass index (BMI) ≥ 35 group; the specific dose associated with this figure was not reported, and a GI breakdown was not provided. ⁴⁶ All Phase 3 safety conclusions must await peer‐reviewed publication.

Survodutide

Evidence level, Phase 2: Probable; Phase 3: Not yet published. Survodutide is a dual GLP‐1/glucagon receptor agonist in the SYNCHRONIZE Phase 3 obesity program. ⁷⁴ In the Phase 2 RCT (n = 387 across doses), AE‐related discontinuation ranged from 16% to 25%, the highest of the three pipeline agents discussed, primarily driven by GI events (Table 7). ⁴⁷ This observation is based on Phase 2 data only and requires Phase 3 confirmation.

Table 7.

Oral investigational and pipeline agents.

Agent (Dose)	Trial [Ref]	Indication	Source quality	Nausea A/Pbo (%)	Diarrhea A/Pbo (%)	Constipation A/Pbo (%)	AE discont. (%)
Oral Semaglutide 25 mg	OASIS 4 19	Obesity	Peer‐reviewed Phase 3 RCT (NEJM 2025;393:1077)	~47/~11	~18/~10	~20/~9	7.0
Orforglipron 36 mg	ATTAIN‐1 21	Obesity	Peer‐reviewed Phase 3 RCT (NEJM 2025 Sep 16)	33.7/~12	23.1/~10	25.4/~9	10.3
Orforglipron 36 mg	ACHIEVE 22	T2D	Peer‐reviewed Phase 3 RCT (NEJM 2025 Sep 16)	16/~6	26/~12	14/~7	8.0
CagriSema 2.4/2.4 mg	REDEFINE 1 44	Obesity	Peer‐reviewed Phase 3 RCT (NEJM 2025;392:2530)	55.0/12.6	NR/NR	30.7/11.6	6.0
CagriSema 2.4/2.4 mg	REDEFINE 2 44	Obesity+T2D	Peer‐reviewed Phase 3 RCT (NEJM 2025;392:2530)	NR/NR	NR/NR	NR/NR	8.4
Retatrutide 12 mg (Ph2)	Ph2 45	Obesity	Peer‐reviewed Phase 2 RCT (NEJM 2023;389:514)	~43/~12	~33/~11	~12/~5	NR
Retatrutide Ph3 TRIUMPH‐4	PROVISIONAL 46	Obesity	PROVISIONAL—press release only (Lilly Dec 2025)	NR	NR	NR	12.1a (BMI ≥ 35)
Survodutide (Ph2 pooled)	Ph2 47	Obesity	Peer‐reviewed Phase 2 RCT (Lancet DE 2024;12:162)	NR/NR	NR/NR	NR/NR	16–25

Note: PROVISIONAL data (press release) are not peer‐reviewed and must not be used for comparative conclusions. Cross‐trial comparisons are descriptive only.

Abbreviations: A/Pbo, active arm/placebo arm; NR, not reported.

^a TRIUMPH‐4 AE discontinuation: specific dose not reported in press release. ~ = approximate values from published text.

CagriSema (cagrilintide + semaglutide)

Evidence level: Established for GI profile within REDEFINE trials. CagriSema combines cagrilintide (an amylin receptor agonist) co‐administered with semaglutide 2.4 mg. In REDEFINE 1 (n = 3417), GI events occurred in 79.6% of active‐arm participants, with nausea in 55.0% and constipation in 30.7%, substantially higher than placebo (12.6% and 11.6%, respectively). ⁴⁴ Despite this high GI event burden, only 6.0% discontinued for AE‐related reasons, suggesting effective clinical management. ⁴⁴

PATIENT‐CENTERED MANAGEMENT OF GI ADVERSE EFFECTS

Effective management centers on three evidence‐informed strategies: slow dose titration, targeted dietary modification, and selective pharmacological intervention. Evidence levels for most specific recommendations are expert consensus, as comparative management RCTs in GLP‐1 RA populations are not available. Multidisciplinary collaboration involving the prescribing clinician, registered dietitian, and patient is essential.

Dose titration strategy

Slow dose escalation, which is advancing no faster than monthly, is the single most robust mitigation strategy, supported by dose‐escalation subgroup data across multiple RCT programs. ³⁶ , ³⁸ , ⁷⁵ Pre‐initiation counseling on the timing of expected GI symptoms and self‐management options improves treatment persistence. ⁵² Dose de‐escalation should be considered before discontinuation in patients with persistent severe GI events; returning to the previously tolerated dose for 4 additional weeks before re‐attempting escalation is a practical expert‐consensus approach.

Dietary and behavioral modifications

Dietary modifications represent first‐line management for upper GI events. Recommended strategies, consistent with current obesity and diabetes nutrition care guidelines, include: smaller, more frequent meals; low‐fat food choices; avoidance of carbonated beverages, fried or high‐fat foods, and spicy foods; slow eating and thorough chewing; remaining upright for at least 30 min after eating; and avoiding eating within 3 h of bedtime. ⁵² , ⁵⁷ , ⁶⁸

For constipation, a stepwise nutritional approach consistent with evidence‐based constipation management guidelines is recommended. ⁵⁷ Adequate fluid intake (minimum 1.5–2 L/day) should be ensured. Gradual titration of soluble fiber supplementation (e.g., psyllium husk, starting at 3–5 g/day and advancing to 10–15 g/day as tolerated) supports stool softening and colonic transit. Insoluble fiber should be avoided in patients with concurrent bloating or nausea. Proactive initiation of fiber supplementation at or before the start of therapy is preferable to reactive initiation once constipation is established. Patients with persistent dietary inadequacy due to GI symptoms should be referred to a registered dietitian for an assessment to ensure nutritional adequacy and prevent unintentional micronutrient deficiency.

For diarrhea, maintaining adequate hydration is the priority. A low‐fat, low‐insoluble‐fiber diet may reduce stool frequency. Infectious causes must be excluded if diarrhea is severe, prolonged, or accompanied by fever. ⁶⁸

Pharmacological management

Antiemetic use should be reserved for patients with persistent nausea impairing oral intake or quality of life despite dietary optimization. Prochlorperazine (5–10 mg orally or buccally) is mechanistically preferable to 5‐HT3 receptor antagonists (e.g., ondansetron) in patients with co‐existing constipation, because 5‐HT3 antagonism may further impair colonic transit; this preference is based on pharmacological reasoning only and has not been tested in a comparative trial in GLP‐1 RA recipients. ²⁴ , ⁶⁸ Where constipation is not a concern, ondansetron (4–8 mg as needed) is an acceptable alternative. Metoclopramide may be considered for refractory nausea, but the risk of tardive dyskinesia limits it to short courses. Ginger supplementation (250–500 mg four times daily) may be offered as a complementary option with modest antiemetic evidence in other contexts. ⁶⁸

For persistent constipation unresponsive to dietary measures, an osmotic laxative (e.g., polyethylene glycol 3350, 17 g/day) is the recommended first pharmacological step, consistent with general constipation guidelines. ⁵⁷ Stimulant laxatives (bisacodyl, senna) may be added if osmotic agents are insufficient. For persistent diarrhea after infectious causes are excluded, loperamide (2 mg after each loose stool, maximum 16 mg/day) is the standard option. ⁶⁸

Peri‐procedural and special considerations

For patients undergoing upper GI procedures or surgeries requiring anesthesia, individualized pre‐procedural assessment is recommended per multi‐society guidance. ⁶⁴ Higher‐risk patients may benefit from a 24‐h clear liquid diet before upper endoscopy regardless of drug continuation status, as this strategy eliminated clinically significant RGV in all OCULUS subgroups. ⁶³ Drug‐holding decisions should be made collaboratively with the proceduralist and anesthesiologist. Patients should be counseled not to discontinue GLP‐1 RAs on their own without prescriber guidance. In patients with chronic kidney disease, dedicated trials of dulaglutide and oral semaglutide demonstrated acceptable GI tolerability, although nausea and vomiting warrant particular attention given the risk of volume depletion. ⁷² , ⁷⁶

LIMITATIONS

This narrative expert review has several important limitations. Evidence selection and weighting reflect expert judgment and are not fully reproducible. No PRISMA study flow, risk‐of‐bias assessment, or meta‐analytic synthesis was performed. Cross‐agent adverse event comparisons in Tables 2, 3, 4, and 7 are from heterogeneous, non‐head‐to‐head trials and should not be interpreted as comparative safety rankings. Retatrutide Phase 3 data are available only from a company press release and are explicitly labeled provisional throughout. The inclusion of non‐peer‐reviewed evidence for emerging therapies is a limitation; all such data are clearly labeled and conclusions based on them are provisional.

CONCLUSION

GLP‐1 receptor agonists and dual GIP/GLP‐1 receptor agonists have transformed the treatment of obesity and type 2 diabetes, producing clinically meaningful reductions in body weight, glycemia, and cardiometabolic risk. GI adverse events, including nausea, vomiting, and altered bowel habits are the most frequently reported adverse effects (AEs) and remain the principal tolerability limitations. These events are generally dose‐dependent, mild to moderate in severity, transient, and most prominent during dose escalation, with rates varying by agent, formulation, dose intensity, and titration schedule. Randomized trial evidence is reassuring for pancreatitis, and weight‐loss mediation is the predominant explanation for biliary disease associations. Careful titration, dietary counseling, and individualized multidisciplinary care can improve tolerability, persistence, and adherence. Significant knowledge gaps remain regarding diarrhea mechanisms, long‐term GI safety, and optimal management of delayed or persistent GI side effects, underscoring the need for prospective trials and standardized real‐world evidence research.

AUTHOR CONTRIBUTIONS

Vikash Singhani: Conceptualization, methodology, investigation, visualization, and writing—original draft. Muhammad Ehsan, Sara Valencia, Ahmed Nadeem, and Eisha Moazzam: Investigation, data curation, validation, and writing—review and editing. Winfield Scott Butsch: Conceptualization, Validation, and writing—review and editing. Samita Garg: Conceptualization, methodology, project administration, supervision, validation, and writing—review and editing. All authors critically reviewed the manuscript, approved the final version, and agree to be accountable for the accuracy and integrity of the work.

CONFLICT OF INTEREST STATEMENT

Samita Garg reports receiving an Investigator Initiated grant from Abbott, research grants with Enterra and Atmo Biosciences, and serving as a speaker/consultant for Novo Nordisk and Ardelyx. W. Scott Butsch is an employee of Novo Nordisk. All other authors have no disclosures to declare.

ETHICS STATEMENT

This article is a narrative expert review based exclusively on previously published literature and publicly available data. No human participants, animals, patient‐identifiable information, or new primary data were involved. Therefore, institutional review board approval and informed consent were not required.