Navigating the Real-World Gastrointestinal Landscape of GLP-1s

The rapid adoption of glucagon-like peptide-1 receptor agonists (GLP-1 RAs) has fundamentally transformed the management of type 2 diabetes mellitus (T2DM) and obesity. By enhancing glucose-dependent insulin secretion, suppressing glucagon, delaying gastric emptying, and signaling satiety centers in the central nervous system, these pharmacotherapies achieve robust glycemic control and significant weight reduction. However, as prescribing expands across primary care, endocrinology, cardiology, and gastroenterology, clinicians increasingly encounter the downstream real-world consequences of GLP-1 RA therapy.

While randomized controlled trials (RCTs) established the efficacy and general safety of these agents, RCT protocols often underrepresent real-world patient complexity, heterogeneous comorbidities, and long-term adherence barriers. A landmark real-world study by Nathani et al. published in Digestive Diseases and Sciences provides essential population-level data evaluating gastrointestinal adverse events (GI-AEs) and healthcare resource utilization among more than 320,000 propensity-matched patients.

Why It Matters

The Scale of Real-World GI Toxicity

In clinical trials, gastrointestinal side effects are frequently characterized as transient, mild-to-moderate events that lead to treatment discontinuation in under 5% to 10% of participants. However, real-world registry data reveals a far more pervasive clinical burden.

Using the TriNetX® global federated health research network (comprising over 120 million patient records), Nathani et al. evaluated 164,987 patients initiating GLP-1 RA therapy matched 1:1 via propensity scores against 164,987 non-exposed controls with T2DM and overweight/obesity (BMI > 25 kg/m²). Within one year of treatment initiation:

• Nearly One-Third Affected: 31.86% of patients on GLP-1 RA therapy experienced at least one documented GI adverse event, compared to 26.05% in the control cohort (HR 1.19, 95% CI \[1.18, 1.21\], p < 0.05).
• Number Needed to Harm (NNH): The overall NNH for any GI adverse event was calculated at 17.2, underscoring that for every 17 patients started on a GLP-1 RA, one additional patient experiences a gastrointestinal complication within 12 months.

Granular Outcome Breakthroughs

The granular breakdown of primary outcomes highlights specific upper GI vulnerabilities that clinicians must proactively address:

1. Nausea and Vomiting: Remained the single most frequent event, occurring in 12.97% of GLP-1 users versus 10.03% of controls (HR 1.24 \[1.21, 1.26\], NNH 34.0). Landmark temporal analysis demonstrated that this risk is heavily concentrated in the first 3 to 6 months following drug initiation, corresponding to dose-escalation phases.
2. GERD and Esophageal Sequelae: De novo GERD occurred in 8.69% of GLP-1 patients (HR 1.18 \[1.14, 1.22\], NNH 58.1), mirroring an increased rate of new proton pump inhibitor (PPI) prescriptions (10.53% vs 8.69%; HR 1.15 \[1.12, 1.19\]). Crucially, GLP-1 RA exposure was associated with downstream mucosal inflammation and metaplasia, including higher hazards for esophagitis (3.10% vs 2.39%; HR 1.24 \[1.19, 1.30\]) and Barrett’s esophagus (0.25% vs 0.18%; HR 1.29 \[1.11, 1.49\]).
3. Gastroparesis: De novo gastroparesis diagnoses were significantly higher in the GLP-1 group (0.68% vs 0.41%; HR 1.57 \[1.43, 1.73\], NNH 367.6). While physiological delays in gastric emptying are intrinsic to the therapeutic mechanism of GLP-1 RAs, symptomatic persistence can transition into clinical gastroparesis.
4. Pancreatitis Discrepancy: Interestingly, while overall acute pancreatitis from all causes did not show a statistically significant difference between groups (0.83% vs 0.86%; HR 0.91 \[0.85, 1.00\]), specific coding for drug-induced pancreatitis demonstrated a substantial hazard elevation (HR 2.85 \[1.82, 4.48\], E-value 5.15). Though absolute incidence was low (0.05%), the high sensitivity E-value signals a strong potential causal relationship.

Healthcare Resource Re-Allocation: The Endoscopy Burden

A critical finding of Nathani et al. is how these adverse events redistribute healthcare utilization. GLP-1 RA use directly increased diagnostic invasive procedures:

• Elevated Diagnostic Endoscopy: 4.43% of GLP-1 patients underwent upper endoscopy (EGD) within 1 year compared to 3.37% of matched controls (HR 1.25 \[1.20, 1.29\], NNH 94.3). Number-of-instances (NI) analysis revealed a 24% increase in cumulative EGD procedures per patient-year (0.051 vs 0.041, p < 0.05).

Conversely, GLP-1 RAs did not increase acute care system collapse:

• Lower Acute Admissions & ER Encounters: Emergency department visits were slightly reduced (21.69% vs 22.20%; HR 0.96 \[0.95, 0.98\]), as were inpatient hospital admissions (24.90% vs 25.48%; HR 0.92 \[0.91, 0.94\]).
• Lower Biliary Interventions: Endoscopic retrograde cholangiopancreatography (ERCP) rates were lower in the GLP-1 cohort (HR 0.72 \[0.61, 0.85\]), as was ERCP specifically performed for gallstone extraction (HR 0.70 \[0.59, 0.85\]). Cholecystitis and choledocholithiasis risks were comparable between cohorts.

This indicates that while GLP-1 RAs generate a noticeable outpatient and diagnostic workload for gastroenterologists, their systemic cardiometabolic benefits likely mitigate emergency room utilization and major inpatient hospitalizations.

Who It Affects

Patient Demographics and Risk Stratification

Understanding the target population of this study allows clinicians to contextualize these findings within their own practice panel.

Population Profile

The matched cohort in Nathani et al. comprised 164,987 adult patients per group with the following baseline characteristics:

• Age: Mean age of 55.4 ± 13.7 years.
• Sex Distribution: 57.0% female, 38.9% male.
• Body Mass Index (BMI): Mean BMI of 37.4 ± 6.6 kg/m² in the GLP-1 cohort and 36.3 ± 6.8 kg/m² in controls.
• Glycemic Control: Mean HbA1c of 8.0 ± 2.1% (GLP-1) versus 7.5 ± 2.0% (control).
• Concomitant Therapies: High baseline metabolic complexity, with 64.2% on metformin, 47.1% on insulin, and 18.0% on SGLT2 inhibitors.

Because propensity matching rigorously balanced baseline HbA1c, BMI, Charlson Comorbidity Index, alcohol dependence, nicotine dependence, and prior GI history, these findings reflect the specific additive contribution of GLP-1 RA therapy in patients with combined type 2 diabetes and obesity.

High-Risk Patient Subtypes

Based on the hazard ratios and clinical mechanisms, several patient phenotypes warrant heightened vigilance prior to and during GLP-1 RA initiation:

1. Patients with Subclinical or Baseline Dysmotility: Patients with long-standing diabetes, peripheral neuropathy, or preexisting slow gastric emptying are at elevated risk for overt gastroparesis (HR 1.57).
2. Patients with Baseline Reflux or Hiatal Hernia: Given the increased risk of de-novo GERD (HR 1.18), esophagitis (HR 1.24), and Barrett’s esophagus (HR 1.29), patients with baseline heartburn or regurgitation require aggressive lifestyle and medical management upon starting GLP-1 therapy.
3. Patients Undergoing Elective Procedures Requiring Anesthesia: Delayed gastric emptying increases residual gastric volume, introducing potential aspiration risks during sedation or general anesthesia.
4. Patients on Rapid Titration Schedules: Because nausea and vomiting peak in the first 3 to 6 months during dose escalations, individuals sensitive to gastrointestinal changes require individualized, slower titration protocols.

What Changes

Actionable Recommendations for Clinical Practice

The insights provided by Nathani et al. necessitate a transition from passive side-effect reporting to structured, proactive patient management. Below are four key practice adaptations for primary care clinicians, endocrinologists, and gastroenterologists.

Protocol 1: Re-structure Patient Counseling & Expectation Setting

Unanticipated GI side effects are the leading driver of GLP-1 RA discontinuation. Counseling should move beyond generic warnings:

• Quantify Expectations: Inform patients that roughly 1 in 3 individuals experience mild-to-moderate GI symptoms within the first year, but emphasize that severe complications requiring ER visits or hospitalizations are rare.
• Timeline Transparency: Clarify that upper GI distress (nausea/vomiting) typically peaks during the first 12 to 24 weeks during dose titration and gradually stabilizes.
• Dietary Counseling: Advise smaller meal volumes, reduced fat intake, avoiding late-night dining, and immediately stopping eating upon reaching satiety to minimize gastric distension.

Protocol 2: Individualized Titration and Symptom Management

• Flexible Titration: Avoid rigid, calendar-based dose escalation. If a patient experiences moderate nausea or reflux upon stepping up a dose, delay the next increment or step down to the previously tolerated dose for an additional 2 to 4 weeks.
• Proactive GERD Management: For patients developing new-onset heartburn or regurgitation, initiate short-term H2-receptor antagonists or PPIs early. Nathani et al. noted a 10.53% rate of new PPI prescriptions in the GLP-1 cohort. Treating reflux promptly may prevent downstream esophagitis.
• Antiemetic Strategy: Short-term, as-needed prescriptions for antiemetics (e.g., ondansetron) during titration steps can prevent treatment abandonment.

Protocol 3: Algorithmic Approach to Diagnostic Workups (EGD Management)

Because GLP-1 RA therapy increases upper endoscopy utilization by 25% (HR 1.25), gastroenterologists and referring primary care providers must establish appropriate diagnostic thresholds:

• Avoid Premature Endoscopy: Avoid ordering EGD for mild nausea or early satiety occurring immediately after drug initiation or dose escalation, as these are physiological consequences of delayed gastric emptying.
• Indications for EGD: Reserve diagnostic EGD for patients presenting with alarming features (dysphagia, odynophagia, unexplained weight loss beyond intended targets, gastrointestinal bleeding, persistent severe vomiting), or refractory GERD symptoms persisting despite high-dose PPI therapy.
• Documenting Gastroparesis: Diagnostic coding for gastroparesis should ideally be confirmed with standardized gastric emptying scintigraphy after holding GLP-1 therapy, rather than relying solely on symptom presentation or endoscopy findings.

Protocol 4: Perioperative and Procedural Hold Protocols

Given the increased risk of gastroparesis (HR 1.57) and residual gastric contents, perioperative safety requires standardized hold times before elective procedures requiring sedation or anesthesia:

• Weekly Injections: Hold weekly GLP-1 formulations for 1 week prior to elective procedures.
• Daily Injections: Hold daily formulations on the day of the procedure.
• High-Risk Considerations: In patients with severe dysmotility or those who remained on GLP-1 RAs without appropriate hold times, point-of-care gastric ultrasound or full liquid/fasting precautions should be implemented to minimize aspiration risk during induction.

Conclusion

The real-world study by Nathani et al. provides essential clinical evidence regarding the gastrointestinal safety profile of GLP-1 receptor agonists in patients with diabetes and obesity. While these therapeutic agents offer remarkable glycemic and cardiorespiratory benefit, approximately 32% of patients experience GI adverse events, driving higher rates of proton pump inhibitor use and diagnostic upper endoscopies. Reassuringly, this GI burden does not translate into increased emergency department visits or hospital admissions.

By adopting flexible titration regimens, offering early symptomatic intervention for reflux, and establishing clear diagnostic pathways for endoscopy, clinicians can effectively manage side effects, improve treatment retention, and optimize long-term health outcomes for their patients.

Reference

1. Nathani P, Hamamah S, Velji J, Patel HK, Nathani R, Radadiya D, Shandilya A, Srinivasan S, Desai M, Repici A, Hassan C, Sharma P. Incidence of Gastrointestinal Adverse Events and Healthcare Resource Utilization in Patients Using Glucagon-Like Peptide-1 Analogs: A Real-World Study. Dig Dis Sci. 2026;71(8):10112-10127. doi:10.1007/s10620-026-10112-7