Glucagon-like Peptide-1 Receptor Agonists Associated Gastrointestinal Adverse Events: A Cross-Sectional Analysis of the National Institutes of Health All of Us Cohort

Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are an injectable class of medications that are well established for use in managing type 2 diabetes mellitus (T2DM). Due to their glucose-dependent mechanisms of insulin secretion augmentation and glucagon suppression, GLP-1 RAs are able to effectively improve glycemic control in T2DM patients without substantially increasing the risk of hypoglycemia [1]. They work by lowering plasma glucose through effects on insulin and glucagon secretion, as well as by slowing gastric emptying, leading to a reduction in plasma glucose levels [2]. The interaction of GLP-1 and its receptors is known to act through the cAMP-PKA pathway, promoting the synthesis of ATP and the release of insulin granules into the blood through exocytosis [1]. These mechanisms contribute to the glucose-lowering effects of GLP-1 RAs, making them valuable in the treatment of T2DM. Moreover, GLP-RAs can also contribute to weight reduction by reducing appetite and the feeling of hunger. This occurs through directly stimulating POMC/CART neurons and indirectly inhibiting neuropeptide Y (NPY) and agouti-related peptide (AgRP) neurons, which can lead to reduced energy intake [3].

GLP-1 RAs comprise several compounds with varying pharmacokinetic properties, duration of action, and clinical effectiveness [1,2]. GLP-1 RAs have also demonstrated effectiveness for weight management and maintenance of weight loss in the treatment of people with overweight and obesity [4,5,6]. They are also used as an adjunct for weight loss in people without diabetes [7]. Some GLP-1 RAs have also been approved and marketed specifically for weight loss based on evidence that they reduce food intake by stimulating insulin and inhibiting glucagon secretion postprandially [4]. Treatment guidelines recommend the use of GLP-1 RAs due to their positive impact on T2DM complication prevention, blood pressure and lipid level improvement, and amelioration of sleep apnea [1,8,9,10].

However, despite these advantages, GLP-1 RAs have been associated with several gastrointestinal (GI) adverse events, though reported differences between specific agents are conflicting [11,12,13]. GI adverse events may range from mild to serious, including abdominal pain, nausea, vomiting, constipation, diarrhea, delayed gastric emptying, and pancreatitis [11]. Recent research has linked the use of GLP-1 RAs for weight loss to an increased risk of pancreatitis, gastroparesis, and bowel obstruction [14]. Under-recognition and delayed management of GI adverse events can exacerbate pre-existing GI and renal conditions [15]. One study found higher reporting odds of GI adverse events with semaglutide versus liraglutide [12], while a systematic review of case reports showed higher rates with liraglutide and exenatide [13]. Another study by Liu et al. reported higher odds of GI adverse events with semaglutide followed by liraglutide and dulaglutide [11]. Most prior studies were based on randomized controlled trials or U.S. FDA passive surveillance data, which may not represent real-world practice. Examining the association between GLP-1 RAs and GI adverse events at the population level is important to establish generalizable evidence. Therefore, this study aimed to examine the association between GI adverse events and patients using GLP-1 RAs in a representative sample from the National Institutes of Health (NIH)’s All of Us research program cohort. Using real-world data on a diverse population can provide additional insights into the comparative safety profile of specific GLP-1 RAs regarding GI tolerability.

This real-world study of over 10,000 patients newly initiated on GLP-1 RAs aimed to evaluate some of the common and rare GI adverse events associated with these agents. We found high rates of abdominal pain, diarrhea, nausea, and other GI adverse events, especially among women and those with comorbidities. Differences existed between specific GLP-1 RAs regarding the risks of adverse events like gastroparesis and nausea and vomiting. Dulaglutide and liraglutide had higher GI adverse events than semaglutide and exenatide. These findings suggest variations in safety profiles between agents that should inform selection, particularly in patients with GI risk factors. Overall, our results from a large, diverse population align with prior evidence and provide further insights into the comparative GI tolerability of GLP-1 RAs in real-world practice [11,12,13,14,15].

Our findings showing higher GI adverse event rates with dulaglutide and liraglutide than semaglutide and exenatide are consistent with those of some prior studies. Liu et al.’s [11] analysis of FDA reports found higher severe GI events with liraglutide and dulaglutide, while Horowitz et al. [16] reported mostly mild side effects across agents, with more diarrhea in liraglutide users than exenatide. However, other studies conflict with our results. For example, Zhou et al.’s [12] pharmacovigilance study indicated that semaglutide had higher rates of multiple GI adverse drug reactions than liraglutide. Additional analyses found semaglutide was associated with the greatest risk for events like nausea and constipation compared to other GLP-1 RAs [11,12,14,17,18,19,20,21]. The reasons for discrepancies likely relate to differences in study design, data sources, and populations. Our real-world analysis of over 10,000 racially/ethnically diverse adults from the NIH All of Us research program, newly initiated on GLP-1 RAs, provides important complementary insights into comparative safety profiles, though findings remain mixed across studies. Further research is critical to clarify inconsistencies between observational studies, precisely identify predictors of GI events with GLP-1 RAs, and provide clarifying confirmatory or novel evidence about the safety profiles of these widely used T2DM medications. While challenging, achieving consistency across real-world analyses would strengthen confidence in the findings and better inform the clinical use of these agents.

Our study also identified that GI adverse event rates with GLP-1 RAs can vary based on several factors. Such factors included the specific drug used, patient demographics, and comorbidities. In this study, we found that compared to semaglutide, dulaglutide and liraglutide displayed higher odds of abdominal pain and nausea and vomiting. Dulaglutide, liraglutide, and exenatide also had significantly increased odds of gastroparesis compared to semaglutide. However, no significant differences were found in GI bleeding risk or pancreatitis between the GLP-1 RAs. In contrast, a real-world disproportionality study based on the FDA adverse event reporting system database found that semaglutide had the greatest risk of nausea, diarrhea, vomiting, constipation, and pancreatitis, while liraglutide had the greatest risk of upper abdominal pain [11]. The differences seen in the results across these studies may stem from variations in aspects such as the study design, the characteristics of the patient samples, and the analytical approaches used for data processing. For instance, the real-world study used data from the FDA adverse event reporting system, which includes reports from healthcare professionals, patients, and manufacturers, and may not be representative of the general population [11]. Furthermore, patient characteristics such as age, sex, race, and comorbidities can also influence the risk of GI adverse events. For example, male sex and increasing age were found to be significantly associated with lower odds of abdominal pain, diarrhea, nausea and vomiting, and pancreatitis compared to women and younger patients.

The high prevalence of GI events and differences in safety profiles between the GLP-1 RAs found in this real-world analysis should inform several aspects of clinical practice. The results highlight the need to consider patient risk factors like sex, age, race, and comorbidities when selecting a specific agent to mitigate adverse events. Clinicians should be aware that dulaglutide and liraglutide may confer greater GI adverse event risk compared than semaglutide or exenatide. Close monitoring for GI events is essential with any GLP-1 RA to ensure early recognition and appropriate management. Findings could eventually impact treatment guidelines or recommendations regarding first-line use or preferential prescribing in high-risk GI patients. Additionally, further investigation is needed through prospective studies confirming comparative safety, examining clinical outcomes related to GI events, and performing head-to-head comparisons of GLP-1 RAs.

Our study has several strengths and some limitations. To the best of our knowledge, this is the first population-based study in the United States (U.S.) evaluating the association between four GLP-1 RAs and GI adverse events. This is the largest and most diverse real-world study, comprising 10,328 people of an under-represented population. The study included participants with other comorbidities, such as CKD and HF, reflecting the general population to inform clinicians’ decision making and the appropriate selection of GLP-1 RAs given the GI risk profile. This is a cross-sectional study of an under-represented population that limits the establishment of causal inference and generalizability. Information obtained through electronic health records (EHRs) may inevitably result in implicit bias and misclassification. Additionally, GI adverse events related to concomitant medication use cannot be excluded. Therefore, future prospective studies with detailed medication profiles are warranted to inform causality, minimize study bias, and provide a comparative evaluation of GLP-1 RAs’ safety and clinical outcomes. Moreover, the All of Us research study participants were not selected based on the probability sampling method. Therefore, the findings of the current study should be interpreted with caution.

The findings of our study highlight that GLP-1 RA-related GI adverse events are common. GI safety profiles vary between agents, with exenatide appearing safer than other GLP-1 RAs, apart from gastroparesis. These results can guide clinicians in the decision making and the selection of GLP-1 RAs accounting for GI risk factors. However, these results should be interpreted with caution given this study’s limitations. Further studies are warranted to inform causality and provide a comparative evaluation of GLP-1 RA safety in relation to concomitant medication use.

The authors would like to acknowledge the AoU research program for providing access to the data. “The All of Us Research Program is supported by the National Institutes of Health, Office of the Director: Regional Medical Centers: 1 OT2 OD026549; 1 OT2 OD026554; 1 OT2 OD026557; 1 OT2 OD026556; 1 OT2 OD026550; 1 OT2 OD 026552; 1 OT2 OD026553; 1 OT2 OD026548; 1 OT2 OD026551; 1 OT2 OD026555; IAA #: AOD 16037; Federally Qualified Health Centers: HHSN 263201600085U; Data and Research Center: 5 U2C OD023196; Biobank: 1 U24 OD023121; The Participant Center: U24 OD023176; Participant Technology Systems Center: 1 U24 OD023163; Communications and Engagement: 3 OT2 OD023205; 3 OT2 OD023206; and Community Partners: 1 OT2 OD025277; 3 OT2 OD025315; 1 OT2 OD025337; 1 OT2 OD025276. In addition, the All of Us Research Program would not be possible without the partnership of its participants”.

Conceptualization, A.S.B. and W.A.A.; methodology, A.S.B.; data curation, W.A.A.; formal analysis, T.M.A.; writing—original draft preparation, A.S.B., T.M.A. and W.A.A.; writing—review and editing, W.A.A.; supervision, A.S.B. All authors have read and agreed to the published version of the manuscript.

The researcher workbench in All of Us employs a data passport model, through which authorized users do not need Institutional Review Board approval.

All the participants involved in the All of Us research program have provided written informed consent.

Data are available through the All of Us database.

The authors declare no conflicts of interest.

This research received no external funding. The article processing charges of this research work were partially covered by the North Dakota State University—RCA Research Support Grant (000365). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.