Background: Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are commonly used diabetes and obesity medications but have been associated with gastrointestinal (GI) adverse events. However, real-world evidence on comparative GI adverse reaction profiles is limited. Objectives: This study aimed to evaluate GI adverse events among GLP-1 RA users and compare semaglutide, dulaglutide, liraglutide, and exenatide safety regarding the GI adverse reaction profile. Methods: This retrospective cross-sectional analysis utilized real-world data on 10,328 adults with diabetes/obesity in the National Institutes of Health All of Us cohort. New GLP-1 RA users were identified, and GI adverse events were examined. Logistic regression determined factors associated with GI adverse events. Results: The mean age of the study population was 61.4 ± 12.6 years, 65.7% were female, 51.3% were White, and they had a high comorbidity burden. Abdominal pain (57.6%) was the most common GI adverse event, followed by constipation (30.4%), diarrhea (32.7%), nausea and vomiting (23.4%), GI bleeding (15.9%), gastroparesis (5.1%), and pancreatitis (3.4%). Dulaglutide and liraglutide had higher rates of abdominal pain, constipation, diarrhea, and nausea and vomiting than semaglutide and exenatide. Liraglutide and exenatide had the highest pancreatitis (4.0% and 3.8%, respectively). Compared to semaglutide, dulaglutide and liraglutide had higher odds of abdominal pain, and nausea and vomiting. They also had higher odds of gastroparesis than semaglutide. No significant differences existed in GI bleeding or pancreatitis risks between the GLP-1 RAs. Conclusions: In this real-world cohort, GI adverse events were common with GLP-1 RAs. Differences in GI safety profiles existed between agents, with exenatide appearing safer than other GLP-1 RAs, except for gastroparesis. These findings can inform GLP-1 RA selection considering GI risk factors. Further studies are needed to evaluate the causal relationship and GLP-1 RA safety with concomitant medication use.

There is a paucity of predictive models for uncontrolled diabetes mellitus. The present study applied different machine learning algorithms on multiple patient characteristics to predict uncontrolled diabetes. Patients with diabetes above the age of 18 from the All of Us Research Program were included. Random forest, extreme gradient boost, logistic regression, and weighted ensemble model algorithms were employed. Patients who had a record of uncontrolled diabetes based on the international classification of diseases code were identified as cases. A set of features including basic demographic, biomarkers and hematological indices were included in the model. The random forest model demonstrated high performance in predicting uncontrolled diabetes, yielding an accuracy of 0.80 (95% CI: 0.79–0.81) as compared to the extreme gradient boost 0.74 (95% CI: 0.73–0.75), the logistic regression 0.64 (95% CI: 0.63–0.65) and the weighted ensemble model 0.77 (95% CI: 0.76–0.79). The maximum area under the receiver characteristics curve value was 0.77 (random forest model), while the minimum value was 0.7 (logistic regression model). Potassium levels, body weight, aspartate aminotransferase, height, and heart rate were important predictors of uncontrolled diabetes. The random forest model demonstrated a high performance in predicting uncontrolled diabetes. Serum electrolytes and physical measurements were important features in predicting uncontrolled diabetes. Machine learning techniques may be used to predict uncontrolled diabetes by incorporating these clinical characteristics.

This study aimed to examine the difference in health-related quality of life (HRQOL) and diabetes-related healthcare events (HCEs) among adults with diabetes who were on metformin, sulfonylurea, insulin, or thiazolidinedione (TZD) monotherapy. The data were sourced from the Medical Expenditure Panel Survey (MEPS). Diabetes patients ≥18 years old who had a complete record of physical component score and mental component scores in round 2 and round 4 of the survey were included. The primary outcome was HRQOL of diabetes patients as measured by the Medical Outcome Study short-form (SF-12v2^TM). Multinomial logistic regression and negative binomial regression were conducted to determine associated factors of HRQOL and HCE, respectively. Overall, 5387 patients were included for analysis. Nearly 60% of patients had unchanged HRQOL after the follow-up, whereas almost 15% to 20% of patients showed improvement in HRQOL. The relative risk of declined mental HRQOL was 1.5 times higher relative to unchanged mental HRQOL in patients who were on sulfonylurea 1.55 [1.1–2.17, p = 0.01] than metformin users. The rate of HCE decreased by a factor of 0.79, [95% CI: 0.63–0.99] in patients with no history of hypertension. Patients on sulfonylurea 1.53 [1.20–1.95, <0.01], insulin 2.00 [1.55–2.70, <0.01], and TZD 1.78 [1.23–2.58, <0.01] had increased risk of HCE compared to patients who were on metformin. In general, antidiabetic medications modestly improved HRQOL in patients with diabetes during the follow-up period. Metformin had a lower rate of HCE as compared to other medications. The selection of anti-diabetes medications should focus on HRQOL in addition to controlling glucose level.