Limited · humanPreprint
This large-scale observational study used an EHR-linked body-composition "digital phenotyping" pipeline — incorporating large language model (LLM)-based data extraction — to compare lean body mass (LBM) changes in routine clinical care among adults initiating GLP-1 receptor agonists (GLP-1RAs). Of 670,422 first-episode GLP-1RA users (456,742 on semaglutide; 213,680 on tirzepatide), 7,965 had paired pre- and post-initiation body-composition measurements analyzable over 12 months. The study found that tirzepatide was associated with greater relative LBM loss than semaglutide at each time point assessed (3, 6, 9, and 12 months), with excess LBM losses ranging from approximately 1.1% to 2.0%. The authors also identified two GLP-1 "metabotypes": a "Depletive" metabotype (>20% total body weight loss with >5% LBM loss), which was more frequent with tirzepatide (10.3%) than semaglutide (6.7%), and a "Prime" metabotype (>10% total body weight loss with preserved LBM). Key limitations include the observational, real-world design; reliance on LLM-extracted EHR data; potential unmeasured confounding; and the relatively small subset with paired body-composition data relative to the overall cohort.
Unknown journal · May 2026DOI ↗ Limited · human
This pharmacovigilance study used the U.S. FDA Adverse Event Reporting System (FAERS) database (Q2 2022–Q2 2025) to examine whether specific GLP-1 receptor agonists (GLP-1 RAs) — exenatide, lixisenatide, liraglutide, dulaglutide, semaglutide, and tirzepatide — differ in their association with alopecia and reproductive or endocrine-related adverse events. Researchers identified 1,276 alopecia-related and 759 reproductive/endocrine-related cases, then applied disproportionality analyses using crude and adjusted reporting odds ratios (cROR/aROR) from logistic regression, controlling for potential confounders. Sensitivity analyses with positive and negative control drugs were used to validate signal robustness. The study found that semaglutide was significantly associated with disproportionate reporting of alopecia (aROR 1.23) and several hormonal conditions, including polycystic ovary syndrome (aROR 6.59) and menstrual abnormalities. By contrast, dulaglutide and tirzepatide showed negative or non-significant associations for several reproductive outcomes. Important limitations include the inherent biases of spontaneous reporting systems (e.g., underreporting, notoriety bias, inability to establish causation), lack of denominator data, and absence of clinical detail. The authors conclude that agent-specific differences in endocrine and dermatologic safety profiles may warrant personalised prescribing decisions and ongoing surveillance.
Diabetes/metabolism research and reviews · May 2026DOI ↗ Limited · human
This pharmacovigilance study examined safety signals associated with potentially counterfeit semaglutide products by conducting a descriptive and disproportionality analysis of Individual Case Safety Reports (ICSRs) submitted to the European pharmacovigilance database EudraVigilance between January 2018 and December 2025. Researchers identified 234 ICSRs linked to suspected counterfeit semaglutide, of which 73.5% involved female patients and 35.5% involved adults. Notably, 89.3% of reported suspected adverse drug reactions (ADRs) were classified as serious. The most commonly reported suspected ADRs were vomiting, nausea, and hypoglycemia. Disproportionality analysis revealed a statistically higher reporting frequency for hypoglycemia, product use in an unapproved indication, malaise, and drug ineffectiveness compared with reports involving non-counterfeit semaglutide. The authors suggest these signals may reflect contamination, incorrect active ingredients, or subtherapeutic dosing in counterfeit products. Limitations include the inherent biases of spontaneous pharmacovigilance reporting (underreporting, lack of denominator data, and inability to confirm causality), reliance on database classifications of "counterfeit," and the descriptive nature of the analysis. The study concludes that pharmacovigilance databases can play a useful role in detecting safety concerns related to counterfeit medicines, and calls for further research.
Frontiers in pharmacology · Apr 2026DOI ↗ Limited · human
This large-scale observational study used a federated biomedical data platform to analyze 135,349 individuals treated with GLP-1 receptor agonists (GLP-1RAs) — specifically semaglutide (Wegovy) and tirzepatide (Zepbound) formulations — to characterize differences between "super responders" (>15% weight loss), "moderate responders" (5–15% weight loss), and a "minimal weight-loss" group. The study found substantial heterogeneity in weight-loss outcomes across patients receiving the same therapies. Notably, super responders to Zepbound showed reduced risk of developing certain comorbidities, including conditions at relative risks as favorable as 0.5 for osteoarthritis (P = .001), while Wegovy super responders showed an association with psoriasis (RR = 2.5, P = .03). The authors conclude that differences in weight trajectories likely reflect a combination of biological, behavioral, and social factors. Key limitations include the observational, retrospective design (which cannot establish causation), the reliance on federated real-world data (subject to coding variability), and lack of randomization. The authors call for prospective studies to develop more individualized weight-loss strategies.
Biology methods & protocols · Apr 2026DOI ↗ Limited · human
This real-world observational study examined weight trajectories in 4,182 patients during the 6 months following their last documented semaglutide or tirzepatide prescription, drawn from a federated health network. The study found that approximately two-thirds of the full cohort showed stable weight or continued weight loss after their final prescription. In a representative subset of 300 patients whose clinical notes were analyzed using a large language model (LLM), treatment discontinuation was confirmed in 119 patients (40%), and of those, 72% did not demonstrate weight regain. The study also noted that exercise counseling was documented more frequently among patients who maintained weight loss compared to those who experienced weight regain (26.2% vs. 14.7%; P = .04). Key limitations include the observational and retrospective design, reliance on documented prescriptions rather than confirmed medication use, potential incompleteness of clinical records, use of an LLM for data curation introducing possible inaccuracies, and the inability to establish causation. The authors acknowledge that further studies are needed to understand the mechanisms behind these real-world patterns of weight maintenance after GLP-1 receptor agonist discontinuation.
Biology methods & protocols · Apr 2026DOI ↗ Limited · human
This cross-sectional study conducted in Saudi Arabia (January–June 2025) investigated the frequency, characteristics, and predictors of hair loss among 254 adults using GLP-1 receptor agonists (GLP-1RAs) — specifically semaglutide (Ozempic), tirzepatide (Mounjaro), liraglutide (Saxenda), or liraglutide (Victoza) — primarily for weight loss. Data were collected via structured questionnaires covering demographics, clinical characteristics, and hair loss details such as timing, severity, and progression. The majority of participants were female (71.3%), with a mean age of approximately 33 years. The study found that overall hair loss prevalence did not differ significantly across GLP-1RA types (p = 0.116); however, severe hair loss was reported significantly more often among Mounjaro (43.4%) and Saxenda (42.9%) users. Female sex and Mounjaro use were identified as notable predictors of hair loss. The authors noted that the hair loss observed was generally non-scarring and potentially reversible, but associated with psychological distress and possible impacts on treatment adherence. Key limitations include the cross-sectional design (precluding causal inference), reliance on self-reported data, the single-country sample limiting generalizability, and the absence of a control group not using GLP-1RAs.
Journal of cosmetic dermatology · Apr 2026DOI ↗ Limited · human
This pharmacovigilance study compared post-marketing adverse event reporting patterns for tirzepatide (a dual GIP/GLP-1 receptor agonist) against semaglutide and liraglutide (GLP-1 receptor agonists) using aggregated Individual Case Safety Reports (ICSRs) from the EudraVigilance database. Researchers applied pairwise disproportionality analyses using reporting odds ratios (RORs) at the System Organ Class (SOC) level, with false discovery rate correction and sensitivity analyses limited to serious and healthcare professional-reported cases. The study found that, compared with semaglutide, tirzepatide was associated with higher relative reporting for immune system disorders (ROR 1.97) and hepatobiliary disorders (ROR 1.71). Compared with liraglutide, tirzepatide showed higher reporting for musculoskeletal (ROR 2.02) and psychiatric disorders (ROR 2.14), but lower reporting for neoplasms (ROR 0.28). Eight SOCs remained significant across all analytical conditions. The authors emphasize that these findings are hypothesis-generating only, as disproportionality analyses cannot establish causality, do not adjust for exposure levels, and are subject to reporting biases inherent in spontaneous pharmacovigilance databases. Confirmation in exposure-adjusted studies is recommended.
International journal of molecular sciences · Mar 2026DOI ↗ Limited · human
This study characterizes PG-102, a bispecific Fc fusion protein that co-activates both GLP-1 and GLP-2 receptors, combining preclinical animal work with a Phase I human trial. In db/db mice (a model of advanced type 2 diabetes with hyperglycemia and catabolic weight loss), PG-102 demonstrated superior and more sustained glycemic control compared to semaglutide or tirzepatide while preserving body weight — an effect attributed to β-cell preservation and enhanced glucose uptake rather than acute insulin stimulation. Mechanistic experiments suggested that dual receptor engagement was necessary for these benefits, with PG-102 also promoting coordinated, delayed receptor internalization compared to monospecific agents. The Phase I randomized, double-blind, placebo-controlled trial enrolled 24 adults with overweight (BMI 25–30 kg/m²) across three dose cohorts, with 18 receiving PG-102 and 6 receiving placebo. The primary endpoint was safety and tolerability. Treatment-emergent adverse events occurred in 83.3% of PG-102 participants and 66.7% of placebo participants; gastrointestinal events were mild to moderate. No serious adverse events or treatment discontinuations were reported. Limitations include the small sample size, single-center design, overweight-only population, and lack of efficacy endpoints in the human portion of the study.
Nature communications · Mar 2026DOI ↗ Limited · human
This pharmacovigilance study analyzed ocular adverse event (AE) reports associated with five GLP-1 receptor agonists (exenatide, tirzepatide, dulaglutide, liraglutide, and semaglutide) using the FDA Adverse Event Reporting System (FAERS) from 2005 to 2024. Ocular AEs comprised 3.61% of all GLP-1 RA-related reports; the median patient age was 63 years, and 62.6% of reports involved female patients. Disproportionality analysis using reporting odds ratios (RORs) identified a statistically significant signal for ocular AEs associated with semaglutide, while exenatide showed a significant annual decline in proportional reporting (-5.15% per year). The authors note that semaglutide's signal may partly reflect its growing market dominance rather than a true differential risk. Key limitations acknowledged by the study include the absence of exposure denominators and comparator groups, vulnerability of FAERS to reporting bias (including under- and over-reporting), and inability to establish causality. The authors characterize the findings as hypothesis-generating and recommend clinician vigilance regarding ocular monitoring, particularly in populations already at elevated risk for diabetic eye disease, while calling for further prospective research to validate these associations and explore potential mechanisms.
Journal of clinical medicine · Mar 2026DOI ↗ Limited · human
This Brazilian retrospective regulatory surveillance study examined patterns of GLP-1 receptor agonist (GLP-1 RA) consumption, pharmacovigilance reports, and product falsification alerts between 2020 and 2024. Researchers triangulated data from three official sources: national pharmaceutical sales records classified by the Anatomical Therapeutic Chemical (ATC) system, adverse event reports submitted to Vigimed (Brazil's national pharmacovigilance platform linked to the global VigiBase), and official counterfeit product alerts. The study found that semaglutide was the dominant GLP-1 RA in the Brazilian market. Consumption was heavily concentrated in higher-GDP regions, revealing a geographic and economic disparity that did not align with regional diabetes prevalence rates. Pharmacovigilance data revealed a notable proportion of reports associated with off-label use, suggesting potential gaps in clinical practice and regulatory oversight. Additionally, documented cases of counterfeit GLP-1 RA products highlighted supply chain vulnerabilities. The authors concluded that rising demand, widespread off-label use, and product falsification collectively require a coordinated regulatory response. Limitations include the retrospective and observational nature of the data, reliance on administrative and surveillance databases, and the potential for underreporting in pharmacovigilance systems.
Diabetes, obesity & metabolism · Mar 2026DOI ↗ Limited · human
This qualitative, multi-method study explored how adults living with obesity make decisions about initiating pharmacotherapy. Researchers used three complementary methods — semi-structured interviews (n=15), a Photovoice study (n=12), and focus group discussions (n=12) — to purposively recruit adults aged 18–70 with a BMI ≥27 kg/m² and at least one obesity-related complication, all of whom were naïve to obesity medications. Through triangulation of findings across all three methods, the study identified five key factors shaping patient preferences: (1) anticipated treatment efficacy, (2) adequacy of information and community support, (3) safety and tolerability profile, (4) treatment burden and lifestyle integration, and (5) logistical and structural barriers. Tirzepatide was the most preferred agent across all methods, largely due to its perceived weight-reduction potential, while semaglutide ranked second, aided by its broader societal familiarity. Perceived efficacy and information trustworthiness emerged as the dominant drivers of preference, with safety concerns consistently moderating enthusiasm. The study is limited by its qualitative design, modest sample sizes, and the fact that all participants were medication-naïve, which may limit transferability to experienced patients. The authors recommend clinicians engage in shared decision-making tailored to individual information needs and social contexts.
Obesity pillars · Mar 2026DOI ↗ Limited · human
This retrospective, claims-based observational study examined real-world persistence, adherence, and weight-loss effectiveness of GLP-1 receptor agonists (semaglutide/Wegovy) and the dual GLP-1/GIP agonist (tirzepatide/Zepbound) among 7,493 adults enrolled in Massachusetts Medicaid (MassHealth) who initiated treatment between July and December 2024. At 6 months, the study found moderate rates of persistence (60.8%, defined as no gap greater than 56 days between fills) and adherence (60.1%, defined as proportion of days covered ≥80%). Male sex and younger age were among the member-specific factors associated with lower persistence and adherence. In a subgroup of highly persistent and adherent members with prior authorization recertification data, both tirzepatide and semaglutide were associated with at least 5% body weight reduction at 6 months. Key limitations include the retrospective claims-based design, which cannot establish causality, the restriction to a Medicaid population limiting generalizability, and the weight-loss outcome being measured only in a select, highly adherent subgroup rather than the full cohort.
Journal of managed care & specialty pharmacy · Mar 2026DOI ↗ Limited · human
This real-world pharmacovigilance study analyzed 40,253 adverse event reports submitted to the FDA Adverse Event Reporting System (FAERS) involving semaglutide, liraglutide, and tirzepatide used for weight loss. Researchers applied disproportionality analysis—using proportional reporting ratios (PRR) and reporting odds ratios (ROR)—to identify safety signals beyond the commonly studied gastrointestinal, renal, and pancreatic effects. The study population was predominantly female (68.6%), with median ages varying by drug. Key findings included that semaglutide showed the strongest disproportionality signals for psychiatric adverse events, including anxiety, depression, and suicidal ideation. Tirzepatide was associated with markedly elevated signals for injection-site reactions and indicators of misuse, such as incorrect dosing and off-label administration. Liraglutide showed a comparatively lower overall risk signal across these categories. Important limitations of this study include the inherent biases of spontaneous reporting systems (underreporting, confounding, lack of denominator data), the inability to establish causality, and the possibility that reporting patterns reflect prescribing trends or media attention rather than true pharmacological risk. The authors conclude that psychiatric monitoring and patient education warrant greater attention in clinical practice.
Pharmaceuticals (Basel, Switzerland) · Feb 2026DOI ↗ Limited · human
This case report describes an 18-year-old woman with no prior medical history who developed euglycemic ketoacidosis (EKA) in association with self-administered semaglutide purchased online without medical supervision. After initiating the medication ten days prior and self-escalating doses, she presented with three days of nausea, intractable vomiting, and reduced oral intake. Laboratory findings revealed a high-anion gap metabolic acidosis (pH 7.24, bicarbonate 14 mmol/L, anion gap 24 mEq/L), markedly elevated β-hydroxybutyrate (5.9 mmol/L), and normal blood glucose (60 mg/dL), meeting criteria for EKA. She was treated with intravenous fluids, dextrose infusion, and supportive care, with clinical recovery and discharge within 36 hours. The authors propose that reduced oral intake combined with GLP-1 receptor agonist-induced gastrointestinal side effects may have triggered a starvation-like ketogenic state. Key limitations include the single-patient design, inability to verify the authenticity or exact composition of the online-purchased product, and lack of confirmed dosing history. The case raises awareness of EKA as a potential complication of GLP-1 receptor agonist use, particularly in unsupervised, non-diabetic individuals using unregulated sources.
Limited · human
This retrospective pharmacovigilance study analyzed FDA Adverse Event Reporting System (FAERS) data from Q1 2022 to Q3 2025 to compare gynecological hemorrhagic event reporting between tirzepatide (a dual GIP/GLP-1 receptor agonist) and semaglutide (a GLP-1 receptor agonist) in female patients. Researchers identified 103,607 female-specific reports—70,768 for tirzepatide and 32,839 for semaglutide—and calculated reporting odds ratios (RORs) using standardized MedDRA queries. The study found that gynecological hemorrhagic events were reported at nearly identical rates (0.60% for tirzepatide vs. 0.62% for semaglutide), with an ROR of 0.97 (95% CI: 0.82–1.14), indicating no statistically significant difference in reporting odds between the two agents. Notably, the authors flagged a substantial difference in reporter source composition: 94.6% of tirzepatide reports originated from consumers versus 53.4% for semaglutide, which may introduce bias and limits direct comparability. The authors concluded that no disproportionate safety signal was detected but emphasized that disproportionality analyses cannot establish causation, and that the consumer-reporting imbalance warrants cautious interpretation. Prospective, controlled studies were recommended to validate these findings.
Limited · human
This retrospective, multi-centre observational study used the TriNetX global healthcare research network to compare outcomes of oral semaglutide versus sitagliptin in patients with heart failure with preserved ejection fraction (HFpEF) and type 2 diabetes (T2D). From a pool of over 800,000 patients with both conditions, 3,470 initiated oral semaglutide and 22,840 initiated sitagliptin between October 2019 and December 2023. After propensity score matching (3,452 patients per group), the study found that the oral semaglutide group had a significantly lower 1-year all-cause mortality rate (4.3% vs. 7.0%, log-rank p < 0.001) and lower rates of hospitalisation compared with the sitagliptin group. Important limitations include the observational, non-randomised design, which introduces potential for residual confounding despite propensity score matching; the reliance on administrative/claims data from the TriNetX platform; and the inability to confirm medication adherence or establish causality. The authors note that prior evidence for semaglutide in HFpEF comes from injectable formulations, and these findings for the oral route remain hypothesis-generating pending randomised controlled trial confirmation.
Diabetes, obesity & metabolism · Jan 2026DOI ↗ Limited · human
This study describes the development and validation of a urine-based liquid chromatography–tandem mass spectrometry (LC-MS/MS) method for detecting semaglutide use in humans. Semaglutide is a GLP-1 receptor agonist originally approved for type 2 diabetes that has gained widespread use as a weight-loss aid. The authors note that semaglutide may confer a competitive advantage in weight-class sports (e.g., wrestling, boxing) by facilitating weight management, raising anti-doping concerns. The proposed method targets two unique urinary metabolites of semaglutide—designated U6 and U7—as biomarkers of exposure. Using only 2 mL of urine, the method achieved a reported limit of detection (LOD) of 50 pg/mL. The authors positioned this approach as a simpler alternative to existing blood-based testing regimens, given urine's relative ease of collection. Limitations include the absence of reported clinical validation data (e.g., pharmacokinetic profiling in dosed volunteers), no information on inter-individual variability, and no discussion of detection windows post-dose. The study is primarily an analytical methods paper rather than a clinical or pharmacological investigation.
Analytical and bioanalytical chemistry · Dec 2025DOI ↗ Limited · human
This retrospective observational study used the TriNetX electronic medical records database to examine cardiovascular outcomes and safety of semaglutide compared with dipeptidyl peptidase-4 (DPP-4) inhibitors in patients with type 2 diabetes (T2D) who had a body mass index (BMI) below 25 kg/m². From a large cohort of 340,721 patients identified between 2018 and 2020, the researchers compared all-cause mortality, cardiovascular outcomes, and adverse events between the two treatment groups. The study found that semaglutide was associated with a lower 3-year risk of all-cause mortality compared with DPP-4 inhibitors in this non-overweight population. These findings are notable because most landmark GLP-1 receptor agonist cardiovascular outcome trials have predominantly enrolled overweight or obese participants, leaving the effects in leaner individuals less well understood. Key limitations include the retrospective, non-randomized design, which introduces potential confounding by indication, residual confounding from unmeasured variables, and reliance on administrative/electronic health record data. Generalizability may also be limited by the database's geographic and demographic composition. The authors attribute differences in outcomes specifically to the semaglutide versus DPP-4 inhibitor comparison within this BMI subgroup.
European heart journal. Quality of care & clinical outcomes · Dec 2025DOI ↗ Limited · human
This study used synthetic target trial emulation and computational predictive modeling to compare amylin-pathway therapies — specifically CagriSema, cagrilintide, and amycretin formulations — for obesity and type 2 diabetes. Following PRISMA 2020 and TARGET framework guidelines, the researchers pooled data from seven randomized controlled trials (N = 5,786 participants) published through September 2025. Rather than analyzing real individual patient data, they reconstructed high-precision synthetic individual patient datasets and applied network meta-analysis, dose-response modeling, virtual head-to-head comparisons, and machine learning. The study reported that synthetic data reconstruction achieved greater than 99% fidelity to source trials, and virtual modeling suggested CagriSema outperformed subcutaneous amycretin at matched timepoints (posterior probability >0.95). Dose-response modeling identified an estimated ED80 for amycretin and benefit-risk analysis suggested a potential therapeutic window in the 10–20 mg subcutaneous range. Machine learning models predicted treatment response with 82–87% accuracy from baseline characteristics. Key limitations include reliance on reconstructed — not real — individual patient data, indirect comparisons rather than direct head-to-head trial evidence, and calibration metrics indicating moderate model uncertainty. The authors suggest these findings may inform future confirmatory trial design.
Metabolism open · Oct 2025DOI ↗ Limited · human
This retrospective observational study used the TriNetX database of de-identified electronic health records (January 2018–December 2020) to compare cardiovascular outcomes between semaglutide and dulaglutide in adults with type 2 diabetes. From a pool of nearly 4.7 million patients with type 2 diabetes, 231,075 semaglutide initiators and 189,103 dulaglutide initiators were identified. Propensity score matching yielded 171,105 patients per group. Over a 3-year follow-up, the study found that semaglutide was associated with statistically significantly lower rates of all-cause death (4.2% vs. 5.6%; HR 0.75, 95% CI 0.72–0.78), acute myocardial infarction (5.2% vs. 5.6%; HR 0.94), stroke (5.8% vs. 6.4%; HR 0.90), and acute heart failure hospitalization (5.3% vs. 6.1%; HR 0.88) compared with dulaglutide. Key limitations include the observational, non-randomized design, which is susceptible to residual confounding despite propensity score matching; reliance on administrative/EHR coding for outcome ascertainment; and the inability to distinguish between semaglutide formulations (oral vs. injectable) or account for dosing differences between agents.
Scientific reports · Jul 2025DOI ↗