Preprints.org logo
Preprint
Article

This version is not peer-reviewed.

Early Signal Detection in GLP-1 Receptor Agonists in Spain: A Comparative Bayesian Disproportionality Analysis in 2024 and 2025

Submitted:

31 July 2025

Posted:

01 August 2025

You are already at the latest version

Abstract
Background: Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are increasingly prescribed for type 2 diabetes mellitus and obesity. Their expanding use, including off-label indications, raises ongoing concerns regarding their evolving safety profiles. Objective: To identify and compare early positive safety signals associated with GLP-1 RAs in Spain during 2024 and 2025 using a Bayesian disproportionality approach adapted from the WHO-Uppsala Monitoring Centre. Methods: Spontaneous adverse drug reaction (ADR) reports submitted to the Spanish Pharmacovigilance System and involving GLP-1 RAs (ATC A10BJ) were analyzed. Reports up to June 2024 and June 2025 were included. A Bayesian Confidence Propagation Neural Network (BCPNN)-based model was used to estimate signal strength. Positive signals were defined as those with a false discovery rate (FDR) < 0.05 and relative risk (RR) ≥ 1. Signals were classified as new, reinforced, diminished, unchanged, or disappeared between the two years. Results: We analyzed 5,322 reports in 2024 and 6,746 in 2025. New signals identified in 2025 included intestinal obstruction (dulaglutide), acute pancreatitis (exenatide), and urticaria at the injection site (liraglutide). Several previously identified signals diminished or disappeared, suggesting dynamic changes in GLP-1 RA risk profiles. Conclusions: This comparative Bayesian pharmacovigilance analysis highlights the evolving safety landscape of GLP-1 RAs. Early signal detection can inform timely regulatory interventions and support safer clinical use.
Keywords: 
GLP-1 receptor agonists
;  
pharmacovigilance
;  
adverse drug reactions
;  
early signal detection
;  
semaglutide
;  
liraglutide
;  
dulaglutide
Subject: 
Medicine and Pharmacology  -   Endocrinology and Metabolism

1. Introduction

Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are a class of incretin-based therapies that mimic the action of endogenous GLP-1, stimulating insulin secretion and inhibiting glucagon release in a glucose-dependent manner. These agents have gained widespread acceptance for the management of type 2 diabetes mellitus (T2DM) due to their efficacy in improving glycemic control, promoting weight loss, and offering cardiovascular protection [1,2].
Several GLP-1 RAs—such as liraglutide, dulaglutide, exenatide, semaglutide, and lixisenatide—have shown superiority over other antidiabetic agents in clinical trials, especially in reducing HbA1c levels and achieving significant weight reduction [3,4]. Notably, large cardiovascular outcome trials (CVOTs) like LEADER, SUSTAIN-6, and REWIND demonstrated cardiovascular benefits beyond glycemic effects, leading to broader therapeutic indications in high-risk populations [5,6,7]. Consequently, their use has expanded rapidly, including off-label use in individuals with obesity without diabetes [8].
However, the growing use of GLP-1 RAs also raises safety concerns, particularly regarding gastrointestinal, pancreatic, thyroid, and renal adverse effects [9,10]. Rare but serious adverse events (AEs)—such as pancreatitis, gallbladder disease, and injection site reactions—have been reported in both clinical trials and post-marketing surveillance [11,12,13]. Moreover, recent real-world studies have highlighted the importance of early detection of adverse drug reactions (ADRs) that may not have been captured during the pre-approval phases [14].
Pharmacovigilance systems, including spontaneous reporting databases, remain essential tools for detecting potential drug safety signals. However, traditional disproportionality methods such as the proportional reporting ratio (PRR) or reporting odds ratio (ROR) may produce false positives due to multiplicity or sparse data [15,16]. Bayesian approaches—such as the Bayesian Confidence Propagation Neural Network (BCPNN) developed by the WHO-Uppsala Monitoring Centre—provide a more robust framework by accounting for uncertainty and prior probabilities [17].
The concept of "early signal detection" refers to the identification of statistically significant drug-event combinations before widespread recognition, potentially enabling earlier regulatory or clinical interventions [18]. The implementation of false discovery rate (FDR) control methods, such as the Benjamini-Hochberg procedure, further improves signal reliability in large datasets with multiple comparisons [19].
This study aims to perform a comparative Bayesian disproportionality analysis of suspected ADRs involving GLP-1 RAs in Spain during the first semesters of 2024 and 2025. By identifying new, reinforced, unchanged, diminished, or disappeared safety signals, this work contributes to the understanding of evolving drug safety profiles and supports timely pharmacovigilance efforts.

2. Results

2.1. Overview of ADR Reports

A total of 5,322 adverse drug reactions (ADRs) associated with GLP-1 receptor agonists (GLP-1 RAs) were reported to the Spanish Pharmacovigilance System in the first half of 2024, increasing to 6,746 reports in the same period of 2025.
This increase may reflect:
  • Higher prescription rates and broader indications;
  • Greater pharmacovigilance awareness among healthcare professionals;
  • Potential real changes in the risk profile of GLP-1 RAs.

2.1.1. Signal Classification by Year

Safety signals were categorized based on their status in 2024 and 2025 using Bayesian disproportionality analysis with false discovery rate (FDR) correction (see Appendix 1: Table A1; Table A2). The following classifications were applied:
  • New signals: drug-event pairs newly detected in 2025;
  • Reinforced signals: previously detected signals with increased risk or statistical strength;
  • Disappeared signals: reduced statistical strength or FDR;
  • Unchanged signals: present with similar strength in both years;
  • Disappeared signals: present in 2024 but not detected in 2025.
The summary of newly identified or altered signals is detailed in Table 1.

2.1.2. Notable New Signals in 2025

Several new positive signals meeting the predefined Bayesian criteria (FDR < 0.05; RR ≥ 1) emerged in 2025:
• Dulaglutide: intestinal obstruction;
• Exenatide: acute pancreatitis and skin mass at injection site;
• Liraglutide: urticaria at the injection site and device administration malfunction;
• Semaglutide: inadequate diabetes control.
These signals may reflect either increased true incidence, expanded patient use, or improved ADR reporting.

2.1.3. Disappeared Signals

Several drug-event combinations detected in 2024 were either absent or statistically weaker in 2025:
• Lixisenatide: dizziness—signal disappeared;
• Liraglutide: minor weight loss—signal diminished.
The disappearance of these signals could indicate changes in clinical use patterns, underreporting, or shifts in the underlying patient population.

2.2. Tables and Signal Summary

All relevant signals and corresponding information from the Summary of Product Characteristics (SmPC) are compiled in Table 1, titled: Table 1. Safety Signal evolution and fact sheet comments for GLP-1 Receptor Agonists between 2024-2025. This table includes:
• The GLP-1 RA involved;
• The preferred term (PT) of the reported adverse event;
• Whether the ADR is described in the corresponding SmPC.
All signals listed in Table 1 were extracted using MedDRA coding and analyzed using Bayesian methods as described in the Methods section.

3. Discussion

This comparative pharmacovigilance study reveals dynamic changes in the safety profile of GLP-1 receptor agonists (GLP-1 RAs) in Spain between 2024 and 2025. The detection of new positive signals—particularly for gastrointestinal and pancreatic adverse events—underscores the importance of continuous post-marketing surveillance in this therapeutic class.
The identification of intestinal obstruction with dulaglutide and acute pancreatitis with exenatide aligns with previous concerns raised in both preclinical and post-marketing reports [9,10,20]. GLP-1 RAs slow gastric emptying, which may theoretically contribute to mechanical or functional obstruction in predisposed individuals [21]. Although these effects are well known, their clinical significance is still being debated, especially as real-world evidence accumulates.
The signal for inadequate diabetes control with semaglutide may reflect inappropriate off-label use or administration errors. This finding is clinically relevant given the increasing popularity of GLP-1 RAs for weight management, sometimes self-administered without medical supervision [8,22]. In this context, improper dosing or skipping injections could lead to subtherapeutic effects or glycemic instability.
Furthermore, several injection-site reactions (e.g., urticaria, bruising, or device malfunction) were newly identified or reinforced in 2025. Although often considered mild, these events can affect treatment adherence, particularly in patients self-injecting long-acting agents [23].
On the other hand, the disappearance or attenuation of some previously detected signals—such as dizziness with lixisenatide—may indicate a reduced use of certain molecules following market withdrawal (as in the case of lixisenatide and exenatide in Spain) or improved risk minimization measures [24].
Our study demonstrates the added value of Bayesian methods, particularly when combined with false discovery rate (FDR) adjustment, in improving signal reliability over traditional disproportionality metrics [17,19]. The use of the Bayesian Confidence Propagation Neural Network (BCPNN) provides a probabilistic framework that is robust to data sparsity and supports regulatory prioritization of signals [18,25].
It is worth noting that some signals correspond to events not described in the official Summary of Product Characteristics (SmPC) at the time of analysis. This suggests the utility of pharmacovigilance data in identifying emerging or evolving ADRs that may not have been observed during clinical development [13,26].

3.1. Strengths and Limitations

The main strengths of this study include:
  • The use of a standardized Bayesian algorithm based on WHO-UMC methodology;
  • Comparison across two consecutive years using real-world data from a national database (see Appendix 1: Table A1; Table A2);
  • Adjustment for multiple testing via FDR, reducing the likelihood of spurious signals.
However, several limitations should be acknowledged:
  • Spontaneous reporting systems are subject to underreporting, missing data, and reporting bias [14,27];
  • Causality cannot be established—signal detection is hypothesis-generating;
  • Changes in the number of users per drug are not available, limiting calculation of true incidence rates.
Future studies using analytical epidemiological designs, such as cohort or case-control studies with prescription databases, are warranted to confirm these preliminary signals [28].

4. Materials and Methods

4.1. Data Source

This study is based on spontaneous reports of suspected adverse drug reactions (ADRs) submitted to the Spanish Pharmacovigilance System for Human Use Medicines (FEDRA®), managed by the Agencia Española de Medicamentos y Productos Sanitarios (AEMPS). Data were extracted from public releases corresponding to reports received up to 30 June 2024 and 30 June 2025.
All included reports referred to drugs within the ATC group A10BJ (GLP-1 receptor agonists), specifically dulaglutide, exenatide, liraglutide, lixisenatide, and semaglutide. Data extraction and preprocessing were performed using R®v3.4.1. R Foundation for Statistical Computing and PhViD® v1.0.8 package for the detection of positive signals [29].
Spontaneous reporting systems are widely used for signal detection and early risk identification, though they are subject to limitations such as underreporting and reporting bias [14,27,30]. Nevertheless, national databases like FEDRA® provide an essential source of real-world evidence for regulatory pharmacovigilance [31].

4.2. ADR Coding and Drug Selection

Adverse events were coded using the Medical Dictionary for Regulatory Activities (MedDRA), specifically the Preferred Term (PT) level. MedDRA is internationally recognized and ensures consistency and comparability in safety signal analysis [32].
GLP-1 RAs included in this study were:
• Dulaglutide (Trulicity®),
• Exenatide (Byetta®, Bydureon®),
• Liraglutide (Victoza®, Saxenda®),
• Lixisenatide (Lyxumia®),
• Semaglutide (Ozempic®, Rybelsus®, Wegovy®).
Drugs that had been withdrawn from the Spanish market by mid-2024, such as exenatide and lixisenatide, were retained for analysis to enable year-to-year comparisons of signal persistence and disappearance.

4.3. Bayesian Disproportionality Analysis

We implemented a Bayesian Confidence Propagation Neural Network (BCPNN) model adapted from the WHO-Uppsala Monitoring Centre (UMC) [17,18,25]. This method estimates the Information Component (IC), a logarithmic metric of disproportionality that accounts for statistical shrinkage and prior probability distributions.
The BCPNN approach is well suited for early signal detection because:
  • It handles sparse data more robustly than frequentist methods;
  • It generates probabilistic outputs, such as credibility intervals;
  • It is less sensitive to extreme values and data volatility [33].
The BCPNN model computes a posterior distribution for each drug-event pair, and signal strength is typically summarized by the IC025, the lower bound of the 95% credibility interval. An IC025 > 0 indicates disproportionate reporting.

4.4. False Discovery Rate and Signal Thresholds

To address the problem of multiple testing—a frequent challenge in pharmacovigilance analyses involving thousands of drug-event pairs—we applied the Benjamini-Hochberg procedure to control the False Discovery Rate (FDR) [19]. Each p-value derived from the Bayesian model was adjusted, and a signal was considered statistically significant if:
• FDR < 0.05, and
• Relative Risk (RR) ≥ 1.
This dual threshold approach ensures that detected signals are not only statistically robust but also clinically meaningful [34].

4.5. Signal Classification

Signals detected in both years were further classified into five categories based on their FDR change over time:
  • New: signal appeared only in 2025;
  • Reinforced: signal was present in both years with increased strength or lower FDR in 2025;
  • Diminished: signal persisted but with reduced statistical strength;
  • Unchanged: signal remained stable;
  • Disappeared: signal was present in 2024 but absent in 2025.
This classification facilitates trend interpretation and regulatory prioritization of evolving safety issues [35].

5. Conclusions

This study provides updated evidence on the evolving safety profile of GLP-1 receptor agonists (GLP-1 RAs) in Spain, applying a Bayesian disproportionality analysis with FDR control to detect early signals of adverse drug reactions (ADRs) in 2024 and 2025. The results highlight newly emerging risks—including intestinal obstruction, acute pancreatitis, and injection-site reactions—as well as the disappearance or attenuation of other signals over time.
The dynamic nature of these signals underscores the importance of continuous post-marketing surveillance, especially as the clinical use of GLP-1 RAs expands beyond their original indications, often to populations not represented in pivotal clinical trials. The appearance of signals related to off-label use and administration errors, such as inadequate diabetes control, suggests a need for greater awareness and patient education regarding proper drug use.
Bayesian pharmacovigilance approaches, particularly when combined with false discovery rate correction, offer a robust framework for early signal detection in real-world data. These methods enhance the reliability of signal prioritization, helping to inform regulatory decisions and guide further epidemiological research.
Future studies should validate these findings using analytical designs such as cohort or nested case-control studies with prescription data. Integrating signal detection into a broader risk management strategy will be key to optimizing the safety and effectiveness of GLP-1 RAs in an increasingly diverse patient population.

Funding

This research received no external funding.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
ADR Adverse drug reaction
SmPC Summary of Product Characteristics

Appendix A

Appendix A.1

Table A1. Bayesian Disproportionality Analysis of spontaneous reports of suspected adverse drug reactions to GLP-1 Receptor Agonists submitted to the Spanish Pharmacovigilance System for Human Use Medicines until June 2024 in Spain.
Table A1. Bayesian Disproportionality Analysis of spontaneous reports of suspected adverse drug reactions to GLP-1 Receptor Agonists submitted to the Spanish Pharmacovigilance System for Human Use Medicines until June 2024 in Spain.
drug event effect (PT) count (N) expected count post.H0 n11/E drug margin event margin FDR FNR Se Sp
semaglutide Off label use 109 55.571 0.000 1.961 1933 153 0.000 0.531 0.001 1.000
dulaglutide Injection site pain 38 12.151 0.000 3.127 1115 58 0.000 0.530 0.003 1.000
exenatide Injection site nodule 10 0.963 0.000 10.386 427 12 0.000 0.530 0.004 1.000
dulaglutide Product dose omission 21 5.866 0.000 3.580 1115 28 0.000 0.530 0.006 1.000
liraglutide Weight decreased mild 39 16.094 0.000 2.423 1713 50 0.000 0.529 0.007 1.000
liraglutide Injection site urticaria 33 12.231 0.000 2.698 1713 38 0.000 0.529 0.008 1.000
exenatide Injection site induration 9 1.284 0.000 7.011 427 16 0.000 0.528 0.010 1.000
dulaglutide Blood glucose increased 29 13.408 0.001 2.163 1115 64 0.000 0.528 0.011 1.000
liraglutide Injection site erythema 41 20.922 0.001 1.960 1713 65 0.000 0.528 0.012 1.000
semaglutide Nausea 190 143.468 0.001 1.324 1933 395 0.000 0.527 0.014 1.000
dulaglutide Injection site haemorrhage 15 4.400 0.001 3.409 1115 21 0.000 0.527 0.015 1.000
exenatide Acute pancreatitis 13 4.172 0.001 3.116 427 52 0.000 0.527 0.017 1.000
dulaglutide Wrong dose administered 17 5.866 0.001 2.898 1115 28 0.000 0.526 0.018 1.000
exenatide Erythema 8 1.605 0.002 4.985 427 20 0.001 0.526 0.019 1.000
liraglutide Drug ineffective 53 32.187 0.003 1.647 1713 100 0.001 0.526 0.021 1.000
exenatide Retching 6 0.642 0.004 9.348 427 8 0.001 0.525 0.022 1.000
exenatide Pancreatitis 11 3.771 0.004 2.917 427 47 0.001 0.525 0.024 1.000
liraglutide Injection site pruritus 29 15.128 0.006 1.917 1713 47 0.001 0.525 0.025 1.000
semaglutide Weight decreased 46 27.967 0.006 1.645 1933 77 0.002 0.524 0.026 1.000
semaglutide Product use for unapproved indication 21 9.443 0.007 2.224 1933 26 0.002 0.524 0.028 1.000
exenatide Skin mass 5 0.401 0.010 12.464 427 5 0.002 0.524 0.029 1.000
exenatide Renal failure 6 1.203 0.011 4.985 427 15 0.003 0.523 0.030 1.000
liraglutide Injection site rash 15 6.116 0.013 2.453 1713 19 0.003 0.523 0.032 1.000
lixisenatide Urticaria 5 1.113 0.018 4.493 126 47 0.004 0.523 0.033 1.000
dulaglutide Decreased appetite 34 21.998 0.021 1.546 1115 105 0.004 0.522 0.034 1.000
semaglutide Diabetes mellitus inadequate control 15 6.538 0.021 2.294 1933 18 0.005 0.522 0.036 1.000
semaglutide Drug intolerance 16 7.264 0.021 2.203 1933 20 0.006 0.522 0.037 1.000
liraglutide Injection site reaction 14 6.116 0.023 2.289 1713 19 0.006 0.521 0.039 1.000
liraglutide Injection site hypersensitivity 12 4.828 0.026 2.485 1713 15 0.007 0.521 0.040 1.000
semaglutide Dyspepsia 39 25.788 0.027 1.512 1933 71 0.008 0.520 0.041 1.000
liraglutide Injection site bruising 10 3.541 0.028 2.824 1713 11 0.008 0.520 0.043 1.000
semaglutide Gastrointestinal disorder 19 10.170 0.032 1.868 1933 28 0.009 0.520 0.044 1.000
semaglutide Product use error 17 8.717 0.034 1.950 1933 24 0.010 0.519 0.045 1.000
exenatide Nodule 4 0.562 0.036 7.122 427 7 0.011 0.519 0.047 0.999
dulaglutide Hypoaesthesia 5 1.048 0.052 4.773 1115 5 0.012 0.519 0.048 0.999
dulaglutide Intestinal obstruction 5 1.048 0.052 4.773 1115 5 0.013 0.518 0.049 0.999
liraglutide Device administration error 7 2.253 0.059 3.107 1713 7 0.014 0.518 0.051 0.999
dulaglutide Accidental overdose 6 1.886 0.059 3.182 1115 9 0.015 0.518 0.052 0.999
semaglutide Overdose 16 9.080 0.065 1.762 1933 25 0.017 0.518 0.053 0.999
liraglutide Injection site swelling 12 6.116 0.066 1.962 1713 19 0.018 0.517 0.054 0.999
lixisenatide Hypoglycaemia 4 1.136 0.068 3.520 126 48 0.019 0.517 0.056 0.999
lixisenatide Dizziness 7 3.315 0.076 2.112 126 140 0.020 0.517 0.057 0.999
exenatide Asthenia 7 3.209 0.076 2.181 427 40 0.022 0.516 0.058 0.999
dulaglutide Limb pain 5 1.467 0.081 3.409 1115 7 0.023 0.516 0.060 0.999
semaglutide Weight increased 19 11.986 0.083 1.585 1933 33 0.024 0.516 0.061 0.998
liraglutide Skin reaction 9 4.184 0.085 2.151 1713 13 0.026 0.515 0.062 0.998
dulaglutide Accidental subtherapeutic dose 4 0.838 0.090 4.773 1115 4 0.027 0.515 0.063 0.998
liraglutide Product quality issue 8 3.541 0.092 2.260 1713 11 0.028 0.515 0.065 0.998
dulaglutide Flatulence 21 14.456 0.095 1.453 1115 69 0.030 0.514 0.066 0.998
liraglutide Device-mediated wrong dose administration 7 2.897 0.099 2.416 1713 9 0.031 0.514 0.067 0.998
semaglutide Upper abdominal pain 35 26.151 0.100 1.338 1933 72 0.032 0.514 0.068 0.998
exenatide Peripheral oedema 3 0.562 0.101 5.342 427 7 0.034 0.514 0.070 0.997
dulaglutide Cholelithiasis 8 3.981 0.103 2.010 1115 19 0.035 0.513 0.071 0.997
dulaglutide Product administration schedule inappropriate 15 9.637 0.105 1.556 1115 46 0.036 0.513 0.072 0.997
dulaglutide Injection site trauma 4 1.048 0.110 3.818 1115 5 0.038 0.513 0.073 0.997
semaglutide Hyperglycaemia 16 10.170 0.112 1.573 1933 28 0.039 0.512 0.075 0.997
exenatide Diabetic ketoacidosis 3 0.642 0.113 4.674 427 8 0.040 0.512 0.076 0.997
lixisenatide Blood glucose increased 4 1.515 0.120 2.640 126 64 0.042 0.512 0.077 0.996
exenatide Pruritus 7 3.691 0.121 1.897 427 46 0.043 0.512 0.078 0.996
lixisenatide Product contamination by body fluid 2 0.071 0.122 28.159 126 3 0.044 0.511 0.079 0.996
lixisenatide Serum triglycerides increased 2 0.095 0.123 21.119 126 4 0.046 0.511 0.081 0.996
lixisenatide Tendonitis 2 0.047 0.124 42.238 126 2 0.047 0.511 0.082 0.996
exenatide Haemoglobin A1c increased 3 0.722 0.125 4.155 427 9 0.048 0.510 0.083 0.995
exenatide Weight decreased 10 6.178 0.128 1.619 427 77 0.049 0.510 0.084 0.995
Relative risk ≥ 1, Number of Monte Carlo simulations NB.MC=10,000. False Discovery Rate (FDR)<0.05. Interpretation of items: N (count): number of couples ‘active ingredient-ADR’ reported; post.H0: posterior probability of null hypothesis; FDR: False Discovery Rate; FNR: False Negative Rate; Se: Sensitivity; Sp: Specificity.

Appendix A.2

Table A2. Bayesian Disproportionality Analysis of spontaneous reports of suspected adverse drug reactions to GLP-1 Receptor Agonists submitted to the Spanish Pharmacovigilance System for Human Use Medicines until June 2025 in Spain.
Table A2. Bayesian Disproportionality Analysis of spontaneous reports of suspected adverse drug reactions to GLP-1 Receptor Agonists submitted to the Spanish Pharmacovigilance System for Human Use Medicines until June 2025 in Spain.
drug event effect count (N) expected count post.H0 n11/E drug margin event margin FDR FNR Se Sp
dulaglutide Injection site pain 45 11.630 0.000 3.869 1171 67 0.000 0.528 0.001 1.000
liraglutide Injection site urticaria 33 10.348 0.000 3.189 1837 38 0.000 0.528 0.003 1.000
liraglutide Weight decreased
(mild not codified separately)		 46 18.517 0.000 2.484 1837 68 0.000 0.527 0.004 1.000
dulaglutide Product dose omission 21 4.860 0.000 4.321 1171 28 0.000 0.527 0.005 1.000
exenatide Injection site nodule 10 0.760 0.000 13.165 427 12 0.000 0.527 0.006 1.000
dulaglutide Blood glucose increased 30 11.804 0.000 2.542 1171 68 0.000 0.526 0.008 1.000
dulaglutide Injection site haemorrhage 17 3.992 0.000 4.258 1171 23 0.000 0.526 0.009 1.000
semaglutide Off-label use 130 82.416 0.000 1.577 3177 175 0.000 0.526 0.010 1.000
liraglutide Injection site erythema 41 18.245 0.000 2.247 1837 67 0.000 0.525 0.011 1.000
exenatide Injection site induration 9 1.013 0.000 8.887 427 16 0.000 0.525 0.013 1.000
dulaglutide Wrong dose administered 19 6.249 0.000 3.040 1171 36 0.000 0.525 0.014 1.000
exenatide Acute pancreatitis 13 3.798 0.000 3.423 427 60 0.000 0.524 0.015 1.000
liraglutide Drug ineffective 56 32.133 0.001 1.743 1837 118 0.000 0.524 0.016 1.000
exenatide Erythema 8 1.393 0.001 5.745 427 22 0.000 0.524 0.018 1.000
liraglutide Injection site pruritus 29 13.343 0.001 2.173 1837 49 0.000 0.524 0.019 1.000
exenatide Retching 6 0.506 0.002 11.849 427 8 0.000 0.523 0.020 1.000
semaglutide Product use for unapproved indication 41 21.664 0.003 1.893 3177 46 0.000 0.523 0.021 1.000
exenatide Pancreatitis 11 3.798 0.004 2.896 427 60 0.001 0.523 0.022 1.000
liraglutide Injection site rash 15 5.446 0.005 2.754 1837 20 0.001 0.522 0.024 1.000
exenatide Renal impairment 6 1.013 0.006 5.924 427 16 0.001 0.522 0.025 1.000
liraglutide Injection site hypersensitivity 13 4.357 0.007 2.984 1837 16 0.001 0.522 0.026 1.000
liraglutide Injection site reaction 15 5.718 0.007 2.623 1837 21 0.002 0.521 0.027 1.000
exenatide Skin mass 5 0.316 0.007 15.799 427 5 0.002 0.521 0.029 1.000
semaglutide Nausea 277 231.705 0.009 1.195 3177 492 0.002 0.521 0.030 1.000
lixisenatide Urticaria 5 0.971 0.012 5.148 126 52 0.003 0.520 0.031 1.000
dulaglutide Decreased appetite 34 21.351 0.013 1.592 1171 123 0.003 0.520 0.032 1.000
liraglutide Injection site bruise 10 3.268 0.018 3.060 1837 12 0.004 0.520 0.034 1.000
exenatide Nodule 4 0.443 0.026 9.028 427 7 0.004 0.519 0.035 1.000
liraglutide Skin reaction 10 3.812 0.030 2.623 1837 14 0.005 0.519 0.036 1.000
liraglutide Injection site swelling 13 5.991 0.031 2.170 1837 22 0.006 0.519 0.037 1.000
semaglutide Product use error 27 16.012 0.034 1.686 3177 34 0.007 0.519 0.038 1.000
dulaglutide Hypoaesthesia 5 0.868 0.037 5.761 1171 5 0.008 0.518 0.040 1.000
semaglutide Dyspepsia 63 46.624 0.039 1.351 3177 99 0.009 0.518 0.041 1.000
semaglutide Weight decreased 58 42.385 0.040 1.368 3177 90 0.010 0.518 0.042 1.000
semaglutide Extra dose administered 27 16.483 0.041 1.638 3177 35 0.011 0.517 0.043 1.000
dulaglutide Accidental overdose 6 1.736 0.044 3.457 1171 10 0.012 0.517 0.044 0.999
dulaglutide Intestinal obstruction 5 1.042 0.045 4.801 1171 6 0.013 0.517 0.046 0.999
liraglutide Product quality issue 9 3.540 0.045 2.542 1837 13 0.013 0.516 0.047 0.999
liraglutide Device administration error 7 2.178 0.048 3.213 1837 8 0.014 0.516 0.048 0.999
lixisenatide Hypoglycaemia 4 0.990 0.050 4.041 126 53 0.015 0.516 0.049 0.999
semaglutide Gastrointestinal disorder 31 20.251 0.050 1.531 3177 43 0.016 0.516 0.050 0.999
dulaglutide Blood glucose abnormal 8 3.298 0.052 2.426 1171 19 0.017 0.515 0.052 0.999
liraglutide Device-related wrong dose administration 8 2.995 0.053 2.671 1837 11 0.018 0.515 0.053 0.999
semaglutide Vomiting 217 190.262 0.063 1.141 3177 404 0.019 0.515 0.054 0.999
dulaglutide Injection site haematoma 7 2.777 0.064 2.520 1171 16 0.020 0.514 0.055 0.999
semaglutide Drug intolerance 21 12.716 0.066 1.652 3177 27 0.021 0.514 0.056 0.999
semaglutide Diarrhoea 159 136.574 0.067 1.164 3177 290 0.022 0.514 0.057 0.999
exenatide Asthenia 7 3.165 0.069 2.212 427 50 0.023 0.513 0.059 0.999
exenatide Pruritus 7 3.165 0.069 2.212 427 50 0.024 0.513 0.060 0.998
dulaglutide Cholelithiasis 9 4.340 0.069 2.074 1171 25 0.025 0.513 0.061 0.998
dulaglutide Accidental subtherapeutic dose 4 0.694 0.070 5.761 1171 4 0.025 0.513 0.062 0.998
semaglutide Diabetes mellitus inadequate control 16 8.948 0.073 1.788 3177 19 0.026 0.512 0.063 0.998
exenatide Peripheral oedema 3 0.443 0.080 6.771 427 7 0.027 0.512 0.064 0.998
lixisenatide Injection site trauma 4 1.270 0.082 3.149 126 68 0.028 0.512 0.066 0.998
dulaglutide Inappropriate schedule of product administration 4 0.868 0.083 4.609 1171 5 0.029 0.512 0.067 0.998
dulaglutide Weight increased 15 9.374 0.086 1.600 1171 54 0.030 0.511 0.068 0.998
semaglutide Dizziness 33 23.547 0.087 1.401 3177 50 0.031 0.511 0.069 0.998
lixisenatide Limb pain 7 3.455 0.088 2.026 126 185 0.032 0.511 0.070 0.998
exenatide Flatulence 10 5.697 0.088 1.755 427 90 0.033 0.510 0.071 0.997
liraglutide Diabetic ketoacidosis 21 14.160 0.091 1.483 1837 52 0.034 0.510 0.072 0.997
dulaglutide Abdominal pain 6 2.430 0.093 2.469 1171 14 0.035 0.510 0.074 0.997
exenatide Malaise 8 4.241 0.095 1.886 427 67 0.036 0.510 0.075 0.997
dulaglutide Haemoglobin A1c increased 21 14.581 0.095 1.440 1171 84 0.037 0.509 0.076 0.997
exenatide Injection site warmth 3 0.570 0.096 5.266 427 9 0.038 0.509 0.077 0.997
dulaglutide Breast cancer 28 20.657 0.098 1.356 1171 119 0.039 0.509 0.078 0.997
liraglutide Drug hypersensitivity 20 13.615 0.104 1.469 1837 50 0.040 0.509 0.079 0.996
semaglutide Chills 63 50.862 0.104 1.239 3177 108 0.041 0.508 0.080 0.996
exenatide Blood triglycerides increased 3 0.633 0.105 4.740 427 10 0.042 0.508 0.081 0.996
liraglutide Product contamination with body fluid 5 1.634 0.109 3.060 1837 6 0.043 0.508 0.083 0.996
liraglutide Tendinitis 5 1.634 0.109 3.060 1837 6 0.044 0.507 0.084 0.996
liraglutide Injection site mass 5 1.634 0.109 3.060 1837 6 0.045 0.507 0.085 0.996
liraglutide Injection site pain 6 2.451 0.116 2.448 1837 9 0.046 0.507 0.086 0.996
lixisenatide Injection site urticaria 2 0.075 0.117 26.770 126 4 0.047 0.507 0.087 0.995
lixisenatide Weight decreased
(mild not codified separately)		 2 0.056 0.117 35.693 126 3 0.048 0.506 0.088 0.995
lixisenatide Product dose omission 2 0.037 0.120 53.540 126 2 0.049 0.506 0.089 0.995
exenatide Injection site nodule 4 1.456 0.121 2.748 427 23 0.050 0.506 0.090 0.995
Relative risk ≥ 1, Number of Monte Carlo simulations NB.MC=10,000. False Discovery Rate (FDR)<0.05. Interpretation of items: N (count): number of couples ‘active ingredient-ADR’ reported; post.H0: posterior probability of null hypothesis; FDR: False Discovery Rate; FNR: False Negative Rate; Se: Sensitivity; Sp: Specificity.

References

  1. Nauck, M.A.; Meier, J.J. Management of endocrine disease: Are all GLP-1 agonists equal in the treatment of type 2 diabetes? Eur J Endocrinol. 2019, 181, R211–R234. [Google Scholar] [CrossRef]
  2. Drucker, D.J. Mechanisms of action and therapeutic application of glucagon-like peptide-1. Cell Metab. 2018, 27, 740–756. [Google Scholar] [CrossRef]
  3. Buse, J.B.; Nauck, M.; Forst, T.; et al. Exenatide once weekly versus liraglutide once daily in patients with type 2 diabetes: a 26-week randomized, open-label trial. Lancet. 2013, 381, 117–124. [Google Scholar] [CrossRef]
  4. Pratley, R.E.; Aroda, V.R.; Lingvay, I.; et al. Semaglutide versus dulaglutide in patients with type 2 diabetes (SUSTAIN 7): a randomised, open-label, phase 3b trial. Lancet Diabetes Endocrinol. 2018, 6, 275–286. [Google Scholar] [CrossRef]
  5. Marso, S.P.; Daniels, G.H.; Brown-Frandsen, K.; et al. Liraglutide and Cardiovascular Outcomes in Type 2 Diabetes. N Engl J Med. 2016, 375, 311–322. [Google Scholar] [CrossRef]
  6. Gerstein, H.C.; Colhoun, H.M.; Dagenais, G.R.; et al. Dulaglutide and cardiovascular outcomes in type 2 diabetes (REWIND): a double-blind, randomised placebo-controlled trial. Lancet. 2019, 394, 121–130. [Google Scholar] [CrossRef]
  7. Marso, S.P.; Bain, S.C.; Consoli, A.; et al. Semaglutide and cardiovascular outcomes in patients with type 2 diabetes. N Engl J Med. 2016, 375, 1834–1844. [Google Scholar] [CrossRef]
  8. Wilding, J.P.H.; Batterham, R.L.; Calanna, S.; et al. Once-weekly semaglutide in adults with overweight or obesity. N Engl J Med. 2021, 384, 989–1002. [Google Scholar] [CrossRef]
  9. Elashoff, M.; Matveyenko, A.V.; Gier, B.; Elashoff, R.; Butler, P.C. Pancreatitis, pancreatic, and thyroid cancer with GLP-1–based therapies. Gastroenterology. 2011, 141, 150–156. [Google Scholar] [CrossRef]
  10. Nauck, M.A.; Quast, D.R.; Wefers, J.; Pfeiffer, A.F.H. GLP-1 receptor agonists in the treatment of type 2 diabetes – state-of-the-art. Mol Metab. 2021, 46, 101102. [Google Scholar] [CrossRef]
  11. Sultana, J.; Giorgianni, F.; Fontana, A.; et al. The safety of GLP-1 receptor agonists: a systematic review. Drug Saf. 2019, 42, 681–698. [Google Scholar] [CrossRef]
  12. Persson, R.; Hedenmalm, K.; Aagaard, L.; et al. Spontaneous reports of pancreatitis associated with GLP-1 receptor agonists. Diabetes Obes Metab. 2020, 22, 409–417. [Google Scholar] [CrossRef]
  13. Petit, C.; Boudiaf, N.; Drouin, J.; et al. Drug-induced acute pancreatitis related to GLP-1 receptor agonists: analysis from a pharmacovigilance database. Diabetes Metab. 2022, 48, 101310. [Google Scholar] [CrossRef]
  14. Ferreira, J.P.; Girerd, N.; Rossignol, P.; Zannad, F. Spontaneous adverse drug reaction reporting in cardiovascular medicine. Eur Heart J. 2018, 39, 3107–3109. [Google Scholar] [CrossRef]
  15. Hauben, M.; Zhou, X. Quantitative methods in pharmacovigilance: focus on signal detection. Drug Saf. 2003, 26, 159–186. [Google Scholar] [CrossRef]
  16. van Puijenbroek, E.P.; Bate, A.; Leufkens, H.G.; et al. A comparison of measures of disproportionality for signal detection in spontaneous reporting systems for adverse drug reactions. Pharmacoepidemiol Drug Saf. 2002, 11, 3–10. [Google Scholar] [CrossRef]
  17. Bate, A.; Evans, S.J.W. Quantitative signal detection using spontaneous ADR reporting. Pharmacoepidemiol Drug Saf. 2009, 18, 427–436. [Google Scholar] [CrossRef]
  18. Norén, G.N.; Hopstadius, J.; Bate, A. Shrinkage observed-to-expected ratios for robust and transparent large-scale pattern discovery. Stat Methods Med Res. 2013, 22, 57–69. [Google Scholar] [CrossRef]
  19. Benjamini, Y.; Hochberg, Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Series B. 1995, 57, 289–300. [Google Scholar] [CrossRef]
  20. Scarpello, J.H.B. The GLP-1 system: physiology and therapeutic applications. Br J Diabetes Vasc Dis. 2001, 1, 224–230. [Google Scholar] [CrossRef]
  21. Nauck, M.A.; Kemmeries, G.; Holst, J.J.; Meier, J.J. Rapid gastric emptying of a solid test meal after administration of exenatide in type 2 diabetes. J Clin Endocrinol Metab. 2011, 96, E884–E886. [Google Scholar] [CrossRef]
  22. Thomas, C.E.; Cadenhead, C.; Donskey, C.J. Off-label and illicit use of GLP-1 receptor agonists for weight loss: a new safety concern. Ann Intern Med. 2023, 176, 261–263. [Google Scholar] [CrossRef]
  23. Sagner, M.; Meier, J.J. Patient-reported barriers to injectable GLP-1 RA use in routine clinical care. Diabetes Ther. 2020, 11, 395–403. [Google Scholar] [CrossRef]
  24. EMA. Withdrawal of the marketing authorization for Lyxumia (lixisenatide) and Byetta (exenatide) in the EU. European Medicines Agency. 2024. [Google Scholar]
  25. Norén, G.N.; Orre, R.; Bate, A.; Edwards, I.R. Duplicate detection in adverse drug reaction surveillance. Drug Saf. 2007, 30, 607–616. [Google Scholar] [CrossRef]
  26. Begaud, B.; Martin, K.; Fourrier-Reglat, A.; et al. Benevolat, usefulness of spontaneous reporting systems for pharmacovigilance: experience and limitations. Therapie. 2002, 57, 569–576. [Google Scholar] [CrossRef]
  27. Lopez-Gonzalez, E.; Herdeiro, M.T.; Figueiras, A. Determinants of under-reporting of adverse drug reactions: a systematic review. Drug Saf. 2009, 32, 19–31. [Google Scholar] [CrossRef]
  28. Hall, G.C.; Sauer, B.; Bourke, A.; Brown, J.S.; Reynolds, M.W.; Casale, R.L. Guidelines for good database selection and use in pharmacoepidemiology research. Pharmacoepidemiol Drug Saf. 2012, 21, 1–10. [Google Scholar] [CrossRef]
  29. Baldé, S.; Dutordoir, M.; Bellet, L.; Thiessard, P. PhViD: Pharmacovigilance Signal Detection [Internet]. R package version 2.2.0; 2019. Available from: https://CRAN.R-project.org/package=PhViD.
  30. Inácio, P.; Cavaco, A.; Airaksinen, M. The value of patient reporting to the pharmacovigilance system: a systematic review. Br J Clin Pharmacol. 2017, 83, 227–246. [Google Scholar] [CrossRef]
  31. AEMPS. Sistema Español de Farmacovigilancia. Available at: https://www.aemps.gob.es/vigilancia-medicamentos-de-uso-humano/farmacovigilancia-en-espana/.
  32. Brown, E.G.; Wood, L.; Wood, S. The Medical Dictionary for Regulatory Activities (MedDRA). Drug Saf. 1999, 20, 109–117. [Google Scholar] [CrossRef]
  33. Norén, G.N.; Sundberg, R.; Bate, A.; Edwards, I.R. A statistical methodology for drug safety signal detection in small databases. Drug Saf. 2008, 31, 525–537. [Google Scholar] [CrossRef]
  34. DuMouchel, W. Bayesian data mining in large frequency tables, with an application to the FDA spontaneous reporting system. Am Stat. 1999, 53, 177–190. [Google Scholar] [CrossRef]
  35. Noguchi, Y.; Tachi, T.; Teramachi, H. Factors associated with the prioritization of safety signals detected from spontaneous adverse event reporting. Pharmaceutics. 2021, 13, 1027. [Google Scholar] [CrossRef]
Table 1. Safety Signal evolution and fact sheet comments for GLP-1 Receptor Agonists between 2024-2025.
Table 1. Safety Signal evolution and fact sheet comments for GLP-1 Receptor Agonists between 2024-2025.
Drug Event Effect (PT) Fact sheet Comments Signal evolution
Dulaglutide Blood glucose abnormal Hypoglycemia in combination with other medications New
Dulaglutide Injection site haematoma Not reported New
Exenatide Renal failure Withdrawn from market in 2024 New
Liraglutide Incorrect dose administered by a medical device Not reported New
Liraglutide Injection site bruise Not reported New
Liraglutide Product quality issue Not reported New
Liraglutide Skin reaction Not reported; skin and subcutaneous tissue disorders reported New
Semaglutide Extra dose administered Not reported New
Semaglutide Diarrhoea Reported as very common New
Semaglutide Off-label use Not reported New
Semaglutide Vomiting Reported as common New
Dulaglutide Decreased appetite Reported as common Reinforce
Dulaglutide Hypoaesthesia Not reported Reinforce
Dulaglutide Accidental overdose Not reported Reinforce
Exenatide Retching Withdrawn from market in 2024 Reinforce
Exenatide Nodule Withdrawn from market in 2024 Reinforce
Liraglutide Injection site rash Not reported Reinforce
Liraglutide Drug ineffective Not reported Reinforce
Liraglutide Injection site swelling Not reported Reinforce
Liraglutide Injection site hypersensitivity Not reported Reinforce
Liraglutide Injection site pruritus Not reported Reinforce
Liraglutide Injection site reaction Reported as common Reinforce
Lixisenatide Hypoglycaemia Withdrawn from market in 2024 Reinforce
Lixisenatide Urticaria Withdrawn from market in 2024 Reinforce
Semaglutide Incorrect technique in product use procedure Not reported Reinforce
Semaglutide Use of product for unapproved indication Not reported Reinforce
Exenatide Asthenia Withdrawn from market in 2024 Diminished
Semaglutide Dyspepsia Reported as common Diminished
Semaglutide Drug intolerance Not reported Diminished
Semaglutide Nausea Reported as very common Diminished
Semaglutide Weight decreased Reported as common Diminished
Semaglutide Gastrointestinal disorder Reported without specification Diminished
Dulaglutide Incorrect dose administered Not reported Unchanged
Dulaglutide Injection site pain Not reported Unchanged
Dulaglutide Blood glucose increased Not reported Unchanged
Dulaglutide Injection site haemorrhage Not reported Unchanged
Dulaglutide Intestinal obstruction Reported, frequency unknown Unchanged
Dulaglutide Dose omission issue with the product Not reported Unchanged
Exenatide Erythema Withdrawn from market in 2024 Unchanged
Exenatide Injection site induration Withdrawn from market in 2024 Unchanged
Exenatide Skin mass Withdrawn from market in 2024 Unchanged
Exenatide Injection site nodule Withdrawn from market in 2024 Unchanged
Exenatide Pancreatitis Withdrawn from market in 2024 Unchanged
Exenatide Acute pancreatitis Withdrawn from market in 2024 Unchanged
Liraglutide Injection site erythema Not reported Unchanged
Liraglutide Minor weight loss Not reported Unchanged
Liraglutide Problem with drug delivery device system Not reported Unchanged
Liraglutide Injection site urticaria Reported as uncommon Unchanged
Dulaglutide Limb pain Not reported Disappeared
Exenatide Renal failure Withdrawn from market in 2024 Disappeared
Liraglutide Injection site bruising Not reported Disappeared
Lixisenatide Dizziness Withdrawn from market in 2024 Disappeared
Semaglutide Inadequate diabetes mellitus control Not reported Disappeared
Semaglutide Overdose Not reported Disappeared
Semaglutide Use of a medicine off-label Not reported Disappeared
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.