Over the last decade, incretin-based drugs, dipeptidyl peptidase-4 (DPP-4) inhibitors and glucagon-like peptide 1 receptor agonist (GLP-1) analogues have emerged as important treatments for patients with Type 2 diabetes mellitus (T2DM). In 2005, Exendin-4 (Exenatide), the world’s first GLP-1 RA, was approved by the U.S Food and Drug Administration (FDA) for the treatment of Type 2 diabetes (T2D). In 2020, the American Diabetes Association (ADA) recommended GLP-1 RAs and sodium glucose co-transporter 2 (SGLT-2) inhibitors as valuable options in treatment with T2D patients and comorbidities in its clinical guidelines for the treatment of diabetes patients.¹⁾

Despite the several therapeutic advantages, safety concerns related to the potential risk of pancreatitis have been raised with some reported cases implying an association between the use of GLP-1 RA agents and increased risk of pancreatitis. In October 2007, the FDA issued its first warning concerning 30 cases of pancreatitis associated with exenatide.²⁾ Until most recent years, health authorities have continued to conduct pharmacovigilance database studies and meta-analysis of randomized controlled trials to assess the potential risk of gallbladder and biliary related disease with the use of GLP-1 RAs, as a clear causative relationship has not been yet definitively established in clinical trials. In fact, clinical trials have been conducted mainly on the Western population, and there confounding factors such as comorbidity, life style factor, genetic characteristics by population, etc. need to be considered.³⁾

To address the gaps in clinical knowledge, this study has been designed to examine the relationship between the use of GLP-1 RAs and gallbladder and biliary disease using real-life pharmacovigilance data. The latest Individual Case Safety Reports (ICSRs) from the international drug surveillance database, VigiBase, will be analyzed. Additionally, since there is insufficient evidence on the intercontinental differences in the risk of biliary disease, we aim to identify whether there are regional differences in disproportional reporting of biliary disease following the use of GLP-1 RA.

Study design and Data source

In this observational pharmacovigilance study, we analyzed the data which have been collected in VigiBase. The VigiBase is the global database of reported potential side effects of medicinal products, managed by the Uppsala Monitoring Centre. It is the largest database of ICSRs in the world, with over 30 million reports of suspected adverse effects of medicines since 1968, originated from various sources, such as physicians, pharmacists, other health care professionals. VigiBase is linked to medical and drug classifications such as Medical Dictionary for Regulatory Activities terms (MedDRA), WHO International Classification of Disease (ICD) and Anatomical Therapeutic Chemical (ATC) classification system. Each ICSR has information on demographic (age group, sex, region), adverse events (seriousness, outcome), and drug information (dosage regimen, route of administration, treatment initiate dates, dechallenge and rechallenge).⁴⁾ The information regarding regions in VigiBase, which is the primary data source for this study, represents to areas where adverse events (AE) have been reported. In VigiBase, regions are categorized into six: African, America, Europe, Eastern Mediterranean, South-East Asia, and Western Pacific region. However, as the purpose of this study to assess the regional differences in biliary disease related to the use of GLP-1 RA, we classified regions into four major regions: America, Europe, Asia, and other regions.

Data extraction and selection

We included all ICSRs for GLP-1 RA agents and all other antidiabetic drugs between April 2005 and January 2023; GLP-1 RA was initially approved in 2005 by FDA. Among all ICSRs related to all other antidiabetic drugs, we retrieved reports from adults (≥18 years). We excluded ICSRs with any of the following characteristics: reports that were not reported as ‘suspected’ on the drug information, unknown or missing sex data; unknown or missing geographic information data; missing preferred terms code data; duplicated reports. Moreover, to improve the validity of included reports, only ICSRs reported as suspected on the drug information. Regarding the follow-up reports with additional information after initial reports, only the latest ones were included. The detailed information about report selection is shown in Fig. 1.

Fig. 1. Study flowchart for report selection of GLP-1 RAs in VigiBase ^®.
Abbreviations: AE, Adverse Event; PT, Preferred Term

Outcomes

The Adverse Events of special interest (AESIs) were defined by the Standardized MedDRA Queries (SMQ), validated standard sets of MedDRA terms based on extensive review, testing, analysis, and expert discussion. Biliary disease cases were identified using MedDRA Preferred terms (PTs) which belong to SMQ ‘Biliary tract disorders’ and ‘Gallbladder-related disorders’. These specific outcomes were selected after reviewing RCT meta-analysis, previous research papers, etc.^3,5-8) The detailed Preferred Terms for outcome definition are listed in the Table S1.

Disproportionality analysis

Disproportionality analysis, which is a validated case-non case method in pharmacovigilance study, was used to examine whether the AESIs were more frequently reported than would be expected with GLP-1 RA analogues compared with AESIs reported for comparator drugs in VigiBase. Disproportionality analysis aims to generate hypothesis on possible causal relations between drugs and adverse effects, followed by clinical assessment of the individual case safety reports. All disproportionality analysis was conducted using AE-pairs defined as pairs of reported drugs and reported adverse events and not the number of reports or ICSRs. In our analysis, cases included reports with the AESIs, with all other AEs as non-cases. The disproportionality analysis was performed separately across 4 regions (America, Europe, Asia, and other regions).

Statistical analysis

Baseline characteristics of reports related to GLP-1 RAs were presented using descriptive statistics across within 4 regions, respectively. Categorical variables were presented using frequencies and proportions (%) and continuous variables using median and interquartile range (IQR). Reporting odds ratio (ROR) and 95% confidence intervals (CIs) were calculated to estimate the extent to which AESI is associated with GLP-1 RA agents vs. all other antidiabetic drugs across the regions. The ROR is based on frequentist method and compares the odds of AE-pairs of cases against the odds of AE-pairs of non-cases. A disproportionality signal was defined when it met the following criteria or threshold: ROR, when the lower end of the 95% confidence interval (CI)≥1 according to Uppsala monitoring Centre (UMC)’s criteria.⁹⁾ In sensitivity analysis, we repeated our main analysis by restricting notifier as physician to improve the validity of our analysis. Also, we explored the distribution of reported biliary tract disorders or gallbladder-related disorders according to MedDRA Preferred terms. Furthermore, we analyzed the annual number of reported cases associated with GLP-1 RA between 2005 and 2022 to identify the reporting trend of AESI across regions.

Data management and analysis were conducted using SAS 9.4 version software program (SAS Inc, Cary, NC), and statistical significance was defined as two-sided p<0.05.

Post-hoc analysis

As a post-hoc analysis, we evaluated the safety signals of GLP-1 RAs using SGLT2 inhibitors and DPP-4 inhibitors as a comparator group, instead of all other antidiabetic drugs in the database. Since both SGLT2 inhibitors and GLP-1 RAs are now recommended as second-line therapy in patients with T2D due to their glucose-lowering effects with additional benefits for cardiorenal disease,10) we need to consider further similar pharmacological effects of these drugs.

Study population and Baseline characteristics of reports

A total of 1,481,822 reports on antidiabetic drugs among adult population (≥18 years) were selected in the VigiBase between April 2005 and January 2023. After processing these raw VigiBase records for exclusion of reports with non-suspected drug as well as duplicated or missing and low-quality reports, we included 392,632 reports related to the use of antidiabetics and 79,715 reports on GLP-1 RA drugs in the analysis (Fig. 1). Among them, it was identified that 757 were classified as cases (America: 454, Europe: 241, Asia: 56, Others: 6), and 78,958 for non-cases (America: 52,359, Europe: 20,966, Asia: 4,910, Others: 723). More than half of all cases with available information were reported from America (59.97%), in the category of 45-64 years of age (53.96%).

Characteristics of reports from America, Europe and Asia were largely similar as a higher proportion of cases were observed among those aged 45-64 years, females, spontaneous reports (Table 1). Reports showed a distinct regional difference in the characteristics of notifiers for cases, indicating that the majority of ICSRs were reported by physicians in Europe (62.24%) and Asia (78.57%), rather than in America (32.82%) and other regions (16.67%). Most of the cases were serious across all regions.

Table 1 Baseline characteristics of all cases and non-cases adverse event reports for GLP-1 RAs and Other antidiabetic drugs in WHO-UMC VigiBase^® from April 2005 to January 2023

Characteristics	America (N, %)		Europe (N, %)		Asia (N, %)		Other regions (N, %)
	Cases	Non-cases	Cases	Non-cases	Cases	Non-cases	Cases	Non-cases
No. of reports	454 (0.86)	52,359 (99.14)	241 (1.14)	20,966 (98.86)	56 (1.13)	4,910(98.87)	6(0.82)	723 (99.18)
Age (years)
18-44	60 (13.22)	4,387 (8.38)	42 (17.43)	2,555 (12.19)	4 (7.14)	1,247 (25.40)	4 (66.67)	334 (46.20)
45-64	245 (53.96)	26,634 (50.87)	107 (44.40)	11,067 (52.97)	17 (30.36)	2,277 (46.37)	2 (33.33)	323 (44.67)
65-74	118 (25.99)	15,665 (29.92)	71 (29.46)	5,544 (26.44)	17 (30.36)	696 (14.18)	0 (0.00)	57 (7.88)
≥75	31 (6.83)	5,673 (10.83)	21 (8.71)	1,800 (8.59)	18 (32.14)	690 (14.05)	0 (0.00)	9 (1.24)
Sex
Male	202 (44.49)	21,093 (40.29)	106 (43.98)	9,217 (43.96)	33 (58.93)	2,071 (42.18)	1 (16.67)	230 (31.81)
Female	252 (55.510	31,266 (59.71)	135 (56.02)	11,749 (56.04)	23 (41.07)	2,839 (57.82)	5 (83.33)	493 (68.19)
Report type
Spontaneous	406 (89.43)	51,292 (97.96)	207 (85.89)	17,079 (81.46)	44 (78.57)	2,164 (44.07)	4 (66.67)	533 (73.72)
Report from study	39 (8.59)	685 (1.31)	34 (14.11)	3,878 (18.50)	12 (21.43)	1,298 (26.44)	2 (33.33)	189 (26.14)
Other	1 (0.22)	295 (0.56)	0 (0.00)	9 (0.04)	0 (0.00)	130 (2.65)	0 (0.00)	1 (0.14)
Unknown	8 (1.76)	87 (0.17)	0 (0.00)	0 (0.00)	0 (0.00)	1,918 (26.84)	0 (0.00)	0 (0.00)
Seriousness*
Yes	416 (91.63)	11,193 (21.38)	194 (80.50)	5,955 (28.40)	49 (87.50)	1,302 (26.52)	4 (66.67)	134 (18.53)
No	36 (7.93)	40,731 (77.79,)	46 (19.09)	14,950 (71.31)	5 (8.93)	3,475 (70.77)	1 (16.67)	543 (75.10)
Unknown	2 (0.44)	435 (0.83)	1 (0.41)	61 (0.29)	2 (3.57)	133 (2.71)	1 (16.67)	46 (6.36)
Notifier
Physician	149 (32.82)	4,968 (9.49)	150 (62.24)	10,301 (49.13)	44 (78.57)	2,038 (41.51)	1 (16.67)	213 (29.46)
Pharmacist	7 (1.54)	1,080 (2.06)	16 (6.64)	2,403 (11.46)	1 (1.79)	314 (6.40)	0 (0.00)	48 (6.64)
Others	298 (65.64)	46,311 (88.45)	75 (31.12)	8,262 (39.41)	11 (19.64)	2,558 (52.10)	5 (83.33)	462 (63.90)

*Fatal, Life-threatening, requiring hospitalization, resulting in significant disability/incapacity, and other medically important conditions

Annual reporting trends in the number of AESIs

The annual number of cases associated with GLP-1 RAs began to gradually increase in 2007. From 2005 to 2016, the highest reporting trend was observed in America, followed by Europe. However, in 2020, Europe started to overtake America’s records with a sharp increase in the reporting rate of cases from 2019 to 2020. In 2022, the number of cases in Europe exceeded that of America with a greater gap compared to 2020. Throughout the entire study period, Asia has shown considerably low reporting rates of cases, in comparison to other countries (Fig. 2).

Fig. 2. Annual cases reporting trends from GLP-1 RAs in VigiBase^® between 2005 and 2022

AESIs (Biliary tract disorders/Gallbladder-related disorders)

We identified 233 cases of biliary tract disorders reported from GLP-1 RA analogues, and for gallbladder-related disorders, 641 cases were included in the study analysis. Disproportionate reporting of biliary tract disorders associated with GLP-1 RAs was only observed in America compared to all other antidiabetic drugs. However, for gallbladder-related disorders, an increased risk of occurrence was identified in most regions to be associated with GLP-1 RAs compared to all other antidiabetic drugs, including America (ROR 2.34; 95%CI 2.06-2.66), Europe (ROR 5.71; 95%CI 4.62-7.05), Asia (ROR 4.12; 95%CI 2.98-5.69), and others (ROR 2.32; 95%CI 0.99-5.43) (Table 2). Regarding the distribution of cases according to the MedDRA Preferred Terms (PTs), the most frequently reported adverse events among biliary tract disorders were jaundice (n=54, 23.18%), jaundice cholestatic (n=49, 21.03%), bile duct stone (n=24, 10.30%), biliary obstruction (n=24, 10.30%) and obstructive pancreatitis (n=23, 9.87%). In addition, the most frequently reported adverse events in gallbladder-related disorders were cholelithiasis (n=316, 49.30%), cholecystitis (n=82, 12.79%), gallbladder disorder (n=70, 10.92%) and cholecystitis acute (n=63, 9.83%), cholecystectomy (n=31, 4.84%) (Fig. 3).

Fig. 3. Frequency of Preferred Terms regarding cases related to GLP-1 RAs between April 2005 to January 2023

Table 2 Signal detection of GLP-1 RAs using disproportionality analysis in WHO-UMC VigiBase^® from April 2005 to January 2023

	Cases of GLP-1 RA	Non-cases of GLP-1 RA	ROR (95% CI)	Signal†
Biliary tract disorders*
America	122	158,002	1.29 (1.05-1.59)	O
Europe	91	47,749	0.64 (0.51-0.79)	X
Asia	19	9,019	1.41 (0.89-2.25)	X
Other regions	1	1,552	0.81 (0.11-5.95)	X
Gallbladder-related disorders*
America	396	157,728	2.34 (2.06-2.66)	O
Europe	195	47,645	5.71 (4.62-7.05)	O
Asia	44	8,994	4.12 (2.98-5.69)	O
Other regions	6	1,547	2.32 (0.99-5.43)	X

Abbreviations: ROR, Reporting Odds Ratio; CI, Confidence Interval; WHO-UMC, World Health Organization-Uppsala Monitoring Centre

*The Adverse Events of interest are defined using Standardized MedDRA Queries (SMQ)

^†Safety signals detection with ROR were assessed as AEs where thresholds of ROR were greater than one.

Sensitivity analysis

Sensitivity analysis was largely consistent with main analysis results. A signal for biliary tract disorders was only found in America (ROR 2.79; 95%CI 1.79-4.36). Similarly, signals for gallbladder-related disorders were consistent across America, Europe, and Asia (Table 3).

Table 3 Signal detection of GLP-1 RAs using disproportionality analysis when restricted to reports by physician in WHO-UMC VigiBase^® from April 2005 to January 2023

	Cases of GLP-1 RA	Non-cases of GLP-1 RA	ROR (95% CI)	Signal†
Biliary tract disorders*
America	26	11,614	2.79 (1.79-4.35)	O
Europe	57	21,012	0.63 (0.48-0.83)	X
Asia	13	2,899	1.65 (0.94-2.90)	X
Other regions	0	358 N/A	N/A
Gallbladder-related disorders*
America	147	11,493	11.31 (8.82-14.49)	O
Europe	119	20,950	7.06 (5.35-9.31)	O
Asia	37	2,875	5.32 (3.71-7.62)	O
Other regions	1	357	1.69 (0.22-13.01)	X

Abbreviations: ROR, Reporting Odds Ratio; CI, Confidence Interval; WHO-UMC, World Health Organization-Uppsala Monitoring Centre

*The Adverse Events of interest are defined using Standardized MedDRA Queries (SMQ)

^†Safety signals detection with ROR were assessed as AEs where thresholds of ROR were greater than one.

Post-hoc analysis

In comparing GLP-1 RAs to SGLT2 inhibitors, we found a higher risk of biliary tract disorders in America (ROR 1.70; 95%CI 1.10-2.65) and Europe (ROR 2.26; 95%CI 1.46-3.50). Similarly, an increased risk of gallbladder-related disorders was identified with GLP-1 RAs compared to SGLT2 inhibitors in America (ROR 2.61; 95%CI 1.93-3.52) and Europe (ROR 4.71; 95%CI 3.10-7.16) (Table S2). However, disproportionate reporting for gallbladder-related disorders was observed solely in Europe when comparing reports from GLP-1 RAs with DPP-4 inhibitors (Europe: ROR 2.53; 95%CI 1.68-3.81) (Table S3).

In this global pharmacovigilance study, we evaluated regional differences in reporting adverse events associated with GLP-1 RA concerning biliary disease. We found higher reporting odds for gallbladder-related disorders in most regions, with a significant reporting risk observed only in America for biliary tract disorders.

Prior studies narrowly focused on specific adverse events associated with GLP-1 RAs and biliary diseases. However, our study took a broader approach, categorizing AESIs as biliary disease, which includes both biliary tract and gallbladder disorders. As for primary study objective, to investigate the regional differences in the occurrence of AESIs was prompted by the hypothesis that pre-existing risks associated with biliary diseases, such as race/ethnicity or medical conditions aside from GLP-1 RA treatment, could potentially impact the occurrence of AESIs.

Furthermore, the rationale behind clustering adverse events according to the SMQ classification was to systematically improve the confirmation of the association between the interested drug and occurrence of AESIs. Dividing AESIs into two adverse event categories in this study was a noteworthy attempt to group GLP-1 RA’s biliary-related safety concerns into a coherent set of related diseases, thereby improving the fundamental understanding of the relationship between the use of the drug and the occurrence of AESIs.

According to the main analysis of this study, a signal was only detected in America for biliary tract disorders, while signals for gallbladder-related disorders were detected in most regions, including America, Europe, and Asia. This suggests that there was regional imbalance for biliary track disorders, whereas no regional disparity was observed among major countries (America, Europe, and Asia) for gallbladder-related disorders. The representative AESIs mentioned to have a higher occurrence when using GLP-1 RAs in prior studies included (acute) pancreatitis, (acute) cholecystitis, cholecystectomy, and cholelithiasis, most of which were classified under the group of gallbladder-related disorders in this study. In the case of biliary tract disorders, where the signal was detected only in America, further investigation through a multinational confirmatory study is deemed necessary by adjusting potential confounding factors that might have caused this imbalance.

One plausible mechanism to consider is the association between obesity and biological formation of gallstones, a well-known cause of acute pancreatitis. Additionally, the fact that America is one of the highest prevalence of overweight or obese individuals among OECD nations may partially explain the observed imbalance in biliary tract disorders identified in this research. In 3 large-scale population-based retrospective cohorts studies indicated that T2DM increases the risk of acute pancreatitis by 1.5 to 3 times, after adjusting for demographic and other risk factors.^11-13) The results of these cohort studies are consistent with the incidence of acute pancreatitis among patients with T2DM in the United Kingdom (UK). After adjusting for age, gender, and known risk factors, the incidence of acute pancreatitis in patients with T2DM is higher than in those without diabetes.¹¹⁾ It is noteworthy that an accompanying trend has been the global increase in T2D14,15) and obesity.^16-20)

The prevalence of obesity was the highest in the United States (36.5% in 2011-2014), while Europe ranked second globally with an average prevalence of 15.9% across EU member states in 2014. In 2020, the percentage of overweight or obese populations among OECD member countries was over 70% in Mexico (74.1%) and the United States (73.1%), while Japan (27.2%) and Korea (37.8%) were below 38%.²¹⁾ T2D and obesity-related clinical factors are identified as risk factors for acute pancreatitis. This suggests a higher risk of acute pancreatitis among patients with T2D compared to the general population.¹³⁾

Obesity promotes gallstone formation, contributing to acute pancreatitis. The differences in gallstone prevalence across worldwide regions would exist due to factors like racial susceptibility, climate, and obesity rates, affecting biliary tract disease. Despite being less common, gallstone-induced pancreatitis is as severe as other causes.^22-24)

In summary, diabetes itself or other potential risk factors such as demographic factors, genetic, obesity-related comorbidities like hypertriglyceridemia and gallstone, lifestyle, or key dietary habits (ex. Western diet, high refined carbohydrates and low in fiber, alcohol abuse) should be necessarily considered when we evaluate the increased risk of biliary disease associated with the use of GLP-1 RAs. We found a higher risk of gallbladder-related disorders related to GLP-1 RAs vs. other all other antidiabetic drugs in America, Europe, and Asia with no geographical differences. Our results align with a UK population-based cohort study that also reported increased bile duct and gallbladder disease risk with GLP-1 RA. Additionally, this study found a two-fold increased risk in cholecystectomy associated with GLP-1 analogs,25) ranking as the 5th most frequently reported AE in our study for gallbladder-related disorders. We performed a post-hoc analysis to investigate and compare the increased risk of AESIs between GLP-1 RAs and SGLT2 inhibitors, considering similar therapeutic effects of two drugs. Our study showed the increased risk of AESIs associated with GLP-1 RAs compared with SGLT2 inhibitors in America and Europe for biliary disease. In previous pharmacovigilance study which has investigated the SGLT2 inhibitors-induced pancreatitis, it was found that there is a possibility of an increased risk of biliary disease associated with the use of SGLT2 inhibitors, and therefore this risk cannot be completely ruled out.²⁶⁾ In line with the controversial safety concern that the use of incretin-based drugs may increase the risk of bile duct and gallbladder diseases, we repeated post-hoc analysis on DPP-4 inhibitors as comparator of GLP-1 RAs. In our study, disproportionality was only found in Europe for gallbladderrelated disorders whereas not in any other regions. These results are not completely consistent with previous population-based cohort study or a case/non-case study from nationwide pharmacovigilance database and meta-analysis of RCTs.^25,27,28) As of now, there’s no definitive conclusion on increased biliary disease risk with GLP-1 RAs, urging caution in prescribing incretin-based drugs to T2D patients.

To our knowledge, this is the first observational pharmacovigilance study exploring the potential ethnic differences in biliary related adverse events associated with the use of GLP-1 RAs. One of the key strengths of this study is the utilization of VigiBase as a source of data, which comprises of reports submitted from countries across the world. This aspect makes it highly beneficial to identify safety issues with medications that may not be detected in a single country. Second, this study generated recent real-world evidence in response to knowledge gaps that have not yet to be established in clinical trials or systematic review and meta-analysis regarding gallbladder or biliary disease associated with GLP-1 RAs treatment. Third, the most significant advantage emphasized in this study is the identification of AESI cases using the SMQ to define terminology in a systematic and integrated manner. This approach could serve as supplementary clinical data for addressing the issue of confusion in signal detection across all marketed GLP-1 RA agents, resulting from the lack of standardization in the criteria and classification or clustering of biliary disease in the warnings and precautions of authorized label for each marketed product. Lastly, we conducted sensitivity analysis restricting the ICSRs which were notified by physicians to examine the robustness of our main findings and the results were largely consistent with the results from main analysis.

However, several limitations should be acknowledged. First, as the inherent limitation of all studies using pharmacovigilance databases, we cannot rule out biases such as underreporting, missing information, potential data duplication, confounding issues, data quality challenges and lack of on the total number of patients exposed to specific drugs. Second, although information on indication, dosage, or duration exposure to interested drugs may influence the risk of AESIs, we could not assess these confounding factors due to lack of data in VigiBase. In recent systematic review and meta-analysis of RCTs, it showed that use of GLP-1 RAs was associated with increased risk of gallbladder or biliary diseases, especially when used at higher doses, for longer durations, and for weight loss.²⁹⁾ We also recognize that our study cannot determine causality between drug and specific adverse events, and it should be noted that RORs are not intended to be used as a measure of real risk value, but rather as an indicator of a potential safety signal. In our study, disproportionate reporting was only observed in America for biliary track disorders associated with the GLP-1 RAs compared with all other antidiabetic drugs. Whereas, for gallbladder-related disorders, the signal for AESIs was detected in America, Europe, and Asia, suggesting no significant regional differences in the reporting probability of these AEs among major countries. These findings are consistent with the result from sensitivity analysis in the study, however, we cannot dismiss the potential confounding factors which might influence to the interpretation of study result. In addition to demographic factors, we must consider other factors that could impact the etiology of biliary disease, such as obesityrelated comorbidities, genetic predisposition, dietary and lifestyle habits. It should also be considered the selective nationwide declarations of certain adverse events due to safety alerts or media attention from regulatory agencies, as this can lead to a bias based on the publicity of specific incidents. A further limitation is that interested drugs are not equally distributed across countries as per the different time a product has been marketed, or reimbursement standard on each GLP-1 RA product in countries when making comparisons between adverse events associated with GLP-1 RAs. Indeed, differences in timing for marketing, reimbursement, and authorization of GLP-1 RA could potentially introduce reporting bias, influencing significant changes in the number of reports across all regions from year to year. In this study, we only used reporting odds ratio, which was calculated using a logistic regression model to adjust for confounders as an indicator of imbalance analysis. The advantage of using reporting odds ratio is that the interpretation of the results is intuitive and relatively easy to understand. Nevertheless, to identify signals more cautiously, it is essential to use a combination of diverse data mining techniques, such as utilizing information components (ICs) to enhance sensitivity and tree-based scan statistics to reduce Type I errors.

Our findings could contribute to a comprehensive understanding of safety profile of GLP-1 RAs associated with the ongoing clinical debate regarding the safety concerns in biliary disease. Study results including real-life data from patient-level can be supportive to guide physician’s medical decisions for the optimal treatment of individual patient under clinical setting and supplementary document in public policy decisions. Due to the nature of spontaneous pharmacovigilance analysis, the results of this study do not provide any conclusive evidence that can be directly applied in clinical practice, and our findings require further validation and should be interpreted with caution. Considering the expectation of expanding indications and increasing usage of GLP-1 RAs worldwide, continuous surveillance by clinicians and pharmacovigilance experts is inevitable. Furthermore, a multinational confirmatory study is required to determine the causality between GLP-1 RAs and biliary disease, validate regional disparities in the association between GLP-1 RAs and biliary disease, and provide additional evidence for the safety of these agents.³⁰⁾

The authors acknowledge the WHO Uppsala Monitoring Center that allowed the authors’ access to VigiBase^®.

This study was conducted with R&D expenses (21153MFDS607) of the Ministry of Food and Drug Safety from 2021 to 2025.

All authors declare that they have no conflict of interest.

The institutional review board of Sungkyunkwan University approved the study (IRB No. SKKU 2024-01-002); the board waived the requirement for obtaining informed consent as this study used anonymized administrative data.

Yoojeong Yoon : Graduate student

Dong Hyuk Kim : Graduate student

Ju-Young Shin : Professor