Understanding the Work of the Pharmacovigilance Risk Assessment Committee (PRAC): A Quantitative Review of the Post-Authorisation Safety Evaluation of Antidiabetic Drugs from 2012 to 2022

Abstract

Background

The Pharmacovigilance Risk Assessment Committee (PRAC) plays a central role in the European Union’s pharmacovigilance system, evaluating drug safety through several procedures and activities. Despite its central role, few studies have quantitatively investigated the PRAC’s activities from a system’s perspective.

Objective

This study aims to map PRAC’s evaluation of safety signals and concerns using antidiabetic products as a case. It characterises the drugs and adverse events involved, analyses the PRAC-led regulatory procedures where the safety signals and concerns were evaluated, and provides a comprehensive review of PRAC meeting minutes.

Methods

From PRAC meeting minutes, we retrieved information on all antidiabetic drug-related adverse events discussed from 2012 to 2022. We identified drug-adverse event evaluations based on the discussion content. These were described by drug classes, System Organ Classes, PRAC procedures, and the evaluation outcomes corresponding to recommendations for regulatory actions. We also analysed the sequence of PRAC-led procedures and activities addressing drug-adverse event pairs across meeting minutes.

Results

A total of 321 drug-adverse event pairs were identified, with 14 pairs associated with drug classes. Second-generation antidiabetic agents, including sodium-glucose transport protein-2 inhibitors, glucagon-like peptide-1 receptor agonists, and dipeptidyl peptidase 4 inhibitors, were the most frequently discussed. Of these, 62 pairs underwent multiple evaluations, resulting in a total of 413 evaluations. In 48% of evaluations, no regulatory action was required. Most evaluations (97%) were concluded in a single procedure, and 66% were concluded in one meeting. Periodic safety update reports accounted for 54% of drug-adverse event evaluations and updates to product information were the most frequent outcome. Signal assessment and prioritisation procedures, while less common, resulted in more diverse recommendations for regulatory action. Referrals were infrequent (N = 5) and were often triggered by the signal assessment and prioritisation procedure.

Conclusions

Periodic safety update reports are the primary source for PRAC evaluations of safety signals although they are not intended for notification of new urgent safety information. Compared with periodic safety update reports, the signal assessment and prioritisation procedure evaluates fewer signals but leads to a wider range of regulatory actions, from risk minimisation measures to referrals. This difference may be attributed to the fact that signals detected in periodic safety update reports are not intended for urgent safety issues, these should be assessed through the signal assessment and prioritisation procedure, as the latter involves real-time signal management, whereas the periodic safety update reports are conducted at predefined intervals.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40264-025-01536-7.

Key Points

We reviewed Pharmacovigilance Risk Assessment Committee meeting minutes spanning 10.5 years to focus on safety signals and concerns related to all antidiabetic drugs, finding that second-generation antidiabetic agents were frequently assessed in the post-authorisation phase.

Over half of the evaluations took place during periodic safety update report procedures, with signal assessment and prioritisation as the second most used procedure for safety discussions, indicating both marketing authorisation holders and regulators play a vital role in safety evaluation.

Updates to the Summary of Product Characteristics and the risk management plan were the main regulatory actions recommended by the Pharmacovigilance Risk Assessment Committee. Additional risk minimisation measures and post-authorisation safety study recommendations were rare.

Introduction

Pharmacovigilance (PV) is essential for the continuous evaluation of the benefit-risk balance of medicinal products after marketing authorisation (MA) [1]. To enhance the PV system in the European Union (EU), the EU PV legislation came into effect in July 2012 as an update to the existing legislation [2]. This legislation led to the establishment of the Pharmacovigilance Risk Assessment Committee (PRAC) at the European Medicines Agency (EMA). The PRAC is integral to the EU PV system, overseeing all aspects of signal and risk management [2, 3].

Signal detection and evaluation are key components of PV and risk management. A safety signal is defined as “information on a new or known adverse event that may be caused by a medicine that warrants further investigation” [4]. Signals can originate from various sources, such as spontaneous reporting systems, clinical trials, scientific literature and observational studies [5]. The PRAC plays a central role in evaluating these safety signals through different procedures and activities. In the post-authorisation phase, the PRAC evaluates safety signals via the signal assessment and prioritisation procedure (Signal Procedure for short). Safety signals are also evaluated in connection with periodic safety update report (PSUR) procedures, referrals, and potential post-authorisation safety studies (PASS) [6]. Following scientific evaluation, the PRAC may recommend that product information, such as the Summary of Product Characteristics (SmPC), be updated to take into account the new safety information; in more severe cases, change the condition for use of the product, or recommend suspension or withdrawal of the product. The latter will generally happen in connection to a referral procedure. It may also propose additional risk minimisation measures (RMMs), such as educational material, the provision of Direct Healthcare Professional Communication (DHPC), or request additional information in the form of a PASS [1, 7, 8].

Studies have shown that spontaneous reports are the predominant sources of new safety signals discussed in the ‘Signal Procedure’ [9, 10]. Important safety issues may require broader discussion and potentially lead to the initiation of a referral procedure [9–11]. One study found that risk management of safety concerns through routine RMMs is the most common outcome of the referral procedures, and approximately 15% of referrals result in the withdrawal or suspension of MA [12, 13]. The PSUR serves as a routine instrument for MA holders to fulfil their mandated responsibility for continuous evaluation of products’ benefit-risk profiles after MA [1, 7]. The PSURs are submitted to regulatory authorities at predefined time intervals, with a frequency of 6 months for newer products to several years for products with more well-established use. The PSUR provides a cumulative overview of all data on the product’s benefits and risks throughout its lifetime and takes into consideration new or emerging information. While PSURs can lead to PRAC recommendations for regulatory actions, such as changes to the SmPC, the PSUR is not intended to serve as a mechanism for informing about significant new safety signals, nor is it meant to be a tool for detecting emerging safety concerns [1, 14].

The PRAC is also responsible for the oversight of the risk management plan (RMP) for centrally authorised medicinal products [15]. Risk management plans are dynamic documents that evolve throughout a medicine’s lifecycle. The RMP outlines the medicine’s safety profile and highlights risks that require further characterisation or management (the ‘safety specification’). It also addresses the planning of studies and PV activities aimed at expanding knowledge of the medicine’s safety and efficacy (the ‘PV plan’). Additionally, the RMP includes the ‘risk minimisation plan’, which covers the planning, implementation and evaluation of RMMs to ensure their effectiveness [1, 15]. The revision of RMP guidelines in 2017 led to fewer identified safety concerns, likely owing to changes in the definition of what constitutes an identified risk [16, 17]. A PASS is conducted after a medicinal product has received an MA to gather additional information on its safety or assess the effectiveness of risk management measures [18]. Thus, it plays a crucial role in providing information about a safety concern, such as causal relationship, seriousness and frequency, and in evaluating the effectiveness of risk management strategies [15, 19]. However, existing research on PASS has mostly examined its role in evaluating RMM effectiveness and the methodological limitations of PASS studies [20, 21].

Despite the comprehensive role of the PRAC in PV, significant gaps remain in understanding how its various procedures contribute to the ongoing assessment of safety profiles for medicinal products. While qualitative studies have provided an overview of PRAC’s role in the EU PV system [1], no quantitative assessment has been conducted on the extent to which these PRAC procedures and activities interact and collectively contribute to the dynamic process of updating a drug’s safety profile. Although quantitative studies of individual PRAC procedures, such as the ‘Signal Procedure’ [7, 9, 10, 12], the referral procedure [12, 13], the PSUR [22, 23] and RMPs [24] have been performed in isolation, there has been little exploration of how these processes interconnect to provide a comprehensive understanding of drug safety.

This study aims to address the existing knowledge gap by mapping how the PRAC evaluates safety signals and concerns in the post-authorisation phase across different procedures and activities. To accomplish this, we conducted a descriptive analysis of PRAC-led procedures and activities addressing safety signal assessments for antidiabetic medicinal products from 2012 to 2022. Specifically, we examined the characteristics of the associated adverse events (AEs), the types of regulatory procedures and oversight activities in which these evaluations were initiated, the sequence of PRAC-led procedures and activities for managing these signals/concerns, and the outcomes of these assessments. Antidiabetic drugs are chosen as the case study because they include both long-established and newer products, utilise diverse administration methods and address a wide range of patients. These features make them well suited for evaluating the PRAC’s activities [25–29].

Method

Data Sources

The PRAC publishes the minutes of its monthly plenary meetings, detailing regulatory procedures and oversight activities in a consistent fixed-headings format that reflects specific types of PRAC procedures and activities. Each regulatory procedure or oversight activity is documented under sub-headings (agenda items) that include the International Non-Proprietary Name(s) (INN(s)) for the active pharmaceutical ingredient(s) (API(s)) it addresses. Each subsection includes background information, discussion and/or a summary of conclusions or recommendations (Fig. S1 in the Electronic Supplementary Material [ESM]). We retrospectively collected all available meeting minutes from July 2012 to December 2022 on the EMA homepage, identifying safety signals/concerns related to antidiabetic drugs.

To identify antidiabetic products marketed in the EU, we used the ‘List of Union Reference Dates and Frequency of Submission of Periodic Safety Update Reports’ (EURD List) [30]. For medicinal products containing combinations of APIs, we consolidated APIs sharing the same PSUR procedure entry using the EURD List. For example, fixed combinations of saxagliptin and metformin were categorised as saxagliptin as they share the same PSUR entry. We then applied the World Health Organization Anatomical Therapeutic Chemical classification [31] to categorise drug classes, grouping insulins by pharmaceutical subgroup (level 3) and other antidiabetic medications by chemical subgroup (level 4) [Table S1 in the ESM].

All PRAC minutes were screened for mentions of the identified antidiabetic products using the R package “PDF Data Extractor” [32] with the INNs serving as keywords. Based on the initial screening, author HL manually screened all minutes that mentioned antidiabetic agents to identify the included PRAC-led procedures and activities that address a safety signal or concern. Sections of PRAC minutes that related to procedures and activities included in this study are listed in Table 1. Identified procedures and activities in the PRAC minutes were excluded if they: (1) did not explicitly identify the AEs assessed; (2) focused on product quality-related AEs; or (3) addressed safety file modifications related to the feasibility of PASS for the given AEs (Fig. 1). Pharmacovigilance Risk Assessment Committee-led procedures and activities related to pre-authorisation, scientific advice and PASS protocols were also excluded.

Table 1.

PRAC procedures and activities included in the study, the table follows the heading structure of the PRAC meeting minutes

Fixed headings in PRAC meeting minutes	Procedure/activity type
2. EU referral procedures for safety reasons: urgent EU procedures	Referral
2.1 Newly triggered procedures
2.2 Ongoing procedures
2.3 Procedures for finalisation
3. EU referral procedures for safety reasons: other EU referral procedures
3.1 Newly triggered procedures
3.2 Ongoing procedures
3.3 Procedures for finalisation
3.4 Re-examination procedures
3.5 Others
4. Signals assessment and prioritisation	Signal assessment and prioritisation (‘Signal procedure’)
4.1 New signals detected from EU spontaneous reporting systems
4.2 New signals detected from other sources
4.3 Signals follow-up and prioritisation
5. Risk management plan	Risk management plan
5.2 Medicines in the post-authorisation phase, PRAC-led procedures
5.3 Medicines in the post-authorisation phase, CHMP-led procedures
6. PSURs	PSURs
6.1 PSURA procedures including centrally authorised products only
6.2 PSURA procedures including centrally authorised products and nationally authorised products
6.3 PSURA procedures including nationally authorised products
6.4 Follow-up PSUR/PSUSA procedures
7. PASS	PASS
7.3 Results of PASS imposed in the marketing authorisation(s)
7.4 Results of PASS non-imposed in the marketing authorisation(s)
7.5 Interim results of imposed and non-imposed PASS submitted before the entry into force of the revised variation regulation
8. Renewals of the marketing authorisation, conditional renewal and annual reassessments	Marketing authorisation renewal
8.1 Annual reassessments of the marketing authorisation
8.2 Conditional renewals of the marketing authorisation
8.3 Renewals of the marketing authorisation
10. Other safety issues for discussion requested by the CHMP or the EMA	Other safety issues
10.1 Safety-related variations of the marketing authorisation
10.3 Other requests
11. Other safety issues for discussion requested by the Member States
11.1 Safety-related variations of the marketing authorisation
11.2 Other requests

In addition, the corresponding headings in Annex I were included in the study. Headings and subheadings not listed here were excluded, as were Annex II, Annex III and explanatory notes. The numbering of headings (but not the heading text) varied slightly across the study period (2012–22). A detailed list of all included and excluded headings and subheadings can be found in Table S2 of the ESM

EMA European Medicines Agency, EU European Union, CHMP Committee for Medicinal Products for Human Use, PRAC Pharmacovigilance Risk Assessment Committee, PASS post-authorisation safety studies, PSURs periodic safety update reports, PSURA periodic safety update report signal assessment

Fig. 1.

Flowchart of analysis unit creation. The 1042 Pharmacovigilance Risk Assessment Committee (PRAC) discussions correspond to 1042 procedures/activities (or agenda items). When removing PRAC procedures/activities (agenda items) excluded from this study (they are listed in the supplement), there are 459 procedures/activities (agenda items) included in the study relating to 321 unique drug-adverse event (AE) pairs. After linking procedures/activities we could identify 413 evaluations (from when a drug-AE pair is first mentioned under one procedure/activity until PRAC concludes [e.g. because specifying an additional assessment]). If then at a later point in time new information emerges on the same drug-AE pair, a new evaluation can commence. APIs active pharmaceutical ingredients, EMA European Medicines Agency, INN international non-proprietary name(s), PASS post-authorisation safety studies

Data Collection and Management

From all identified PRAC regulatory procedures and oversight activities involving antidiabetic products, we gathered information on the type of procedure or activity (Table S2 in the ESM). Based on Good Pharmacovigilance Practice (GVP), we categorised them as either ‘signal management procedures’ (Referral, Signal assessment and prioritisation, PSURs, Other safety issues, MA renewal) [6, 14] or risk management oversight activities (RMP, PASS) [15, 19]. We also noted the date of PRAC meetings at which each specific procedure or activity was addressed.

We recorded AEs assessed using Medical Dictionary for Regulatory Affairs Preferred Terms (version 26.1), alongside their corresponding primary System Organ Classes. We used the Medical Dictionary for Regulatory Affairs important medical events list to identify AEs that are medically important regardless of severity [33]. Furthermore, we documented the outcomes of each procedure, categorised as no action needed, recommendation for regulatory action or request for additional information. For procedures/activities recommending regulatory actions, we specified the nature of the actions, including updates to the product information (SmPC updates with/without patient leaflet or label updates), initiation of a referral procedure, revisions to the RMP, requests for a PASS or suspension/withdrawal. A comprehensive summary of information collected from eligible PRAC procedures/activities is provided in Table S3 in the ESM.

The evaluations of an antidiabetic drug-related AE may span several PRAC meetings and be addressed through various types of PRAC procedures and activities, i.e. initiated in one meeting and concluded in another. We combined regulatory procedures and oversight activities for the same antidiabetic drug-related AE into a single evaluation if they shared the same EMA procedure number or unique study identifiers, or if they referred to previous PRAC meetings.

For procedures and activities lacking consistent EMA procedure numbers or study identifiers, we used additional descriptive information mentioned in the minutes, such as reference to study protocol numbers, study design, specific drugs and AEs, to determine whether two subsequent PRAC procedures for a given API belonged to the same evaluation. To reconcile the inconsistencies in study identifiers mentioned in the minutes, we consulted multiple databases, including the European Union Drug Regulating Authorities Clinical Trials Database (EudraCT) [34], the HMA-EMA Catalogue of real-world data studies [35] and ClinicalTrials.gov. [36]

Data Analysis

This study employed descriptive statistics to characterise PRAC’s evaluation of antidiabetic agents. For analysis, we created the following analysis units.

Drug-AE Pair

From PRAC procedures/activities, we identified all unique drug-AE pairs, defined as a combination of an INN or drug class with an AE. A PRAC procedure/activity can address multiple drug-AE pairs. For eligible PASS procedures that did not explicitly mention AEs, we used AEs identified in the PASS study title, if available.

Drug-AE Evaluation

We defined a drug-AE evaluation as the period from the first mention of a drug-AE pair in the PRAC minutes until the associated safety signal or concern was concluded by the PRAC. When a drug-AE pair was discussed multiple times across different PRAC plenary meetings, these instances were linked and treated as a single evaluation. The linking process made use of information, such as drug-AE pairs, EMA procedure numbers and other relevant data as previously described. A single drug-AE evaluation may encompass multiple types of PRAC procedures and activities or repeated occurrences of the same procedure/activity type across several PRAC meetings. Moreover, the same drug-AE pair may be evaluated more than once (in separate sequences that could not be linked). For examples of what is considered as one drug-AE evaluation, refer to Table S4 in the ESM.

Each evaluation was assigned an ID, with sequence numbers given to each procedure/activity in the evaluation based on their appearance in PRAC meeting minutes. The “initial procedure/activity” of a drug-AE evaluation was defined by the procedure/activity type where the drug-AE pair first appeared in the minutes. One drug-AE evaluation might involve multiple recommendations for regulatory actions. For drug-AE evaluations without any recommendations for regulatory action, these were defined as “no regulatory action needed”.

A descriptive analysis was employed to investigate the characteristics of the drug-AE pairs, including drug class, System Organ Class, medical importance and special populations if relevant. We report on the “initial procedure/activity” of the drug-AE evaluations and the recommendations for regulatory actions resulting from each evaluation. We also reported the specific procedure/activity type under which each recommendation was made. For a drug-AE evaluation where the drug-AE pair was addressed in different PRAC procedures/activities, we provided an overview of the sequence of PRAC procedures/activities that were part of the drug-AE evaluation. All analyses were conducted in R (version 4.2.3).

Results

During the study period, antidiabetic drug APIs were mentioned 4276 times in the PRAC minutes, covering 1042 procedures/activities related to AEs. Based on eligibility criteria, 459 PRAC procedures/activities were included, referring to 321 unique drug-AE pairs. By linking the relevant content of each PRAC procedure/activity, 413 drug-AE evaluations were identified, some spanning multiple PRAC meetings (Fig. 1).

Characteristics of Antidiabetic Drug AEs

Among the 321 drug-AE pairs analysed, 14 were associated with drug classes, and 307 were linked to individual INNs (Table 2, Table S5 in the ESM). Of these, AEs related to sodium-glucose transport protein-2 inhibitors (N = 80), dipeptidyl peptidase 4 inhibitors (N = 70) and glucagon-like peptide-1 receptor agonists (N = 90) were the most frequently evaluated. Adverse events were most often related to “skin and subcutaneous tissue disorders” (N = 44), “neoplasms” (N = 42) and “gastrointestinal disorders” (N = 36). Pemphigoid (N = 7) and angioedema (N = 8) were the most frequently discussed skin disorders, primarily associated with second-generation antidiabetic agents. Detailed information under each System Organ Classes is provided in Table S5 in the ESM.

Table 2.

Characteristics of unique drug-AE pairs discussed in the Pharmacovigilance Risk Assessment Committee (2012–22)

	Overall	Insulin and GLP-1RA	SGLT-2i	DPP-4i	GLP-1RA	Insulin	Othera	Biguanides
	N (%)	N (%)	N (%)	N (%)	N (%)	N (%)	N (%)	N (%)
Earliest MA year within drug class		2014	2012	2007	2006	1995	1974	1959
Total number of drug-AE pairs (N)	321	9	80	70	90	25	37	10
AE system organ class
Skin and subcutaneous tissue disorders	44 (13.7)	2 (22.2)	13 (16.2)	12 (17.1)	5 (5.6)	5 (20.0)	5 (13.5)	2 (20.0)
Neoplasms benign, malignant and unspecified (including cysts and polyps)	42 (13.1)		6 (7.5)	7 (10.0)	18 (20.0)	4 (16.0)	7 (18.9)
Gastrointestinal disorders	36 (11.2)	1 (11.1)	5 (6.2)	12 (17.1)	17 (18.9)		1 (2.7)
Metabolism and nutrition disorders	29 (9.0)	1 (11.1)	8 (10.0)	1 (1.4)	6 (6.7)	3 (12.0)	8 (21.6)	2 (20.0)
Renal and urinary disorders	27 (8.4)		13 (16.2)	5 (7.1)	8 (8.9)		1 (2.7)
Hepatobiliary disorders	23 (7.2)	2 (22.2)	3 (3.8)	4 (5.7)	10 (11.1)	1 (4.0)	2 (5.4)	1 (10.0)
Infections and infestations	19 (5.9)		12 (15.0)	3 (4.3)	4 (4.4)
Cardiac disorders	17 (5.3)		1 (1.2)	8 (11.4)	3 (3.3)	2 (8.0)	3 (8.1)
Immune system disorders	13 (4.0)		3 (3.8)	2 (2.9)	4 (4.4)	4 (16.0)
Musculoskeletal and connective tissue disorders	13 (4.0)		4 (5.0)	8 (11.4)			1 (2.7)
General disorders and administration-site conditions	11 (3.4)			1 (1.4)	5 (5.6)	3 (12.0)	2 (5.4)
Investigations	11 (3.4)	3 (33.3)	3 (3.8)	1 (1.4)	4 (4.4)
Other System Organ Classb	36 (11.2)		9 (11.2)	6 (8.6)	6 (6.7)	3 (12.0)	7 (18.9)	5 (50.0)
Important medical event	169 (52.6)	3 (33.3)	44 (55.0)	40 (57.1)	52 (57.8)	8 (32.0)	18 (48.6)	4 (40.0)

AE adverse event, DPP-4i dipeptidyl peptidase 4 inhibitors, GLP-1RA glucagon-like peptide-1 receptor agonist, MA marketing authorisation, SGLT-2i sodium-glucose transport protein-2 inhibitor

^aOther antidiabetic drugs include Thiazolidinediones (ATC: A10BG), Sulfonylureas (ATC: A10BB), Sulfonamides (heterocyclic) (ATC: A10BC) and Other blood glucose-lowering drugs, excluding insulins (ATC: A10BX)

^bOther system organ classes include: Blood and lymphatic system disorders, Ear and labyrinth disorders, Endocrine disorders, Eye disorders, Injury, poisoning and procedural complications, Nervous system disorders, Psychiatric disorders, Reproductive system and breast disorders, Respiratory, thoracic and mediastinal disorders, Social circumstances, Surgical and medical procedures, Vascular disorders

Important medical events comprised 53% of all drug-AE pairs, including diabetic ketoacidosis (N = 8) and pancreatic carcinoma (N = 10). Only 11% (N = 35) of drug-AE pairs concerned specific sub-populations, often involving concomitant drug use (N = 14), comorbidities (N = 8) or elderly patients (N = 8) [data not shown].

Characteristics of Evaluations

A total of 413 drug-AE evaluations were conducted based on the 321 drug-AE pairs (Table 3), with 19% (N = 62) of drug-AE pairs undergoing more than one evaluation during the study period. The canagliflozin-pancreatitis pair was evaluated six times within 6 years as part of the PSUR single assessment, with the third evaluation recommending pancreatitis to be added as an important potential risk in the RMP for all canagliflozin-containing products. In other cases, PSURs could also result in outcomes such as SmPC updates or continued routine monitoring with no regulatory action needed.

Table 3.

PRAC procedures or oversight activities initiating each drug-AE evaluation related to antidiabetic drugs in the PRAC (2012–22)

	Total number of drug-AE evaluations	Signal management (N = 277)					Oversight activities (N = 136)
		PSURs	Signal procedure	Referral	Other safety issue	MA renewal	RMP	PASS
Procedure initiating evaluation, n (%)	413	224 (54.2)	44 (10.7)	2 (0.5)	2 (0.5)	5 (1.2)	81 (19.6)	55 (13.3)
Number of PRAC plenary meetings in one evaluation
1 meeting, n (%)	273	204 (74.7)	10 (3.7)		2 (0.7)	4 (1.5)	37 (13.6)	16 (5.9)
2–4 meetings, n (%)	130	19 (14.6)	31 (23.8)			1 (0.8)	44 (33.8)	35 (26.9)
> 4 meetings, n (%)	10	1 (10.0)	3 (30.0)	2 (20.0)				4 (40.0)
Drug class of drug-AE evaluation
GLP-1RA, n (%)	123	68 (55.3)	16 (13.0)	1 (0.8)		2 (1.6)	21 (17.1)	15 (12.2)
SGLT-2i, n (%)	110	64 (58.2)	5 (4.5)				21 (19.1)	20 (18.2)
DPP-4i, n (%)	84	40 (47.6)	15 (17.9)	1 (1.2)	2 (2.4)	1 (1.2)	13 (15.5)	12 (14.3)
Other antidiabetica, n (%)	47	26 (55.3)	6 (12.8)				9 (19.1)	6 (12.8)
Insulin, n (%)	30	7 (23.3)	2 (6.7)			2 (6.7)	17 (56.7)	2 (6.7)
Biguanides, n (%)	10	10 (100.0)
Insulin and GLP-1RA, n (%)	9	9 (100.0)
Evaluation of important medical event, n (%)
Yes	227	121 (53.3)	24 (10.6)	2 (0.9)	2 (0.9)	2 (0.9)	37 (16.3)	39 (17.2)
No	186	103 (55.4)	20 (10.8)		3 (1.6)		44 (23.7)	16 (8.6)

One drug-AE evaluation included one or more occurrences of the procedures/activities at different PRAC plenary meetings where several assessment rounds are required. The evaluation could occur in the same procedure/activity or transit between different procedures/activities. Each drug-AE pair could have multiple drug-AE evaluations

AE adverse event, DPP-4i dipeptidyl peptidase 4 inhibitors, GLP-1RA glucagon-like peptide-1 receptor agonist, MA renewal renewals of the marketing authorisation, conditional renewal and annual reassessments, Other safety issue other safety issues for discussion requested by the Committee for Medicinal Products for Human Use or the European Medicines Agency, PASS post-authorisation safety studies, PSURs periodic safety update reports, RMP risk management plan, SGLT-2i sodium-glucose transport protein-2 inhibitor, Signal Procedure signal assessment and prioritisation procedure

^aOther antidiabetic drugs include Thiazolidinediones (ATC: A10BG), Sulfonylureas (ATC: A10BB) and Other blood glucose-lowering drugs, excluding insulins (ATC: A10BX)

Two thirds (N = 277) of drug-AE evaluations were initiated with ‘signal management procedures’, with the PSUR initiating 54% (N = 224) of all evaluations, accounting for 81% (224/277) of signal management-initiated evaluations. In contrast, ‘Signal Procedure’ addressed 11% (N = 44) of all of the evaluations, representing 16% (44 of 277) of signal management-initiated evaluations. Additionally, 136 evaluations were related to risk management, with 60% (N = 81) involving RMP and 40% (N = 55) involving PASS. Two thirds (N = 273) of PRAC evaluations were concluded within one meeting, 80% of these related to ‘signal management procedures’ and 39% of these related to risk management activities. Of the evaluations concluded within one PRAC meeting, 75% were related to the PSUR procedure.

Sequence of Between PRAC Procedures/Activities Within Drug-AE Evaluations

Most evaluations (94%) were resolved within the same procedure type, with two thirds discussed in only one PRAC meeting (Table 4). Evaluations involving RMP and PASS more often spanned two meetings compared to those in ‘signal management procedures’. In 26 drug-AE evaluations, the drug-AE pairs were addressed in more than one type of PRAC procedure/activity before finalisation, often extending the evaluation period. This was particularly the case for those involving PASS and referral procedures. The evaluation involving multiple PRAC procedure/activity types most commonly occurred following ‘signal management procedures’, with the PSUR (N = 10) and the ‘Signal Procedure’ (N = 8) being the main triggers. The PRAC conducted five referrals for antidiabetic medications during the study, addressing issues including glucagon-like peptide-1 receptor agonists and dipeptidyl peptidase 4 inhibitors with pancreatitis, metformin with lactic acidosis, and sodium-glucose transport protein-2 inhibitors with diabetic ketoacidosis and limb amputations. One evaluation, focused on saxagliptin and acute kidney injury (see Table S6 in the ESM for details), was initiated by a PASS procedure. Within this evaluation, the appearance of the drug-AE pair alternated between PASS procedures and ‘Signal Procedures’.

Table 4.

Overview of the sequence of PRAC procedure/activity types involved in drug-adverse event evaluations related to antidiabetic drugs in the PRAC (2012–22)

		Evaluations		Plenary meetings (minutes)
		n	(%)	Median	[IQR]
Total		413		1	[1, 2]
Evaluations finalised within the same procedures/activities type		387	93.7%	1	[1, 2]
Signal management procedures	PSURs	214	51.8%	1	[1, 1]
	Signal Procedure	35	8.5%	2	[1, 2]
	MA renewal	2	0.5%	1	[1, 1]
	Other safety issue	5	1.2%	1	[1, 1]
Oversight activities	RMP	79	19.1%	2	[1, 2]
	PASS	52	12.6%	2	[1, 3]
Evaluations across 2 procedure/activity types		22	5.3%	3	[2, 4]
Signal management procedures	PSURs → RMP	8	1.9%	3	[2.75, 3]
	Signal Procedure → PSURs	3	0.7%	2	[2, 3]
	Referral → RMP	2	0.5%	6	[6, 6]
	Signal Procedure → Other safety issue	2	0.5%	2	[2, 2]
	Signal Procedure → PASS	2	0.5%	4.5	[4.25, 4.75]
	PSURs → Other safety issue	1	0.2%	2
Oversight activities	PASS → RMP	2	0.5%	2.5	[2.25, 2.75]
	RMP → PASS	2	0.5%	4	[4, 4]
Evaluations across more than 2 procedure/activity types		4	1.0%	10.5	[6.25, 15.75]
Signal management procedures	PSURs → Referral → RMP → PSURs	1	0.2%	7
	Referral → Signal Procedure → Other safety issue → Referral → RMP → PASS → PASS	1	0.2%	21
	Signal Procedure → Referral → RMP → Other safety issue → RMP → PASS	1	0.2%	14
Oversight activities	PASS → Signal Procedure (parallel with further analysis in PASS) → Signal Procedure	1	0.2%	4

Evaluations finalised within one procedure or activity type could involve one or several occurrences of the same regulatory procedure at different PRAC plenary meetings. Additionally, evaluations may undergo multiple discussions before a transition occurs

IQR interquartile range, Other safety issue other safety issues for discussion requested by the Committee for Medicinal Products for Human Use or the European Medicines Agency, PASS post-authorisation safety study, PRAC Pharmacovigilance Risk Assessment Committee, PSUR periodic safety update report, RMP risk management plan, Signal Procedure signal assessment and prioritisation procedure

Evaluation Outcomes and Recommended Regulatory Actions

Among 413 drug-AE evaluations, nearly half (N = 197, 47.7%) concluded with no regulatory action (Table S7 in the ESM). Of the 216 evaluations that resulted in recommendations for one (N = 204) or more (N = 12) regulatory actions, the most frequent recommendations relate to updates to the SmPC (N = 90) or RMP (N =83) or both (N = 42) (Table 5). Recommendations for additional RMMs or PASS were infrequent, comprising less than 2% of all evaluations (details are listed in Table S8 in the ESM). In three cases, the PRAC evaluation led to a referral; descriptions of these complex cases are provided in Table S6 in the ESM. Finally, no withdrawal recommendation was observed during the study period.

Table 5.

Recommendations for regulatory action(s) of drug-AE evaluations related to antidiabetic drugs in the Pharmacovigilance Risk Assessment Committee (2012–22)

	Total number of evaluation	Total number of recommendation	No regulatory action	SmPC update	RMP update	aRMM	PASS	Referral
Total, n (%)	413	467	197 (42.2)	132 (28.3)	125 (26.8)	8 (1.7)	2 (0.4)	3 (0.6)
Drug class of INN-AE evaluation
GLP-1RA, n (%)	123	136	57 (41.9)	48 (35.3)	29 (21.3)		2 (1.5)
SGLT-2i, n (%)	110	124	58 (46.8)	28 (22.6)	31 (25.0)	5 (4.0)		2 (1.6)
DPP-4i, n (%)	84	95	34 (35.8)	28 (29.5)	33 (34.7)
Other antidiabetic druga, n (%)	47	57	29 (50.9)	15 (26.3)	13 (22.8)
Insulin, n (%)	30	33	8 (24.2)	5 (15.2)	18 (54.5)	2 (6.1)
Biguanides, n (%)	10	13	9 (69.2)	1 (7.7)	1 (7.7)	1 (7.7)		1 (7.7)
Insulin and GLP-1RA, n (%)	9	9	2 (22.2)	7 (77.8)

An evaluation could conclude single or concurrent regulatory action recommendations. For evaluations with concurrent recommendations, we counted each type of regulatory action separately. For example, if an evaluation recommended both an RMP update and an SmPC update, we counted one instance under ‘RMP update’ and one under ‘SmPC update’. If no regulatory action was identified, the evaluation was classified as ‘no regulatory action’

The total number of recommendations is larger than the evaluation. The percentage uses the total number of evaluations as the denominator. Therefore, the sum of the percentage is over 100%.

AE adverse event, aRMM additional risk minimisation measures, DPP-4i dipeptidyl peptidase 4 inhibitors, GLP-1RA glucagon-like peptide-1 receptor agonist, INN international non-proprietary name(s), PASS post-authorisation safety studies, RMP risk management plan, SGLT-2i sodium-glucose transport protein-2 inhibitor, SmPC summary of product characteristics

^aOther antidiabetic drugs include Thiazolidinediones (ATC: A10BG), Sulfonylureas (ATC: A10BB), and Other blood glucose-lowering drugs, excluding insulins (ATC: A10BX)

Figure 2 highlights the 401 evaluations where recommendation(s) for regulatory action were given in connection to one type of PRAC procedure/activity or resulted in ‘no regulatory action’. Over half (55%, N = 220) of these outcomes originated from PSURs, with SmPC updates as the primary regulatory action. However, the majority of PUSR evaluations resulted in no recommendation for regulatory action (138/220). Other major sources included RMP (N = 84) and ‘Signal Procedure’ (N = 35), the latter frequently recommending SmPC updates (36%). Risk management plan updates predominantly originated in RMP sections (85%), either alone (N = 46) or in combination with other updates, such as the SmPC (N = 22). All referral recommendations came from evaluations with multiple conclusions (see details in Table S9 in the ESM).

Fig. 2.

Distribution of recommendation for regulatory action and the Pharmacovigilance Risk Assessment Committee procedure type in which the recommendation was given, for 401 drug-AE evaluations the recommendations were only provided in a single meeting (and not scattered across several meetings) or where the evaluation did not result in a recommendation for regulatory actions. This figure includes 401 evaluations where recommendation(s) for regulatory action were given in connection to one Pharmacovigilance Risk Assessment Committee procedure/activity type or resulted in ‘no regulatory action’. The x-axis listed the signal management procedures or oversight activities in which the regulatory recommendations were given. The y-axis represents the number of evaluations. Different colours correspond to distinct recommendations. aRMM additional risk minimisation measures, MA renewal renewals of the marketing authorisation, conditional renewal and annual reassessments, Other safety issue other safety issues for discussion requested by the Committee for Medicinal Products for Human Use or the European Medicines Agency, PASS post-authorisation safety studies, PSURs periodic safety update reports, RMP risk management plan, SmPC summary of product characteristics

Discussion

This quantitative study analysed the process of post-authorisation safety assessment within the PRAC, focusing on AEs evaluated by the PRAC for antidiabetic medications over 10.5 years. From when the PRAC commenced its work in July 2012 until the end of 2022, 413 drug-AE evaluations were identified, corresponding to 321 unique drug-AE pairs. Most evaluations were finalised within a single procedure or activity, though the use of multiple processes/activities within one evaluation was observed. The PSURs emerged as the absolute most prominent procedure for PRAC to assess post-authorisation safety, frequently resulting in SmPC updates and recommendations to update the RMP.

The PSUR is a tool for continuous post-authorisation assessment of a medicine’s risks and benefits during routine medical practice and long-term use [14]. Our analysis found that PSURs addressed a greater number of AEs than the ‘Signal Procedure’, which is surprising, as “A PSUR is not intended, in the first instance, for notification of significant new safety or efficacy information or to provide the means by which new safety issues are detected.” (GVP IIV, Section VII.B.1.) [14]. However, this finding might reflect that while the PSURs are not the appropriate tool for urgent safety issues, such as those with a potentially major impact on the risk-benefit balance, patients’ or public health, and which require urgent regulatory action, new signals may also be included in PSURs. This occurs if MA holders complete their assessment of a signal detected through continuous monitoring of EudraVigilance within 6 months before the PSUR submission [6]. Additionally, as newer antidiabetic drugs are the primary source of safety signals and PSURs are more frequently required for newer medicines, the high representation of newer antidiabetic drugs may also contribute to the higher number of drug-AE pairs addressed in PSURs.

In addition, this study did not assess whether the signals were already reflected in the SmPCs at the time of a PRAC assessment, limiting our ability to determine if the signals in PSURs were new or updates to known AEs. Additionally, we did not examine which sections of the SmPC were recommended for updates as a result of PRAC safety evaluations. Our finding that PSURs frequently result in PRAC-recommended updates to the SmPC aligns with those of Ebbers et al., who explored the origin of safety-related type II variation for biopharmaceuticals prior to the 2012 implementation of the EU Pharmacovigilance Regulation [25]. While our findings aligned with those of Ebbers et al. regarding the role of PUSRs in addressing non-urgent safety issues, it is important to note that their study focused on DHPC and safety-related type II variations, conducted during the pre-PRAC period. In contrast, our study focused on PRAC’s work during the post-marketing phase. Furthermore, although most type II variations are led by the Committee for Medicinal Products for Human Use, the PRAC leads the “type II variations concerning clinical safety to update the product information and/or the Risk Management Plan, as a follow-up to a previous PSUR procedure or following a previous PRAC assessment of a signal” [37].

A high number of signals were observed in sodium-glucose transport protein-2 inhibitors, dipeptidyl peptidase 4 inhibitors and glucagon-like peptide-1 receptor agonists, compared with other antidiabetic agents. This likely reflects the more intensive monitoring of new products, including more frequent PSUR submissions [1, 29]. Nevertheless, PSURs appear to be an important tool for both older and newer drug classes, as the majority of safety signals for antidiabetic medicines (except for insulin) discussed by the PRAC emerged in connection to the PSUR procedure. This underscores PSUR’s role in the continuous evaluation of the risk-benefit balance of medicinal products [1, 23, 25, 38], and its potential use as a tool for signal prioritisation and evaluation, even though this is not its primary purpose [14].

Consistent with previous studies, our findings indicate that SmPC updates are the most common regulatory action recommended by the ‘Signal Procedure’ [7, 9, 10]. Of the 35 recommendations originating from the ‘Signal Procedure’, 13 were related to SmPC updates, including nine stand-alone SmPC updates and four combined with RMP updates (Fig. 2). Consequently, approximately 37% of conclusions from the ‘Signal Procedure’ involved SmPC updates, aligning with the findings of Farcaş et al. (37%) [9]. However, our result is lower than in other studies [7, 10], likely because of differing definitions. While those studies combined patient leaflet changes and SmPC updates under the broader category of product information updates, our study defined SmPC updates when PRAC meeting minutes explicitly documented the amendment of product information without or without mention of corresponding updates to the package leaflet. Our findings also support previous research identifying DHPCs as the most prevalent additional RMMs [39], prior to the GVP module XVI, revision 3 re-categorised DHPCs as communication tools rather than additional RMM. Finally, it is important to note that PRAC’s recommendations are advisory; it is the Committee for Medicinal Products for Human Use that reviews the new knowledge related to the risk-benefit balance and enforces actions to implement these recommendations [40].

To our knowledge, this is the first study to investigate how the PRAC procedures/activities work as a system within a specific therapeutic area. While it remains uncertain whether the antidiabetic drug field adequately represents the PRAC’s activities across other therapeutic areas, our findings generally align with those of Potts et al. [7], which gave an overview of PRAC evaluations across all therapeutic areas. Moreover, we lack information on whether drug-AE pairs were included in the RMP at MA or added to safety profiles before 2012. This limits our ability to assess how many drug-AE evaluations with PRAC’s no-action recommendations were already included and monitored based on available evidence.

Despite covering a comprehensive period, our study is limited by its follow-up time, especially regarding signals detected toward the end of the study. For example, a potential risk of tubulointerstitial nephritis linked to the interaction between empagliflozin and lithium was noted in the November 2021 PRAC meeting minutes, but further information was unavailable at the study’s conclusion, preventing monitoring of the progress. In addition, the sequence of procedures/activities within a given drug-AE evaluation was defined when the drug-AE pair appeared under a given procedure/activity in the PRAC meeting minutes. Relying only on what is documented in the publicly available PRAC meeting minutes does not necessarily provide a full understanding of how a drug-AE pair is assessed.

Given that the PRAC meeting minutes are a summary intended for laypersons, our study faces limitations arising from the ambiguity of information presented in these minutes. Additionally, the lack of historical records of known AEs in the SmPC hindered a detailed data collection on PASS and PSUR updates. The absence of historical RMP files on the EMA website further complicated data collection. Furthermore, linking procedures/activities of one evaluation across different PRAC procedures/activities and meeting minutes is challenging because of varying EMA procedure numbers for the same case, inconsistent study identifiers or non-functional hyperlinks. Identifying PASS objectives was challenging, as study identifiers were typically provided as protocol codes, which made it difficult to link to relevant databases (such as the HMA-EMA Catalogue of real-world data studies). As transparency was a legislative goal implemented in 2011, further studies are needed to comprehensively analyse transparency issues.

Conclusions

This study examines how the PRAC evaluates safety signals and concerns across different procedures and activities in the context of antidiabetic drugs. Periodic safety update reports are the main source of safety signal evaluation used by PRAC for post-marketing signal management. These evaluations frequently result in SmPC update recommendations. Compared with PSURs, signal assessment and prioritisation procedures generate a wider range of regulatory actions, including a variety of RMMs and referrals. This difference may suggest that PSURs are not intended for urgent safety issues.

Supplementary Information

Below is the link to the electronic supplementary material.

Declarations

Funding

Open access funding provided by Copenhagen University. No external funding was received for the preparation of this article.

Conflicts of Interest/Competing Interests

Christine E. Hallgreen is an employee of the Copenhagen Centre for Regulatory Sciences (CORS), Department of Pharmacy, University of Copenhagen. CORS is a cross-faculty university-anchored institution involving various public (Danish Medicines Agency, Copenhagen University) and private stakeholders (Novo Nordisk, Lundbeck, Ferring Pharmaceuticals, LEO Pharma) as well as patient organisations (Rare Diseases Denmark). The centre is purely devoted to the scientific aspects of the regulatory field and has a patient-oriented focus, and the research is not related to a company-specific product or directly company related. In the last 3 years, CORS and Christine E. Hallgreen have received funding from Novo Nordisk for projects not related to this study. Per Sindahl is employed by the Danish Medicines Agency and a guest researcher at CORS. At the time of the project, Per Sindahl was also a PhD fellow at the Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences, Faculty of Science, Utrecht University. Morten Andersen is an employee at the Department of Drug Design and Pharmacology, University of Copenhagen. His Professorship and the Pharmacovigilance Research Centre were funded by a grant from the Novo Nordisk Foundation to the University during 2016–22. He has previously participated in research projects funded by AstraZeneca, H. Lundbeck & Mertz, Janssen, Novartis, Merck Sharp & Dohme and Pfizer, with grants paid to the institutions where he was employed, and personally has received fees from Atrium, the Danish Pharmaceutical Industry Association, for leading and teaching pharmacoepidemiology courses. Haoxin Le is a PhD student at CORS, and has no conflicts of interest that are directly relevant to the content of this article. The authors declare no other known competing financial interests or personal relationships that could have appeared to influence the work reported in this article.

Ethics Approval

Not applicable.

Consent to Participate

Not applicable.

Consent for Publication

Not applicable.

Availability of Data and Material

Data used for this article are available at 10.5281/zenodo.13991282.

Code Availability

Not applicable.

Authors’ Contributions

HL: writing (review and editing), writing (original draft), visualisation, methodology, investigation, formal analysis, data curation, conceptualisation. PS: writing (review and editing), investigation. MA: writing (review and editing), conceptualisation, supervision. CEH: writing (review and editing), validation, supervision, resources, project administration, methodology, investigation, conceptualisation. All authors have read and approved the final version.