Naming and managing the safety concerns
Safety concerns are a defined set: the important identified risks, the important potential risks and the missing information. One set of definitions, with a test for what makes a risk identified, potential or important and for what counts as missing information, serves the development safety update report (DSUR), the risk management plan (RMP) and the periodic safety update report (PSUR) alike. The second revision of good pharmacovigilance practice (GVP) Module V redefined the set for the RMP alone, tying it to the need for further characterisation and management, so a concern can leave the plan while the PSUR still carries it, and companies need a strategy for the safety concerns across the regions a global report goes to.
What a safety concern is
Safety concerns are a defined set of important identified risks, important potential risks and missing information, a concept already part of Volume 9A, the EU pharmacovigilance guideline that the GVP guidelines replaced in 2012. They are defined during clinical development and carried forward into the post-marketing phase. The definitions are set out in an appendix to the GVP guidance. Although they are well defined there, in practice safety concerns are frequently the subject of interactions with health authorities.
Across the life cycle, the DSUR carries the important identified and potential risks, while the RMP and the periodic benefit-risk evaluation report carry the important risks and the missing information. The safety concerns are central to all three documents and have become major drivers of their content and of the resources needed to write them. Before 2012 safety concerns had very little impact on the writing and management of pharmacovigilance documents; that changed with the introduction of the DSUR in 2011 and of the GVP modules on the RMP and the PSUR in 2012. GVP Module V describes how the safety concerns are managed and supplies the RMP template, and the RMP is the document in which the writer sets out what will be done about them.
The three documents answer four questions in order: what is known about the medicine's safety, which are its risks; what is unknown about its safety that could affect the benefit-risk balance, which is its missing information; how the risks can be managed or minimised; and how acceptable the risks are in the context of the benefit. The benefit-risk balance needs to remain positive, but a medicine with serious risks can still be acceptable in the context of the underlying disease it is meant to cure.
The definitions applied alike to the DSUR, the RMP and the PSUR, and before 2018 the safety concerns presented in the RMP were identical to those presented in the PSUR; the second revision of GVP Module V, released in 2017, brought a new definition of safety concerns that applies only to the RMP. The change caused much confusion for the teams and writers who work on the three documents. The second revision of GVP Module V gives more specific wording and clarifications for the definitions of identified and potential risks, missing information and important risk, and adds guidance for a pragmatic approach to applying them.
| Term | GVP Module V Revision 1 | GVP Module V Revision 2 |
|---|---|---|
| Identified risk | An untoward occurrence for which there is adequate evidence of an association with the medicinal product of interest | Undesirable clinical outcomes for which there is sufficient scientific evidence that they are caused by the medicinal product |
| Potential risk | An untoward occurrence for which there is some basis for suspicion of an association with the medicinal product of interest but where this association has not been confirmed | Undesirable clinical outcomes for which there is scientific evidence to suspect the possibility of a causal relationship with the medicinal product, but where there is currently insufficient evidence to conclude that this association is causal |
| Important identified risk and important potential risk | An identified risk or potential risk that could have an impact on the benefit-risk balance of the product or have implications for public health | The RMP should focus on the important identified risks that are likely to have an impact on the risk-benefit balance of the product. An important identified risk to be included in the RMP would usually warrant:● Further evaluation as part of the pharmacovigilance plan (e.g. to investigate frequency, severity, seriousness and outcome of this risk under normal conditions of use, which populations are particularly at risk);● Risk minimisation activities: product information advising on specific clinical actions to be taken to minimise the risk, or additional risk minimisation activities.The important potential risks to be included in the RMP are those important potential risks that, when further characterised and if confirmed, would have an impact on the risk-benefit balance of the medicinal product |
| Missing information | Gaps in knowledge about a medicinal product, related to safety or use in particular patient populations, which could be clinically significant | Gaps in knowledge about the safety of a medicinal product for certain anticipated utilisation (e.g. long-term use) or for use in particular patient populations, for which there is insufficient knowledge to determine whether the safety profile differs from that characterised so far |
Identified and potential risks
An identified risk is an untoward occurrence for which there is adequate evidence of an association with the medicinal product, such as a side effect listed in the product information. To be called identified, a risk will likely have been demonstrated in non-clinical studies and later confirmed by clinical data, observed in well-designed clinical trials that show a causal relationship, or documented after marketing by a great number of spontaneous reports in which causality is strongly reported.
A potential risk is an untoward occurrence for which there is some basis for suspicion of an association with the medicinal product but where the association has not been scientifically confirmed. Examples could be a non-clinical safety finding that has not been observed or resolved in clinical studies; an adverse event seen in clinical trials whose data do not suggest a causal relationship; a signal arising from spontaneous reporting in post-marketing surveillance; or an event known to be associated with other products of the same class, or scientifically expected from the nature of the product.
For the RMP, the second revision clarifies that risks should be identified through adverse clinical outcomes that are caused by use of the product, for identified risks, or that might be caused by it, for potential risks. If off-label use in children, for example, is considered an important potential risk, the adverse clinical event potentially associated with it should be defined.
What makes a risk important
Importance is a defined term, and only an important risk is a safety concern. An important risk is an identified or potential risk that has or could have an impact on the benefit-risk balance of the product or could have implications for public health. By one rule of thumb, a risk affects the benefit-risk balance if the product is no longer given to the patient when the risk is too high and so outweighs the benefit, which covers anything for which the doctor withdraws the medicine when the patient complains of the side effect.
Impact on public health is probably a rather abstract criterion, and in one experienced writer's account it seldom comes up. Its definition turns on absolute risk in relation to the size of the target population, and an example might be anything consumed by very many patients, such as an over-the-counter medicine with frequent side effects that many patients then need treated, or likely a vaccine. That writer had yet to come across a product whose important risk rested on impact on public health.
Many factors should be accounted for when deciding whether a risk is important: its impact on the individual and on public health, its medical seriousness, its preventability and its reversibility. A risk that appears as a special warning or a contraindication in the product information may also be important. Product information carries a great many special warnings and contraindications, so teams have to consider carefully which of them to promote to safety concerns, and GVP Module V gives examples. A contraindication in patients with another disease, such as severe renal or hepatic impairment, usually qualifies, and off-label use may also be an important risk. An account of signal management goes further, holding, with the ICH guidance on the periodic benefit-risk evaluation report as its reference, that any risk included in the contraindications or in the warnings and precautions should be considered important.
The DSUR is about significant safety issues rather than any safety finding, so only important risks should be included in it, and the writer has to ask whether the risks can really be considered important. A safety result relevant to a first-in-human study, such as one from non-clinical studies, is presented among the safety concerns only if the risk is important.
For the RMP, the second revision links the importance of a risk not only to its impact on the benefit-risk balance but also to the need for further characterisation and management: for important identified and potential risks, the need for further evaluation through pharmacovigilance activities, and for important identified risks, the need for management through risk minimisation measures. In the account of one who was closely involved in the revision, the concept of an important risk did not change with it, but it was revisited and explained in line with the principles of the International Council for Harmonisation (ICH), and the revised guidance clarifies the focus of the plan with regard to safety concerns, strengthening a risk-proportionate approach. As the first revision already specified, the RMP should still focus on the risks relevant to the benefit-risk balance of the product. It should include only the risks that have an impact on the benefit-risk balance and that would usually warrant further evaluation in the pharmacovigilance plan, risk minimisation activities, or both; the plan is now a more risk-proportionate document that focuses only on the safety concerns that need further characterisation and management. Among the important risks, those that could affect the benefit-risk balance, a subset may now be considered more important, for example the risks that need further characterisation through additional pharmacovigilance activities or management through additional risk minimisation measures.
One explanation separates four terms by analogy with crime:
- Signals are petty thieves who want to be criminals, although very few succeed.
- Adverse drug reactions have adequate evidence to be sentenced for a causal relationship, sometimes only on circumstantial evidence.
- Risks are the damage the reaction causes, such as a broken window or a broken nose.
- Important risks are more severe vandalism, where probation, a rehabilitation programme or both are deemed necessary.
A risk is important if its clinical outcome, such as an infection caused by the adverse drug reaction neutropenia, is a clinically relevant change severe enough to affect the benefit-risk balance.
Two rules of thumb carry these principles into a plan:
- Discussing the benefit-risk impact in Module SVII.3 clarifies whether a risk is important.
- A safety concern is either important enough for additional risk management activities or not important enough to be included in the RMP.
Missing information
Missing information is a critical or significant gap in knowledge about the medicinal product related to safety, or about its use in particular patient populations not investigated in clinical trials, where the gap could be clinically significant. Such populations, likely to be exposed within the target population but usually not investigated in clinical trials, include pregnant women and patients with hepatic impairment. Whether a population excluded from the clinical trials becomes missing information is a matter of scientific and medical judgement, and not every one does.
For the RMP, the second revision puts the focus on a potentially different safety profile in certain situations or populations compared with the known safety profile. Including missing information in the RMP now needs a scientific rationale, so an absence of data alone, as for pregnant women, does not automatically make a population a potential safety concern.
A safety concern removed from the plan
The second revision of GVP Module V puts weight on the evidence that supports identifying important risks and missing information, and Module SVII of the RMP now includes sections to discuss the evidence for defining, reclassifying or removing safety concerns. Under the principle of risk proportionality, knowledge about a product's safety profile is expected to increase across its life cycle, and the list of safety concerns to change with it as a causal association with the product is confirmed or refuted; the pharmacovigilance activities and risk minimisation measures can change too, as when studies are newly planned or completed or when risk minimisation measures are integrated in clinical practice or shown to be ineffective. The revision gives guidance on the post-authorisation removal of safety concerns and encourages holders to revise the list of concerns and the associated pharmacovigilance activities and risk minimisation measures critically during the post-marketing phase.
When knowledge of the product's safety increases and pharmacovigilance activities or additional risk minimisation measures are no longer needed, safety concerns might be removed from the plan or reclassified, and the revision sets out these changes for each kind of safety concern. An important identified or potential risk is removed from the plan when no further evaluation is needed or possible in the pharmacovigilance plan and its risk minimisation measures have become part of established clinical practice or no additional measures are needed. An important potential risk is reclassified as an important identified risk when a causal association is confirmed, and removed from the plan when one is rejected. Missing information is removed when sufficient new data are available, or when there is no reasonable expectation of further characterisation through pharmacovigilance activities. A lack of data over time might itself be a reason for removing important potential risks and missing information. An identified risk with no further evaluation in the pharmacovigilance plan or additional risk minimisation activities could now be removed, and a potential risk with no additional studies or other activities to investigate whether it is actually an identified risk would also be removed. An important potential risk to be included in the RMP usually requires further evaluation as part of the pharmacovigilance plan, which is part III of the RMP.
In one example, an RMP carries diarrhoea, severe skin reactions, keratitis and pancreatitis as safety concerns, and because the risks are well characterised and adequately reflected in the summary of product characteristics (SmPC), their removal from the RMP can be suggested, for instance during a renewal assessment procedure. Its missing information is the paediatric population under 18 years and patients with severe renal impairment, and the SmPC has adequate wording on them, for example that there is no relevant use in the authorised indication, so that the potential for these patients to receive the product is very low. Where the SmPC covers them, as it covers renal impairment, they can be removed as missing information.
Since the second revision of GVP Module V, assessors have asked some marketing authorisation holders to revise their list of safety concerns accordingly, and other holders have proposed removals themselves, for instance when submitting the PSUR. The revision led to a well-received reduction in the number of safety concerns presented in plans and in the workload of writing, updating and assessing them. In 2020 feedback from the European Medicines Agency (EMA) was inconsistent: sometimes safety concerns were removed without hesitation, and sometimes the agency asked for concerns to remain in the plan although no additional pharmacovigilance or risk minimisation activities were attached to them. Many plans then still carried safety concerns that did not strictly meet the revised approach, which posed issues in particular for companies writing plans for generic products. Holders and assessors might need further training to ensure that only risks with additional pharmacovigilance or risk minimisation activities are included. The revised guidance as generally understood, and individual opinions shared by members of the Pharmacovigilance Risk Assessment Committee, clearly point to a critical review of the list of safety concerns during a medicinal product's life cycle, but in 2017 it remained open whether holders will consider the available evidence sufficient for such a review and whether assessors will agree with the changes proposed. Once implemented, the clarifications about safety concerns will hopefully lead to a smoother update process and, in the long run, fewer important risks that have to be managed in the plan.
Whether the periodic report keeps it
When a holder is asked to remove safety concerns from the EU RMP, the question arises how to handle them in the PSUR, in the DSUR and in the countries outside the EU that receive the RMP, since a global company usually uses the EU RMP for other countries as well. The RMP and the PSUR ask different questions: of the RMP, whether there are activities to investigate or minimise the risk; of the PSUR, whether the risk has any impact on the product's benefit-risk balance. The safety concerns of the PSUR are still defined as they were when GVP was introduced in 2012, in the GVP annex of definitions, as the risks that affect the benefit-risk balance. So simply removing an important risk from the plan does not justify removing it from the report.
Even when the EMA itself asks for a concern to be taken out of the EU RMP, it should stay in the PSUR unless the agency specifically asks for it to be removed from the PSUR as well. The PSUR is the instrument for communicating with the authority about how the safety concerns will be managed in future, and one experienced writer reported in 2020 that a document the EMA had just released confirmed the approach of keeping such a concern in the PSUR. It may help to discuss the strategy for managing the safety concerns in the PSUR with colleagues in the team, because the same PSUR is used in other countries too. The agency might ask for an important identified risk removed from the plan to be kept in the report, either as a monitoring topic or in section 16 as an important risk, while in some cases routine pharmacovigilance without inclusion in the report could be enough. If new findings later triggered a re-evaluation and recategorisation of such risks, the plan would be updated.
Safety concerns are not to be removed from a PSUR lightly. As a general rule, key risks are not removed. Key risk is not a defined term, but teams know what a product's key risks are, such as bleeding for an anticoagulant, and such a risk is likely not to be removed from the PSUR. In 2020 companies' experience of removing safety concerns from the PSUR still differed, and assessors in different countries did not give consistent feedback; removing missing information that has no impact on the benefit-risk balance was considered fine, with some positive experience.
Where the change goes in the PSUR
The PSUR presents the post-marketing evaluation of the safety concerns in section 16. The handling of a concern removed from the RMP runs through three sections of the PSUR and turns on the reporting period, because the PSUR is a periodic document. Section 16.1 presents the safety concerns at baseline, meaning as they stood at the beginning of the reporting period, so a change made to the EU RMP after that date removes nothing from it. Section 16.3 is where new information collected during the interval goes, and there the writer may note that a concern was removed, or give the reasons for proposing its removal. Section 16.4 carries the evaluation of the safety concerns, and a concern stays there until the agency agrees to its removal, with the rationale for proposing it or, where the health authority asked for it, a statement that the concern will not be presented in the following PSURs. Once the agency has agreed, the following PSUR need not present the concern at all, unless the timing means it is still among the baseline concerns in 16.1, in which case it appears there but no longer in 16.4.
Missing information removed from the EU RMP after the start of the reporting period stays in 16.1 for that period. A concern that some other country, rather than the EMA, asked to have removed stays in 16.4. A concern reclassified during the period, such as a potential risk that became identified, is presented in 16.1 and 16.4 as it stood at baseline, with the reason for the change, and if it has stopped being important it is removed in the next PSUR. A risk that is no longer classified as important is unlikely to affect the benefit-risk balance, so its removal from the PSUR can then be proposed.
Where approval of the RMP takes longer than the PSUR period, the change goes into the PSUR at the approval stage rather than at submission, because the agency may always disagree. Changing anything in the PSUR before the EMA has approved the change is not recommended.
One set of safety concerns across the regions
The RMP is an EU document and the PSUR is a global one: a single PSUR is written and submitted in the EU and may also go to the US Food and Drug Administration (FDA) and to many other countries. So the change in the EU definition of safety concerns for the RMP raises questions for global document management, such as how to handle different local versions and whether removing a concern from the RMP means removing it directly from the PSUR. The revised definition has implications not only for the other pharmacovigilance documents but also for the management of safety concerns worldwide. Its more stringent approach may leave the list of safety concerns in an EU RMP differing from the list in the global PSUR. Removing, adding or reclassifying a safety concern in the EU RMP under the revised guidance affects the RMPs submitted to other countries and the benefit-risk profile in the PSUR, so an authoring team should carefully consider the potential impact on other documents when making the change.
The EU RMP refers to the safety concerns approved by one health authority and describes the risk minimisation measures in the local product information, so copying its Module SVII into section 16 of a global PSUR is not enough. The report should also include safety concerns defined by authorities outside the EU, and where countries list the same risk under different categories, identified in one and potential in another, the deviation has to be taken into consideration and described. A holder may also keep its own list representing its view of the product's benefit-risk profile worldwide. Companies need strategies for managing the safety concerns across regional and global pharmacovigilance documents.
Where risks differ between regions, the ICH PSUR guidance gives an example, and it is good to keep the PSUR global, presenting every risk relevant in any country it is submitted to. The summary table of safety concerns then keeps the overview, and its footnotes say which risk applies where, which keeps the overview clear for the local health authorities.
A company should have a global PSUR strategy, ideally held by a function with oversight of where PSURs are submitted and required and of what the company's local entities need. That oversight minimises the number of PSURs and of variant versions submitted to local health authorities. The main benefit of such a strategy is identical safety concerns in the PSUR everywhere. A local health authority may always want a different set, though one experienced writer had not met such a case, so the same set should be possible in all PSURs, at least in the countries that accept the ICH format of the periodic benefit-risk evaluation report. One way to use the same PSUR in many countries is to harmonise the international birth date with local birth dates proactively: where a product is first approved in the EU and six months later in another country, the company approaches that country's authority to align the data lock point of its PSUR with the international birth date.
Guidance on safety concerns in the periodic report
Since the second revision of Module V there have been outstanding questions about how safety concerns are expected to evolve in a PSUR when they are revised in, or removed from, the RMP. In September 2020 the periodic-report guidance, GVP Module VII, had not been updated since the second revision of Module V, which left an apparent disconnect between the criteria that apply to the two documents and no clear guidance on managing safety concerns between them. The update of Module VII that the agency announced when it issued the revision of Module V had been put on hold, while holders and agencies were in active dialogue about the re-categorisation of safety concerns and the interactions between RMPs and PSURs. The experience gained since the revision of Module V should provide a good basis for revising the periodic-report guidance. In early 2020 the revision of Module VII, though still announced, had been delayed a couple of times, with no indication of when it would be released.
There was some ICH and EU guidance on presenting regional deviations in the list of safety concerns in a PSUR, but none that said unambiguously how to categorise in the report a concern removed from the RMP: as a monitoring topic, as a risk not considered important, or as an important risk. In early 2020 one experienced writer thought that an explanation document the EMA had just released already covered the main topics of the revision of Module VII, and hoped it would bring some clarity on how to manage in PSURs the safety concerns removed from the EU RMP. Whether the EU list really should serve as the minimum for global reports, as the explanatory note to the periodic-report module indicates, was left open. These questions are practical rather than philosophical: they arise from holders' actual interactions with authorities, their later PSUR submissions and possible RMP updates.
One experienced writer expected the revision of Module VII to settle on the concept that a safety concern stays in the PSUR if it has an impact on the benefit-risk balance, with the assessor usually making that call, at least in the EU, and expected the EMA to remain hesitant about removing safety concerns from PSURs. An RMP may then have no safety concerns at all, because none of them carry additional activities, while the PSUR still carries several.