Skip to main content

Reference Safety Information (RSI) for Clinical Trials- Part III

Posted by: , and , Posted on: - Categories: Compliance matters, Good clinical practice


Since RSI blog posts part I and II were released, the MHRA GCP inspectorate has continued to see non-compliance in this key aspect of pharmacovigilance. Whilst additional guidance was released by the Heads of Medicines Agencies, Clinical Trials Facilitation and co-ordination Group (CTFG), we are still seeing unreported suspected unexpected serious adverse reactions (SUSARs) and absence of adequate risk mitigation measures due to incorrect use of the RSI.

The MHRA GCP inspectorate completed a pilot of office-based inspections (OBI) focusing on clinical trials’ pharmacovigilance and the RSI. Two organisations were selected for the pilot: we identified critical findings in both. The RSI OBIs have now been incorporated into our routine inspection programme.

Since 1 January 2019, the MHRA GCP inspectorate has identified 8 organisations with critical findings when reviewing RSI on inspection. Given the number of non-compliances observed, we thought it would be helpful to outline continuing issues and how compliance can be improved.

It is important that organisations understand that the CTFG Q&A on RSI should be read in conjunction with CT-1, CT-3 and the national legislative requirements in member states. In the UK this is currently the SI 2004/1031 (as amended).

It is important to note that although the transition period ended on 31 December 2020, the UK will still accept the provisions in CT-1 and CT-3 guidance in so far as they relate to the UK as a sovereign regulator. Therefore, it is our expectation that they are followed as this is our interpretation as to how the legislative requirements can be met and this will help in terms of harmonisation and running global trials.

What is the RSI?

As a starting point, it is worth remembering for clinical trials what the RSI is and what it is used for. The purpose of the reporting obligation to national competent authorities (NCAs) is to:

  • make them aware of SUSARs
  • collect information on the safety profile of an investigational medicinal product (IMP)
  • take appropriate actions to protect the safety of the trial participants

The RSI is used for determining the expectedness of a Serious Adverse Reaction (SAR). If the serious event is considered related to the IMP and the serious reaction is not included in the RSI, then this becomes a SUSAR and must be reported to the MHRA (and Research Ethics Committee for cases originating in the UK) as per statutory timelines. Inclusion of a SAR as an expected event in the RSI needs to be accompanied by an assessment of the benefit-risk profile of the IMP and ongoing trials. It is important to decide whether the existing measures are sufficient to mitigate the risk of the occurrence of that reaction or if additional actions need to be implemented (including changes to the eligibility criteria, additional safety monitoring, drug discontinuation criteria). Upon receipt of an RSI, the competent authorities need to decide whether the list of expected reactions and associated risk minimisation measures are acceptable for each trial.

The RSI is not used for investigator decision-making in clinical trials. It is used by the sponsor for identifying and reporting SUSARs and by the competent authority to determine the continuing risk-benefit of the trial. Therefore, the MHRA must approve the RSI and any changes via a substantial amendment, so that the trial participants’ safety is assured. By not informing the NCA of RSI changes, assessors are prevented from making an informed decision about the clinical trial authorisation (for example, if this can continue or if any additional changes are required to protect trial participants).  The expectedness assessment is not a medical decision. The RSI should use MedDRA Preferred Terms (PT) which can be compared with the MedDRA PT the SAR has been coded to. Often, this can be automated by programming expected terms into the safety database, which will then auto-label the event based on a pre-specified list. However, the decision to include a SAR in the RSI section is based on medical assessment. There must be a good rationale why a SAR is expected (for example, biological, plausibility, temporal association) and appropriate risk mitigation measures have to be in place. This rationale should be documented and provided in the covering letter to the MHRA when there is a substantial change to the RSI.

Common findings

Below is a list of common findings which have been identified on inspection and tips on how to avoid them:

CTFG Q&A Impact Assessment

Organisations had until 1 January 2019 to ensure that their procedures and RSIs - for example, Investigator Brochures (IBs) - were in compliance with the CTFG Q&A on RSI. We have seen delays in the implementation of this. Whilst an impact assessment on the quality system is often performed, there is a lack of impact assessment on the safety profile of the IMP and whether there have been any unreported SUSARs as a result. It is expected that this retrospective review is performed. When reviewing SARs during this impact assessment, organisations should use the RSI that had been approved by the MHRA for that trial by the time of onset of the SAR. If unreported SUSARs are identified, this should be submitted as a potential Serious Breach to the MHRA GCP Inspectorate for review and assessment. Further guidance on how to report a Serious Breach can be found here.

Note: for the purposes of this impact assessment, if upon review of the CTFG Q&A on RSI your organisation updated the RSI section of the IB in line with the guidance (such as removal of non-serious or fatal events) and this was subsequently approved, there is no requirement to assess the impact of this update as part of your impact assessment. As mentioned above, this impact assessment should use the RSI that had been approved for the trial by the MHRA at the time of onset of the SAR.

Onset date

As per CT-3 (55), the version of the RSI at the moment of occurrence of the SUSAR applies. We continue to see the RSI version applied from case receipt date and not the onset date. This is incorrect. SARs may be reported late by the investigator or identified late through monitoring. Therefore, the onset date should be used to determine which RSI version is applicable for expectedness assessment, and this should not change when follow-up information is received. This should be the RSI that has been approved by the MHRA and implemented by the organisation by the time of onset of the SAR.

Comparator IMPs

As per the Interpretation section of SI 2004/1031 (and article 2(d) of the Directive 2001/20/EC), an IMP is any active or placebo being tested or used as a reference in a clinical trial and includes medicinal products with a marketing authorisation. The sponsor of a trial has a duty to report SUSARs relating to each IMP used in a clinical trial, therefore, also relating to the comparator(s).

We have seen that the UK legislation and CTFG guidance have only been applied to products which are owned by the sponsor and not to all IMP products including comparators. This has resulted in the under-reporting of SUSARs for comparator products (IMPs) on a trial. It is expected that the comparator SmPCs (summary of product characteristics) are reviewed periodically. This is necessary not only to check for updates to the RSI, but in general to make sure that changes to the SmPC are acknowledged to protect the safety of the trial participants. For example, there may be changes to contra-indications or warnings that would necessitate a protocol amendment in relation to inclusion or exclusion criteria, or participants may need to be informed and consented if new safety information comes to light.

Fatal and Life-Threatening (LT) SARs

The CTFG Q&A on RSI clarified the requirements already in legislation and guidance, including that fatal and life-threatening (LT) SARs should not be considered expected (unless explicitly stated in the RSI and approved by the NCA). For example:

  • Regulation 2 of UK SI 2004/1031 (as amended) states: ‘unexpected adverse reaction’ means an adverse reaction the nature and severity of which is not consistent with the medical product in question
  • Article 2(p) of Directive 2001/20/EC states: unexpected adverse reaction’: an adverse reaction, the nature or severity of which is not consistent with the applicable product information (e.g. investigator’s brochure for an unauthorised investigational product or summary of product characteristics for an authorised product)’
  • CT-3 (50) states ‘reports which add significant information on the specificity, increase of occurrence, or severity of a known, already documented SAR, constitute unexpected events.’

This is not a new requirement since the CTFG Q&A on RSI came into effect. If your organisation had previously been considering these events expected (without MHRA approval), then this is a non-compliance and the events should be re-assessed as unexpected and unblinded for SUSAR determination and reporting.

Where procedures have been updated to reflect this requirement, we have seen that this issue can still occur where manual review of these events is required. For instance, if auto-labelling is used, manual review of LT and fatal events is required where the auto-labelling cannot also take into consideration the severity. Where this manual review does not take place there have been instances where fatal and LT SARs have wrongly been considered expected and this has not been picked up during case review. An impact assessment reviewing this requirement retrospectively has rarely been completed in our inspection experience which has resulted in the inspector identifying unreported SUSARs.

RSI implementation date

We have observed on inspection that control of the RSI implementation has not been performed in accordance with the legislation, resulting in the implementation of a substantial amendment prior to NCA approval. The quality management system should clearly define when the RSI will be implemented (following approval by the competent authority). There should be a clear change control and tracking process which can demonstrate when the RSI was approved (by each competent authority) and implemented by those making the expectedness assessment. Where you have more than one trial using the same IMP, this can become tricky and so robust tracking is required to ensure the RSI is not implemented on a trial prior to approval by the competent authority. This also requires effective communication between various departments in an organisation such as Regulatory Affairs, Pharmacovigilance and Clinical Operations. Trial sponsors are reminded that the RSI must be approved at a trial level. Therefore:

  • before an RSI can be used to assess expectedness of serious reactions occurring in a UK-only trial it must be approved by the UK competent authority
  • before the RSI can be used in a trial being conducted in both the UK and European countries it must be approved by both the MHRA and all the competent authorities of the EU member states.

Therefore, if an RSI is applicable to both trials A and B and the RSI was approved by the MHRA for trial A; it does not mean that this RSI can be used for trial B before the MHRA has approved its use for trial B.

As per CTFG Q&A for RSI, a substantial change to the RSI (such as IB) must be approved by each NCA prior to implementation (for example, to determine expectedness). The IB can be sent to investigators for information following approval of the substantial amendment by the MHRA, Research Ethics Committee (REC) and Health Research Authority (HRA), in other words, you do not have to wait until all NCAs approve the RSI before sending to investigators, as this would follow the national substantial amendment process. However, if the IB is sent for information to the principal investigator (PI) or chief investigator (CI) ahead of release of the RSI for use in determining expectedness in the trial, this process should be well controlled in the event that PIs or the CI are making this assessment on behalf of the sponsor (as is sometimes the case in non-commercial trials)

RSI for a licensed product

The RSI for a licensed product can be the RSI section in the IB (for instance, if the IMP is being used outside of the marketing authorisation indication or if you are wanting to use one RSI across a global trial). However, it is not acceptable to copy and paste section 4.8 of the SmPC into the RSI section of an IB. If the best document to support the conduct of the trial is the IB, the RSI section should be written in compliance with safety reporting requirements for clinical trials (a list of serious adverse reactions considered expected for safety reporting purposes).

Section 4.8 of the SmPC is not written to comply with the safety reporting requirements of clinical trials. If there is a good rationale why the SmPC can be used to support the conduct of the trial, the MHRA will accept that Section 4.8 of the SmPC is used as the RSI despite its limitations (that it includes adverse reactions rather than serious adverse reactions).

It is therefore expected that there might be differences between section 4.8 of the SmPC and the RSI section of the IB (with any differences documented with a rationale). Where reactions are considered expected in the RSI, there should be risk mitigation measures included in the protocol (such as patient monitoring and dose modifications).

Lack of efficacy and disease progression

SARs due to lack of efficacy or disease progression should not be considered expected, unless this has been approved as part of the trial protocol or is listed in the RSI. Lack of efficacy is not per se an adverse event, but can cause an adverse event/reaction. If the latter happens the investigator should assess the AE secondary to lack of efficacy in the same manner as any other AE/SAE. The protocol should clarify what approach will be taken regarding disease progression. The protocol could, for example, state that disease progression is recorded as an SAE only if it is considered more severe than expected in the trial population or it is recorded only if it is considered related to the study drug. This approach is acceptable if approved by the competent authority. If there is no specific wording, it is expected that disease progression, if serious, is treated as any other SAE occurring in the trial.


Where death is reported as an outcome of a SAR, this should be considered a fatal event and expedited within the reporting timelines as SUSARs, as fatal events are to be considered unexpected unless explicitly stated in the RSI and approved by the competent authority. We have seen SARs in database listings with a fatal outcome incorrectly assessed as expected, with the organisation indicating this was due to disease progression (due to other information available in the narrative of the SAE report). However, in these instances there should be appropriate training of investigators on pharmacovigilance requirements for the trial, including SAE reporting and causality assessments. If a death is due to disease progression this means that a serious adverse event occurred, the event was disease progression and the outcome is fatal. Therefore, this is not a SAR and cannot be a SUSAR. If the investigator believes that disease progression was caused by the IMP, then disease progression is a SAR, the outcome is fatal and a SUSAR report is required (unless fatal disease progression is a SAR included in the RSI). If conflicting information is provided by the investigator the Sponsor should follow up via a query to ensure accurate reporting.

MedDRA terms and updates

As per CTFG Q&A on RSI, it is recommended that the expected reactions included in the RSI are described using MedDRA Preferred Terms (PTs). When assessing the expectedness of a SAR, the sponsor needs to refer to the specific PTs included in the RSI. Any SAR which adds additional information about the specificity and/or severity of an expected reaction must be considered unexpected. For example, if the RSI includes acute tubular necrosis, a serious adverse reaction of chronic tubular necrosis is unexpected. If the RSI includes hepatitis, a serious adverse reaction of fulminant hepatitis is unexpected. In addition, a process must be in place to assess whether MedDRA updates have an impact on the RSI (and this assessment should be documented). In case changes to the RSI are needed following a MedDRA update, a substantial amendment must be submitted and approved before the changes can be implemented. The list of expected SARs included in the RSI must be the same as the one included in any software used by the sponsor to assess expectedness.

Drug Safety Update Reports (DSURs)

The CTFG Q&A on RSI has clarified how to align your RSI update with your DSUR reporting period and submission. However, we continue to observe findings in this area. For example:

  • the RSI used for the DSUR listings is not the same RSI in place at the start of the reporting period
  • more than one RSI was used by the organisation during the reporting period, but this was not evident in the DSUR or the cover letter, therefore, the MHRA assessors would not be aware of this additional RSI

Please remember that for the purpose of writing the DSUR for trials conducted in the UK, as well as other EU countries, you need to use the RSI that was approved at the beginning of the reporting period by both the MHRA and European member states! If your RSI was updated during the reporting period, approved and implemented for expedited reporting, then you will need to assess the impact of this change on your line listings in the DSUR. We have seen that the SUSARs listed in the DSUR are based on the extract from the safety database, which may have expectedness assessments based on a version of the RSI different from that approved at the beginning of the reporting period. The DSUR demonstrates the safety profile of your IMP over a year and the RSI is fixed at the start of the reporting period in order to set a baseline for review of all safety data received in comparison with this. If the data submitted is based on different RSI versions, then this impedes the competent authority’s assessment of the risk benefit profile of the IMP.


Whilst this is not strictly to do with the RSI, it does have an impact on expedited reporting of SUSARs. We have noticed that a conservative approach to causality has not been maintained by organisations when reviewing SARs. For example, when the investigator causality assessment is missing or not known, it is expected that a conservative approach is applied and the event treated as related until informed otherwise (regardless of the company causality assessment). Therefore, the sponsor should be actively following up with investigators to obtain a causality assessment before the reporting deadline.

Frequency of expected SARs

The frequency of the expected SARs should be indicated using Council for International Organizations of Medical Sciences (CIOMS) categories.

The categories are:

  • very common
  • common
  • uncommon
  • rare
  • very rare

A frequency of ‘not known’ is not deemed acceptable because it does not allow assessment of whether the new SAR report represents an increased frequency and therefore is unexpected. If the frequency cannot be calculated, the number of serious adverse reactions should be provided.

RSI for biosimilars/generics

As a general rule, all SARs caused by a biological medicinal product whose biosimilarity to the reference product has not been proved should be considered unexpected. However, if the product is a potential generic and is a small molecule that will be administered via the same route as the reference product, there might be a good rationale why the SARs expected for the reference product should be considered expected for the potential generic even before similarity is demonstrated. The rationale should be provided at the time of the clinical trial authorisation (CTA) application. Each case will be assessed on an individual basis.

Tips to improve compliance

Submission of RSI updates

Once a substantial amendment has been submitted, the competent authority has 35 days to approve the amendment unless a Grounds for Non-Acceptance is raised sooner. Did you know that you can submit the RSI update as a substantial amendment for all your trials conducted in the UK in one go? You will need to list in the covering letter all the trials the substantial amendment refers to and this will allow the assessor to review the impact for all trials in one go, and issue an approval in one go too. Our assessors have reviewed an RSI update for over 70 trials in one single application! Bulk submissions of IB amendment affecting more than one trial are encouraged.

You can also submit more than one amendment at a time. So, you do not need to wait for your protocol/IMPD update to be finalised before submitting the RSI update. Parallel submissions can take place as long as you are clear in the covering letter what other submissions have been made and how these are linked (for example, the risk mitigation measures which have been included in the protocol following an RSI update).

Tracking of RSI updates

Have a clear change control procedure and green light for when an RSI update can be implemented. Ensure there is documentation and tracking to support the process (which should be retained for inspection).

Perform an Impact Assessment

Assess the impact of CTFG Q&A on RSI on your quality system and on your clinical trial safety data to determine if there has been any under-reporting of SUSARs. If you have identified any, these should be reported as a serious breach. It is recommended that you map out the process from end to end and include all stakeholders and departments impacted. If your organisation uses an IB template, it is recommended that this is also reviewed to ensure the RSI section is in accordance with the CTFG Q&A on RSI.

Compliance reviews/metrics

Organisations collate compliance metrics and review for case processing, but we have seen in commercial organisations that this focuses on post-marketing, and often omits the requirements for interventional clinical trials. Therefore, it is recommended that clinical trial safety reporting metrics, including review of submissions to RECs, are included in your compliance reviews/oversight mechanisms. For a comprehensive assessment this should include both initial and follow-up reports in the metrics.

DSUR review

Quality control and review of the listings in the DSUR to ensure that they are in accordance with the RSI at the start of the reporting period.

Investigator Brochures

Remember that by waiting for RSI approval before sending an IB to investigators, you are not withholding urgent safety information from them. If there is an import safety signal or risk which investigators need to be informed of there are provisions in the regulations for this: a Dear Investigator Letter or Urgent Safety Measure should be considered in these circumstances.

Trial-specific RSI

It is possible to have a trial-specific RSI, which can be particularly useful when there are multiple IMPs. For example, it is possible to review the SmPC of each IMP and include in the trial-specific RSI only the serious adverse reactions which can be considered expected in your trial population. This may be useful in long-running trials and trials with multiple IMPs. It is also possible to have one RSI for different brands or molecules of IMP under the same ATC (Anatomical Therapeutic Chemical) code.

Ask the MHRA

If you have any questions, you can always contact us via

To summarise, there continues to be non-compliance when it comes to RSI for clinical trials. It is hoped that by following the recommendations in this post, safety reporting compliance will improve so both the regulator and sponsor can fulfil their obligations to protect patients.


Don’t miss the next post,  sign up to be notified by email when a new post comes out


Sharing and comments

Share this page