The EU and the FDA (as well as other HAs) have required sponsors/NDA holders/Marketing Authorisation Holders (MAHs) to monitor medical literature for adverse drug reactions for many years. The EMA has finally put forth the details of its new system to monitor medical literature for MAHs.
In the US, the requirements for the post-marketing monitoring of medical and scientific literature are found in 21CFR314.80. They require the submission of serious unexpected 15-day reports and in periodic reports derived from literature.
Similarly, the EMA requires literature reports as expedited reports and in PSURs. Most companies have developed systems of varying efficiency and complexity to do this and/or have outsourced the literature review.
As promised by the EMA, the EU will begin monitoring literature reports of selected products authorized in the EU for the MAHs. Regulation 1235/201 of 15 December 2010 requires the EMA to monitor selected medical literature for reports of suspected ADRs related to active substances. The EMA is publishing a list of the substances that must be monitored, will enter them into the EudraVigilance database and prepare a detailed guide on literature monitoring.
The detailed guide was published as a draft in June 2014 and finalized 12 May 2015. It came into effect in July 2015.
Also published is a spreadsheet of the ~3296 substances to be monitored (In fact, these represent various different salts and combinations of active substances. Unique substances are about 400.). They will also monitor certain herbal products.
Interestingly, in the guide document, the EMA notes that “The total number of substance groups included in the literature-monitoring service is based on the Agency’s allocated budget.”
The EMA started monitoring the first 50 of these in July and has said it will expand to the full list perhaps by September 2015.
Sources of Medical Literature
In a separate document, the EMA notes the sources they use for medical literature. This includes Embase, which is a privately held database by Elsevier. It covers some 8500 journals including all MEDLINE journals (see the fact sheet).
In addition, they also use journals covered by EBSCO, a private US company (which also includes MEDLINE with full text), International Pharmaceutical Abstracts covering 800 journals and The Allied and the Complementary Medicine Database. These sources cover Medical Subject Headings (MeSH).
See the URLs for further information on the databases and journals covered.
Comment: This is a very extensive listing covering the journals and sources that are most likely to have information on marketed drugs and some clinical trial drugs also. They won’t cover all small, national, local or certain less common languages.
These sources are searched every day except Saturday and Sunday by the EMA. In addition, a monthly search is done focusing on drug therapy and pharmaceutical information including complimentary medicine and alternative treatments.
Search strategies are created using search strings which the Guide says are published on the EMA website under “MLM Search Strategies”. The Guide gives some high level details on how they create these strategies and strings. Presumably this will be posted sometime in the future as I could not find it on the EMA website for the preparation of this posting.
Comment: This is a critical point. How the searches and strings are constructed and kept up to date will determine whether all the appropriate citations in the literature are found.
Screening & Scientific Review and Assessment
The results are reviewed and assessed within one calendar day of the search. The EMA looks for serious reports that contain valid Individual Case Safety Reports (ICSRs) using criteria outlined in Module VI:
- suspected adverse reactions originating from spontaneous reports and solicited reports in humans;
- special situations such as use of a medicinal product during pregnancy or breastfeeding, use of a medicinal product in a paediatric or elderly population, reports of off-label use, misuse, abuse, overdose, medication errors and occupational exposure with suspected adverse reactions;
- lack of therapeutic efficacy;
- suspected adverse reactions related to quality defects or falsified medicinal products;
- suspected transmission via a medicinal product of an infectious agent.
They only include SUSARs in the EEA and in third countries as well as suspected non-serious adverse reactions occurring in the EEA. The hits from the search are then analyzed using a library management tool to see if they quality for ICSR reporting. If not, they are removed and archived.
A duplicate check is done and those records which may refer to potential ICSRs are collected and databased for processing. See the Guide for further details on how this is done.
Comment: Since they do not search for non-serious suspected ARs that occur outside the EEA (e.g. the US, Canada) this may not cover literature searches required by other health agencies around the world. Be sure to verify that this is sufficient if you only rely on the EMA search for your literature searches.
Processing of Confirmed ICSRs
For literature reports of confirmed cases which can generate ICSRs, a full text of the citation is obtained and, if not in English, translated into English though it is not specified how/who does this or how long this takes).
The article is reviewed and the number of valid ICSRs is determined and seriousness/non-seriousness is noted.
An ICSR is then created along with a case narrative for serious cases. No narrative is prepared for non-serious ICSRs.
Causality assessments and relatedness are also done.
Timing: SUSARs from the EEA or elsewhere are prepared within 7 calendar days from day zero. Non-serious EEA cases within 21 calendar days. Day zero is defined as the date on which the Agency’s service provider becomes aware of a record containing the minimum information for a valid ICSR. If the full text is not available in sufficient time, an ICSR with the minimum information available is prepared as an initial record. See the Guide for more details on this including a flow chart. The Guide says that there are detailed SOPs and Work Instructions on the EMA website for this.
The ICSRs are sent to the European health authorities within one calendar day.
Companies (MAHs) can access and download the ICSRs from EudraVigilance. A list of available cases is published daily in the “MLM ICSRs” section of EudraVigilance though it does not appear to be there yet. It is hoped that the actual text of the article (and English language translation) will be easily available in EudraVigilance
Processing of Potential ICSRs
These are processed similarly but are archived if found not to be classified as a valid ICSR. The outcome of the screening of the full text article can also be found in the file “MLM ICSR”.
Follow-Up of Individual Cases
The EMA will/has put in place a process to ensure that follow up is done with the author(s) of the publication as necessary to obtain supplementary detailed information for the scientific evaluation of the case as outlined in Module VI.
The Guide notes on page 16:
“In principle, one attempt to follow-up with the primary author is made for suspected serious adverse reactions based on a risk-based approach. This refers to individual cases, where the outcome is not known, where pre-defined clinical information is missing as regards important medical events or for both and for individual cases where not all of the minimum reporting criteria are available for a valid ICSR. The MedDRA Important Medical Events (IME) list is published by the Agency at the EudraVigilance website”.
Follow up attempts are recorded in the file “MLM ICSRs” and published daily in the EudraVigilance website. ICSRs are created within 7 calendar days of receipt of the new information and day zero is the date of receipt of any new information.
Comment: The follow up process is interesting. Only one attempt is made (and they don’t say how) on a “risk-based approach” which presumably means someone may judge a case to be sufficient based on the publication alone and not do follow up. This process has always been tricky for sponsors and MAHs. How much is enough? How many attempts should be made? Is email enough or should telephone calls be done (by a medical professional or even an MD?)? Perhaps this is explained in detail in the SOPs and Work Instructions. I have not checked them personally.
In my experience most companies make two attempts on serious cases to get follow up information and sometimes these attempts are not “passive” (e.g. emails) but medical professional phone calls. This usually is done for all serious cases not just “risk-based” ones. Often the publication, unlike spontaneous cases, does contain enough data to make a valid safety analysis. However, sometimes the publication is not focused on the safety of the product but on some other aspect such as efficacy and the safety aspects are described only briefly. This means that some follow up needs to be done. See the Guide for more details and a flow chart on how the EMA does this.
The Guide describes the high level details for the Quality Management System (QMS) that they are putting in place and which companies and MAH’s must also have. See the document for the details.
There will be an English language (only) service desk open during working hours.
Although not covered in this Guide there are fees for this service and they are summarized in a fee guide. Unfortunately it is not totally clear how an MAH can calculate the fees as the fee of 67 euros is based on “chargeable units” which are not easy to calculate.
There is a long section of the EMA website devoted to the literature monitoring. This site also has some 30 links to other sites and documents including several videos.
EMA is putting together a very complex system for thousands of products with daily updates! I am not aware of pharmaceutical companies that do this on a daily basis for such a large volume of products. How this will be tracked and quality assured remains to be seen.
This is, arguably, more complex than the systems put in place by companies. There are several reasons for this: EMA covers more products than any one company, they have multiple stakeholders including the other health agencies in the EEA/EU as well as the companies and, presumably, the public who probably will be able to access the data.
The burden seems to be on the companies now to go the site, probably every working day, as it would not be wise to have serious ICSRs available to the public before the company.
It’s not clear how the company will handle those cases where they disagree with the causality and/or expectedness. This is very important since it is likely that there will be some disagreements by the MAH on causality. It is presumed the EU will do the expectedness analysis against the EU labeling (SmPC). For companies with different labels (particularly in the US) they may be obliged to do additional expectedness reviews against other labels and the Core Safety Information (CSI).
It is also not clear whether the company can or should do follow up on their own even if the EMA has already done so. It is highly likely that for a “critical” drug the company will try to do its own follow up. It is also likely that the company will not want or be able to wait to see if the EMA is successful in getting follow up. The EMA will discover that follow up does not rapidly flow in after the request. It can take weeks or more, especially if the case is still on-going.
How the case will be handled then remains to be seen if the company’s new version of the case differs from that of the EMA. Will the EMA update its database (EudraVigilance) with a company’s version? It is possible two or more companies may do follow up in addition to the EMA if both make the drug. How will these multiple versions be handled by the EMA and EudraVigilance?
What will companies do if this is not a sufficient mechanism for searching for cases required for other health agencies (e.g. FDA)? How will clinical trial case reports (ICSRs) be handled for a product which is marketed for one indication and being studied for another (as yet unapproved) indication?
Conclusions, Comments and Suggestions
This is “Version 1.0” of the new system. There will be many major and minor bugs to work out. One wonders if the EMA has made a wise choice here. Have they fixed something that wasn’t really broken. I’m sure a lot will happen in the next year or two.
Here are some suggestions:
- Anyone involved in literature searches for pharma companies must pay careful attention to this as it evolves. Study the documents referenced here in great detail, read the SOPs, get into EudraVigilance and understand what is happening.
- See which of your drugs are covered and which are not. It is not totally clear all the remaining 350 products will be covered by September 2015.
- Check the website every day!
- Do not stop doing literature searches even if they seem to be duplicative of the EMA’s searches. See how this plays out and see if both searches pick up the same cases or whether some are missed. Compare your search strings to the EMA’s if they are available.
- New SOPs and Work Instructions will be needed.
- Figure out whether you need to keep doing searches for other HA’s (e.g. FDA).
- Figure out your follow up strategy with the authors. Are you willing to wait weeks or more for EMA’s revised ICSRs after follow up?