Bart’s Corner: Medical Literature Review
Having done several audits/mock inspections in the past several months, one of the items that seems to still be confusing to companies is the area of searching the medical literature for safety information.
First, the easy part.
For drugs in clinical trials but not yet on the market, the situation is relatively straightforward. The sponsor must do continuous and periodic searches of the medical and scientific literature for information on the drug, similar drugs, class effects, etc. This information is used to update the benefit/risk analysis, the Investigator Brochure (IB), report expedited cases to FDA and other health agencies (HAs), inform investigators and IRBs/ECs, etc. If the drug is totally controlled by the sponsor such that no other company or individual is doing studies outside of those set up by the company, the likelihood of many or any publications except those from the sponsor is small. That is, the company is in control of all publications in most instances. Any publications found are also reported in the Annual Safety Report/DSUR, etc.
The trickier situation is for marketed drugs (including those also in clinical trials at the same time). In this situation the company/sponsor is not in full control of publications. The drug may be studied by any company or individual by simply purchasing the drug on the market. The company owning the drug may or may not be aware of ongoing studies, safety reports and other publications on the drug in question. If there are multiple manufacturers or Marketing Authorization (MA)/NDA holders or if the product is generic there may be many publications.
The question then arises about what is to be sought in the medical literature and how are publications that are found to be handled.
One of the earliest references is in ICH E2D, dating from 1996, which states:
“MAH is expected to regularly screen the worldwide scientific literature by accessing widely used systematic literature reviews or reference databases. The frequency of the literature searches should be according to local requirements or at least every two weeks. Cases of ADRs from the scientific and medical literature, including relevant published abstracts from meetings and draft manuscripts, might qualify for expedited reporting. A regulatory reporting form with relevant medical information should be provided for each identifiable patient…The regulatory reporting time clock starts as soon as the MAH has knowledge that the case meets minimum criteria for reportability.”
The FDA regulations are minimal in terms of detail. They simply state that cases and safety information from the medical literature must be sought and reported as expedited reports if they meet the criteria and are to be included in periodic reporting.
The EMA, however, has published detailed information on its requirements which expand upon and change the E2D proposals. The EMA information is found in Module 6 of the Good PV Practices Guidelines on page 10 stating the principles and in more detail in VI. Appendix 2:
The concept is summarized as follows:
“Marketing authorisation holders are therefore expected to maintain awareness of possible publications through a systematic literature review of widely used reference databases (e.g. Medline, Excerpta Medica or Embase) no less frequently than once a week… In addition, marketing authorisation holders should have procedures in place to monitor scientific and medical publications in local journals in countries where medicinal products have a marketing authorization…Reports of suspected adverse reactions from the scientific and medical literature, including relevant published abstracts from meetings and draft manuscripts, should be reviewed and assessed by marketing authorisation holders to identify and record ICSRs originating from spontaneous reports or non-interventional post-authorisation studies.
If multiple medicinal products are mentioned in the publication, only those which are identified by the publication’s author(s) as having at least a possible causal relationship with the suspected adverse reaction should be considered by the concerned marketing authorisation holder(s). Valid ICSRs should be reported…One case should be created for each single patient identifiable…”
More details are given in Appendix 2. Key points are briefly summarized here:
When to Start and Stop Searching
The company must do worldwide literature searching starting at submission of the MA, including the period between submission of the dossier and authorization (approval) of the MA. The searches must continue as long as the MA is “active”. It is specifically stated that this applies whether the product is marketed or not. The goal is to identify Individual Case Safety Reports (ICSRs) and any possible changes to the benefit-risk profile for reporting to HAs and inclusion in PSURs.
Where to Look
A major database covering large numbers of journals should be used. The Module specifically names Medline, Embase and Exerpta Medica. One or more may be used. Other databases may be used if they meet the criteria for global, comprehensive journal coverage. Specialized databases are also to be used if appropriate. Local journals that are not covered in the major databases should also be tracked where possible. Published abstracts and draft manuscripts should also be reviewed if possible. Meetings where information is presented should be included. Although attendance at such meetings is not obligatory for the company, if they have representatives there, they should seek out the abstracts and publications, posters etc.
The module describes generically the nature of database searches making the point that the search should be optimized for each database based on how that database is constructed. Sometimes a broad search suffices (e.g. product name and active substance name) but in other situations more terms are needed to refine the search.
The search should be for the active substance, not just the brand name. It may be appropriate to include searches for excipients, other salts, alternate spellings etc. It may be appropriate to exclude searches for alternate routes of administration or pharmaceutical forms if not relevant. However, the search should include those articles where route or form is not specified. Searches should also include off-label use, medication error, misuse, abuse, overdose and occupational use.
Selection of Search Terms
The appropriate search terms should be used including outcomes (e.g. death), pregnancy and other less obvious terms that may help find safety reports.
Limits to a Search
Limits should be carefully applied and must be justified. One obvious limit is the date range if searches are done weekly. Cases in all languages should be sought not just in English.
Good record keeping is obligatory. The construction of the search, the results of the search and the date of the search must be kept.
The search output should usually include more than just the title of the article. Abstracts and more detail should be obtained in the search results where possible.
Review and Selection of Articles
An information specialist trained in PV or a PV professional should review the outputs and select articles in a systematic way. Demonstrable due diligence must be done. Although not stated directly, an SOP or procedural document for this should be in place. Quality Control checks should also be in place.
Searches should be done not just to pick up cases for expedited reporting but also for finding articles that will be included in the PSUR.
All articles that are “likely” to contain relevant information should be obtained and reviewed. The urgency of doing this should be proportional to the content of the material and to any actions that might be taken as a result of the material.
If another company’s product is clearly named and identified in a publication, that instance may be excluded. However, if no product is named, the article should be included. Other exclusion criteria include:
- Individual case safety reports that originate in a country where a company holds a marketing authorization but has never commercialized the medicinal product;
- ICSRs which are based on an analysis from a competent authority database within the EU. The reporting requirements remain for those ICSRs which are based on the analysis from a competent authority database outside the EU;
- Literature articles, which present data analyses from publicly available databases or, which summarize results from post-authorization studies. This type of literature article describes adverse reactions, which occur in a group of patients with a designated medicinal product with the aim of identifying or quantifying a safety hazard related to a medicinal product, and aggregated data on patients are often presented in tables or line listings. The main objective of those studies is to detect/evaluate specific risks that could affect the overall risk-benefit balance of a medicinal product.
The regulatory clock starts when anyone in the organization or third parties with contractual arrangements becomes aware of the article. This is usually the day of the search and is printed on the output. For articles that have been ordered, day zero starts when a valid case is available. Organizations are encouraged to obtain and review articles promptly.
As with other ICSRs, duplicates should be sought in the database to avoid reporting the same case more than once.
This is permitted. Any such arrangement should be done with a written contract with each out-source organization. The contract should detail which functions will be handled by the third party. The ultimate responsibility still remains with the sponsor/company however. The third party should be “assessed” to be sure they are capable of handling the functions out-sourced. The word “audit” is not used but some level of assurance that the company is capable of the work should be made. Day zero starts when the third party gets the information not when the third party supplies it to the originating company.
Where operational in the HA, electronic reporting, with the article attached, should be done. If not set up, a paper copy of the article should be supplied. The Appendix does not discuss the language issue, but in general, translation into English of an article (or at least the relevant sections) not in English should be done.
Comment: This then is a detailed set of instructions on how to handle literature searching. It represents “best practices” and most probably should be used for all searches, including those done for the FDA. For companies with many products, products with wide usage, products with many manufacturers or generics, this may turn into a formidable undertaking and require one or more PV or medical experts to evaluate the articles. The timing can be tight if an expedited report is found. For marketed drugs the EU requires weekly searches.
Companies must do medical literature review. They do not have a choice. At some point the EMA may start its promised literature searching supplying the documents to the companies. Whether this will be sufficient for US requirements and company signaling remains to be seen.