US Food and Drug Administration


1401 Rockville Pike

                                                                Rockville, MD 20852-1448




To:             Committee Members, VRBPAC

From:        CBER Office of Biostatistics and Epidemiology

Re:            February 27, 2007 Meeting Topic 1: Pharmacovigilance plan for an H5N1 Inactivated Influenza Vaccine Manufactured by Sanofi Pasteur


General Considerations for Pandemic Influenza Vaccine Safety Monitoring


Limited safety data will be available for pandemic influenza vaccines prior to use. The threat of a swine influenza pandemic in 1976-77 led to a mass vaccination effort in the United States. Unfortunately, the vaccine was subsequently found to cause rare cases of Guillain Barré Syndrome (GBS), a severe paralyzing neurologic condition. In the event of a pandemic and a subsequent mass vaccination program, it seems imperative that the United States be well prepared to detect a problem that occurred once before, and that the public would expect no less. Robust safety monitoring systems need to be in place. The timely identification and evaluation of serious adverse events is important for at least three reasons. First the morbidity to the individual may be severe, representing an important safety issue. Second, unwarranted fear of adverse reactions can lead to lack of confidence in the vaccine and unnecessary disruption of a beneficial vaccination program. Third, new and timely information that quantifies the incidence of serious vaccine adverse effects, if any, can facilitate benefit-risk analysis of continued vaccine use.


Post-licensure monitoring of vaccine safety involves identification of possible adverse effects of vaccination, followed by evaluation of these “signals” for a possible causal link with vaccines. Although an ideal system would include a national registry of all vaccinees, with common identifiers allowing researchers to determine the vaccination status of adverse event (AE) cases, such a system is currently unavailable in the United States. Systems currently available to conduct vaccine safety monitoring include the Vaccine Adverse Event Reporting System (VAERS) and the Vaccine Safety Datalink (VSD). Adaptation of these systems for use during a pandemic may be required. Other large automated databases of encounter and/or claims data that might be useful for pandemic influenza vaccine safety monitoring include those of the Centers for Medicaid & Medicare Services, Veterans Affairs, Department of Defense, and private insurance groups not currently part of the VSD. The FDA and CDC are evaluating these systems for their potential to contribute to the monitoring of pandemic influenza vaccine safety.


Any plan for monitoring pandemic influenza vaccine safety must account for a number of factors, including stage of the pandemic, the population receiving the vaccine, and the strategy for vaccine distribution. For example, during the pre-pandemic period no disruptions to normal government, social and business functions are anticipated. However, during a pandemic period, it is possible that widespread and severe morbidity and mortality could cause large scale absences from work and changes to the operation of the health care system. Under such circumstances, safety monitoring might also need to adapt including modifying some safety monitoring activities from those routinely conducted pre-pandemic. FDA has not yet decided on what specific modifications, if any, will be made to adverse event reporting guidelines.


Another factor to consider is that there may be different priority groups (e.g., the elderly, health-care workers, etc.) for vaccination at different phases of the pandemic and vaccine safety considerations in the different groups may vary. VAERS is a national spontaneous reporting system that covers all groups. However, other systems that utilize large health databases may or may not adequately capture certain demographic groups. To facilitate early detection of any potential vaccine safety problems, it is important to assess whether existing systems (e.g., VSD) adequately cover all demographic groups, and particularly those who may be prioritized for early vaccination. This includes assessing whether sample size (both overall and for relevant demographic strata) is adequate to evaluate uncommon serious adverse events. Possible gaps could potentially be addressed by expansion of safety assessment capability using additional health databases (government and/or private databases). The additional databases could also provide complementary data to assess consistency of results and facilitate decision-making.


An additional important consideration that may affect the utility of certain health databases is the strategy for vaccine distribution. If a pandemic influenza vaccine were to be distributed through usual health care channels then systems currently in place for vaccine safety monitoring may be adequate. Generally, influenza vaccine is administered in a variety of settings including public health departments, medical clinics, nursing homes, work places, pharmacies, and others. Whether or not the vaccination is captured in a particular database depends on a variety of factors including whether the provider participates in a particular insurance plan, whether a claim for the vaccine and/or vaccine administration procedure is submitted, or whether a vaccination encounter record is created. A shift in financing or delivery of vaccination (e.g., if public health departments play a greater than usual role, initially or at any time, in distributing a pandemic vaccine) could affect the rate of capture of vaccinations in certain databases. Although large automated databases are helpful in identifying rare adverse events that may go undetected otherwise, they are limited in capturing vaccinations that occur outside of the particular system.


Thus, the populations that will require monitoring, and systems available for monitoring, may be different during different stages of the pandemic. Moreover, our capacity to monitor AEs will also depend on the vaccine distribution system chosen. A robust and flexible AE surveillance system, with collaboration among government agencies and vaccine manufacturers, will be needed to address the range of possibilities that an influenza pandemic presents.


Pharmacovigilance Planning


Beginning in 2005, FDA has requested that vaccine manufacturers submit a Pharmacovigilance Plan with their Biologics License Application (BLA) according to the FDA and International Conference on Harmonization E2E guidelines for Pharmacovigilance Planning ( This ensures adequate time for review and development of these plans and improved integration of the pharmacovigilance plan review, with the review of other safety data submitted with the BLA. In addition, sponsors are encouraged to develop a plan to monitor safety of their pandemic influenza vaccines in the event of a pandemic influenza outbreak (Guidelines for Clinical Data Needed to Support the Licensure of Pandemic Influenza Vaccines-Draft Guidance March 2006). The development of a detailed pharmacovigilance plan is an important step in ensuring that adequate safety monitoring will be conducted post-licensure. In general, a pharmacovigilance plan should provide a summary of ongoing safety issues (e.g., important identified risks, potential risks, and/or missing information), outline routine pharmacovigilance practices (systems and processes for collecting and reporting adverse events, preparation of reports for regulatory authorities, continuous monitoring of the safety profile, and other requirements deemed necessary), an action plan for safety issues, and a summary of actions to be completed including milestones.


In addition to routine pharmacovigilance activities (e.g. submission of spontaneous adverse event reports to VAERS), safety surveillance and/or observational epidemiological studies are often required as post-licensure commitments to investigate either populations not studied pre-licensure or rare events requiring vaccine utilization in very large populations. Furthermore, sponsors may wish to enhance the safety database of their pandemic influenza vaccine by conducting safety studies prior to a pandemic influenza outbreak (e.g., among individuals identified by public health authorities for pre-pandemic vaccination) (Guidelines for Clinical Data Needed to Support the Licensure of Pandemic Influenza Vaccines-Draft Guidance March 2006). It is highly recommended that sponsors work closely with the FDA and CDC to develop and conduct studies to monitor vaccine safety. A well-coordinated effort among all parties, including government and vaccine manufacturers, will help ensure that safety issues are being monitored and addressed effectively and in a timely manner.


Review of Sanofi Pasteur’s Pharmacovigilance plan


A “H5N1 Influenza Virus Vaccine Risk Management Plan” was submitted to FDA by sanofi pasteur on January 26, 2007. The plan outlines the pharmacovigilance activities that sanofi pasteur has proposed to conduct during the pre-pandemic and pandemic periods.


Sanofi pasteur has proposed the following routine pharmacovigilance activities for the pre-pandemic period:

  • Expedited reporting in accordance with international regulations.  Reporting of all serious adverse events within 15 calendar days from receipt and unexpected serious adverse events occurring in a clinical trial that result in death or are life threatening within 7 calendar days to the Health Authority (HA). All adverse event reports are entered into a validated Global Pharmacovigilance System database (ARIS-g).
  • Production of Periodic Safety Update Reports (PSURs) every 6 months for 2 years and then annually.  
  • All pre-natal exposure reports and outcomes of pregnancy are captured within the ARIS-g with pregnancy reports being followed until outcome is known.
  • Signal detection and analysis. When a safety signal is identified, a medical and scientific analysis is initiated by the Product Safety Officer (PSO). All signals that represent a safety issue are further evaluated with a proposed action plan, resources and estimated target date for completing the actions.
  • Preparation of Ad-Hoc reports upon identification of a potential safety issue and/or upon a HA request or other healthcare customer.


Sanofi pasteur has proposed that the routine pharmacovigilance plans for the pre-pandemic period be modified during the pandemic period, specifically:

  • Only significant safety issues will be reviewed and evaluated on an ongoing basis whereas non-urgent activities will be minimized and addressed after the pandemic.
  • Expedited reporting will be the basis of safety evaluation. A common collection form should be used by all parties. All serious adverse events, deaths, life-threatening events, and adverse events of special interest (AESI) need to be reported. AESI include anaphylaxis, Guillan-Barré Syndrome, Bell’s palsy, neuritis, convulsions, and syncope. A system needs to be developed to triage the consumer reports based on severity and agreed upon case definitions. Reporting should be done ASAP and no later than 15 calendar days, with more precise timetables being implemented. Rapid and open communication between sanofi pasteur and Authorities/Public Health Services is essential. Electronic communication should be established.
  • PSURs will not be submitted or alternatively may be submitted focusing on only serious adverse events, deaths, life-threatening events and AESI. An aggregated PSUR will be submitted when the pandemic is declared finished.
  • Safety evaluation will be based on signal detection taking into account the particular circumstances of a pandemic situation, and its potential impact on assessment of the benefit/risk profile. Signal analysis will be similar to the pre-pandemic period. It is important that all parties come to a set o predefined criteria which would potentially trigger the suspension of the vaccination campaign.
  • Although sanofi pasteur does not propose to perform a Phase 4 study, they indicate that cohort studies should be considered to bridge safety and efficacy data in populations not studied to date. Because the cohort studies would not be of sufficient size to detect rare events; studies to detect rare events should be coordinated by national or international public health agencies.
  • Sanofi pasteur also recommends that a true efficacy study be organized by public health agencies.
  • Vaccine failure reporting will not be requested for routine pharmacovigilance activities. Case reports of vaccine failures will however be recorded in the safety database.


In reviewing the submitted plan, FDA makes the general observation that sanofi pasteur has proposed several changes to the requirements for adverse event reporting under FDA regulations, including as stated in sanofi pasteur’s “H5N1 Influenza Virus Vaccine Risk Management Plan”: a)“Routine pharmacovigilance practices will cease with the official announcement of the pandemic by the WHO and special PV practices proposed in §3.3 could be implemented by the appropriate authorities” , b) changes to the requirements for submission of periodic adverse experience reports, c) “safety evaluation will be based on signal detection taking into account the particular circumstances of a pandemic situation, and its potential impact on assessment of the benefit/risk profile”, d) vaccine failure will not be requested for routine pharmacovigilance activities, and e) they do not propose to perform any Phase 4 study. While FDA acknowledges that during an influenza pandemic there might be a need to modify some safety monitoring activities from those routinely conducted pre-pandemic, the FDA has not yet decided on what specific modifications, if any, will be made to adverse event reporting guidelines. It is possible that increased attention to adverse event reporting, or to reporting of specific adverse events, might be feasible and required.


FDA also suggests a) clarification of the statement “common collection form should be used by all parties”. It is not clear whether sanofi pasteur is proposing the development of a new collection form or is referring to the VAERS form. Also, it is not clear who are the “parties” that is being referred to in this statement, b) possible editing of the AESI list to include other adverse events, c) modifying the statement “A system needs to be developed for the triage…” to “A system should be developed  for the triage..”, d) that it may not be feasible to come up with a complete apriori list of criteria that would potentially trigger the suspension of the vaccination campaign, and e) discussion of some of the theoretical constellation of events that might trigger suspension of a vaccination program under §4 as well as any other potential risk minimization plans the Health Authorities might use in conjunction with vaccinations.


In addition, FDA notes that sanofi pasteur recommends that epidemiological studies of safety and efficacy be conducted but does not volunteer what role they might play in such studies. FDA agrees that such studies will likely be important and that collaboration among government agencies and sanofi pasteur in conducting such studies would be beneficial. Particularly, it would be desirable for sanofi pasteur to commit to working both with the CDC and FDA to fill gaps in the collection and analysis of important safety, and possibly effectiveness, data in a pandemic situation.