Vaccine adverse event reporting: the importance of follow-up

Vaccines are exceedingly safe, but, as with all drugs, there are possible side effects. Most vaccine side effects are mild and transient. These include local injection site reactions such as pain, swelling and erythema. More serious side effects, such as Guillain Barré syndrome after the 1976 swine influenza vaccine, have been reported and receive much public attention [1]. It is also recognized that new or idiosyncratic side effects may occur when a vaccine is given to large numbers of people. For these reasons, and also to maintain public confidence in vaccine safety, it is necessary to monitor possible adverse events during the lifetime of a vaccine.

In the USA, the Vaccine Adverse Event Reporting System (VAERS) was set up to receive reports of possible adverse effects involving US licensed vaccines. VAERS was established by the National Childhood Vaccine Injury Act (NCIVA) of 1986 and is a cooperative effort of the US Food and Drug Administration (FDA) and the Centers for Disease Control and Prevention (CDC). The actual operation of the database is contracted to a private company. While VAERS was set up primarily with pediatric vaccines in mind, reports are accepted for all vaccines. VAERS is a voluntary, so-called passive, reporting system. That is, VAERS depends upon the initiative of individuals to report possible adverse events. The NCIVA requires physicians and other healthcare providers to submit reports of adverse events defined in the injury table [101] and in the contraindication section of the product label. Manufacturers are required by law to submit reports of all possible adverse events involving their licensed vaccine(s) that they learn about by any means [2]. Physicians and other healthcare workers may contact a manufacturer or report directly to VAERS. VAERS has numerous well known strengths: it is a national system; it can be used to detect rare events; it has mutliple reporting routes; and it is relatively simple to use. Disadvantages of passive surveillance such as under- reporting, incomplete and missing data, and reporting bias are also well known [3].

In reports from manufacturers and non-manufacturers, the information received at the FDA is often incomplete and sometimes erroneous. There are numerous reasons for this such as transcription errors, misunderstanding of exactly what the form is asking and forgetfulness. Individuals sometimes contact a manufacturer for specific product-related information and may be unwilling or unable to provide additional information about the adverse event. Nevertheless, if a possible adverse event is mentioned, the company is required to report that information to VAERS. Manufacturers are also required to follow-up reported events and forward any new information to VAERS. However, additional information received at VAERS may not be complete or timely.

For all of these reasons, follow-up by FDA, CDC and VAERS contractor staff is necessary to determine the facts of a case. That report is then the impetus and starting point for an investigation of the facts.

When to follow-up

It is not possible to follow-up all reports received at VAERS. Approximately 17,000 reports were received in 2004. At the FDA, serious reports are reviewed by a physician within 24 h of data entry. Serious reports are classified by the US Code for Federal Regulations as those which report a death, hospitalization, life-threatening illness, prolongation of hospitalization, permanent disability, or some other event which may reasonably lead to hospitalization, permanent disability or life-threatening illness [4]. Based on the judgment of the reviewing physician, any report may be selected for immediate, initial follow-up. Reporters of serious adverse events receive a letter acknowledging receipt of the report, and follow-up letters are sent 60 days and 1 year after receipt of the initial report. This ‘60-day letter’ requests information on the resolution of the adverse event and any key information that was missing from the original report. Most follow-up is carried out by the contractor’s staff of nurses, who request pertinent medical records and administer prepared questionnaires for some predetermined conditions. The FDA physician may sometimes personally conduct further inquiry in a focused manner. Due to the severity of the event, 100% of domestic death reports receive follow-up by the contractor’s staff. Another reason for focused follow-up activity includes special attention to the safety of vaccines that are newly released to the market. Vaccination campaigns involving expanded use or indication of a vaccine, such as smallpox vaccine administration in the US to the military and selected civilians, merit follow-up. Factors that may be used in prioritizing follow-up of individual reports are shown in box 1.

In general, follow-up includes inquiry of medical history, details of the event being reported, and prognosis or recovery. In addition to cases of childhood deaths, information regarding the mother’s pregnancy and family histories is reviewed. Death certificates or autopsy reports, when available, are used to ascertain the cause of death. Further details concerning questions guiding an inquiry are available in a paper by Silvers and colleagues [5] and from Brighton Collaboration case-definition descriptions, discussed hereafter [6–8]. It is known that initial reports often contain errors of various types, so all follow-up questionnaires include verification of the details in the initial report. For example, causes of death mentioned in an initial report are significantly changed in 25% of cases after follow-up [9].

Follow-up activities may involve FDA field offices, CDC or other government agencies. State health departments are intimately involved in immunization programs and often receive reports of possible vaccine adverse events. Such reports are then transmitted to VAERS. In exceptional cases, a state health department may have a role in a follow-up activity.

How to improve the initial report?

Obviously it would be extremely helpful if the initial adverse event report contained more higher quality information. Several approaches are being used to improve reporting. First, the FDA has posted on its website a guidance entitled, ‘How to complete the VAERS form’ [102]. Second, web-based electronic reporting has been implemented. The electronic form contains prompts to guide the reporter [101]. Third, several studies demonstrated the value of case definitions in classifying vaccine adverse events [10,11]. The most ambitious and complete attempt to improve possible vaccine-related adverse event reporting is the Brighton Collaboration [12]. This collaboration brings together a voluntary, international group of experts to craft a case definition of an adverse event and to specify what information should be gathered to define that event. A goal is to develop such documents for 100 adverse events related to vaccine safety. To date, approximately 20 documents have been completed. Some of these, intussusception, cellulitis and nodule at the injection site, have been published in the journal Vaccine [6–8]. Globally accepted case definitions would make data from different countries and studies more comparable for use in assessing vaccine safety.

Examples of special follow-up situations

Any day in the Division of Epidemiology at the Center for Biologics Evaluation and Research may bring a report of a situation that demands immediate attention through VAERS, or often by telephone or email from a reporter or another government agency. Although such reports are, fortunately, rare, the Division of Epidemiology must be prepared; prepared means having people who are especially knowledgeable about vaccines, adverse events and medicine, and how to efficiently obtain additional information and evaluate that information. Three recent examples of reports received during the 2003–2004 influenza vaccine season that required immediate attention are described briefly:

Twenty reports in one practice of a severe, stinging sensation on inoculation. This report was received in a telephone call from a nurse in the practice. A return call to the physician found that the number of reactions was actually six, the sensation was not severe, and there was no local erythema, induration or tenderness. The state epidemiologist’s office was not aware of any other similar reports involving that vaccine lot.

Three reports of anaphylaxis from one Veteran’s Administration (VA) hospital, an unusual number if true. This report was disseminated widely by email at the same time as it was emailed to the FDA. A review of the VAERS database did not find any similar reports involving that lot. Follow-up calls determined that there was only one case of anaphylaxis. The other two cases referred to in the original call were, in fact, pneumonia and chronic obstructive pulmonary disease. Furthermore, inquiries made to other VA hospital locations found that they had not had a similar experience. It was concluded that the initial report of one likely case of anaphylaxis was not unusual, given the large number of people vaccinated. This relieved concerns that could have lead to an interruption in influenza vaccination in the VA system.

Two reports of ruptured aortic aneurysm following influenza vaccination from a small city. This report was received in the evening at the FDA’s emergency number. The information was also disseminated in that small city by word-of-mouth. Follow-up determined that one 65-year-old man only had died of a ruptured aorta approximatley 2 weeks after receiving influenza vaccine. This is unlikely to be related to the vaccine.

The first step in investigating these reports was to consult the VAERS database for any similar reports. FDA physicians immediately sought additional information by telephone, hospital records and laboratory results, to be obtained as soon as possible.

These reports were received early in the influenza season, in November. As of May 2005, no further reports of clusters similar to these three examples have been received. However, it is prudent to remain on alert for any future reports with similar features.

The three situations described point out that firstly, information may come from routes other than VAERS; secondly, the initial report may not be complete; and thirdly, rapid action is important as information that could adversely impact immunization programs may be spreading. To accomplish this, an experienced, well-trained team must be in place.

Follow-up for surveillance summaries

Most surveillance summaries involving VAERS will involve some follow-up of reports. The reasons for this include:

The initial follow-up may not have included information specific to the study plan

The report may not have been followed up previously because it is classified as a nonserious report

The report may have been received before the current enhanced follow-up activities were in place

The surveillance summary may require follow-up of 100% of all reports received within a time period, only a subset of the reports selected by a search of relatively nonspecific coding terms, or by review of submitted report forms.

Three recent surveillance summaries from the FDA are described briefly to show some of the information that can be obtained from a focused follow-up.

A postlicensure review of adverse event reports received for 7-valent pneumococcal conjugate vaccine involved follow-up of reports of seizure. This review found that in 98 reported seizures in children, 79 (80.6%) had a prior history of seizure or fever at the time of the postvaccinal seizure [13]. The other 19 vaccinees had neither fever reported with the seizure nor a medical history that might account for the seizure. After 2 months, 12 of these 19 children did not have additional seizure activity and were not receiving treatment for seizures. The authors concluded that the limitations of VAERS precluded attributing the reported seizures to 7-valent pneumococcal vaccine, but the vaccine may have played a role in some cases.

In a review of Stevens–Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) reports submitted to VAERS, the initial query of the VAERS database for reports coded SJS, TEN, dermal necrosis and serious erythema multiforme yielded 131 reports [14]: 35 reports were selected for follow-up after initial review revealed that the majority of coded reports were not SJS/TEN. Additional information was sought by telephone to decide if the reports met the case definition and if there were any obvious triggers for the onset of disease. One definite and five probable cases of SJS/TEN met the definition of SJS/TEN after immunization without obvious triggers. Thus, it was possible to conclude that SJS and TEN might very rarely be caused by vaccination, but controlled studies would be needed to confirm this hypothesis.

A third example of late follow-up is a study of hypotensive–hyporesponsive episodes (HHE) after vaccination [15]. An initial search of 40,000 VAERS reports received during a 3-year period using a computer algorithm yielded 435 possible reports of HHE. After attempts to follow-up all of the possible reports, it was found that 215 (49%) met the case definition. In addition, it was found that 93% of the children with HHE had received a pertussis-containing vaccine, and 98.6% of the children returned to their original state of health within a median time of 6 h. In the 1–3 years following the HHE and the end of the study, 66.5% of the children had one or more subsequent vaccinations or vaccine components with-held because of the HHE event.

Patient confidentiality

Adverse event reporting, and especially follow-up activities, must respect patient privacy laws. These laws vary from country to country and may impact government agencies and drug manufacturers differently. In the USA, the applicable law is the Health Insurance Portability and Accountability Act, HIPAA [16]. This act contains a specific exception concerning the release of protected information to public health authorities, such as the FDA and CDC.

Summary

Follow-up of possible vaccine adverse events is usually necessary to obtain a complete and accurate description of the event. An organized approach and experienced staff are necessary due to the volume of reported events. The information that can be obtained by follow-up is vital in investigating acute concerns, such as possible lot-related issues, for future focused reviews of newly licensed vaccines or specific adverse events, as well as for other purposes.

Acknowledgements

The author thanks all those individuals who submitted reports to VAERS. The VAERS working group is: Robert Chen, John Iskander, Elaine Miller, Scott Campbell, Vitali Pool, Penina Haber, Sean Shadomy, Alena Khromova at the CDC; Robert Ball, Jane Woo, Soju Chang, Marthe Bryant-Genevier, Ann McMahon, Dale Burwen, Hector Izurieta, Miles Braun, Phillip Perucci at the FDA; Vito Caserta, Geoffrey Evans at the Health Resources and Services Administration.

Possible vaccine adverse events should be reported to the VAERS program (Tel: [phone] or www.vaers.hhs.gov), the pharmaceutical company or both.

•	Severity of event
•	Frequency of event
•	Newly recognized event
•	Time since immunization
•	Current national interest
•	Organized immunization campaign
•	Research project
•	New vaccine