by M. Miles Braun, MD, MPH
The Vaccine Adverse Event Reporting System (VAERS) was started in 1990 as a national passive surveillance system for adverse events occurring after vaccination1, 2. VAERS is administered jointly by the Centers for Disease Control and Prevention's National Immunization Program and the Food and Drug Administration's Center for Biologics Evaluation and Research.
About 11,000 reports are received by VAERS each year:10-15% involve hospitalization, permanent disability or considered life-threatening, about 2% involve deaths. Reports of adverse event are accepted by VAERS regardless of whether the vaccination is shown to have caused the adverse event.
Reporting to VAERS is voluntary and anyone may report vaccine adverse events. Reports are made by physicians, nurses, vaccine manufacturers, parents, persons affected by adverse events and others. The simple one-page reporting form may be found at the end of the Physicians Desk Reference or may be ordered by telephoning 1-800-822-7967.
Additional information about VAERS is available at http://www.fda.gov/cber/vaers.html, and the VAERS form may be downloaded or printed from this website.
The most salient positive attributes of VAERS are described below.
Vaccines recently licensed in the United States are highly scrutinized in VAERS. Some of the reasons for this include:
- Prelicensure studies include thousands to tens of thousands of vaccine administrations. However, because adverse events occur in less than one per 20,000 vaccinations, it is possible that none will be detected before licensure.
- Individuals in subgroups, such as those with certain medical conditions, may not have been represented in the prelicensure trials.
- The numerous possible combinations of vaccines (as well as drugs and biologics) that may be concomitantly administered with the new vaccine may not have been represented in the prelicensure trials.
Several vaccines have been licensed since the inception of VAERS including vaccines against hepatitis B (infant indication), hepatitis A, varicella, and acellular pertussis vaccines. Postlicensure evaluations using VAERS data of hepatitis B vaccine in infants3 and of acellular pertussis vaccine for the 4th and 5th pertussis doses4 have confirmed the safety of the vaccines for the general population; other such postlicensure evaluations are ongoing.
Timeliness And Availability of Data. Reports to VAERS are rapidly coded and computerized. After computer entry, the reports are available for analysis the following day by scientists at both the Centers for Disease Control and the Food and Drug Administration. The general public may access VAERS data (except for personal identifiers) by acquiring computerized datasets through the National Technical Information Service (Tel. 703-487-4650). Certain data requests may also be filled in accordance with the Freedom of Information Act (Tel. 800-827-2000).
Lot Surveillance. Largely because of the timeliness of the data, the Food and Drug Administration uses VAERS in its ongoing surveillance of all vaccine lots in use in the United States. Serious reports received by VAERS are reviewed weekly. Serious non-fatal reports and death reports are tabulated by vaccine lot and, when concerns are raised about a particular lot, reporting rates are calculated (taking into account the amount of time the lot has been in use). To date, such surveillance using VAERS has identified no lot-specific vaccine problems.
Rare Disease Registry Useful For Case Series. Adverse events following vaccination are rare and some are extremely rare, occurring in about 1-10 per 1,000,000 vaccinations or less. Examples of such rare adverse events include anaphylaxis and Guillain-Barré syndrome. For such rare events, a national surveillance system such as VAERS is especially useful to gather enough cases for useful study. In certain circumstances, such case series may serve as the "case" component of case-control studies that search for causal factors in vaccine reactions.
The Vaccine Adverse Event Reporting System also serves as a registry of rare adverse events following immunization. Case series of VAERS reports may be compiled to elucidate the events' characteristics, etiologies, and possible preventive measures. For example, a case series of reports to VAERS of thrombocytopenia following immunization with attenuated measles virus-containing vaccines demonstrated that the severity of this thrombocytopenia could be more severe than was previously recognized5. Another study from VAERS was the first case series to study syncope after immunization. The syncope study demonstrated the very close temporal association between syncope and vaccination, the greater reporting of this adverse event among children and young adults, and, perhaps most importantly, the rare occurrence of severe injuries such as skull fracture6. Alopecia following vaccination was first described in a case series from VAERS7. The evidence for this vaccine-adverse event association was strengthened by the demonstration of "positive rechallenge" in several of the cases.
When 2 or more administrations of a vaccine (or drug) to the same person are followed by the same adverse event, as occurred with several cases in the alopecia-vaccination case series mentioned above, positive rechallenge may be said to occur. Positive rechallenge represents stronger evidence than the simple temporal association of a single adverse event occurrence, and therefore positive rechallenge represents a useful pharmacoepidemiologic indicator when analyzing case series information.
Geographic and Demographic Coverage. VAERS is a national reporting system which covers vaccinations occurring in the entire United States - in the public and private sectors, as well as the military. A descriptive epidemiologic overview of VAERS8 demonstrated that reporting rates were significantly higher in some states than others and that the states with highest reporting rates had higher proportions of reports from the public sector. In addition to domestic reports, international reports of serious vaccine adverse events involving vaccines licensed in the U.S. are also submitted to VAERS by vaccine manufacturers.
Seriously mistaken conclusions can be made by interpreting VAERS data without taking the important limitations of VAERS into account.
Accuracy of Reports. VAERS is a passive surveillance system, and the large number of reports to VAERS precludes checking the specifics of the vast majority of reports, especially the less serious ones. Few reports to VAERS include full medical record documentation, and some doubtless include errors. In addition, the VAERS forms often have missing data; even simple data such as age, sex, vaccination date and adverse event onset date are missing for about 11%, 8%, 10% and 14%, respectively, of domestic VAERS reports.
Underreporting. Underreporting is an inherent problem of passive surveillance systems, including VAERS. The degree of underreporting varies according to the adverse event. For example, one study estimated that 68% of cases of vaccine-associated polio are reported to VAERS, but only 4% of MMR-associated thrombocytopenia are reported9. This variability in undereporting can make it hazardous to assume that the relative frequencies of adverse events in VAERS reflects their relative rates of occurrence. In addition, for new products on the market, increased reporting of adverse events may occur; this has been termed the "Weber effect"10.
Overreporting. Adverse events may be included in the VAERS database that are not accurate descriptions of the event that occurred, and erroneous diagnoses may be reported. For example, a case of simple fainting after vaccination may be incorrectly reported as anaphylaxis, a potentially life-threatening condition.
Overreporting may also result from reports that describe adverse events for which a definitive diagnosis has not yet been reached. For example, if the reporter writes on the reporting form a diagnosis of "rule out meningitis", that VAERS report will be computer coded as meningitis. Unless specific follow-up is received indicating that meningitis was ruled out, the meningitis coding term will remain associated with the computerized VAERS report.
Adverse Event Incidence And Trends. As a result of the overreporting and underreporting issues described above, drawing conclusions from VAERS data about how many adverse events occurred in the United States must be done extremely carefully. Similarly, because the factors affecting overreporting and underreporting may vary over time, drawing conclusion from observations of changes, or lack of changes, in the frequency of adverse events in VAERS over time (trends) may be hazardous. In addition, numbers of adverse events reported to VAERS will at least partially reflect the number of doses of vaccine administered. For example, VAERS reports reflect the current trend of a decline in usage of whole-cell pertussis-containing vaccines and a simultaneous increase in the use of acellular pertussis vaccines; however, it would be incorrect to draw conclusions about the relative safety of these vaccines from such trend data. Finally, the number of doses of vaccine administered according to age is not part of the VAERS database, and therefore actual rates of adverse events are not strictly calculable.
Causal vs. Temporal Association. To encourage complete reporting to VAERS reporters are not expected, or asked to, make a determination about whether the adverse event they report is caused by vaccination. Although certain adverse events are clearly due to vaccination, such as injection site reactions, a large proportion of adverse events cannot be clearly causally linked to vaccination; the most that can be concluded is that the adverse event followed vaccination. This lack of clear causality is particularly common among serious adverse events and deaths, and is largely due to the fact that there are no laboratory tests or clinical signs that can determine whether the majority of vaccine adverse events were caused by vaccination. These difficulties can lead investigators, when there is sufficient interest, to mount epidemiologic studies to determine whether vaccination is associated with specific adverse events.
For example, hundreds of reports of Sudden Infant Death Syndrome (SIDS) have been made to VAERS since 1990. Because SIDS occurs in an age range during which vaccination also occurs, the occurrence of SIDS in temporal association with vaccination may simply reflect chance. Indeed, the possibility of a link between SIDS and DTP vaccination was assessed by a committee of the Institute of Medicine11. The committee evaluated epidemiologic studies of this issue and came to the conclusion that the evidence "does not indicate a causal relation" between SIDS and DTP vaccine.
Many steps occur to assure the safety of vaccines--starting with laboratory testing, animal studies, and several phases of clinical testing in humans. Subsequent to a vaccine's licensure by FDA, each lot of vaccine goes through pre-release testing. VAERS is the repository of reports of adverse events associated with these vaccines. VAERS reports are monitored by FDA and CDC. Despite its important limitations, VAERS' multiple useful aspects make it an important component of the federal vaccine safety effort.
3. Niu MT, Davis DM, Ellenberg S. Recombinant hepatitis B vaccination of neonates and infants: emerging safety data from the Vaccine Adverse Event Reporting System. Pediatr Infect Dis J 1996;15:771-6.
4. Rosenthal S, Chen R, Hadler S. The safety of acellular pertussis vaccine vs whole cell pertussis vaccine. Arch Pediatr Adolesc Med 1996;250:457-60.
5. Beeler J, Varricchio F, Wise R. Thrombocytopenia following immunization with measles vaccines: review of VAERS reports (1990-1994). Pediatr Infect Dis J 1996;15:88-90.
6. Braun MM, Patriarca PA, Ellenberg SS. Syncope after immunization. Arch Ped Adolesc Med 1997;151:255-9.
7. Wise RP, Kiminyo K, Salive ME. Hair loss after routine immunizations. JAMA 1997;278:1176-8.
8. Braun MM, Ellenberg SS. Descriptive epidemiology of adverse events following immunization: reports to the Vaccine Adverse Event Reporting System, 1991-1994. J Pediatr 1997;131:529-35.
9. Rosenthal S, Chen R. The reporting sensitivities of two passive surveillance systems for vaccine adverse events. Am J Public Health 1995;85:1706-9.
10. Weber JCP. Epidemiology of adverse reactions to non-steroidal antiinflammatory drugs. Advances in Inflammation Research 1984;6:1-7.
11. Institute of Medicine. Adverse Effects of Pertussis and Rubella Vaccines. Howson CP, Howe CJ, Fineberg HV, eds. Washington:National Academy Press, 1991.