Unvaccinated Controls

Discussion arose about the issue of the lack of unvaccinated controls in the studies under consideration. Some participants felt that a true control group in a study of vaccines and adverse events would consist of never-vaccinated children (whereas now, control groups often consist of children who had not been vaccinated recently). It would then be possible to look at the frequencies of adverse events and disease in these children in comparison with the frequencies in those who had been vaccinated. This might be particularly relevant in studying adverse events with long latencies from the time of vaccination. There was concern by some that unvaccinated individuals are so different from vaccinated individuals that these differences cannot be controlled for; that is, completely unvaccinated individuals might possess other health or socioeconomic traits if they are from a population that does not believe in traditional medical treatments, or they might represent a limited gene pool if the population consists of a longstanding, religious sect. One participant argued that comparison with unvaccinated controls might make vaccines look more beneficial than they are, because those who are not vaccinated are more likely to be underprivileged or have more diseases or disorders for other reasons. The decreased incidence of SIDS following immunization with DPT is an example of this (Institute of Medicine, 1991). Others noted that there are individuals who would be willing to participate, or have their children participate, as unvaccinated controls, and it was suggested that statistical methods could be used to address the problem of the self-selection of the group. Still others believed that the number of unvaccinated subjects is too low for detection of rarely occurring adverse events.

One participant noted that religious groups that decline vaccination have occasionally been studied during acute outbreaks of disease and speculated that it might be possible to work with them in attempting to find the background rates of particular adverse events in unvaccinated children. It was noted that the possible problem of a restricted gene pool would have to be taken into consideration, particularly for certain adverse events.

Background Incidence Rates

The need for more complete information on the background incidence rates of the adverse events was discussed. It was suggested that vital statistics data or other studies of infant mortality be exploited for ecologic correlations, that is to look for changes in mortality in periods before and after new vaccines are introduced. Some participants thought that this was a fairly crude approach to the subject and that it should be looked at in great depth. It was noted that a vaccine would have to account for a large proportion of deaths (and there is no reason to believe that any vaccine does) to show a difference using crude mortality data. In addition, one participant thought that there would be too much seasonal and other variation for the infant mortality approach to be helpful unless a concurrent, controlled study were to be done.

It was noted that the question of background incidence rates was also of concern to the IOM vaccine-adverse event committees (Institute of Medicine, 1991, 1994a). It may be possible to derive reasonably accurate rates for such conditions as GBS and multiple sclerosis in adults, but when considering such conditions as encephalopathy and encephalitis in children, the conditions are diagnosed so differently among various studies that the terms become almost meaningless. People must agree to use the same diagnostic criteria for the same conditions. Much education is necessary to get good background information. Some felt that this was possible and that having standard definitions then assumes great importance. It was noted that even neurologists have not reached consensus on the definitions of encephalopathy and encephalitis. However, in response to the question of whether swine flu vaccine can cause GBS, a group of neurologists was brought together to develop a definition that has been generally accepted (Asbury and Cornblath, 1990; Asbury et al., 1978).

Adverse Events with Long Latencies

Previous discussions examined the potential utility of case reports in assessing causality. This is particularly true for assessing the causalities of adverse events that occur shortly after vaccination. Although anaphylaxis, for example, can occur without an obvious cause, the relation to vaccination in an individual case is easy to determine if it occurs within minutes of exposure to the antigen. In contrast, problems of assessing causality for adverse events with long latencies from the time of exposure are many. The many discussions in the published literature regarding the British National Childhood Encephalopathy Study (NCES), including a recently released IOM report by the committee to Study New Research on Vaccines (Institute of Medicine, 1994), show the lack of agreement in the scientific community on how best to address these questions.

A participant noted that events with long latencies have traditionally been studied by the case-control method (as was done in the NCES) and that this will probably continue to be the primary way in which they are investigated. The selection of controls, however, is very important. If the coverage rate for the vaccine is high, the required sample size could be large.

It was suggested that it might be useful to identify children who participated in pre-marketing trials and follow them over the longterm. This is not currently done routinely, but it might be a worthwhile effort, even though it would be difficult. Another participant commented that the unexposed group in a randomized trial will, if the vaccine is licensed and in general use in a few years, most likely become vaccinated. In terms of looking at long latencies, there is not much difference between those who received a vaccine 18 years earlier and those who received it 21 years earlier.

Subacute sclerosing panencephalitis (SSPE) was mentioned as an example of the difficulty of dealing with long-latency events. A participant commented on SSPE and multiple sclerosis; in the 1960s, both were believed to be associated with measles disease. Since the advent of the measles vaccine, there has been a 90 percent decrease in the incidence of SSPE, whereas there has been no change in the incidence of multiple sclerosis. A participant commented that the question now is not whether the measles vaccine can prevent SSPE; obviously it can, by preventing disease from wild-type measles virus infection. Still in question is whether there are rare cases of SSPE caused by the vaccine. Although technical advances are making it possible to isolate virus from the brains of patients with SSPE and type the RNA as being of the vaccine strain or the wild-type, the failure to isolate vaccine-strain virus does not prove that the vaccine cannot cause SSPE. As was articulated by the IOM committees (Institute of Medicine, 1991, 1994a) it is impossible to prove a negative. However, typing an isolated virus from a patient with SSPE as vaccine strain would be important data to suggest that measles vaccine can cause SSPE. This is hypothetical at present.

Multifactorial Etiologies

Adverse events that are multifactorial in etiology, for example, insulin-dependent diabetes mellitus, provide particular challenges. In these instances, the vaccine might serve as an additional stimulus to a genetic background that predisposes a child toward the condition. It was mentioned that as more and more genes for diseases are identified, more conditions that fall into this category may be identified.

Once causality between a condition and a vaccine is established, epidemiologic methods could be used to study why some people who received, for example, influenza vaccine got GBS and others did not. What is different about those who develop the disease or condition? But even if one could identify risk factors for adverse events, it may not be feasible or practical to screen for them prior to immunization.

A participant commented that there are a number of suspicions about predisposing conditions that might be exacerbated by vaccines (for example, arthritis and hepatitis B vaccine). Such examples should be defined and looked at in LLDBs. It was suggested that general research projects designed primarily to study the pathophysiology or genetics of conditions such as diabetes might include vaccine issues as part of their study designs.

Need for Basic Research

A participant stressed that it is important to understand the mechanisms of the adverse events. It would be a mistake to address the problem of safety only from the moment of licensure. Basic laboratory research on possible immunopathologic or infectious sequelae of vaccines, in addition to surveillance and epidemiologic studies, is needed.

Problems inherent in all the research strategies (case reports, passive surveillance, LLDBs, and clinical trials) include the implications of not using unvaccinated children or adults as controls, a lack of information on the background incidence rates of the adverse reactions, long latencies for some reactions, and reactions with multifactorial etiologies. Basic research on both vaccines and the pathophysiologies of adverse reactions should be continued.