Beneficial and adverse side effects of vaccination: perception in developing countries

From the time that vaccination first came into existence, the mere evaluation of a specific effect (i.e., vaccine efficacy), which would appear to be relatively straightforward, has been, and remains, a matter of much debate [1]. A noteworthy example was the controversy between Pearson and Wright in the British Medical Journal in the early 20th century [2]. Current controversy has now generally centered around the adverse side effects that might result from a temporal association between vaccination and a disease [3]. Beyond the evident beneficial effects of vaccines, however, some favorable side effects are also known [4,5].

Reported adverse effects are more numerous than favorable effects, but the question is, why? Good health is the normal condition of humans, whereas illness is less easily accepted, leading us to tend to search for a ‘supernatural’ scientific explanation when there is no clear-cut etiology. It should be noted that adverse effects associated with vaccination principally concern illnesses with unexplained etiology (multiple sclerosis, autism, sudden infant death syndrome, diabetes mellitus, paralytic poliomyelitis, simian virus 40 infection, Gulf war syndrome, chronic inflammatory arthritis) [6]. Moreover, the initial adverse effects reported following hepatitis B vaccination did not mention multiple sclerosis [7]. The refuting of falsely reported adverse effects requires a great deal of time and a number of studies: examples include those linking hepatitis B vaccine to multiple sclerosis [8,9] and measles, mumps and rubella (MMR) vaccine to autism [10,11], among others.

Beliefs differ in developing countries. Among those living in sub-Saharan Africa, a particular vaccination is not necessarily associated with the prevention of a specific disease; for example, polio vaccine might be assumed to protect against malaria, diarrhea or pneumonia [12]. In general, vaccination is well accepted in developing countries. For example, in Burkina Faso, vaccination resembles a local incision technique practiced by herbalists, which infuses health into children’s bodies [13]. Conversely, in 1997 in Uganda, a national vaccination campaign was followed 4 months later by a malaria epidemic characterized by very high mortality; the following year, immunization coverage dropped from 98 to 67% because the epidemic was believed to be linked to the vaccinations [12]. In 2003, the belief that the polio vaccine caused infertility and transmitted HIV stopped the polio eradication campaign in Nigeria and spread the virus to countries previously free of the disease [14].

When vaccination is associated with mortality

Observations on vaccination and overall mortality in countries with generally high mortality rates have triggered debate on their favorable or unfavorable outcome. The initial aim of an Expanded Program of Immunization (EPI) was to reduce infant mortality by targeting a specific disease or cause of mortality. Cohort studies had noted a decrease in overall nonspecific mortality after a measles vaccination campaign [15].

Conversely, side effects might include higher mortality for vaccinated children: indeed, in Guinea-Bissau, the combination of one dose of diphteria, pertussis and tetanus and polio vaccines seemed to increase mortality [16], although most recent studies have observed lower mortality [17–21]. The question of causal inference remains open.

Central to the debate is the quality of information concerning exposure to the vaccines and comparability of the outcome. Likewise, the method used is of importance. Most studies were performed according to a cohort design, replicating the initial observations from Guinea-Bissau, and none adopted either case control or its developmental designs, such as nested case-control and case-only studies [22]. Randomized controlled trials, in which only some children would be vaccinated could not be used to study the association between vaccines and general mortality for evident ethical vaccination, although the invesitigators could choose to vaccinate children at different ages so as to determinate the most favourable calendar. Thus, observational, case-control and follow-up studies are the only means for assessing whether a particular vaccine has an unfavorable impact upon or an unexpected benefit in terms of mortality.

Much of the literature deals with comparisons between randomized controlled trials and observational studies. Depending on the authors, randomized controlled trials and observational studies are presumed to either give [23] or not give [24] comparable results. Comparisons of case-control and follow-up studies have been less frequently carried out [25]. However, the debate concerning experimental and observational studies should include case-control strategies and not simply follow-up strategies.

Concerning mortality following vaccination, follow-up should be used for assessment of vaccine status, and case-control strength should be used to compare cases and controls, in order to overcome the intrinsic difficulties linked to the fact that the exposure status (vaccinated vs unvaccinated) is not random and might be related to confounding factors. Much can be gained by not restricting our considerations to the field of vaccines, but instead opening them up to include general aspects of pharmacoepidemiology, a field in which, as with environmental epidemiology, methodologic development is foremost [26].

Nevertheless, finding relationships in data from observational studies is a complex process. A hypothesis had been advanced that, “Bacillus Calmette–Guérin (BCG) and measles vaccines provide protection against death from infections other than tuberculosis and measles, while diphtheria, tetanus, pertussis (DTP) does not provide such protection. Live vaccines may activate the innate immune system, while killed vaccines with alum adjuvants do not” [27]. Confirmation of this hypothesis is not easy, as most vaccinated children receive several vaccines, and it is thus difficult to assess the association between a particular vaccine and mortality. In particular, the reduction of mortality among measles-vaccinated children did not take into consideration other vaccines, and since the measles vaccination was the most recent, the comparison was made between vaccinated and nonvaccinated children. Several studies have found that BCG has a greater effect when immunization takes place during the first month of life, which implies that we need to examine more closely the relationship between the vaccination calendar and mortality in less developed countries [18]. Among DTP-vaccinated children, studies in the USA, the UK and France have concluded that there is a reduction in sudden infant death syndrome mortality, whereas studies in Guinea-Bissau and Benin have found an increase in general mortality [16]. Priority should be given to evaluating the differences between vaccinated and nonvaccinated children, which is associated with other variables, in particular, the general level of care given by the mother.

Studies on this subject place stress on acceptance [28] or refusal of vaccination for improving vaccination coverage, and not on differences that may influence confounding factors. “More qualitative research is needed to provide relevant knowledge about psychosocial and behavorial variables. The methodologic challenge is to find ways to examine the complexity in the analysis of data” [29].

Allegation of negative effects have a mostly negative impact on the level of immunization. Outbreak of epidemics following drastically reduced vaccination coverage [30] resulting in higher morbidity [31] and mortality [32] have been reported. However, such allegations should be openly and properly adressed in order to maintain and improve research on vaccines, a less profitable field than that of drug production. The vaccine industry represents only 2% of the drug industry and the number of manufacturers is decreasing, while there is a great need for new vaccines against tropical diseases.