Indian Pediatrics 2001; 38: 453-460
Pneumococcal Conjugate Vaccine– Relevance for Developing Countries
It has long been recognized that immunity to pneumococcal disease is conferred by immunity to the polysaccharide capsule of the pneumococcus. In general, this is sero-type specific, with at least 90 serotypes of pneumococcus presenting a daunting target for vaccine developers. However, the serotype problem is simplified by the fact that most pneumococcal disease is caused by a relatively small number of serotypes, and most of the serotypes can be grouped together into small, cross reacting groups(6). The latter is the basis for the system of naming serotypes in which a number designates the serogroup number and letters define members of the group (e.g., 6B, 9V, 19F). Where no cross-reacting serotypes exist, the serotype is known by a number only (e.g., 1, 2, 5). The first attempt to construct a vaccine from pneumococcal polysaccharide was under-taken in 1911, and by the time of the first introduction of penicillin considerable progress had been made(7). The dramatic curative properties of penicillin convinced the medical world that there was no need for a pneumococcal vaccine and for two decades there was little activity in the field. In the 1960s the work of Dr. Robert Austrian and colleagues in South Africa revived interest in the field, demonstrating the protective efficacy of a 14 valent polysaccharide vaccine in South African mine workers(8). This led to further refinement of the vaccine and in 1983 the 23 valent pneumococcal polysaccharide vaccine was licensed. That vaccine is used to a variable extent in Western countries to protect adults and high-risk children, but doubts remain about its efficacy for preventing pneumonia and its safety in HIV infected individuals.
The role of the polysaccharide vaccine in young children has never been clarified. A study by Riley and colleagues in young children (6 months to 3 years of age) in the highlands of Papua New Guinea demons-trated a significant reduction in all cause mortality and pneumonia related mortality in recipients of the polysaccharide vaccine(9). These findings have never been accepted by the international scientific community because of perceived flaws in the design and analysis, but those flaws have never been clearly presented. In 1989 plans to conduct a formal study of pneumococcal polysaccharide vaccine in 3 doses in Gambian infants were abandoned because of the demonstration of poor immunogenicity in Gambian infants(10), uncertainty about the Papua New Guinea findings and the expectation that a superior protein-polysaccharide conjugate vaccine would be available imminently. No further studies have been conducted on the use of the polysaccharide vaccine in infants, and the strategy is not used anywhere in the world, not even in Papua New Guinea.
Part of the reason for the optimistic view of the prospects for a pneumococcal conju-gate vaccine, even before a prototype had been developed, was based in the spectacular success of the conjugate vaccine against Haemophilus influenzae type b (Hib)(11). That vaccine, which was licensed for use in infants in the USA in 1990, has virtually eliminated Hib disease wherever it has been introduced with a combination of high level individual protection and an unexpectedly high degree of herd immunity. Despite this, a decade later Hib conjugate vaccines are still not widely used in the developing world (outside of the Americas) because of a combi-nation of poor understanding of the epidemio-logy of Hib disease and high prices demanded by the multinational vaccine producers. Market forces have now reduced the price of the vaccine and the Global Alliance for Vaccines and Immunization is endeavoring to make the vaccine available to the poor children of the world. Poor understanding of the burden of disease remains a barrier, particularly in Asia(12).
In the developing world Hib causes disease, predominantly in two forms, menin-gitis and pneumonia. If a lumbar puncture is undertaken at the appropriate time and good bacteriological methods employed, Hib can be isolated from the cerebrospinal fluid and a definitive diagnosis of Hib meningitis made. The diagnosis of Hib pneumonia is much more difficult. Indirect methods are un-reliable, and the only definitive diagnosis is by the isolation of Hib from blood, lung fluid or pleural fluid. In Gambia a series of pneumonia etiology studies had suggested that 5-10% of severe pneumonia was due to Hib(13). To the surprise of the investigators the trial of PRP-T Hib vaccine in Gambian infants prevented 20% of severe pneumonia in vaccine recipients indicating that at least half of the burden of Hib pneumonia was hidden to conventional methods(14). Similar principles apply to the understanding of the burden of pneumococcal pneumonia and this will be an important outcome of the current generation of pneumococcal vaccine trials (15,16).
By the mid 1990s at least four companies were working on pneumococcal conjugate vaccines. For a carrier protein, most used the same protein as they had in their Hib vaccines. Thus, one company used the outer membrane protein of Neisseria meningitidis Group B, another concern used the mutant diphtheria toxin known as CRM197 and a third used both diphtheria and tetanus toxoids(17). For each producer an important question was what serotypes to include in the vaccine. Analysis of serotypes causing disease around the world showed that in most parts of the world 9-11 serotypes covered most of the pneumococci causing disease(18,19). Spatial and temporal fluctuations limited the conclusions that could be drawn from individual studies(20), but gradually the concept of a family of regional vaccines covering different groups of serotypes was abandoned in favor of a global vaccine that could be used all over the world. The exception to this rule was the USA. In that country serotypes 1 and 5 are not important, whereas in most parts of the rest of the world they are important pathogens. As a consequence the first pneumococcal conjugate vaccine covering 7 serotypes (4,6B,9V,14, 18C,19F,23F) which was licensed for use in infants in the USA in February 2000, lacks those important serotypes and thus cannot be considered an optimal vaccine for global use.
By the end of the year 2000 three producers had products suitable for global use (covering 9 or 11 serotypes) in advanced stages of development. The 7-valent Pnc-CRM won the race for US licensure, although the high price of the vaccine ($58 per dose for a 4-dose schedule) appears to be inhibiting the use of the vaccine in the US. This is despite favorable recommendations from the US Immunization Practices Advisory Committee (ACIP) that should eventually see pneumo-coccal vaccines used universally for US infants. This licensure was based on the excellent efficacy against invasive pneumo-coccal disease demonstrated in the Northern California study(21). In that study the main outcome of interest was invasive pneumo-coccal disease although in fact most cases were febrile bacteremia episodes (febrile infants with a positive blood culture but no focus of infection). At the time of completion of the California study another study was underway among the native American population of Arizona. That study employed cluster randomization in order to examine herd effects. Unfortunately, recruitment in the study was stopped before sufficient cases had accrued to evaluate efficacy against invasive disease in this high-risk population. Results of the analysis for radiological pneumonia should be out soon.
Meanwhile two other studies are underway with the 9-valent Pnc-CRM, which contains serotypes 1 and 5. Enrolment was completed in a study in Soweto, South Africa in September 2000 and results should become available in 2001. That study should demonstrate the ability of the vaccine to prevent radiological pneumonia and possibly invasive pneumococcal disease in an African community with a high prevalence of HIV infection. Another study with the 9-valent Pnc-CRM vaccine started in The Gambia in late 2000. That study is being conducted in a rural African population with a high infant mortality rate. The study is of particular importance because it addresses the key public health questions that will define the utility of the vaccine: effect on all cause mortality, effect on clinical and radiological pneumonia and effect on culture proven pneumococcal disease. It is anticipated that results of that study will not be available for at least 5 years.
Meanwhile two vaccine manufacturers have 11-valent vaccine in advanced stages of development. These cover serotypes 1,3,4,5, 6B,7F,9V,14,18C,19F and 23F. The vaccine produced by one of them uses both tetanus toxoid and diphtheria toxoid as carrier proteins, while the vaccine produced by the other uses the conserved Haemophilus influenzae protein, Protein D. An efficacy trial of the former vaccine began in the Philippines in July 2000. It will be several years before the effectiveness of these vaccines is established.
For many years researchers have been looking for pneumococcal protein antigens that could provide protection against all serotypes. Such vaccines would have the advantages of being immunogenic in early infancy, relatively inexpensive to produce and suitable for production in large volumes for the developing world. Most attention has been placed on pneumococcal surface protein A (PspA), but promising results have also been found with pneumococcal surface adhesin A (PsaA) and a genetically modified pneumo-lysin(22-25). So far only PspA has entered phase 1 trials in humans and early results indicate that the sera of vaccinated adults protect experimental animals against pneumo-coccal challenge. Many experts in the field believe that these vaccines offer the best hope of controlling pneumococcal disease amongst children in the developing world.
Up to now the research agenda for pneumococcal conjugate vaccines has been largely dictated by the needs of industry. Thus a four-dose schedule involving a primary series of three doses and a booster has been evaluated, even though data from the Californian study suggest that fewer doses may also be efficacious. There is a need for independent investigators to investigate simpler, inexpensive regimens that may be more suitable for developing countries. Studies investigating such regimens should take into account two important factors. Firstly, a substantial proportion of the serious pneumococcal disease occurring in develop-ing countries occurs in early infancy, often too early to be prevented by a vaccine given with DTP(26). Secondly, there is considerable evidence that infants primed with pneumo-coccal conjugate vaccine can elicit a booster response with the inexpensive polysaccharide vaccine, even when the booster is given as early as six months of age(27,28). There are two approaches to covering infants in early infancy. Maternal immunization with a pneumococcal polysaccharide vaccine has been tried and is still under evaluation. Including a neonatal dose in an infant regimen is the other potential strategy. Investigation of both strategies has been recommended by WHO(29). Since their first licensure for use in infants in 1990, the use of Hib conjugate vaccines in developing countries has been greatly impeded by lack of understanding about the burden of Hib disease, especially Hib pneumonia. Now, a decade later pneumococcal conjugate vaccines are being introduced, yet our understanding of the burden of pneumococcal disease is less than we had with Hib disease in 1990. Particularly lacking is information about the burden of pneumonia, a proporation of which will be pneumococcal in origin. Previous attempts to address this have been confounded by lack of a clear case definition(30). Data on the burden of pneumonia are required from all areas, yet it is essential that the data are comparable and can be related to the findings of the field trials of pneumococcal vaccines currently under-way. A group at WHO in Geneva is currently working on the standardization of radiological definitions to be used in pneumococcal vaccine trials in the expectation that this will be accepted as the standard measure of culture negative pneumonia burden. Thus, any future pneumonia burden studies need to express their results using the same definitions, so that estimates of the vaccine-preventable burden of pneumonia can be calculated using the results of the vaccine trials.
Another important component of the global burden of pneumococcal disease is otitis media. The pneumococcus is known to be an important cause of otitis media(31) yet remarkably little is known about the epidemiology and etiology of otitis media in developing countries. It is likely that this leads to substantial short-term and long-term morbidity that is generally not appreciated.
After the euphoria that followed the publication of results from the California pneumococcal vaccine trial, a couple of potential problems have arisen that will need to be specifically addressed as the field moves forward. In the Californian study the two invasive disease failures were both serotype 19F as were all six culture positive otitis failures(21). Furthermore, in the Finnish otitis media study the efficacy of the same vaccine for the prevention of acute otitis media due to any vaccine serotypes was 57%, (95% confidence interval 44-67%) while the efficacy against acute otitis media due to serotype 19F pneumococcus was only 25% (95% CI-14, 51)(32). These findings are despite the apparently good immunological response to that serotype, raising questions about the validity of antibody titers as surrogate markers of immunity following pneumococcal conjugate vaccination. Of greater concern is the issue of serotype replacement(33). There has long been a theoretical concern that the use of a pneumococcal vaccine containing only some of the 90 serotypes will open the way for colonization and disease by non-vaccine serotypes rushing to fill the niche vacated by the vaccine types. This phenomenon was found during studies of pneumococcal carriage following conjugate vaccination in Gambia and South Africa(34,35). The first demonstration of serotype replacement in disease was found in the Finnish otitis media study in which vaccine recipients had more disease caused by pneumococcal serotypes not included in the vaccine(32). At this stage it is not known whether this will occur with pneumonia cases, potentially offsetting or negating efficacy against pneumonia. It is essential that these issues be addressed carefully in subsequent studies or vaccine introduction exercises.
As penicillin resistance among pneumo-cocci is largely restricted to the serotypes included in the conjugate vaccines, pneumo-coccal vaccination has the potential to reduce the prevalence of penicillin resistant pneumo-cocci in a community. This has already been demonstrated in South Africa(34). However, the impact of pneumococcal vaccination on antimicrobial resistance could go far beyond that. Most antibiotic resistance among strains of S. pneumoniae and H. influenzae stems from exposure of the organisms to (usually orally administered) antibiotics while colonizing the nasopharynx of young children who being empirically treated for (usually viral) respiratory infections. The approach of physicians to empiric management of a young child with fever or cough and fast breathing should be altered if the child is known to have had Hib and pneumococcal vaccines. The case for routine broad spectrum antibiotics becomes quite weak, and many more cases could be managed without antibiotics if the risk of Hib or pneumococcal disease is very small. Similarly, the empiric management of serious conditions like osteomyelitis and meningitis should be affected by knowledge of the child’s vaccination status. Thus the introduction of Hib and pneumococcal conjugate vaccines into a community provides the ideal opportunity for a drive to enforce the judicious use of antibiotics(36).
Initial results with pneumococcal conju-gate vaccines are encouraging, but it is not possible to predict the true utility of these vaccines in the developing world until the results of studies focusing on impact on pneumonia and mortality in developing countries are made available. In the mean time studies are urgently needed to define the vaccine preventable burden of pneumococcal disease in countries considering pneumo-coccal vaccination, and to evaluate regimens with pneumococcal conjugate vaccine that may be more suited to the needs of developing countries. Even if all these issues are addressed and the results prove to be favorable from the point of view of the vaccine, pricing and supply issues are likely to limit the use of pneumococcal conjugate vaccines in the developing world for some time to come. These issues need to be addressed aggressively to ensure that, once available, a lifesaving pneumococcal vaccine is made available to the children most in need, the poor. Meanwhile it is essential that development and evaluation of the new generation of pneumococcal protein vaccines proceed at full speed.
Competing interests: None stated.