, Volume 65, Issue 2, pp 229–255

Advances in Pneumococcal Vaccines

Advantages for Infants and Children


    • Department of Clinical Sciences South Bristol, Institute of Child HealthUniversity of Bristol, Level 6, UBHT Education Centre
  • Adam Finn
    • Department of Clinical Sciences South Bristol, Institute of Child HealthUniversity of Bristol, Level 6, UBHT Education Centre
Review Article

DOI: 10.2165/00003495-200565020-00005

Cite this article as:
Bernatoniene, J. & Finn, A. Drugs (2005) 65: 229. doi:10.2165/00003495-200565020-00005


The introduction of Haemophilus influenzae type b (Hib) vaccine into the universal immunisation schedules of many industrialised countries and the subsequent remarkable decline in the incidence of invasive Hib disease has further highlighted the impact of invasive pneumococcal diseases. Streptococcus pneumoniae is now the leading cause of bacterial meningitis in children in many settings and a leading cause of vaccine-preventable bacterial disease in children worldwide.

The currently marketed 23-valent pneumococcal polysaccharide vaccine provides large serotype coverage at a relatively low cost. However, it is not efficacious in young children. Pneumococcal conjugate vaccines (PCVs) are highly effective in preventing invasive disease in infants and young children, with favourable safety and immunogenicity profiles. These vaccines have also shown efficacy in reducing cases of non-invasive disease (i.e. otitis media), nasopharyngeal acquisition of vaccine-specific serotypes of S. pneumoniae, and protection against pneumococcal disease caused by resistant strains. However, PCV contains a limited number of pneumococcal serotypes and, given adequate ecological pressure, replacement disease by non-vaccine serotypes remains a threat, particularly in areas with very high disease burden. Furthermore, although capsular-specific antibodies have been shown to be highly protective, it remains unclear what concentration of these serotype-specific antibodies protect against disease and, more recently, it has become clear that opsonic activity and avidity of these antibodies are more critical determinants of protection than concentration. Therefore, monitoring disease burden and defining immune correlates of protection after widespread use of conjugate vaccines are crucial for the evaluation of these new generation vaccines. Furthermore, a need exists to develop pneumococcal vaccines with lower cost and larger serotype coverage.

Development of one or more protein vaccines that might be easier and, thus, less expensive to manufacture, and which might provide protection against multiple serotypes, is in progress. This article reviews the current state of pneumococcal disease and pneumococcal vaccines in clinical use.

Streptococcus pneumoniae is the leading bacterial cause of infection worldwide, ranging from common infections such as otitis media to life-threatening invasive infections such as pneumonia, sepsis and meningitis.[13] Approximately 2.6 million children <5 years of age die annually of pneumonia, predominantly in the developing world: approximately one-half of these deaths are attributable to pneumococci either solely or in conjunction with a viral respiratory infection, malnutrition or HIV infection.[4,5]

S. pneumoniae is a Gram-positive, facultative bacterium that is commonly encapsulated. Pneumococcal capsular polysaccharides form the basis of serogroup and serotype classifications. There are currently over 90 recognised serotypes.[6] In addition to the capsule, a number of protein antigens are either exposed on the surface or released from the pneumococcus (figure 1). The capsule is a major virulence factor, and antibodies directed against it facilitate opsonisation of the bacteria for phagocytosis and are thus critical in protection against invasive disease.
Fig. 1

Diagram of pneumococcal surface. CbpA (PspC) = choline-binding protein A; lyt A = autolysin; PsaA = pneumococcal surface adhesin A; PspA = pneumococcal surface protein A; TA = teichoic acid.

Asymptomatic nasopharyngeal carriage of pneumococci is widely prevalent among young children in both developed and developing countries,[7] and can lead both to spread of the pathogen to other susceptible individuals or development of disease.[8,9] Described pneumococcal nasopharyngeal carriage rates in children attending daycare centres vary from 17.9% to 50% or higher in developed countries (figure 2).[817] In developing countries, including Gambia,[17,18] Pakistan[19] and the Philippines,[20] rates can be higher still.
Fig. 2

Prevalence of Streptococcus pneumoniae nasopharyngeal carriage in children attending daycare centres in different countries.[817]

S. pneumoniae is responsible for a broad spectrum of invasive and non-invasive diseases, including 25–50% of all cases of bacterial otitis media.[21,22] In the UK, it is estimated that 14% of children have acute otitis media (AOM) during the first year of life, 18% in the second and 12% in the third; S. pneumoniae may be responsible for 180 000–350 000 episodes of disease annually in children aged <5 years.[23] In one 10-year AOM study, S. pneumoniae was the most common bacterial isolate (figure 3).[24] In recent Finnish[25,26] and multicentre studies,[22,27] S. pneumoniae was cultured from middle ear fluid in 18–30% of children with AOM, suggesting that pneumococcal AOM is perhaps the most common reason for paediatric outpatient consultations.
Fig. 3

Distribution of micro-organisms in 2807 ear cultures from patients with acute otitis media in the US.[24]

Invasive pneumococcal disease (IPD) is associated with significant mortality: >1 million children die annually of pneumococcal disease worldwide.[5] The incidence of IPD varies with age and is highest in children <2 years of age (table I). The reported incidence rate of IPD per 100 000 children in England and Wales is 39.7 for those <1 year of age, 14.5 for those <5 years and 6.6 for those <15 years.[28,29] In the US the reported incidence is 166.9 per 100 000 population in children <2 years.[30] S. pneumoniae is one of the most common causes of bacterial meningitis in children >3 months of age. According to the Centers for Disease Control and Prevention, the annual incidence of pneumococcal meningitis in the US in children <5 years of age is 2.5 cases/100 000.[5] Worldwide, pneumococcal meningitis results in 75 000 deaths annually. In developing countries 12–50% of individuals with all types of acute bacterial meningitis die, compared with 5% in developed countries.[31,32] Pneumococcal meningitis usually has higher case fatality rates (21–37%) than meningitis caused by Haemophilus influenzae (2%) or Neisseria meningitides (3–10%).[3134] In a meta-analysis of bacterial meningitis outcomes in developed countries, S. pneumoniae also caused more neurological sequelae than either H. influenzae or N. meningitidis.[35] Of the children with pneumococcal meningitis reviewed in the meta-analysis, 17% developed mental retardation, 14% a seizure disorder, 28% deafness, and 12% spasticity or paresis. The case fatality rate was 15%. Despite effective antimicrobial therapy, the associated morbidity (20–30%) and case fatality rates have changed very little in the past 30 years.[31,36,37] Rates of non-bacteraemic pneumococcal pneumonia in young children are estimated to be two to ten times those of invasive disease (table I). However, the true incidence of pneumonia in children in the community is not known.
Table I

Incidence of invasive pneumococcal disease in children in selected countries

A protein-polysaccharide conjugate vaccine (PCV7; Prevnar®) was licensed for use in infants and young children in US and is now licensed for use in a number of other countries, including those in the EU. Clinical trials and post-licensure surveillance indicate that this conjugate vaccine is highly efficacious against IPD. Effects on mucosal infection are more modest, with reductions in nasopharyngeal colonisation and otitis media due to vaccine serotypes[5058] counterbalanced, at least to some extent, by increases in non-vaccine serotypes.[50,5355,59,60] Notwithstanding this, important beneficial herd immunity effects have recently been observed in the US.[61] The possibility of further changes in the epidemiology of pneumococcal disease following the introduction of pneumococcal conjugate vaccines (PCVs) is a reason to continue active surveillance of disease.

In this review we discuss the current epidemiology of paediatric pneumococcal disease and the development, application and efficacy of new pneumococcal vaccines in paediatric medicine.

1. Pneumococcal Vaccines

1.1 Composition

Current pneumococcal polysaccharide vaccines (PPV23) contain capsular polysaccharides of 23 pneumococcus serotypes,[62] which were chosen on the basis of the relative distribution of the individual serotypes that cause around 90% of invasive pneumococcal infections.[6368]

The 7-valent pneumococcal conjugate vaccine (PCV7) includes seven purified capsular polysaccharides of S. pneumoniae, each coupled to a non-toxic diphtheria protein analogue (cross-reactive material; CRM) CRM197. Each 0.5mL dose contains approximately 2μg each of capsular polysaccharide from serotypes 4, 9V, 14, 19F and 23F, and oligosaccharide from 18C, 4μg of serotype 6B, 20μg of the carrier protein CRM197 and 0.125mg of aluminium phosphate adjuvant. The 9-valent vaccine (PCV9) contains two additional conjugated saccharides, types 1 and 5; and the 11-valent vaccines (PCV11) contain another two, types 3 and 7F. Several carrier proteins have been utilised in development of conjugate vaccines, including H. influenzae protein D, N. meningitidis outer-membrane protein complex (OMPC), a non-toxic variant of diphtheria toxin (CRM197), diphtheria and tetanus toxoids (table II).
Table II

Safety and immunogenicity studies on pneumococcal conjugate vaccines in children

The serotypes included in the PCV7 are responsible for 69–79% of reported IPD cases in children aged <5 years in England and Wales, and the 9-valent and 11-valent vaccines cover 77–87% and 82–91%, respectively.[28,29] In other European countries, such as Denmark, Finland, Germany, Spain, Sweden and Norway, 60–80% of paediatric IPD cases are caused by serotypes included in the PCV7.[40,110114] Overall, PCV7 covers 22–84% of recent paediatric isolates, depending on geographic area.[29,115117] However, serotype prevalence is known to vary not only with place but also over time.

1.2 Immunology of Systemic Protection

1.2.1 Pneumococcal Polysaccharide Vaccine (PPV)

The PPV23 is not effective in children <2 years old, the most vulnerable age group for IPD. After receiving one dose of PPV, infants develop good antibody responses to types 3, 4, 8, 9N and 18C; intermediate responses to types 1, 2, 7F, 19F and 25; and poor responses to types 12, 14, 23F, and 6A and 6B.[118121] All immunoglobulin classes are included in these responses, although among IgG subclasses, IgG2 and IgG4 predominate.[122] The antibody levels decrease rapidly in a few months after immunisation in small children[123,124] and a second dose elicits no anamnestic response.[118,125]

In 1978, Borgono et al.[124] reported data from several studies of PPV including groups of children immunised once (with a 12-valent PPV) at 6–24 months of age (group I), at 3–5 months and then again 6 months later at 9–11 months of age (group II), and at 3 months of age and again 21 months later at 2 years (group III). At least 50% of group I infants showed significant antibody increases to ten of the 12 capsular types which were highest against types 3, 4, 8, 9 and 56, and lowest against types 6, 12, 14 and 23. Primary responses were smaller for most capsular types in group II infants than in older infants and declined rapidly for those types in which the best antibody responses had occurred. Antibody responses after revaccination were more consistent in group III (2 years) than group II (9–11 months).

In subsequent studies in infants and children the variable immunogenicity of the different serotype capsular polysaccharides and the transient nature of antibody responses of all isotypes (IgG, IgA and IgM) have been confirmed with IgA responses being, if anything, the least reliable and most transient.[118,119,121,125128] Studies from different regions suggest there are minor differences in the immunogenicity of certain serotypes between populations, although the broad patterns are consistent. These antigens always tend to induce poor responses in young children, which improve with age throughout childhood.[129]

Children at high risk of pneumococcal infection may respond initially to PPV but again their responses are usually short-lived.[130132] Revaccination 1 year later leads to good antibody responses in about one-half of renal disease patients but levels decline over the following few months.[131] Revaccination of children with sickle cell disease at 5 years of age is followed by poor serotype-specific antibody responses in most and, in those who respond, almost one-half show a ≥2-fold decline in antibody levels 10–15 months later.[130] Older children and young adults may respond better than younger children to primary vaccination and revaccination.[133,134]

There are several studies demonstrating successively smaller antibody responses on repeated administration of doses of certain polysaccharide antigens in both adults and children, most notable meningococcal group C polysaccharide.[135138] This phenomenon has been called hyporesponsiveness. The mechanism of this effect is unknown, although theories include the possibility that antigenic stimulation ‘exhausts’ the B memory cell pool while failing adequately to replace them. There are some limited data in both adults[135,137,138] and children[87,124,137] that suggest that similar hyporesponsiveness effects can be seen with at least some serotype responses with PPVs. In the context of the apparent surprising lack of efficacy of PPV in certain settings,[60,139,140] this raises the theoretical concern that PPV could induce transient protection, at least in older children, followed by a state of partial immunological paralysis and, thus, enhanced susceptibility which might simply be worsened by repeated doses. More evidence is urgently needed to address this theoretical problem.

1.2.2 Antibody Responses to Pneumococcal Conjugate Vaccines (PCVs)

In young children, immune responses to polysaccharide antigens can be greatly enhanced by coupling the polysaccharide antigen to a protein carrier that can be processed and presented to T cells bearing specific receptors for the protein complex. T cells exposed to peptides derived from the protein carrier promote vigorous antigen-specific B-cell proliferation, affinity maturation and immunological memory.[141,142]

The immunogenicity of PCVs has been evaluated in many clinical trials (table II), showing the conjugates to be immunogenic in all paediatric age groups[26,79,89,103,143] and to prime for memory responses in early life.[26,78,144] Five- to ten-fold rises in titres are not unusual when postimmunisation anticapsular antibody titres are compared with the preimmunisation levels,[95,96,145] even to serotypes 6B and 23F, which are poorly immunogenic in infants when given as purified polysaccharides.[146] Pneumococcal-CRM conjugates, which have been most extensively studied, act as typical T-cell-dependent antigens; generally, low antibody responses are seen after the first dose and substantial antibody responses are seen after subsequent injections. Antibody concentrations achieved after the primary immunisation series usually decline during subsequent months; however, a dose of either PCV or PPV early in the second year of life produces a marked and rapid increase in antibody levels in children who were primed with conjugate vaccine in infancy.[78,85]

The IgG antibodies induced by vaccination are functionally active, as demonstrated by their high avidity and opsonophagocytic activity (OPA) in functional opsonophagocytosis assays.[70,87,97,108,141,143,146149] In infants, serum antibodies are almost entirely of subclass IgG1, but after booster immunisation usually consist of both IgG1 and IgG2 subclasses.[73,143,147,148,150] Avidity increases over the months after priming and further after the fourth dose,[143,148] indicating that these conjugate vaccines induce B-cell affinity maturation.

Primary pneumococcal vaccination begins in early infancy because IPDs occur mainly in young children. Among infants, three doses are needed to elicit significant antibody responses to polysaccharides 6B, 14 and 23F.[53,71,74,82,84,88,89] For all of the other types included in the 7-valent CRM197 conjugate vaccine, antibody levels are also higher after three than after two doses.[82,94] Furthermore, because the antibody avidity and functional activity are highest in children who received three injections, it has been suggested that three doses are required to achieve optimal protection.[85] A booster vaccination later in childhood induces a brisk high avidity antibody response.[146148] However, Käyhty and colleagues[151] have recently shown that two doses PCV7 at 3 and 5 months induce good immunological memory, as indicated by high antibody concentrations after the third dose given at 12 months. It remains to be confirmed whether a smaller number of priming doses is sufficient, or even preferable, for the induction of immunological memory[102,109] and for satisfactory protection.

Much work has been done on standardising the laboratory techniques for the determination of antibody concentration in serum; however, it is not known what concentrations of specific antibody confer protection. It also likely that different concentrations (and probably different antibody isotypes) confer protection against the different forms of non-invasive (mucosal) and IPD.[152]

Determination of antibody concentrations by ELISA has become the standard technique to measure immune responses to PCV and PPV. This is similar to the situation with H. influenzae type b (Hib) vaccines and contrasts with that for meningococcal group C conjugate vaccines where functional complement-dependent bactericidal assays have became the standard serological correlates of protection following the early studies of Gold et al.[153155]

Since antibody concentrations measured by ELISA do not distinguish between antibodies with low and high functional affinity and may not give an accurate or consistent picture of the level of protection achieved by an individual in a group, and since opsonin-dependent phagocytosis is considered a major host defence against encapsulated bacteria such as pneumococci, OPA has been proposed as an alternative measurement of the functional activity of vaccine-induced antibodies. Several studies have evaluated OPA responses in selected populations. In a study in adult volunteers, the increase in OPA was 3-fold after the PPV but 59-fold after the conjugate vaccine.[156] Opsonic activity has also been shown to increase significantly in healthy infants following immunization, and the results of OPA and ELISAs were strongly correlated.[70] PPV23 produced a significant increase in antibody concentrations to common paediatric serotypes as measured by ELISA in children with sickle cell disease, but either failed to produce significant increases in OPA or elicited more modest increases compared with PCV.[90] PCV elicited both increases in antibody concentration and increases in OPA to the common paediatric serotypes in children with sickle cell disease. In this and perhaps other populations, OPA assays may be a better measure of protection.[134]

The number and complexity of conjugate or other vaccines in use in the universal primary schedule continues to increase. It is becoming clear that multiple effects simultaneously influence the immunogenicity of individual antigens: increasing the immunological characteristics of the capsular polysaccharide antigens (which vary), the protein carriers and the adjuvant used, and interference effects between vaccine antigens. Other factors, such as age at immunisation, gestational age at birth, maternal status, maternally derived antibodies, presence of breast-feeding and other undetermined host genetic factors probably also affect individuals’ responses

Most clinical trials conducted with PCV so far have used tetanus or diphtheria toxoids or closely related proteins. This is simply because these antigens have been extensively used without adverse effects for many decades and are highly immunogenic in young infants. However, as the number of conjugates in use rises, this may lead to problems — particularly since immunisations using tetanus and diphtheria toxoids are administered at the same times. However, it is a theoretical concern that it may prove necessary to find more alternative carriers to avoid such complications.

The optimal carrier may be different for different serotypes. In one study, pneumococcal polysaccharide conjugated to CRM197 and tetanus toxoid elicited higher avidity responses to 6B and 23F than OMPC and diphtheria toxoid conjugates,[148] while no vaccine specific differences were found for serotypes 14 or 19F. In another study, a diphtheria conjugate evoked a better response to types 3, 9V and 14, while a tetanus conjugate evoked a better type 4 response and responses to types 6B, 18C, 19F and 23F did not differ.[98] In a third study, diphtheria conjugates induced better mucosal responses, while tetanus conjugates were more efficient in inducing systemic responses.[157]

Concomitant primary immunisation with diphtheria-tetanus-pertussis (DTP), Hib conjugated with CRM carrier (HbOC) and PCV (CRM197) can enhance antibody responses to diphtheria and H. influenzae antigens.[53,84] When PCV7 was combined with HbOC vaccine in the same injection, pneumococcal polysaccharide antibody responses were reduced, while responses to Hib increased.[97] When PCV7 was given concomitantly with DTaP (acellular pertussis) and HbOC in the booster series, antibody responses were reduced against all of the vaccine antigens.[96,103] Tetanus toxoid PCV given concomitantly with Hib-tetanus conjugate inhibited Hib- and tetanus-specific antibody responses,[158] and concomitant administration of Hib and pneumococcal diphtheria conjugates similarly resulted in a decreased Hib response.[100] The clinical importance of these observations is not known but they are strongly suggestive of carrier-specific effects.

The valency may have an influence on the immunogenicity of vaccines, but its influence can be difficult to assess by comparing data from different trials as the amount of antigen often varies between vaccines. The differences in antibody response might also be due to characteristics of the study population. However, some comparisons are possible. First, there is a trend towards lower serum antibody concentrations following diphtheria conjugates (types 3, 6B, 14 and 18C) in Finnish and (type 14) in Israeli infants, when the concentrations induced by PCV11 are compared with the concentrations induced by 4- and 8-valent diphtheria-conjugated vaccines.[77,78,98,101] Antibody concentrations for tetanus conjugates (types 4, 9V, 19F and 23F) in Finnish infants after immunisation with PCV11 are comparable with those after immunisation with 4-and 8-valent tetanus-conjugated vaccines.[79] Different formulations of PCV11 that use both tetanus and diphtheria toxoids have been evaluated in many countries, including Israel, Finland, Iceland and the Philippines (table II). For most of the serotypes, the immunogenicity of PCV11 is comparable with the immunogenicity of other PCVs that use other carrier proteins.[108,159]

1.3 Immunology of Mucosal Protection

Hib conjugate vaccines reduce Hib nasopharyngeal carriage rates,[160] and induce IgA and IgG antibodies to Hib capsular polysaccharide in saliva.[161] Data from the study of H. influenzae conjugate vaccine[162] suggest that conjugate vaccine does not eliminate bacteria that have already colonised the nasopharynx but that the vaccine seems to prevent the acquisition of carriage of the new serotypes to which immunity has been induced.

Mucosal immune responses to PCV7 mediate the effects of the vaccine on nasal carriage and, thus, its influence on the epidemiology of pneumococcus through herd immunity effects. They also underlie the prevention of AOM, a mucosal infection, by the vaccine.[55] Mucosal immunity may also play a part in protection against IPD. We have previously shown that PCV7 induces both salivary IgG and IgA responses in children to the capsular antigens in the vaccine in UK children given three doses at 2, 3 and 4 (or 5, 6 and 7) months of age, followed by a dose of PPV23 at 13 months.[163] Salivary IgA and IgG responses to primary immunisation were generally poor. However, IgA mean concentrations at age 5 months were higher in the treatment groups than in control subjects for serotype 14 only. At age 14 months, substantial 4- to 5-fold increases in salivary IgA concentrations for serotypes 4, 9V, 14 and 19F in the treatment groups were observed. This contrasts with very poor salivary IgA responses (1.2- to 1.4-fold increases) to serotypes 1 and 5, and suggests that PCV7 primed for mucosal IgA memory responses for the above PCV7 serotypes. Furthermore, salivary IgA correlated with salivary secretory component, suggesting that the salivary IgA was locally produced in secretory form. Specific salivary IgG concentrations decreased at age 5 months in both PCV7-primed and control infants. After booster, large 4- to 11-fold rises in IgG levels were observed in both treatment groups for serotypes 4, 18C, 19F and 23F. This contrasts with mediocre 1.3- to 2.2-fold rises for serotypes 1 and 5, suggesting that IgG memory responses were detected in saliva for the above serotypes (4, 18C, 19F and 23F) in PCV7-primed infants. These results suggest that the mucosal immune system may have been successfully primed with PCV7 for most PCV7 serotypes.

Several studies reporting mucosal responses to PCVs have been conducted in Finland.[76,101,164] Overall, both the IgA and IgG mucosal responses observed appear weaker than in our study and, unlike ours, these studies failed to show evidence of mucosal immune response.[164] The responses were more consistently seen (16 months: IgA 12.5–80.5% detectable, IgG 4.4–60.9%) in one Finnish study when mucosal responses to six of the serotypes in PCV7 were studied using a similar protocol, but comparing conjugate and polysaccharide boosters.[98] Overall, the immune responses to PCV7 in our UK study appear to have been larger and more consistently present, but there are several possible reasons for this, including different assays, different vaccine schedules and different populations. Several groups have shown that PCV7 reduces nasal carriage rates of vaccine serotypes substantially following a three-dose priming course,[52,53] as well as during the second year of life or after boosting.[5052] This suggests that, even in the absence of detectable antibody, infants may be primed for a protective response to prevent or eliminate carriage. However, we have recently shown that by the age of about 3 years, by when carriage rates have fallen in both groups, the effects of vaccination in infancy were insufficient to show significant effects on carriage of vaccine serotypes when compared with unvaccinated controls.[165] Soininen et al.[166] studied the natural development of antibodies to pneumococcal capsular polysaccharides of types 1, 6B, 11A, 14, 19F and 23F, and its association with pneumococcal carriage and AOM. The study showed that contact (carriage or AOM) with serotypes 11A and 14 was associated with increased antibody concentrations. Children with contact with serotypes 6B, 19F and 23F had levels similar to those in children without contact. There was no difference in antibody increase in children without known contact with pneumococcus and in children with contact with heterologous serotypes. Moreover, antibody concentrations were equal after carriage or AOM.

The induction and effector sites of these mucosal immune responses are unclear. The polysaccharide antigens might be absorbed into the body and transported to upper respiratory tract lymphoid tissue. Injected Hib polysaccharide antigen has been shown to be dispersed in the body.[167,168] Other possibilities are that either antigen-presenting cells or B cells or both are stimulated in the lymphoid tissues near the antigen injection site and then migrate to the upper respiratory mucosa.[169,170]

Although it seems reasonable to conclude that these antibodies modify carriage, since capsular is the only pneumococcal element used in the vaccines, it is not clear how they work or what the relative importance of the different isotypes is. It has also been shown that capsule-specific polymeric serum IgA and breast milk-derived secretory IgA elicited by both natural infection and immunisation supports killing of pneumococci by complement and phagocytosis.[171,172] However, the relevance of this mechanism in vivo is unclear and other effects, such as IgA-mediated adherence to proteins and IgG-mediated killing, are also likely to operate. Furthermore, naturally occurring anti-pneumococcal protein antibody production has been induced in blood[173,174] and in the upper respiratory mucosa[175] in unvaccinated children — presumably in response to colonisation — and these responses may also play a role in regulating carriage. Therefore, further studies are needed to investigate mechanisms of action of these mucosal antibodies.

1.4 Efficacy Studies

1.4.1 PPV in Children

PPVs have little or no efficacy against AOM in young children.[126,176,177] Infants aged <6 months gain no protection, although the vaccine may transiently reduce AOM in children aged 6–23 months and 2–6 years by 52% and 73%, respectively.[126] However, this protection lasts only for 6 months[125,126,178] and revaccination provides no additional benefit.[125,179] A recent study from The Netherlands[60] also shows that giving a PPV vaccine after priming with PCV is not effective in prevention of AOM in children >1 year of age with recurrent AOM. In this study, overall nasopharyngeal carriage of pneumococci was not affected by pneumococcal vaccination, because of a concurrent significant increase in non-vaccine serotypes. Booster vaccination with PPV23 did not seem to prevent carriage of serotypes not included in the conjugate vaccine. No significant effect on carriage of vaccine serotypes was observed in another non-randomised controlled study of 2- to 5-year-old children who had received three doses of PCV7 in infancy and PPV23 at 13 months.[165] Other studies in otitis-prone children <2 years of age gave mixed results.[125,180,181]

In contrast, PPV was shown to reduce mortality from acute lower respiratory tract infections that commonly affect children in Papua New Guinea.[182184] In a study of >7000 children pneumococcal vaccination reduced mortality by 59% in children of all ages (p = 0.008; 95% CI 19%, 79%), and by 50% in children vaccinated at ≤2 years of age (p = 0.043; 95% CI 1%, 75%). However, vaccination was only marginally protective against moderate-to-severe disease and did not protect against mild illness.[184] More recently, investigators using an indirect cohort method showed that pneumococcal vaccination of 2- to 5-year-old children was 62% effective (95% CI 35%, 78%) in preventing IPD due to vaccine serotypes.[185] However, in view of data suggesting induction of immune hyporesponsiveness following repeated doses of certain polysaccharide vaccine antigens that may be evident as early as 18 months after receipt of the last dose of the polysaccharide vaccine antigen[87,186] (see section 1.2.2), it would be of interest to have data on post immunisation vaccine efficacy over longer periods.

The main recommendations for use of PPV in children concern those aged ≥2 years at higher risk of pneumococcal infection (see section 1.7); however, no efficacy studies of these vaccines in such children have been reported.

1.4.2 PCV versus Invasive Disease

Three studies have reported the effect on IPD of the PCV7 conjugated to CRM197. In the NCKPT (Northern California Kaiser Permanente Trial),[56] efficacy against vaccine serotype disease was 97.4% (95% CI 82.7%, 99.9%) among children who had received four doses of either the control or PCV7; 39 cases caused by vaccine serotypes occurred in control children and one case, caused by vaccine serotype 19F, occurred in a child who had received PCV7. The vaccine was 89.1% (95% CI 73.7%, 95.8%) efficacious against all IPD regardless of serotype. Serotype-specific efficacy was found for serotypes 19F, 14, 18C and 23F; too few cases occurred to evaluate the other vaccine serotypes. Furthermore, the level of antibody achieved by 95% of vaccines was 0.5 μg/mL in the NCKPT.[56] A subsequent analysis of the trial showed that the vaccine was also efficacious against IPD in preterm and low birthweight infants.[187]

The second, cluster-randomised trial[188] enrolled 8292 infants from Navajo and White Mountains Apache children. The trial used the same vaccines and vaccination schedule as the NCKPT, except that older children were also enrolled and received two doses of the study vaccine or control vaccine. Eight cases of pneumococcal bacteraemia of vaccine serotype were seen in controls and two in fully immunised children, giving an efficacy, controlling for community randomisation, of 76.8% (95% CI 9.4%, 95.1%). In the intention-to-treat analysis 14 cases arose in the control group and two in the vaccinated group for an efficacy of 86.4% (95% CI 40.3%, 96.9%). An overall reduction in all cases of IPD (both vaccine and non-vaccine serotypes) of 50% was observed. The lower efficacy against IPD compared with that observed in the NCKPT reflects a substantial number of cases caused by non-vaccine serotypes in this community. Disease caused by non-vaccine serotypes, especially serotype 12F, had increased in these Native American communities before the initiation of the conjugate vaccine study.[189] Since randomisation was by community, this study also measures indirect vaccine effects providing protection against carriage or disease among nonimmunised individuals by reducing transmission of the organism within the community.

The efficacy of PCV9 in children with and those without HIV infection was evaluated in a randomised, double-blind study in South Africa.[190] At 6, 10 and 14 weeks of age 19 922 children received PCV9 conjugated to a CRM197 and 19 914 received placebo. Among children without HIV infection, the vaccine reduced the incidence of a first episode of IPD due to serotypes included in the vaccine by 83% (95% CI 39%, 97%). Among HIV-infected children the efficacy was 65% (95% CI 24%, 86%). The protective efficacy of the vaccine against all serotypes of IPD in all children was 50%.

1.4.3 PCV Effect on Carriage of Pneumococcus

Several investigators have explored the effect of PCV on nasopharyngeal carriage (table III). The first published study evaluated the effect on carriage of two or three doses of a pentavalent PCV, followed by PPV in infants in Gambia.[50] Twenty months after primary immunisation there was a reduction in carriage of vaccine serotypes in children who had received PCV CRM197 in infancy and PPV23 at age 18 months. There was no difference in all vaccinated groups with regard to overall carriage. However, children in the group who received three doses of PCV carried non-vaccine-type pneumococci significantly more often than children in the unvaccinated group (77% vs 43%, p = 0.002). Later prospective randomised studies from Israel[51,52] showed a similar reduction in carriage of vaccine type strains after immunisation of infants and toddlers with PCV conjugated to OMPC or tetanus (diphtheria) toxoid carriers (table III).
Table III

Pneumococcal conjugate vaccine effects on nasopharyngeal carriage in vaccinated and control infants and toddlers

Three studies have evaluated the effect of PCV9 with a CRM carrier on carriage in infants[53,103] and toddlers.[54] In all three studies there was no significant difference between overall carriage rates in children who received PCV and controls, but in two[53,54] apparent drops in vaccine-serotype carriage and compensatory rises in non-vaccine type rates reached statistical significance. A recent study on PCV administration to infants followed by PPV at 13 months showed no significant effects of vaccination on rates of nasal carriage of pneumococcus or serotype replacement in children by 2–5 years of age.[165]

Overall, primary immunisation of young children with PCV, and subsequent PCV or PPV boosting, is associated with reductions of carriage rates of vaccine types by about 50% and is often associated with rises in detection rates of non-conjugate vaccine serotypes, at least for 12–18 months. Serotype replacement is thought to be principally due to acquisition of these serotypes, but could, to some extent, be due to the phenomenon called ‘unmasking’ and, thus, detection of non-vaccine strains already present when vaccine serotypes are selectively eliminated.

1.4.4 PCV Efficacy versus Acute Otitis Media

Five completed studies have evaluated the efficacy of PCVs against AOM.[55,56,60,200,201] In the first large, double-blind efficacy trial in >38 000 infants in the NCKPT, children were randomised to receive the PCV7 vaccine or the meningococcal C conjugate vaccine at 2, 4, 6 and 12–15 months. The pneumococcal vaccine was shown to decrease AOM visits by 8.9%, AOM episodes by 7% and recurrent episodes by 9.3%. It also showed a 20% reduction in ventilatory tube placement.[56]

An aetiological specific study of AOM from Finland evaluated the efficacy of the same vaccine. In this study, 1662 infants were randomised to receive PCV7-CRM or Hepatitis B (HepB; control vaccine) at 2, 4, 6 and 12 months of age.[55] Aetiological diagnosis of AOM was made by tympanostomy. There was a 57% (range 44–67%) reduction in the incidence of disease caused by pneumococcal serotypes included in the vaccine. Prevention of AOM irrespective of aetiology was 6% (range 4–16%). However, there was an increase of 34% in episodes caused by non-vaccine serotypes.

The third study to evaluate AOM was also conducted at the same time in Finland by the same group of investigators and used similar methods.[58,200] In this efficacy trial, 1666 infants were randomised to receive PCV7-OMPC or HepB at 2, 4 and 6 months of age, and at 12 months the children received PCV7-OMPC or PPV23. PCV7-OMPC efficacy was 56% (95% CI 44%, 66%). The serotype-specific efficacy ranged from 37% for 19F to 82% for 9V. The two boosters seemed to provide equal protection against AOM, but PPV23 induced markedly higher antibody concentrations. The efficacy of PCV7-OMPC was comparable with that of the recently licensed PCV.

The effect of a PCV on the occurrence of AOM and other upper respiratory tract infections was also evaluated in a double-blind, randomised controlled study performed in Israel. 264 toddlers aged 12–35 months at enrolment were randomised to receive either a PCV9-CRM197 or a control vaccine (conjugate meningococcus C vaccine) and were followed for an average of 22 months. Investigators showed an insignificant apparent reduction in episodes of AOM (17% fewer episodes; 95% CI −2%, 22%). The apparent magnitude of any reduction was somewhat larger in children <36 months (23% fewer episodes; 95% CI −3%, 42%) than in children ≥36 months (12% fewer episodes; 95% CI −20%, 35%).

In a recent prospective, randomised controlled trial in The Netherlands, PCV7 effectiveness on AOM episodes and nasopharyngeal carriage in children aged 1–7 years with documented recurrent AOM was evaluated.[60] 383 enrolled children with two AOM episodes per year were followed up for AOM episodes for 18 months after completion of the vaccination scheme. The children were randomised to receive either PCV7 followed by PPV23, or hepatitis A or B vaccines. Children aged 12–24 months in the pneumococcal vaccine group were immunised with PCV7 twice, followed 6 months later by PPV23. Children aged 25–84 months in the pneumococcal vaccine group received one dose of PCV7, followed 7 months later by PPV23. The results show that PCV vaccination combined with PPV did not prevent AOM in this group. Although vaccination did reduce nasopharyngeal carriage of the PCV7 serotypes, including serotype 6B, overall nasopharyngeal carriage of pneumococci was not affected by pneumococcal vaccination because of a concurrent significant increase in non-vaccine serotypes (11, 15 and 16). Booster vaccination with PPV23 did not seem to prevent carriage of serotypes not included in the conjugate vaccine.

PCV7 has demonstrated efficacy for prevention of serotype-specific pneumococcal otitis; however, studies from Finland have reported an increase in incidence of disease caused by non-vaccine serotypes,[55] although surveillance to date in the US has not reported any rises in the incidence of non-vaccine type invasive disease.

1.4.5 PCV Efficacy versus Pneumonia

Currently, several trials are evaluating efficacy against pneumonia; results from only two studies have been published to date.[190,202] In the NCKPT,[202] children who received PCV7 with CRM carrier had 4.3% (95% CI −3.5%, 11.5%) fewer first episodes of clinically diagnosed pneumonia than the comparison group, according to the per protocol analysis. The vaccine reduced episodes of pneumonia in which a chest radiograph was obtained by 9.8% (95% CI 0.1%, 18.5%) and episodes with a ‘positive’ radiographic reading (defined as infiltrates beyond the perihilar area or with consolidation or empyema) by 20.5% (95% CI 4.4%, 34.0%). The greatest vaccine effect on radiographically confirmed pneumonia was a reduction of 32.2% in the first year of life and a 23.4% reduction in the first 2 years, but only a 9.1% reduction in children >2 years of age.

Klugman et al.[190] evaluated the effect of PCV9 in children with and without HIV infection. Among children without HIV infection, the vaccine reduced the incidence of first episodes of radiologically confirmed alveolar consolidation by 20% (95% CI 2%, 35%) in the intention-to-treat analysis and by 25% in the per-protocol analysis (i.e. among the fully vaccinated group). They also found that the vaccine offered significant protection (17%) against pneumonia in HIV-positive children.

1.4.6 PCV Effects on Antimicrobial Resistance

Use of pneumococcal vaccine also has the potential to reduce substantially incidence of drug-resistant pneumococcal infections. A reduction in carriage rates of penicillin-, multi- and cotrimoxazole-resistant strains has been shown in studies from Israel.[53,201,203] The trial in Israel using the PCV9 (CRM carrier protein) that evaluated AOM also assessed the effect of vaccination on antimicrobial use among children attending daycare centres.[201] During 5556 child-months of follow-up, at least one illness (85% involved the respiratory tract) occurred during 2542 (45.8%) child-months and 755 illness episodes resulted in antibacterial therapy. The pneumococcal vaccine group had 7% (95% CI 2%, 12%) fewer illnesses and 15% (95% CI 3%, 25%) fewer courses of antibacterials than the meningococcal vaccine recipients. Another recent double-blind controlled trial of healthy daycare centre attendees aged 12–35 months in Israel was conducted to determine the effect of PCV9 on the carriage of antibacterial-resistant pneumococci of healthy daycare centre attendees aged 12–35 months.[204] Antibacterial resistance was found mainly in the five serotypes included in the PCV (6B, 9V, 14, 19F and 23F) and in two related serotypes (6A and 19A). In the PCV9 vaccinated group, a clear and significant reduction in carriage rates of vaccine serotypes and the related serotype 6A and an increase in the carriage rates of non-vaccine serotypes were observed. In parallel, a significant decrease in carriage rate of antibacterial-resistant pneumococci was observed most marked in <36-month-olds.

1.5 Post-Implementation Effectiveness

Black et al.[205] have evaluated the impact of the introduction and routine use of PCV7 on the epidemiology of IPD. The incidence of IPD caused by vaccine serotypes before the licensure and routine use of PCV ranged between 51.52 and 98.15 cases per 100 000 person-years in children <1 year of age and fell to 9.35 after introduction of vaccine. The incidence in children <2 years of age was 81.67–113.80 before introduction and 38.22 cases per 100 000 person-years after introduction of the vaccine into the general population. These reductions in disease rates were greater than expected and were due to vaccine coverage alone in each group. Decline in IPD after the introduction of PCV7 was assessed by Whitney et al.[61] using data from 1998 through 2001 (population, 16 million). The rate of IPD dropped from an average of 24.3 cases per 100 000 persons in 1998 and 1999 to 17.3 per 100 000 in 2001. The largest decline was in children <2 years of age. In this group, the rate of disease was 69% lower in 2001 than the baseline rate (59.0 vs 188.0 cases per 100 000, p < 0.001); the rate of disease caused by vaccine and vaccine-related serotypes declined by 78% (p < 0.001) and 50% (p < 0.001), respectively. Since PCV7 was introduced, indirect vaccine effects providing protection against carriage or disease among nonimmunised individuals by reducing transmission of the organism within the community were measured.[205207] Because the PCV reduces the carriage of vaccine-related strains by vaccinated children, it was hypothesised that the introduction of PCV7 in young children could affect transmission of pneumococcus to adults and reduce the risk of invasive disease in adults. In fact, such a reduction has been observed, most prominently in two age groups: those 20–40 years of age, and those ≥60 years of age who experience significant morbidity and mortality from pneumococcal disease. This study also involved surveillance of pneumococcal invasive disease caused by non-vaccine serotypes. Each isolate was then classified as a vaccine serotype, as a cross-reacting serotype for isolates of the same serogroup but not the same serotype as those in the vaccine, or as a serogroup unrelated to any represented in the vaccine. No increase in disease incidence was observed for possibly cross-reacting serotypes or nonvaccine serotypes.[205,206]

1.6 Safety Data

In the NCKPT efficacy trial,[56] PCV7 vaccination resulted in less frequent local reactions than DTwP (whole cell pertussis), but more frequent local reactions than DTaP and the control vaccine. Fever (≥38°C) ≤48 hours after vaccination was more common among children who received PCV7 concomitantly with DTaP than among those who received the control vaccine. This difference was statistically significant after each dose of vaccine in the primary series but not the fourth dose. Rates of fever >39°C were substantially greater among those who received PCV7 after dose two of the primary series (2.5% vs 0.8%). Febrile seizures after vaccination were slightly more common in the PCV7 group; however, the majority of events occurred when DTwP was administered concurrently with PCV7. Of the eight PCV7 recipients who had a seizure, seven had received concurrent DTwP vaccine. A total of 32 children who were originally enrolled in the study had died; however, none of the deaths were reported by investigators to be related to the vaccine.

In another trial performed in 302 infants immunised at 2, 4 and 6 months of age, the rate of injection site reactogenicity was less for the PCV7 than for the combined whole DTP/Hib and the HepB vaccine sites.[96] In another study of 212 infants who were randomised to receive either PCV7 or an investigational conjugate meningococcal group C vaccine at the same time as they received other vaccines in the routine schedule, the reported rates for any local reaction at 2, 4 and 6 months of age were 36%, 43% and 35% for PCV site versus 56%, 58% and 46% at the combined DTP/Hib injection site. No severe local reactions occurred, and the rate and severity of reactions did not increase with subsequent doses of the PCV. Furthermore, other clinical trials performed since 1992, 7-, 9- and 11-valent PCVs have also been demonstrated to be safely used with only minor local adverse reactions limited primarily to the injection site.[56,9597,103] No serious adverse events have been attributed to these vaccines in studies to date. With respect to local reactions, 7-, 9- and 11-valent PCVs are significantly less reactogenic than DTwP or DTwP/HbOC,[9597,103,108] but significantly more reactogenic than DTaP, MenC (serogroup C meningococcal conjugate vaccine containing CRM197) or inactivated polio vaccine.[56]

A postmarketing, open-label, noncomparative, passive, observational trial with the primary objective of expanding the safety profile of PCV7-CRM is currently ongoing in the NCKP population. As of December 2000, 22 020 subjects younger than 120 days of age have received at least one dose of PNCRM7; of these subjects, 15 546 have received two doses, 9245 have received three doses and 85 have received four doses. A preliminary interim report included evaluation of preselected events (e.g. fever-related diagnoses, seizure events) during four postvaccination periods (days 0–2, 0–7, 0–14 and 0–30) compared with a control period (days 31–60) in hospital, emergency room and clinic settings. During this time period, more fever was noted for subjects in the emergency room setting after receiving the first dose and the primary series for the 0–2 days postvaccination period compared with the control period (31–60 days postvaccination). These results are consistent with the prelicensure findings. Vaccination with PNCRM7 has not been associated with an increased rate of seizures for any of the four postvaccination time periods compared with the control period.[208]

1.7 Current US and UK Recommendations

After licensure, PCV7 was recommended for routine use in infancy in the US by the Advisory Committee on Immunization Practices (table IV).[209,210] The vaccine is licensed for use in healthy children aged 2 months to 2 years of age, as a course of four doses, to prevent meningitis and septicaemia,[56] and is the first of several new pneumococcal vaccines designed for young children to become available.[211] In the US, it is recommended that all children aged ≤23 months should be vaccinated with PCV7 using the schedule shown in table IV. Vaccination is also offered to older children who have underlying medical conditions or are immunocompromised and at risk of pneumococcal disease (table V).
Table IV

US recommended schedule for vaccination using 7-valent pneumococcal conjugate vaccine for children by age at time of first vaccination and previously vaccinated infants

Table V

Summary of US recommendations for the use of 7-valent pneumococcal conjugate vaccine (PCV7) among infants and children

In contrast, current UK recommendations are to immunise children at enhanced risk of pneumococcal infection (with asplenia or severe splenic dysfunction, chronic renal disease or nephritic syndrome, immunodeficiency or immunosuppression, chronic heart, lung or liver disease, and diabetes mellitus) with PPV if aged ≥2 years and only with PCV if they are aged <2 years, followed by PPV on or after their second birthday.[212] Others have advocated use of such PCV plus PPV schedules in older children at high risk.[213] This kind of strategy is unlikely to have a significant impact on overall disease as most occurs in previously healthy children.

Urgent and careful investigations to inform recommendations for the use of PPV in children, whether alone or in combination schedules, following PCV are needed and it is hoped that cost will not prohibit wide use of PCVs in countries outside the US in the near future.

2. Pneumococcal Protein Vaccines

The development and licensure of PCV7 has been a major breakthrough in the control of IPD and other conjugates may soon follow PCV7. However, their usefulness is limited by restricted capsular serotype distribution, multiple-dose regimens and high cost.[211] Over the past decade, several pneumococcal protein virulence factors have been described. If these proteins are essential for virulence and pathogenesis, it is possible that antibodies raised against them could form the basis of third-generation pneumococcal vaccines.

The surface of S. pneumoniae is covered with proteins that are covalently and non-covalently attached to the cell wall (figure 1). Currently, several candidate pneumococcal proteins are under study, including pneumolysin, pneumococcal surface protein A (PspA), pneumococcal surface adhesion A (PsaA) and choline-binding protein A (CbpA, also referred to as PspC, SpsA or Hic).[214218] Of these pneumococcal protein antigens, pneumolysin, PspA and PsaA have been shown to contribute to the virulence of pneumococci and to be produced by virtually all clinical isolates.[219,220] CbpA is likely to play a role in nasopharyngeal colonisation and appears to be expressed by most, if not all, isolates. Preliminary studies in mice have shown that immunisation with these proteins can protect against infection with multiple serotypes of S. pneumoniae[221] and/or prevent nasopharyngeal carriage.[222,223] It has been shown that pneumococcal carriage and infections induce salivary and serum antibodies to PsaA, pneumolysin and PspA in children.[173175,224]

As S. pneumoniae is a mucosal pathogen colonising the nasopharynx, mucosal immunisation may prove a more effective way to protect against mucosal carriage than parenteral immunisation. Therefore, to induce herd immunity against pneumococci it will be necessary to induce protection against carriage. PsaA and CbpA may be good vaccine candidates against carriage as they are associated with pneumococcal adherence.[217,225] As PspA and pneumolysin are known to elicit protective immunity against bacteraemia and pneumonia in mice,[221,226] their inclusion in a mucosal vaccine might promote prevention of invasive disease as well as carriage. Zhang et al.[175] recently conducted a study to determine whether these proteins are good mucosal immunogens in humans. The data suggest that CbpA could be proved to be an immunogenic mucosal vaccine. Moreover, CbpA induces transcription and release of several chemokines by human respiratory epithelial cells.[227] Recent evidence suggests that chemokines may influence antigen-specific immune responses.[228] This aspect of immunological regulation by prospective protein vaccine antigens needs further study.

In a recent phase I trial,[222,229] a recombinant PspA vaccine was shown to be safely used and immunogenic in healthy adults, including broadly cross-reactive serum antibodies. Further human studies of protein-based vaccines are awaited.

3. Conclusion

To date, studies have demonstrated that PCVs can be safely used and are immunogenic and efficacious in protecting young infants and children against various types of IPD. The serotypes included in the PCV7 are responsible for 60–80% of reported IPD cases in many European countries.[28,29,40,110114] The 9- and 11-valent vaccines cover 77–87% and 82–91%, respectively.[28,29] An additional factor that is important to the efficacy of PCV is its ability to prevent nasopharyngeal colonisation, which result in a decrease in transmission of the organism and herd immunity. A significant reduction in colonisation by vaccine-type pneumococci was seen in the vaccinees compared with control groups; however, non-vaccine serotypes were found significantly more frequently in vaccinees, suggesting a degree of replacement in serotype distribution in vaccinees instead of a simple decrease in pneumococcal colonisation.[5054,59,230] Therefore, the possibility that the use of PCV could lead to an increase in disease caused by non-vaccine serotypes is a potential concern[55] and surveillance of clinical isolates should continue after the introduction of these vaccines, which in the future may require reformulation. Wider use of these vaccines could be encouraged if the vaccines prove to be efficacious with a lower number of doses and, thus, lower costs.

It is certain that pneumococcal vaccines are needed in routine vaccination programmes in developing countries. However, low serotype coverage, replacement phenomena and high costs may prove problematic. The development of one or more protein vaccines may prove easier and less expensive. Delivery of such vaccines directly to the mucosa of the nasopharynx is a highly desirable goal.


No sources of funding were used to assist in the preparation of this manuscript. The authors have no conflicts of interest that are directly relevant to the content of this review.

Copyright information

© Adis Data Information BV 2005