Impact of pneumococcal conjugate vaccines on nasopharyngeal carriage and invasive disease among unvaccinated people: Review of evidence on indirect effects

Highlights

• Systematic review for indirect effect of PCV on vaccine-type disease and correlation with NP carriage.

• We identified studies from 13 countries.

• Vaccine-type IPD and carriage in non-targeted populations consistently decreased.

• VT-IPD reductions occurred contemporaneous with NP reductions.

• Indirect PCV impact on VT-IPD mediated by carriage has been significant.

Abstract

Background

Invasive disease due to Streptococcus pneumoniae remains an important worldwide cause of morbidity and mortality, particularly in young children and the elderly. The development and use of pneumococcal conjugate vaccines (PCVs) have had a dramatic impact on rates of vaccine-type invasive pneumococcal disease (IPD) not only in the pediatric population targeted for vaccination but in non-vaccinated age-groups as well. This indirect effect is directly mediated by a reduction of vaccine-type nasopharyngeal carriage and thus transmission by vaccinated children. Current PCV licensing procedures do not take into consideration nasopharyngeal carriage impact, and thus the indirect effect. This review summarizes the evidence for the indirect effect of PCV on vaccine-type disease and its correlation with changes in carriage among unvaccinated populations, to assess the basis for inclusion of carriage in the PCV licensing process.

Methods

Randomized controlled trials, surveillance and other observational studies published between 1994 and 2013 were systematically identified from global, regional and review databases and conference abstracts. We included as primary evidence, studies in non-vaccinated groups addressing changes in both vaccine-type IPD and carriage between pre- and post-PCV introduction periods; studies missing one of these four components were included as supporting rather than primary evidence.

Results

We identified studies from 14 countries, nearly all developed countries. Vaccine-type IPD and carriage in non-targeted populations consistently decreased after PCV introduction, with the magnitude of decrease growing over time. Where IPD and carriage were observed in the same population, VT-decreases occurred contemporaneously. These relationships held true across age-groups and between indigenous and non-indigenous populations in the US and Australia.

Conclusions

Indirect PCV impact on VT-IPD and VT-carriage has been significant. Impact on carriage should be considered for inclusion in the PCV licensure process as a predictor of indirect effects.

Introduction

Streptococcus pneumoniae caused over 500,000 estimated deaths among children under 5 years of age globally in 2008. [1] Adults, primarily the elderly and immunosuppressed, also suffer a high burden of mortality and morbidity from this pathogen [2]. In all age-groups there is a disproportionate burden of disease among those who live in the developing world or have limited access to treatment [3].

In 2000 the first pneumococcal conjugate vaccine (PCV) was licensed in the United States. It included the seven most common serotypes causing invasive pneumococcal disease (IPD) among young children in North America [4]. Unlike pure polysaccharide vaccines that generate a T cell-independent, antibody-mediated response, conjugate vaccines engage T-cell-mediated immunity, stimulating serotype-specific antibody production and immunologic memory, providing protection beginning in infancy against disease from included serotypes.

The basis for licensing the first PCV product was clinical efficacy against vaccine-serotype (VT) IPD demonstrated through randomized, double-blind, clinical trials of infants [5], [6]. Experience in the prior decade with Haemophilus influenzae type b (Hib) conjugate vaccine demonstrated decreased Hib oropharyngeal and nasopharyngeal (NP) carriage in vaccinated children, reducing transmission to and disease in unvaccinated children; this is termed the indirect or herd effect. Because of the Hib vaccine experience, early PCV studies evaluated the impact on pneumococcal NP carriage as an indicator of the potential for indirect protection. Numerous studies showed that PCV reduces NP acquisition of VT carriage among vaccinated subjects [7]. However, PCV also increases the colonization prevalence of non-vaccine serotypes (NVTs) – a phenomenon termed serotype replacement – leaving overall pneumococcal carriage prevalence virtually unchanged.

PCV introduction into the routine pediatric immunization schedule in the United States and other countries has resulted in near-elimination of VT-IPD not only in infants (the age-group targeted for vaccination), but also in the unimmunized general population [8]. This indirect protection is a critical component of the vaccine's public health impact. In the United States, it accounted for 69% of all IPD cases prevented in the first three years of licensure [9] and a 44–63% absolute decrease in pneumococcal pneumonia admissions in adults [10].

PCVs have now been incorporated into routine childhood immunization in 96 countries. Another 51 countries, many in the developing world, plan to introduce PCV in the coming years [11]. With demand growing, multiple manufacturers are developing PCV products; licensing authorities have had to determine what data should support such licensure and be required for post-licensure monitoring. Disease endpoint trials are now difficult or impossible to conduct because of ethical considerations in placebo-control comparisons and sample size requirements in head-to-head trials. Licensure approaches are therefore anchored on correlations of immunogenicity to IPD protection established in the randomized controlled trials, and immunogenicity non-inferiority measures in new PCV products [12].

Although this approach has a strong scientific basis and is accepted by the European Medicines Evaluation Agency, the United States Food and Drug Administration, and the World Health Organization (WHO), it lacks a crucial component: impact of pneumococcal vaccines on NP carriage among both the vaccinated and unvaccinated, and consequent effects on disease among the unvaccinated as well as the fully or partially vaccinated. NP effects may also prove an essential component of the licensing approach for novel non-polysaccharide pneumococcal vaccines such as those based on pneumococcal proteins. Not only do vaccine products merit consideration from this perspective of impact on carriage, so do vaccine schedules; the number of primary-series doses and addition of a booster dose may affect the magnitude of the indirect effect. We posited the causal chain in the indirect effect paradigm as follows (Fig. 1):

• PCV decreases VT-carriage prevalence and density in vaccinated individuals. Reduction in prevalence is achieved by reductions in acquisition rates and density, rather than reductions in duration of VT carriage [13], [14], [15].

• These changes result in decreased VT-carriage transmission to, and carriage among, unvaccinated (and vaccinated) populations.

• Because individual NP carriage acquisition is the prerequisite for pneumococcal disease, reductions in VT-carriage among vaccinated and unvaccinated age groups cause a reduction in VT-IPD in these same populations.

Evidence for the first link in this chain and for individual carriage as a precondition for pneumococcal disease is addressed elsewhere [16]. Here we evaluate the evidence for the links between PCV use and reductions in both VT-carriage and VT-IPD in non-target age-groups.

The most compelling evidence for this link is from studies (community-randomized trials or pre- and post-PCV observational studies) simultaneously examining rates of VT-carriage and VT-IPD in non-targeted groups, with and without PCV. Also relevant are studies examining PCV-associated changes in IPD or carriage alone. Others that provide secondary supporting evidence for the validity of the causal chain include studies comparing VT-IPD or NP carriage rates in non-targeted age-groups in early vs. mature post-introduction periods (time-series analyses); those comparing these rates pre- and post-introduction in populations which are predominantly non-targeted but include some targeted individuals (“mixed” populations); and those which compare pre- and post- introduction rates of all-type (AT) IPD in non-target age-groups without distinguishing VT from NVT disease.

We performed a comprehensive review of studies meeting each of these descriptions to assess the evidence for the importance of NP carriage as a component of licensure of new pediatric pneumococcal vaccine products.