This analysis is part of a larger project describing the impact of PCV dosing schedules on invasive pneumococcal disease (IPD), immunogenicity, nasopharyngeal carriage, pneumonia and indirect effects.^11–14 Details on the literature search terms and methods used in this systematic review are described elsewhere (see Methods Appendix¹⁵). In brief, a systematic literature review was performed to collect all available English language data published from January 1994 to September 2010 (supplemented post hoc with studies from 2011) on the effect of various PCV vaccination schedules among immunized children on immunogenicity, nasopharyngeal colonization, IPD, pneumonia and on indirect effects among unvaccinated populations. Articles published in 14 databases, from ad hoc unpublished sources and abstracts from meetings of the International Symposium on Pneumococci and Pneumococcal Disease (1998–2010) and the Interscience Conference on Antimicrobial Agents and Chemotherapeutics (1994–2010), were searched. We included all randomized controlled clinical trials (RCTs), nonrandomized trials, surveillance database analyses and observational studies of any PCV schedule on one or more outcomes of interest. Studies were included for abstraction if pneumococcal polysaccharide vaccine (PPV23) was used as a booster dose, but not as a primary dose. Titles and abstracts were reviewed twice and those with relevant content on 1 of the 5 outcomes (immunogenicity, carriage, invasive disease, pneumonia and indirect effects) underwent full review using a standardized data collection instrument. We did not search non-English language literature because of the low likelihood they would have relevant data for this project. Details on the search methods are provided in the Methods Appendix.¹⁵

Citations recovered through the literature search went through several stages of independent review to determine their eligibility, as described (see Methods Appendix¹⁵). Citations meeting inclusion criteria were categorized on an outcome-specific basis into “study families,” where each family included abstracts or publications generated from a single protocol, population, surveillance system or other data collection system relevant to that outcome. Investigators identified primary data from the individual studies making up each study family for inclusion in the analysis. The primary data were selected as the most current and complete data available for that study family. In some cases, these data were drawn from >1 publication within a family. We also defined “study arms” as a group of children distinguished by immunization schedule or PCV product.

We abstracted core information on the following: number of children in a “study arm;” PCV manufacturer, valency and conjugate protein; co-administered vaccines; country; age at each dose and date of study and publication. Additional data abstracted for pneumonia included specific endpoints, case definitions, study design, study population and incidence rates or percent change.

This article presents the data on the direct effects of PCV on pneumonia in children of an age targeted for vaccination. As studies included a variety of case definitions for endpoints, findings were grouped by endpoint according to the following categories: clinical pneumonia (including lower respiratory tract infections and acute respiratory tract infections), radiologically confirmed pneumonia, pneumococcal pneumonia (including bacteremic pneumonia) and empyema.

We included data published during or after 1994 from clinical trials, surveillance database analyses and observational studies of PCV schedules on immunogenicity, IPD, nasopharyngeal carriage, pneumonia and indirect effects. We included all licensed and unlicensed PCV products (denoted as PCV with a number indicating the valency, eg, PCV7). We excluded studies with vaccination series beginning after 12 months of life, as well as observational studies that only reported data before or after PCV introduction but not for both periods. Unless ≥50% vaccination coverage was documented, observational studies were also excluded if vaccination was only available through the private sector or to high-risk groups. Studies that only provided incidence rates during the year of vaccine introduction, or did not specify a period, were excluded.

We defined a primary series as either 2 or 3 doses received before 7 months of age. A booster dose was defined as a dose of PCV or PPV23 received after 9 months of age and after the completion of a primary series. A complete series was defined as the primary series plus any booster doses implemented in a population; examples of this include a 2-dose primary series with or without a booster (2+1, 2+0) or a 3-dose primary series with or without a booster (3+1, 3+0).

Studies evaluating impact on pneumonia following PCV introduction used a variety of methods; the variety prevented us from performing a formal meta-analysis. Therefore, we conducted descriptive analyses of the amount and variability of the data and of the magnitude of the change in the pneumonia outcomes observed for each dosing schedule type. We also performed subanalyses to evaluate various endpoints related to pneumonia. Studies reporting only qualitative data with no ability to determine magnitude of impact were excluded from analysis.

For observational studies reporting pneumonia incidence over time, we calculated percent change as: (baseline incidence —post-PCV introduction incidence)/baseline incidence. Baseline incidence was defined as the mean of all data points reported before PCV introduction. When annual data on postintroduction incidence were available, we calculated percent change using the data point given for each year reported. When only the average postintroduction incidence rate over a period of years was provided, we calculated percent change from baseline to the reported rate and assigned it to the median year of the date range provided. When possible, incidence rates during the year of introduction were excluded from these calculations. We conducted all analyses using SAS 9.3 (SAS Institute Inc., Cary, NC).

Of 12,980 citations reviewed, we identified 106 pneumonia outcome citations that met initial criteria for further evaluation (Fig. 1). After further review, 81 citations met inclusion criteria for full data abstraction; of these, 39 studies were excluded from analysis because they contained duplicate data of included studies or reported changes in pneumonia risk only qualitatively so magnitude of impact could not be assessed. Of the 42 included citations, 20 evaluated clinical pneumonia, 13 radiologically confirmed pneumonia, 16 pneumococcal pneumonia and 9 all-cause empyema; however, case definitions varied widely for each endpoint.^16–57

Almost all (n = 39, 93%) citations of pneumonia were published during or after 2004. Most citations were from North America (n = 23, 55%), Europe (n = 9, 22%) or Oceania (n = 5, 12%), with the remaining 5 from Africa (n = 3, 7%), Asia (n = 1, 2%) and Latin America (n = 1, 2%). Although many studies focused on the general population of children, 6 citations focused on high-risk groups (ie, children with HIV or indigenous populations). Thirty-seven citations evaluated PCV7 and only one study evaluated PCV10⁵⁷ (Table 1).

We identified only 2 studies, both observational, that compared the effectiveness of different PCV dosing schedules within the study itself. One study directly evaluated the impact of 2 versus 3 primary PCV doses against clinical pneumonia incidence in a general pediatric population.²⁸ This propensity-score-matched, case-cohort study conducted in the United States evaluated the rate of hospitalizations and ambulatory visits for lower respiratory tract infections and found that children who received 3 primary PCV doses had fewer ambulatory visits and hospitalizations up to the point of receipt of a booster dose (9.5 admissions per 1000 children) than those who only received 2 primary doses [17.3 admissions per 1000 children; rate difference = 7.8 cases per 1000 children (95% confidence interval (CI): 0.8–14.8)]. This difference disappeared after the booster dose was administered [23.2 admissions per 1000 children vs. 20.9 admissions per 1000 children for 3+1 vs. 2+1, respectively; rate difference = −2.3 cases per 1000 children (95% CI: −14.8 to 9.3)]. This difference between 2 and 3 primary doses was seen for children born in the 2002 birth cohort, but not for children born in 2003; the authors hypothesized that by 2003, 3 years after introduction of PCV7, herd effects had lessened the difference in risk between the 2 groups. The other study directly comparing dosing schedules, a retrospective cohort conducted among Australian Indigenous infants,⁵³ evaluated risk of clinical and radiologically confirmed pneumonia after each of 3 PCV7 primary doses plus 1 PPV23 booster (3+PPV23) but did not find evidence of reduced risk for either endpoint by number of doses.

Among 5 studies evaluating the impact of PCV on populations at high risk for pneumococcal disease, 2 (3 citations) used a 3+0 schedule and 3 used a 3+1 schedule. Two RCTs evaluated the impact of PCV7 among a high-risk population using a 3+0 schedule (Table 2).^18,22,26 One trial, conducted in South Africa, found a 13–15% efficacy against clinical and radiologically confirmed pneumonia in children with HIV. The other clinical trial, from Papua New Guinea, found PCV7 to be 18% (95% CI: 4–31%) efficacious against clinical pneumonia in neonates.¹⁸ We identified 1 RCT⁵⁶ and 2 observational studies^20,52 from the United States and Australia that evaluated a 3+1 (3+PPV23 for Australian Indigenous) schedule in indigenous populations. The RCT, conducted among a population of American Indians in the United States, showed no efficacy against the first episode of radiologically confirmed pneumonia (authors' data, per protocol vaccine efficacy: –8.0%, 95% CI: –37.0% to 14.9%); however, only inpatient pneumonia cases were included in this analysis unlike other RCTs. One of the observational studies found a trend of declining incidence for clinical pneumonia in Australian Indigenous children; however, this finding was not significant (P = 0.13), and study investigators speculate the lack of sufficient follow-up time as a possible reason.⁵² The other observational study, evaluating empyema in a US Alaskan Native pediatric population, found no change in empyema-associated hospitalizations following PCV introduction and rates remained higher than those for children in the general US population.²⁰ Study investigators did note an apparent increased rate in empyema due to S. pneumoniae and, in particular, episodes due to nonvaccine serotypes, which could explain the lack of change in overall empyema rates.