Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Effectiveness of pneumococcal vaccines in preventing pneumonia in adults, a systematic review and meta-analyses of observational studies

  • Myint Tin Tin Htar ,

    Contributed equally to this work with: Myint Tin Tin Htar, Anke L. Stuurman

    Myint.TinTinHtar@pfizer.com

    Affiliation Pfizer: Vaccines Clinical Epidemiology, Pfizer Inc, Paris, France

  • Anke L. Stuurman ,

    Contributed equally to this work with: Myint Tin Tin Htar, Anke L. Stuurman

    Affiliation P95 Epidemiology and Pharmacovigilance Consulting and Services, P95, Leuven, Belgium

  • Germano Ferreira,

    Affiliation P95 Epidemiology and Pharmacovigilance Consulting and Services, P95, Leuven, Belgium

  • Cristiano Alicino,

    Affiliation Department of Health Sciences (DiSSal), University of Genoa, Genoa, Italy

  • Kaatje Bollaerts,

    Affiliation P95 Epidemiology and Pharmacovigilance Consulting and Services, P95, Leuven, Belgium

  • Chiara Paganino,

    Affiliation Department of Health Sciences (DiSSal), University of Genoa, Genoa, Italy

  • Ralf René Reinert,

    Affiliation Pfizer: Vaccines Medical Development and Scientific Clinical Affairs, Pfizer Inc, Paris, France

  • Heinz-Josef Schmitt,

    Affiliation Pfizer: Vaccines Medical Development and Scientific Clinical Affairs, Pfizer Inc, Paris, France

  • Cecilia Trucchi,

    Affiliation Department of Health Sciences (DiSSal), University of Genoa, Genoa, Italy

  • Thomas Vestraeten,

    Affiliation P95 Epidemiology and Pharmacovigilance Consulting and Services, P95, Leuven, Belgium

  • Filippo Ansaldi

    Affiliation Department of Health Sciences (DiSSal), University of Genoa, Genoa, Italy

Abstract

Introduction

S. pneumoniae can cause a wide spectrum of diseases, including invasive pneumococcal disease and pneumonia. Two types of pneumococcal vaccines are indicated for use in adults: 23-valent pneumococcal polysaccharide vaccines (PPV23) and a 13-valent pneumococcal conjugate vaccine (PCV13).

Objective

To systematically review the literature assessing pneumococcal vaccine effectiveness (VE) against community-acquired pneumonia (CAP) in adults among the general population, the immunocompromised and subjects with underlying risk factors in real-world settings.

Methods

We searched for peer-reviewed observational studies published between 1980 and 2015 in Pubmed, SciELO or LILACS, with pneumococcal VE estimates against CAP, pneumococcal CAP or nonbacteremic pneumococcal CAP. Meta-analyses and meta-regression for VE against CAP requiring hospitalization in the general population was performed.

Results

1159 unique articles were retrieved of which 33 were included. No studies evaluating PCV13 effectiveness were found. Wide ranges in PPV23 effectiveness estimates for any-CAP were observed among adults ≥65 years (-143% to 60%). The meta-analyzed VE estimate for any-CAP requiring hospitalization in the general population was 10.2% (95%CI: -12.6; 33.0). The meta-regression indicates that VE against any-CAP requiring hospitalization is significantly lower in studies with a maximum time since vaccination ≥60 months vs. <60 months and in countries with the pediatric PCV vaccine available on the private market. However, these results should be interpreted cautiously due to the high influence of two studies. The VE estimates for pneumococcal CAP hospitalization ranged from 32% (95%CI: -18; 61) to 51% (95%CI: 16; 71) in the general population.

Conclusions

Wide ranges in PPV23 effectiveness estimates for any-CAP were observed, likely due to a great diversity of study populations, circulation of S. pneumoniae serotypes, coverage of pediatric pneumococcal vaccination, case definition and time since vaccination. Despite some evidence for short-term protection, effectiveness of PPV23 against CAP was not consistent in the general population, the immunocompromised and subjects with underlying risk factors.

Introduction

Streptococcus pneumoniae can cause a wide spectrum of diseases, and is the leading cause of community-acquired pneumonia (CAP) [1]. In adults CAP is the most common type of pneumococcal disease [2]. Pneumococcal CAP (pCAP) can either be invasive bacteraemia or nonbacteremic (non-invasive). Two types of vaccines are indicated for use in adults in Europe: 23-valent pneumococcal polysaccharide vaccines (PPV23) [3, 4] first licensed in 1983, and a 13-valent pneumococcal conjugate vaccine (PCV13) which was first licensed in 2011 [3, 4].

There is evidence that pneumococcal vaccines are effective in preventing invasive pneumococcal disease (IPD) in adults. Several meta-analyses of randomised controlled trials (RCTs) have described clinical efficacy of PPV23 against IPD ranging from 10% to 80%. The clinical efficacy of PCV13 was 75% for vaccine-type IPD and 52% for any-cause IPD in adults aged 65 years or older in a recent RCT (CAPiTA trial)[5].

The absolute burden of vaccine preventable disease is much higher for CAP than for IPD [1, 6, 7]. The ability of pneumococcal vaccines to protect against CAP in adults is key in evaluating the overall public health benefits of the vaccination program. efficacy [8]. However, available clinical efficacy estimates of pneumococcal vaccines against CAP widely vary and are highly influenced by study settings and case definitions. The meta-analyzed clinical efficacy of PPV23 ranged from -11% to 74% for pCAP and from -13% to 29% for any-cause CAP [2, 911]; the clinical efficacy of PCV13 was 31% for pCAP and 5% for any-cause CAP [5].

A recent review estimated the real-world effectiveness of PPV23 for CAP among adults (>50 years) as 17% (95% CI: −26% to 45%) and 7% (95% CI: −10% to 21%) for cohort and case–control studies, respectively [11]. There was no real-world data available for PCV13 due its recent use in adults. To date, no systematic literature review and meta-analysis has been exclusively conducted for post-marketing observational studies evaluating the real-world effectiveness of pneumococcal vaccines against CAP in the general adult population, in immunocompromised and in immunocompetent patients.

We aimed to systematically review the literature on observational studies that assessed pneumococcal vaccine effectiveness (VE) against CAP (any-CAP, pCAP or nonbacteremic pCAP) in adults (see File 1 for PICO questions), with specific emphasis on any-CAP requiring hospitalization in the general population.

Methods

Identification of studies

MEDLINE (Pubmed), LILACS and SCIELO were searched to identify peer-reviewed articles published between January 1st 1980 and October 30th 2015 (date of search), in English, French, Spanish, Portuguese, Dutch, German and Italian. The search string consisted of terms for pneumococcal vaccines combined with terms for effectiveness (see S1 File for details). In addition, a grey literature search was performed through major public health organization websites and targeted search terms using Google. The protocol is available (S2 File).

Inclusion criteria

We included published peer-reviewed observational studies conducted in adults (16 years or older), of any design (case-control, cohort, indirect cohort design (Broome method), test-negative case-control, and screening-method) of any pneumococcal vaccines indicated and used for adults (PPV23 or PCV13) that reported VE results on the protection for any clinically relevant outcome other than IPD; and that focused on direct effects [12]. Studies reporting exclusively on IPD and exclusively on the impact of pneumococcal vaccination program in adults were excluded. VE estimates comparing receipt of both PPV23 and influenza vaccine with the receipt of either influenza vaccine only or neither vaccine were excluded.

Selection and data collection

Two reviewers (AS, GF) independently screened all titles and abstracts. The full-text of selected papers was reviewed and data were extracted (AS). For quality control results for 10% of the papers were extracted in duplicate (GF). An extensive hand search was conducted, based on the reference lists of relevant papers retrieved from the search, to identify additional studies. Data extraction was performed using MS Excel and Access. A complete list of data extraction items is available (S1 File).

Clinical outcomes

Pneumonia was considered as community-acquired unless otherwise reported. CAP was classified into: any-CAP, pCAP, and nonbacteremic-pCAP. No a priori case definitions for CAP, pCAP or nonbacteremic-pCAP were set for this review, the case definitions used by the respective authors were accepted. Within each pneumonia type, estimates were further classified by healthcare setting (any setting, ambulatory, outpatient, hospitalized), and ‘mortality’.

Effect measures

Estimates of VE expressed as percentage (%) as well as their confidence intervals (CIs) were derived from the effect measures, odds ratios, relative risks, hazard ratios and incidence rate ratios using VE = [1-effect measure]x100. VE estimates >0% suggest a protective effect.

Quality assessment

Study limitations and risk of bias were assessed using the Newcastle-Ottawa assessment scale for cohort and case-control studies with scores ranging from 0 to 9 [13]. The following criteria were assessed: selection and comparability of study populations, and exposure (in case-control studies) or outcome (in cohort studies).

Population

The studies were classified according to study population: general population, immunocompromised patients, and patients with underlying risk factors (other than immunosuppression). General populations consisted of populations from primary care centers, health maintenance organizations and insurance companies regardless of individual health status.

Meta-analyses

We conducted meta-analyses for VE against any-CAP requiring hospitalization. Study heterogeneity was investigated by the chi-squared test for heterogeneity, for which p-values <0.05 indicate a significant amount of heterogeneity, and quantified using the I2 statistic with low, moderate and high levels of heterogeneity corresponding to I2 values of 25%, 50% and 75%, respectively. When I2 > 25%, a random-effects model was applied to obtain the pooled effect estimate [14].

Potential sources of study heterogeneity were explored by conducting stratified meta-analyses and by using meta-regression. The potential sources of heterogeneity we explored include: usage of pediatric pneumococcal vaccine (no PCV use, PCV available in private market, PCV in national immunization program), age (<65 years, ≥ 65 years, no specific age group described), study design (cohort or case-control) and maximum time since vaccination (<60 months, ≥ 60 months). The maximum time since vaccination refers to the greatest possible duration between the occurrence of the outcome and the vaccination by design of the study. R2 index was used to quantify the proportion of variance explained by the covariates in the meta-regression.

The influence of inclusion of a study on the results of the meta-analyses was assessed using study deletion diagnostics including Cook’s distance and DFBETAs. Publication bias was assessed visually using funnel plots. One additional post-hoc sensitivity analysis was carried out excluding all but the most recent of studies with potentially overlapping study populations. All analyses were performed in R version 3.2.0, using the metafor package for meta-analyses [15, 16].

Results

The search retrieved 1159 unique articles. Grey literature and public health websites searches retrieved no additional reports. After selection and hand search, 61 studies reported on VE against any clinically relevant outcome other than IPD in adults (Fig 1).

thumbnail
Fig 1. Flowchart of selection procedure.

The flowchart was based on the flowchart from the PRISMA group [17]. *Invasive pneumococcal disease (n = 11), not a primary research article (n = 7), vaccine or vaccinated group not of interest (n = 5), study protocol or rationale without results (n = 4), study design not of interest (n = 4), data from same source was presented in another article (n = 4), inadequate comparison group (n = 3), no relevant information (n = 3), data in children (n = 2), safety endpoint (n = 1), impact (n = 1). **Outcome other than any community-acquired pneumonia (CAP), pneumococcal CAP (pCAP) or nonbacteremic pCAP (n = 18), no vaccine effectiveness estimate provided (n = 7), estimates comparing receipt of both PPV23 and influenza vaccine with the receipt of influenza vaccine only (n = 2), CAP data from same source was presented in another article (n = 1).

https://doi.org/10.1371/journal.pone.0177985.g001

Of these, 33 studies on CAP reported at least one VE estimate. By study design, there were 20 cohort studies, 11 case-control studies, one case-cohort study and one study with self-controlled risk windows. By etiology 29 studies reported VE for any-CAP, 8 for pCAP and 5 for nonbacteremic-pCAP across all study populations. No studies reported effectiveness of PCV13 in adults. Newcastle-Ottawa quality scores ranged from 4 to 7 and from 3 to 8 for cohort and for case-control studies (S1 Table).

Any-CAP

Overall, 15 studies reported VE estimates for CAP in the general population, 9 in the immunocompromised, and 8 in patients with underlying risk factors. 7 studies were conducted by one research group in primary care centers in Tarragona (Spain) and therefore overlapping study populations cannot be excluded. Estimates with detailed study description are reported in S2 Table.

Any-CAP in general population.

For protection against CAP hospitalization in the general population, there were 13 studies and 18 VE estimates (Fig 2). Several studies were conducted in Spain (n = 4) for which most estimates only included adults >65 years. Among adults aged >65 years, the VE estimates ranged from -143% [18] to 60% [19].

thumbnail
Fig 2. VE with most adjusted estimates for any-CAP in the general population, comparing PPV23 vaccinated with unvaccinated.

https://doi.org/10.1371/journal.pone.0177985.g002

The overall meta-analyzed VE for any-CAP hospitalization was 10.20% (95%CI -12.62; 33.01). The between-study heterogeneity was high (I2 = 99.24%, p < 0.01). The study by Tsai [19] was found to be influential (DFBETA = 1.46). After exclusion of this study, the meta-analyzed VE was 5.14% (95%CI: -2.09; 12.37, I2 = 71.66%).

Based on the meta-regression, usage of pediatric pneumococcal vaccine and maximum time since vaccination were both significant (p<0.01) and explained part of the heterogeneity (R2 = 77.4%; I2 = 41.2%;) (S3 Table). However, these results should be interpreted with caution since two studies (Tsai et al. [19] and Christenson et al. [20]) were found to be influential.

In the stratified meta-analysis by pediatric PCV use, the meta-analyzed VE was -6.31 (95%CI: -15.78; 3.17, I2 = 60%,), 9.01 (95%CI: −0.62; 18.64, I2 = 43%) and 29.40 (95%CI: -0.78; 59.59, I2 = 96%) when pediatric PCV was part of the national immunization program, when pediatric PCV was not available and when PCV was available in the private market only (Fig 3). In the stratified meta-analysis by maximum time since vaccination, the meta-analyzed VE was 32.6% (95%CI: -5.9; 71.1, I2 = 99%) and 2.4% (95%CI: -5.4; 10.1, I2 = 65%) when the time since vaccination was less than 60 months and 60 months or more, respectively. (Fig 4). There was no evidence of publication bias.

thumbnail
Fig 3. Meta-analysis, stratified by availability of pediatric pneumococcal vaccine (PCV): VE with most adjusted estimates for hospitalization due to any-CAP in the general population, comparing PPV23 vaccinated with unvaccinated.

# weights from the random-effects model; *>70% for at least one dose; **45–64 years; ***65+ years.

https://doi.org/10.1371/journal.pone.0177985.g003

thumbnail
Fig 4. Meta-analysis, stratified by maximum time since vaccination: VE with most adjusted estimates for hospitalization due to any-CAP in the general population, comparing PPV23 vaccinated with unvaccinated.

*weights from the random-effects model; **45–64 years; ***65+ years.

https://doi.org/10.1371/journal.pone.0177985.g004

Among three studies evaluating VE against any-CAP in any healthcare setting, VE ranged from -143% (95%CI: -1501; 63) [21] to 69% (95%CI: 48; 81) [22]. Neither of two studies presenting VE for any-CAP in outpatient setting was statistically significant [23, 24]. Among 5 studies evaluating VE against death due to any-CAP, estimates ranged from -4% (95% CI: -69; 36)[25] to 93% (95% CI: 92; 94) [19].

Any-CAP in immunocompromised patients and patients with underlying risk factors.

S1 and S2 Figs summarize the VE estimates for any-CAP in immunocompromised populations and populations with underlying risk factors, respectively.

Among 9 studies in immunocompromised patients VE estimates ranged from -88% (95% CI: -349; 21) [26] to 78% (95%CI: 46; 91) [27] across age groups, CD4 counts, severity and healthcare settings. VE against any-CAP in HIV-infected individuals decreased with increasing viral load [28], but no effect of CD4 count was observed [28, 29].

Among 8 studies among patients with underlying risk factors, the VE against any-CAP ranged from -338% [18] to 72% [30] (S2 Fig). Four studies were conducted in patients with chronic respiratory and/or COPD aged 65 years or older [3134]. These yielded 7 VE estimates, of which only two reached statistical significance [31, 33]. Neither of the two VE estimates against death due to any-CAP was statistically significant [30, 31].

Pneumococcal CAP

There were 8 studies evaluating the VE of pCAP, five studies were from Tarragona region, Spain (S4 Table). For pCAP, the identification of etiologic agent was made from blood or other sterile site culture, sputum or other respiratory samples culture, or urinary antigen test or from clinical diagnosis codes. The proportion of bacteremic pCAP cases among all pCAP cases varied from 11 [25] to 33% [35, 36] in the general population, 58% in the HIV-infected adults study [37] and 20% in the study conducted in chronic respiratory disease patients [38].

Pneumococcal CAP in general population.

Four studies evaluated the VE for pCAP in the general population. All four studies were in subjects aged 50 years or older and three were conducted in Tarragona. The VE estimates for pCAP irrespective of severity and setting were 45% (95%CI: 12; 66) [36] and 53% (95%CI: 33; 68) [35] among those aged 65 years or older. The estimates were slightly higher during the influenza season (60% and 61%) than out of the influenza season (49% and 37%) [35, 36]. The VE for hospitalization due to pCAP varied from 32% (95 CI: -18; 61) to 51% (95% CI: 16; 71) [25].

Pneumococcal CAP in immunocompromised patients and patients with underlying risk factors.

Among two studies in immunocompromised patients, the VE estimates for pCAP were 77% (95%: -6; 92) in HIV-infected patients [37] and 71% (95%CI: 24; 89) in possibly immunocompromised patients (all causes combined) [35]. Among 4 studies identified in patients with underlying risk factors, the VE estimate for pCAP irrespective of severity and setting was 29% (95%CI: -39; 63) among those aged 50 years or older with chronic respiratory disease [38] and 41% (95%CI: 10; 61) for those with various chronic disease [35]; VE for pCAP hospitalization among those aged 65 years or older was 24% (95%CI: -90; 70) among those with chronic respiratory disease [31] and 38% (95%CI: -5; 70) among those with various risk factors [39]. None of VE estimates in patients with chronic respiratory disease reached statistical significance. The VE point estimates in patients with chronic respiratory diseases was lower outside the influenza season (versus influenza season), albeit statistically non-significant.

Nonbacteremic-pCAP

Out of four identified studies conducted in the general population, three studies were from the same research group and geographical area in Tarragona, Spain. Study characteristics and VE estimates for nonbacteremic pCAP are presented in S5 Table. In these studies, potential CAP cases were initially identified based on ICD-9 codes with different case definitions for each study. Nonbacteremic-pCAP was defined when CAP was accompanied by a positive sputum culture for S. pneumoniae (but negative/not performed blood culture) and/or a positive Binax-NOW urinary antigen test [25, 35, 40], or only a positive Binax-NOW urinary antigen test [36]. In addition, two of these studies also used laboratory records to identify cases missed by ICD codes [35, 36]. One study in the US based the definition on a typical clinical syndrome of pneumonia and radiographic confirmation, sputum culture positive and at least one negative blood culture result for S. peumoniae [41].

The VE estimates for nonbacteremic-CAP of any severity or setting in the general population aged 50 years or older, ranged from 39% (95% CI: -6; 65) [42] to 42% (95%CI: 14; 61) [35]. The VE was -1% (95% CI: -86; 45) for nonbacteremic-pCAP in the US adult (18 years or older) study [41] irrespective of severity and setting and -3% (95% CI: -53; 31) for nonbacteremic-pCAP hospitalization in Spanish study [25]. The VE was higher (54%) during the influenza season compared to outside the influenza season (44%) [35].

No studies in immunocompromised patients were identified and only one study in patients with chronic respiratory disease was reported. The VE estimate was 34% (95% CI: -34; 67) [38].

Vaccine effectiveness and time since vaccination

In general, lower VE was observed in studies with longer maximum time since vaccination (at least 60 months post-vaccination) [28, 35]. A Spanish prospective cohort study (CAPAMIS) showed how VEs varied according to the time since vaccination [25]. The VE increased and became more significant after excluding the patients vaccinated more than 60 months ago from the analysis. The VE estimates were 25% (95%CI: 2; 42), 51% (95% CI: 16; 71), 62% (95% CI: -68; 91) and 48% (95% CI: 8;71) in the analysis exclusively including the recently vaccinated patients (less than 60 months before study start, compared to those never vaccinated) for any-CAP, pCAP, bacteremic-pCAP and nonbacteremic-pCAP, respectively. The VE estimates were 8% (95%CI: -20; 29), 32% (95%CI: -18; 61), 53% (95% CI: -145; 91) and 29% (95% CI: -27; 61) for any-CAP, pCAP, bacteremic-pCAP and nonbacteremic-pCAP, respectively, in the analysis regardless of time since vaccination (i.e. comparing those vaccinated any time before study start to those never vaccinated).

Sensitivity analysis

Since three out of 13 studies were conducted in health care centers from Tarragona region in Spain, a sensitivity analysis was carried out excluding all but the most recent of studies with potentially overlapping study populations. Particularly, the meta-analyses were repeated excluding Dominguez et al. 2010 [43] and Vila-Corcoles et al. 2006 [36, 44]. The overall meta-analyzed VE for any-CAP hospitalization was 7.52% (95%CI -17.71; 32.76, I2 = 99.36%).

Discussion

This literature review exclusively describes pneumococcal VE in protecting against overall CAP in adults using observational studies. No studies reported effectiveness of PCV13 in adults, likely due to the recent PCV13 introduction and recommendation for adult population. Analyses were first stratified by populations (general population, immunocompromised patients and patients with underlying risk factors) and health care setting. Although availability of pediatric vaccine and time since vaccination partially explained between-study heterogeneity, the remaining heterogeneity remained substantial. Indeed, the included studies were heterogeneous in different aspects, including study populations, case definitions, case ascertainment, healthcare practices and pneumococcal serotype epidemiology, in particular variations in distribution of serotypes over time and across regions.

A wide range of VE estimates for any-CAP in adults was found, regardless of population type and healthcare setting. Overall, PPV23 was not consistently demonstrated to be effective in protecting against any-CAP in the general population, although some evidence for short-term vaccine effectiveness exists. In fact, the meta-analyzed VE for hospitalized CAP in the general population was 10.20% (95%CI -12.62; 33.01). The VE among the general population was lower than the meta-analyzed VE estimated from RCTs, 27% (95%CI: 6; 44) and 28% (95% CI: 7; 44) in two meta-analyses [2, 9], and higher than -10% (95% CI: -30; 7) in a recent meta-analysis [45]. On the other hand, our results are similar to another review based on observational studies with VEs of 17% (95%CI: -26; 45) and 7% (95%CI: -10; 21) for cohort and case-control studies respectively [11].

The VE of PPV23 for any-CAP in immunocompromised patients or patients with underlying risk factors (COPD and chronic respiratory disease) remains controversial. In our review, the VE ranged from -88% to 78% in immunocompromised patients (mostly HIV-infected patients) and from -338% to 72% in patients with underlying risk factors. These findings confirmed those obtained from meta-analyses of RCTs conducted among these patients [2, 911].

As expected, VE estimates for pCAP in all populations were consistently higher than against any-CAP [25, 31, 3539, 46]. This review revealed positive VE estimates for pCAP, 45% and 53% in adults aged 65 years or older in the general population [35, 36]. Our VE estimates for nonbacteremic pCAP ranged from 39 to 42% in adults aged 50 years or older in the general population. In meta-analyses of RCTs, pCAP is often split into presumptive pCAP (nonbacteremic) and definite pCAP (bacteremic). Our findings for nonbacteremic pCAP were similar to two meta-analyzed efficacy estimates for presumed pCAP based on RCTs, 54% (95% CI: 16; 75) and 36% (4; 57) [2, 9].

The PPV23 effectiveness depends on time since vaccination. In the CAPAMIS study, the effectiveness estimates became higher and significant after excluding the patients vaccinated for more than 60 months ago [25]. Our meta-analyzed VE was 2% (95%CI: -5;10) in studies where the maximum time since vaccination was 60 months or more while it was 33% (95%CI: -6; 71) in patients vaccinated less than 60 months ago. These results are consistent with immunological findings where the initial rise in antibody titers declined over time, reaching to approximate pre-vaccination baseline after 4–7 years of vaccination [47, 48]. The lack of efficient memory induction by PPV23 would explain these waning protection [49].

Specific precautions are necessary to interpret the effect of pneumococcal vaccine in adults in populations with different approaches to pediatric pneumococcal vaccination. In the US and Israel, the pediatric PCV uptake was high and this likely led to reduce vaccine-serotype S. pneumoniae circulation, which in turn could have led to a lower proportion of vaccine serotype cases among any-CAP cases in the general population (herd protection), resulting in a lower VE estimate [17, 50]. Simultaneously, all studies in the category “PCV in national immunization program” also had a maximum time since vaccination of more than 60 months, likely lowering the VE estimate further. The effectiveness estimate in settings with no PCV was positive and higher than the estimate for “PCV in national immunization program”, and the estimate was higher still in settings with “PCV available in private market”. It is expected that the estimate for “PCV available in private market” (with low PCV uptake at the time of the study) is higher than that for “PCV in national immunization program” (with high PCV uptake). However, it is counterintuitive that the estimate for “PCV available in private market” is higher than the one for settings with no PCV; this warrants further research that takes into account detailed information on factors such asthe actual pediatric vaccination coverage and time since vaccination.

Limitations

We restricted the meta-analysis to any-CAP requiring hospitalization in the general population due to the limited number of studies reporting on the other outcomes.

Heterogeneity in case ascertainment and case definitions between the studies would influence the final results. CAP diagnosis was sometimes based on ICD-9 codes only (e.g. ICD-9 480–487 [51] or ICD-9 codes 481, 486, 482.9, 485, V1261 [52]); on ICD-9 codes complemented by a medical record check and a chest radiograph for confirmation [36]; or on signs and symptoms and a chest radiograph [43]. However, almost all studies required radiographic and laboratory-confirmation for a diagnosis of pCAP (including nonbacteremic pCAP).

A further potential source of bias may be the misclassification of invasive diseases as non-invasive diseases. In studies from Tarragona, non-bacteremic pCAP was defined as clinical and radiological pneumonia with negative blood culture and/or not performed blood culture, and sputum culture positive for S. pneumonia and/or positive Binax [35, 40]. However, the proportion of patients with non-performed culture and prior antibiotics before the diagnosis was unknown. Moreover, most pCAP diagnoses in the observational studies involved identification of S. pneumonia from either sterile or non-sterile site culture. Including isolates from sterile culture in the definition is a possible explanation for the fact that VE for pCAP in observational studies is close to VE for IPD in RCTs. Heterogeneity can exist in the extent of misclassification between the studies. Nonetheless, the proportion of invasive diseases among pCAP was known for few studies [3538].

The definition of maximum time since vaccination differed between studies and could have introduced additional heterogeneity. The meta-analysis and meta-regression of results stratified by maximum time since vaccination attempted to partially overcome this limitation.

Although the most adjusted estimate was considered, and many studies adjusted for the effect of influenza vaccination, this was not always taken into account. The influenza vaccine is known to be effective in reduction of any-CAP and hence, combined vaccination with pneumococcal vaccine might overestimate the VE attributed to pneumococcal vaccine in the general population [53, 54].

An inherent limitation of observational studies is that vaccines are channeled to certain populations. For example, frailer patients (e.g. older, more co-morbidities) with a higher baseline risk for CAP, hospitalization and death may be more likely to be targeted for vaccination, therefore possibly underestimating the effect of pneumococcal vaccines. Simultaneously, the healthy vaccinee effect could result in an overestimation of the VE. To minimize the influence of these limitations, most included studies adjusted for the potential confounders using propensity scores or multiple regression between vaccinated and non-vaccinated patients.

Study quality was assessed using the Newcastle Ottawa score. While important aspects of study design are rated using this tool, it did not take into account the rationale and the context (e.g. local S. pneumoniae epidemiology, healthcare practices) in which a study was initiated and performed. Missing information in the publication lowers the overall score.

Conclusions

Wide ranges in PPV23 effectiveness estimates for any-CAP were observed, with a higher, albeit not statistically significantly, effectiveness in more recently vaccinated populations and in settings with low-to-moderate paediatric PCV vaccine uptake. Overall, the effectiveness of PPV23 against CAP has not been consistently demonstrated in the general population (including the elderly), the immunocompromised and subjects with underlying risk factors. This lack of consistency may be related to a great diversity of study populations, circulation of S. pneumoniae serotypes, coverage of pneumococcal pediatric vaccination.

Pneumococcal vaccination programs, both adult and pediatric programs as an integrated public health intervention are likely to impact the proportion of CAP caused by S. pneumoniae and thus influencing the effectiveness of the vaccines evaluated. Monitoring of adult pneumococcal VE should continue in the context of increasing PCV13 use in adults and increasing vaccination coverage in the pediatric population.

Supporting information

S4 File. References for all included studies.

https://doi.org/10.1371/journal.pone.0177985.s004

(DOCX)

S1 Fig. VE with most adjusted estimates for any-CAP in immunocompromised populations, comparing PPV23 (or other Pneumococcal vaccine NOS) vaccinated with unvaccinated.

https://doi.org/10.1371/journal.pone.0177985.s005

(TIF)

S2 Fig. VE with most adjusted estimates for any-CAP in patients with underlying risk factors, comparing PPV23 (or other Pneumococcal vaccine NOS) vaccinated with unvaccinated.

https://doi.org/10.1371/journal.pone.0177985.s006

(TIF)

S1 Table. Quality assessment of included studies using the Newcastle-Ottawa Score.

https://doi.org/10.1371/journal.pone.0177985.s007

(DOCX)

S2 Table. Study characteristics and results for any-CAP.

https://doi.org/10.1371/journal.pone.0177985.s008

(DOCX)

S3 Table. Results of the meta-regression: Vaccine effectiveness against CAP requiring hospitalization.

https://doi.org/10.1371/journal.pone.0177985.s009

(DOCX)

S4 Table. Study characteristics and results for pneumococcal CAP.

https://doi.org/10.1371/journal.pone.0177985.s010

(DOCX)

S5 Table. Study characteristics and results for nonbacteremic pneumococcal CAP.

https://doi.org/10.1371/journal.pone.0177985.s011

(DOCX)

Author Contributions

  1. Conceptualization: MTTH GF ALS TV.
  2. Data curation: ALS GF.
  3. Formal analysis: GF KB.
  4. Funding acquisition: MTTH.
  5. Investigation: ALS GF KB MTTH.
  6. Methodology: ALS GF MTTH KB.
  7. Project administration: GF MTTH.
  8. Resources: MTTH AS GF KB.
  9. Software: GF KB.
  10. Supervision: GF MTTH.
  11. Validation: GF.
  12. Visualization: GF KB.
  13. Writing – original draft: ALS MTTH.
  14. Writing – review & editing: MTTH ALS GF CA KB CP RRR HJS CT TV FA.

References

  1. 1. Welte T, Torres A, Nathwani D. Clinical and economic burden of community-acquired pneumonia among adults in Europe. Thorax. 2012;67(1):71–9. pmid:20729232
  2. 2. Moberley S, Holden J, Tatham DP, Andrews RM. Vaccines for preventing pneumococcal infection in adults. Cochrane Database Syst Rev. 2013;1:CD000422.
  3. 3. Merck. PNEUMOVAX® 23—Highlights of prescribing information 2015 [09 October 2015]. Available from: https://www.merck.com/product/usa/pi_circulars/p/pneumovax_23/pneumovax_pi.pdf.
  4. 4. Pfizer. PREVNAR 13—Highlights of prescribing information 2015 [09 October 2015]. Available from: http://labeling.pfizer.com/showlabeling.aspx?id=501.
  5. 5. Bonten MJ, Huijts SM, Bolkenbaas M, Webber C, Patterson S, Gault S, et al. Polysaccharide conjugate vaccine against pneumococcal pneumonia in adults. New England Journal of Medicine. 2015;372(12):1114–25. pmid:25785969
  6. 6. European Centre for Disease Prevention and Control SE. Surveillance of invasive bacterial diseases in Europe, 2012. Stockholm: ECDC, 2015.
  7. 7. Feikin DR, Scott JA, Gessner BD. Use of vaccines as probes to define disease burden. Lancet. 2014;383(9930):1762–70. pmid:24553294
  8. 8. Saadatian-Elahi M, Horstick O, Breiman RF, Gessner BD, Gubler DJ, Louis J, et al. Beyond efficacy: The full public health impact of vaccines. Vaccine. 2016.
  9. 9. Huss A, Scott P, Stuck AE, Trotter C, Egger M. Efficacy of pneumococcal vaccination in adults: a meta-analysis. CMAJ. 2009;180(1):48–58. PubMed Central PMCID: PMC2612051. pmid:19124790
  10. 10. Schiffner-Rohe J, Witt A, Hemmerling J, von Eiff C, Leverkus FW. Efficacy of PPV23 in Preventing Pneumococcal Pneumonia in Adults at Increased Risk—A Systematic Review and Meta-Analysis. PLoS One. 2016;11(1):e0146338. pmid:26761816
  11. 11. Kraicer-Melamed H, O’Donnell S, Quach C. The effectiveness of pneumococcal polysaccharide vaccine 23 (PPV23) in the general population of 50 years of age and older: A systematic review and meta-analysis. Vaccine. 2016.
  12. 12. Halloran E, Longini IM, Struchiner CJ. Design and analysis of vaccine studies. New York: Springer; 2010.
  13. 13. Wells GA, Shea B, O'Connell D, Peterson J, Welch V, Losos M, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses Ottawa: Otawa Hospital Research Institute; [cited 2015 8 October 2015]. Available from: http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp.
  14. 14. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled Clinical Trials. 1986;7(3):177–88. pmid:3802833
  15. 15. R Development Core Team. R: a language and environment for statistical computing, ed. R Foundation for Statistical Computing. Vienna, Austria2013.
  16. 16. Viechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat Softw. 2010;36(3):1–48.
  17. 17. Moore MR, Link-Gelles R, Schaffner W, Lynfield R, Lexau C, Bennett NM, et al. Effect of use of 13-valent pneumococcal conjugate vaccine in children on invasive pneumococcal disease in children and adults in the USA: analysis of multisite, population-based surveillance. The Lancet Infectious Diseases. 2015;15(3):301–9. pmid:25656600
  18. 18. Vila Corcoles A, Ochoa Gondar O, Hospital Guardiola I, Marin Canseco ML. [Chronic obstructive pulmonary disease and tobacco dependency: main risk factors in pneumonia in the over-65s]. Aten Primaria. 2003;31(4):272.
  19. 19. Tsai YH, Hsieh MJ, Chang CJ, Wen YW, Hu HC, Chao YN, et al. The 23-valent pneumococcal polysaccharide vaccine is effective in elderly adults over 75 years old—Taiwan's PPV vaccination program. Vaccine. 2015;33(25):2897–902. Epub 2015/05/06. pmid:25936662
  20. 20. Christenson B, Hedlund J, Lundbergh P, Ortqvist A. Additive preventive effect of influenza and pneumococcal vaccines in elderly persons. The European respiratory journal. 2004;23(3):363–8. Epub 2004/04/07. pmid:15065822
  21. 21. Song JY, Lee JS, Wie SH, Kim HY, Lee J, Seo YB, et al. Prospective cohort study on the effectiveness of influenza and pneumococcal vaccines in preventing pneumonia development and hospitalization. Clinical and vaccine immunology: CVI. 2015;22(2):229–34. Epub 2014/12/30. PubMed Central PMCID: PMCPmc4308868. pmid:25540271
  22. 22. Gable CB, Holzer SS, Engelhart L, Friedman RB, Smeltz F, Schroeder D, et al. Pneumococcal vaccine. Efficacy and associated cost savings. Jama. 1990;264(22):2910–5. Epub 1990/12/12. pmid:2232086
  23. 23. Vila-Corcoles A, Ochoa-Gondar O, Hospital I, Ansa X, Vilanova A, Rodriguez T, et al. Protective effects of the 23-valent pneumococcal polysaccharide vaccine in the elderly population: the EVAN-65 study. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America. 2006;43(7):860–8. Epub 2006/08/31.
  24. 24. Jackson LA, Neuzil KM, Yu O, Benson P, Barlow WE, Adams AL, et al. Effectiveness of pneumococcal polysaccharide vaccine in older adults. The New England journal of medicine. 2003;348(18):1747–55. Epub 2003/05/02. pmid:12724480
  25. 25. Ochoa-Gondar O, Vila-Corcoles A, Rodriguez-Blanco T, Gomez-Bertomeu F, Figuerola-Massana E, Raga-Luria X, et al. Effectiveness of the 23-valent pneumococcal polysaccharide vaccine against community-acquired pneumonia in the general population aged >/ = 60 years: 3 years of follow-up in the CAPAMIS study. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America. 2014;58(7):909–17. Epub 2014/02/18.
  26. 26. Hung CC, Chen MY, Hsieh SM, Hsiao CF, Sheng WH, Chang SC. Clinical experience of the 23-valent capsular polysaccharide pneumococcal vaccination in HIV-1-infected patients receiving highly active antiretroviral therapy: a prospective observational study. Vaccine. 2004;22(15–16):2006–12. Epub 2004/05/04. pmid:15121313
  27. 27. Guerrero M, Kruger S, Saitoh A, Sorvillo F, Cheng KJ, French C, et al. Pneumonia in HIV-infected patients: a case-control survey of factors involved in risk and prevention. Aids. 1999;13(14):1971–5. Epub 1999/10/08. pmid:10513657
  28. 28. Teshale EH, Hanson D, Flannery B, Phares C, Wolfe M, Schuchat A, et al. Effectiveness of 23-valent polysaccharide pneumococcal vaccine on pneumonia in HIV-infected adults in the United States, 1998–2003. Vaccine. 2008;26(46):5830–4. Epub 2008/09/13. pmid:18786586
  29. 29. Guevara RE, Butler JC, Marston BJ, Plouffe JF, File TM Jr., Breiman RF. Accuracy of ICD-9-CM codes in detecting community-acquired pneumococcal pneumonia for incidence and vaccine efficacy studies. Am J Epidemiol. 1999;149(3):282–9. pmid:9927225
  30. 30. Wagner C, Popp W, Posch M, Vlasich C, Rosenberger-Spitzy A. Impact of pneumococcal vaccination on morbidity and mortality of geriatric patients: a case-controlled study. Gerontology. 2003;49(4):246–50. Epub 2003/06/07. pmid:12792160
  31. 31. Ochoa-Gondar O, Vila-Corcoles A, Ansa X, Rodriguez-Blanco T, Salsench E, de Diego C, et al. Effectiveness of pneumococcal vaccination in older adults with chronic respiratory diseases: results of the EVAN-65 study. Vaccine. 2008;26(16):1955–62. pmid:18343541
  32. 32. Nichol KL. The additive benefits of influenza and pneumococcal vaccinations during influenza seasons among elderly persons with chronic lung disease. Vaccine. 1999;17:S91–S3. pmid:10471189
  33. 33. Nichol KL, Baken L, Wuorenma J, Nelson A. The health and economic benefits associated with pneumococcal vaccination of elderly persons with chronic lung disease. Archives of internal medicine. 1999;159(20):2437–42. Epub 2000/02/09. pmid:10665892
  34. 34. Hechter RC, Chao C, Jacobsen SJ, Slezak JM, Quinn VP, Van Den Eeden SK, et al. Clinical effectiveness of pneumococcal polysaccharide vaccine in men: California Men's Health Study. Vaccine. 2012;30(38):5625–30. Epub 2012/07/14. pmid:22789510
  35. 35. Vila-Corcoles A, Salsench E, Rodriguez-Blanco T, Ochoa-Gondar O, de Diego C, Valdivieso A, et al. Clinical effectiveness of 23-valent pneumococcal polysaccharide vaccine against pneumonia in middle-aged and older adults: a matched case-control study. Vaccine. 2009;27(10):1504–10. Epub 2009/01/28. pmid:19171174
  36. 36. Vila-Corcoles A, Ochoa-Gondar O, Hospital I, Ansa X, Vilanova A, Rodriguez T, et al. Protective effects of the 23-valent pneumococcal polysaccharide vaccine in the elderly population: the EVAN-65 study. Clin Infect Dis. 2006;43(7):860–8. pmid:16941367
  37. 37. Lopez-Palomo C, Martin-Zamorano M, Benitez E, Fernandez-Gutierrez C, Guerrero F, Rodriguez-Iglesias M, et al. Pneumonia in HIV-infected patients in the HAART era: incidence, risk, and impact of the pneumococcal vaccination. Journal of medical virology. 2004;72(4):517–24. Epub 2004/02/26. pmid:14981752
  38. 38. Vila-Corcoles A, Ochoa-Gondar O, Rodriguez-Blanco T, Gutierrez-Perez A, Vila-Rovira A. Clinical effectiveness of 23-valent pneumococcal polysaccharide vaccine against pneumonia in patients with chronic pulmonary diseases: a matched case-control study. Human vaccines & immunotherapeutics. 2012;8(5):639–44. Epub 2012/05/29.
  39. 39. Hung IF, Leung AY, Chu DW, Leung D, Cheung T, Chan CK, et al. Prevention of acute myocardial infarction and stroke among elderly persons by dual pneumococcal and influenza vaccination: a prospective cohort study. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America. 2010;51(9):1007–16. Epub 2010/10/05.
  40. 40. Ochoa-Gondar O, Vila-Corcoles A, Rodriguez-Blanco T, de Diego-Cabanes C, Hospital-Guardiola I, Jariod-Pamies M, et al. Evaluating the clinical effectiveness of pneumococcal vaccination in preventing myocardial infarction: The CAPAMIS study, three-year follow-up. Vaccine. 2014;32(2):252–7. pmid:24262314
  41. 41. Musher DM, Rueda-Jaimes AM, Graviss EA, Rodriguez-Barradas MC. Effect of pneumococcal vaccination: a comparison of vaccination rates in patients with bacteremic and nonbacteremic pneumococcal pneumonia. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America. 2006;43(8):1004–8. Epub 2006/09/20.
  42. 42. Vila Corcoles A, Ochoa Gondar O, Salsench Serrano E, Hospital Guardiola I, Vilanova Navarro A, Raga Luria X. [EVAN-50 study: Effectiveness of polysaccharide pneumococcus vaccine in preventing pneumococcal infections in the over-50 population]. Atencion primaria / Sociedad Espanola de Medicina de Familia y Comunitaria. 2006;38(5):299–303. Epub 2006/10/06.
  43. 43. Dominguez A, Izquierdo C, Salleras L, Ruiz L, Sousa D, Bayas JM, et al. Effectiveness of the pneumococcal polysaccharide vaccine in preventing pneumonia in the elderly. The European respiratory journal. 2010;36(3):608–14. Epub 2010/01/16. pmid:20075048
  44. 44. VC A, OG O, HG I, MC M. L, GO I, ÁL M. Efectividad de la vacuna antineumocócica en pacientes mayores de 65 años. Effectiveness of pneumococcal vaccine in patients aged 65 years or older. Medifam. 2003:61–8.
  45. 45. Schiffner-Rohe J, Witt A, Hemmerling J, von Eiff C, Leverkus FW. Efficacy of PPV23 in Preventing Pneumococcal Pneumonia in Adults at Increased Risk—A Systematic Review and Meta-Analysis. PloS one. 2016;11(1):e0146338. PubMed Central PMCID: PMCPMC4711910. pmid:26761816
  46. 46. Wiemken TL, Carrico RM, Klein SL, Jonsson CB, Peyrani P, Kelley RR, et al. The effectiveness of the polysaccharide pneumococcal vaccine for the prevention of hospitalizations due to Streptococcus pneumoniae community-acquired pneumonia in the elderly differs between the sexes: results from the Community-Acquired Pneumonia Organization (CAPO) international cohort study. Vaccine. 2014;32(19):2198–203. Epub 2014/03/13. pmid:24613522
  47. 47. 23-valent pneumococcal polysaccharide vaccine. WHO position paper. Wkly Epidemiol Rec. 2008;83(42):373–84. pmid:18927997
  48. 48. Dransfield MT, Harnden S, Burton RL, Albert RK, Bailey WC, Casaburi R, et al. Long-term comparative immunogenicity of protein conjugate and free polysaccharide pneumococcal vaccines in chronic obstructive pulmonary disease. Clin Infect Dis. 2012;55(5):e35–44. PubMed Central PMCID: PMC3491850. pmid:22652582
  49. 49. Clutterbuck EA, Lazarus R, Yu LM, Bowman J, Bateman EA, Diggle L, et al. Pneumococcal conjugate and plain polysaccharide vaccines have divergent effects on antigen-specific B cells. J Infect Dis. 2012;205(9):1408–16. PubMed Central PMCID: PMC3324398. pmid:22457293
  50. 50. Lexau CA, Lynfield R, Danila R, Pilishvili T, Facklam R, Farley MM, et al. Changing epidemiology of invasive pneumococcal disease among older adults in the era of pediatric pneumococcal conjugate vaccine. Jama. 2005;294(16):2043–51. pmid:16249418
  51. 51. Ansaldi F, Turello V, Lai P, Bastone G, De Luca S, Rosselli R, et al. Effectiveness of a 23-valent polysaccharide vaccine in preventing pneumonia and non-invasive pneumococcal infection in elderly people: a large-scale retrospective cohort study. The Journal of international medical research. 2005;33(5):490–500. Epub 2005/10/15. pmid:16222881
  52. 52. Leventer-Roberts M, Feldman BS, Brufman I, Cohen-Stavi CJ, Hoshen M, Balicer RD. Effectiveness of 23-valent pneumococcal polysaccharide vaccine against invasive disease and hospital-treated pneumonia among people aged >/ = 65 years: a retrospective case-control study. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America. 2015;60(10):1472–80. Epub 2015/02/12.
  53. 53. Chan MC, Lee N, Ngai KL, Wong BC, Lee MK, Choi KW, et al. A "pre-seasonal" hospital outbreak of influenza pneumonia caused by the drift variant A/Victoria/361/2011-like H3N2 viruses, Hong Kong, 2011. J Clin Virol. 2013;56(3):219–25. Epub 2012/12/04. pmid:23201458
  54. 54. Chan TC, Fan-Ngai Hung I, Ka-Hay Luk J, Chu LW, Hon-Wai Chan F. Effectiveness of influenza vaccination in institutionalized older adults: a systematic review. J Am Med Dir Assoc. 2014;15(3):226 e1–6.