The effectiveness of PPV is controversial. Despite numerous studies, contradictory results have been reported and several meta-analyses have been inconclusive to date. Excluding earlier trials in younger adults,[21, 22] prospective RCTs have failed to demonstrate a significant protective effect of the vaccination. However, several case-control and cohort studies have reported a considerable protective effect (ranging between 40-80%) in preventing IPD in different populations [23–28].
In general, meta-analyses of RCTs concluded that PPV did not prove any significant protective effect whereas those meta-analyses which also included observational studies in their analysis concluded that PPV could be considerably effective in preventing IPD (although its efficacy could be low or null among immunocompromised, high-risk and older adults) [7–17].
In the present study, a significant adjusted vaccine effectiveness of 72% (46-85) against overall IPD has been found. The effect of the vaccine in preventing vaccine-type IPD was greater, with an estimated effectiveness of 77% (40-92).
In our study, vaccination was significantly effective as for people 60-79 years as well as for older people 80 years or more. In addition, vaccination appears significantly effective for high-risk people and subjects with possible immunocompromise. Our findings fit with a prior case control study that showed an effectiveness of 66% against all bacteraemic pneumococcal pneumonia and 76% against vaccine-type infections.
Our results differ with those reported in the most recent meta-analysis by Huss et al, who concluded that PPV would be not efficacious against either IPD or pneumonia. However, in the present authors opinion, with regard to IPD this conclusion exceeds the scope of the evidence from the meta-analysis, given that pneumococcal bacteraemia was a rare event in the included RCTs (with only 44 cases of pneumococcal bacteraemia from 6 studies including 32 770 participants) and confidence intervals (CI) of vaccine efficacy were extremely wide.
In contrast, our findings fit with those reported in the latest Cochrane systematic review, which included 10 RCTs involving 35,483 participants assessing this outcome, showing a vaccine efficacy of approximately 74% (95% CI: 56% to 85%) against all IPD. In the mentioned Cochrane review, the pooled vaccine effectiveness based on 7 observational studies reached 52% (95% CI: 59% to 63%), which is also in agreement with the result observed in the present study.
Our data is also in agreement with data found in a prior cohort study among elderly people in the same geographical area, which pointed to an effectiveness of 40% against overall IPD. Our result also accords with vaccine effectiveness against IPD reported in the case-control studies of Shapiro and Clemens (67%), Sims et al (70%), Farr et al (81%) and Dominguez et al (72%).
In the present study, stratified analyses according to age subgroups showed that vaccination was significantly effective for all people 60 years or older. We did not found evidence that vaccination effectiveness decreases with increasing age considering that in our study vaccine effectiveness was 68% (26-86) in people 60-79 years and it reached 71% (9-91) in people 80 years or more.
According to risk strata of the patients, in our study the pneumococcal vaccination was significantly effective in risk strata 1 (immunocompromised subjects) and 2 (immunocompetent subjects with any high-risk condition). Vaccination did not emerge significantly effective among the subgroup of patients in risk strata 3 (immunocompetent subjects without high-risk conditions); however, in order to interpret this result, it must be noted that our study was clearly underpowered to detect a possible benefit of vaccination in this subgroup considering the low vaccine coverage observed among these patients (27.8%) and the small number of case-control sets that contributed to this analysis (only 18 sets). Surprisingly, we did not find lower vaccine effectiveness among immunocompromised people in risk stratum 1. However, in order to interpret this result, it must be noted that this stratum included patients with possible (but not confirmed) immunocompromise and a few patients assigned to this risk stratum had conditions associated with severe immunocompromise (see footnote in table 5).
A limitation of case-control studies is their observational nature, which can lead to bias and confounding. We took care to avoid bias in selection of controls, by using rigorous methods to select control subjects and matching cases to controls as closely as possible according to the main risk factor and further adjusting for underlying diseases in the multivariable regression models. It must be noted that, in our study, case patients had a significant lower coverage of influenza vaccination than control subjects, and this could have contributed to a higher rate of pneumococcal infection among case subjects. Thus, considering influenza as a possible confounder, we made stratified analysis on vaccine effectiveness within influenza and noninfluenza periods and adjusting for influenza vaccine status. Nevertheless, a residual confounding on the estimates of vaccine effectiveness can not be completely excluded, perhaps because of undiagnosed illnesses, differences in functional status of the patients or other epidemiological factors not considered in this study.
The study's major strengths were that it was population-based, and the study population (people 60 years or older with and without chronic illness) was representative and large enough to evaluate the most specific outcome (IPD due to vaccine serotypes) related with pneumococcal vaccination in adults. In addition, vaccine effectiveness was estimated adjusting for important covariables such as age, underlying conditions and influenza vaccine status. Thus, although it was not a RCT, it provides an adequate basis for assessing the potential benefit of the 23-valent PPV in preventing IPD among people 60 years or older, with and without high-risk conditions.