eTable. Multivariable logistic regression analysis of factors predicting pneumococcal vaccination. Propensity score model.
Vila-Corcoles A, Ochoa-Gondar O, Rodriguez-Blanco T, de Diego C, Satue E, for the EPIVAC Study Group. Ineffectiveness of Pneumococcal Vaccination in Cardiovascular PreventionThe CAPAMIS Study. JAMA Intern Med. 2013;173(20):1918-1920. doi:10.1001/jamainternmed.2013.6901
The benefits of the 23-valent pneumococcal polysaccharide vaccine (PPV23) are controversial.1 Recently, based on the association between pneumonia and acute thrombotic events, it has been suggested that pneumococcal vaccination may protect patients from coronary and/or cerebrovascular attacks. Several retrospective studies have investigated this issue with conflicting results.2
We designed a 3-year prospective cohort study, known as CAPAMIS,3 with the major aim of evaluating PPV23 effectiveness against community-acquired pneumonia (CAP), acute myocardial infarction (AMI), and stroke among the general population 60 years or older. An interim analysis covering the first 12-month follow-up has been reported.4 This Research Letter reports the final results of the study at the end of 36-month follow-up.
The study design has been described elsewhere.3 Briefly, this is a closed, population-based prospective cohort study including 27 204 individuals 60 years or older assigned to 9 primary care centers (PCCs) in Tarragona, Spain, who were prospectively followed from December 2008 until November 2011. The PCC’s electronic clinical record system was used to classify cohort members by their pneumococcal vaccination status as well as to identify baseline characteristics and underlying conditions of the cohort at study start. Primary outcomes were hospitalization for CAP, AMI, and ischemic stroke.
Community-acquired pneumonia-related episodes of AMI/stroke were considered when they occurred within 30 days after CAP diagnosis. Outcomes were initially identified on the basis of hospital discharge International Classification of Diseases, Ninth Revision, diagnosis codes for pneumonia (codes 480-487.0), myocardial infarction (code 410), and ischemic stroke (codes 433, 434, 436, and 437). Hospital clinical records were reviewed by a physician for all presumptive cases; only physician-verified diagnoses were included.
Pneumococcal vaccination status was determined by a review of the PCC’s clinical records, which contain specially designated fields for pneumococcal and influenza vaccinations (virtually all of them are administered at the PCCs in the Spanish health system). At the beginning of the study, cohort members were classified as vaccinated against pneumococcus if they had received at least 1 dose of PPV23 within 5 years before the study started. Across the study period, pneumococcal vaccination status was a time-varying condition given that some individuals received PPV23 after the study start.
Cox regression models were used to assess the association between having received PPV23 and the time to the first outcome. We modeled separate analyses: first, to adjust for potential confounding by indication, we estimated PPV23 effectiveness adjusted by propensity score for vaccination (eTable in Supplement); second, we performed classical multivariate analysis adjusted for age, sex, influenza vaccination, and underlying conditions.
Cohort members were followed for a total of 76 033 person-years, of which 29 065 person-years corresponded to immunized individuals. Vaccinated persons were slightly older, had a higher proportion of influenza vaccination, and had more comorbidities than nonvaccinated individuals (Table 1).
A first episode of CAP occurred in 566 persons; AMI in 359 (16 were CAP-related episodes); and ischemic stroke in 343 (4 were CAP-related episodes). Incidence rates (per 1000 person-years) were 7.51 for CAP (7.19 vs 7.71 among vaccinated and unvaccinated individuals, respectively), 4.75 for AMI (4.74 vs 4.75), and 4.53 for ischemic stroke (4.60 vs 4.49).
In the primary analyses, PPV23 vaccination did not substantially alter the risk of CAP (hazard ratio [HR], 0.94 [95% CI, 0.80-1.12]; P = .51), AMI (HR, 1.01 [95% CI, 0.82-1.25]; P = .92), or ischemic stroke (HR, 1.03 [95% CI, 0.83-1.28]; P = .81). After propensity and multivariate adjustments, PPV23 vaccination remained unassociated with the outcomes studied (Table 2).
Our data do not support a cardiovascular protective role of PPV23 and refute prior data reporting a protective effect in our interim analysis.4 Our findings show an increased short-term risk of acute thrombotic events among patients with CAP (2.8% and 0.7% of them developed AMI or ischemic stroke within 30 days after a CAP diagnosis, respectively), but vaccination did not reduce these risks.
The major strengths of this study were its population-based design and the validation of outcome events by checking clinical records. The main limitation is that vaccination status was not randomized. We used rigorous methods to adjust for the propensity for vaccination and further adjusted for underlying conditions in multivariate analyses; however, as with all observational studies, a residual confounding cannot be completely excluded. Those individuals who had received PPV23 more than 5 years ago were considered to be unvaccinated. It is possible that some of these individuals could have a certain degree of antibody immunity that was not considered; however, a possible bias is unlikely considering that vaccination did not seem to be effective in more recent vaccinated subjects.
In conclusion, we found that the PPV23 does not provide any clinically relevant benefit against overall CAP, AMI, or stroke among the general population old than 60 years. More effective antipneumococcal vaccination strategies (eg, using conjugated and/or protein-based pneumococcal vaccines)5,6 for adults are needed.
Corresponding Author: Dr Vila-Corcoles, Institut Catalá de la Salut, Prat de la Riba 39, Tarragona 43001, Spain (firstname.lastname@example.org).
Published Online: May 27, 2013. doi:10.1001/jamainternmed.2013.6901.
Author Contributions: Dr Vila-Corcoles had full access to all the data in the study and had final responsibility for the decision to submit for publication.
Study concept and design: Vila-Corcoles and Ochoa-Gondar.
Acquisition of data: de Diego and Satue.
Analysis and interpretation of data: Vila-Corcoles, Ochoa-Gondar, and Rodriguez-Blanco.
Drafting of the manuscript: Vila-Corcoles, Ochoa-Gondar, de Diego, and Satue.
Critical revision of the manuscript for important intellectual content: Vila-Corcoles, Ochoa-Gondar, and Rodriguez-Blanco.
Statistical analysis: Rodriguez-Blanco.
Obtained funding: Vila-Corcoles and Ochoa-Gondar.
Administrative, technical, and material support: de Diego and Satue.
Study supervision: Vila-Corcoles and Ochoa-Gondar.
Conflict of Interest Disclosures: None reported.
Funding/Support: Funding for the study was provided by a grant from the “Fondo de Investigación Sanitaria” of the Instituto de Salud Carlos III (FIS 09/00043) of the Spanish Health Ministry.
EPIVAC (EPIdemiology and VACcinology) Study Group Members: The following persons, investigators of the EPIVAC Study Group, collaborated in this study: Frederic Gomez Bertomeu, MD, PhD, Xavier Raga Luria, MD, PhD, Empar Valdivieso López, NP, Elisabet Salsench Serrano, PhD, Cristina Torrente Fraga, MCS, Manel Jariod Pamies, MD, PhD, Beatriz Gonzalez Lamban, MS, Antonia Gutierrez Perez, PhD, and Angel Vila Rovira, MS.
Correction: This article was corrected on October 10, 2013, to correct the study group name.