Correspondence
Pneumococcal Vaccination in Older Adults
N Engl J Med 2003; 349:712-714August 14, 2003
- Article
To the Editor:
Jackson et al. (May 1 issue)1 demonstrate that the polysaccharide pneumococcal vaccine did not reduce the incidence of pneumonia among older adults — a finding that is consistent with the results of the previous blinded, randomized, controlled trials.2 The authors assert, however, that vaccination should nevertheless continue, because a study3 calculated that it is cost effective in preventing pneumococcal bacteremia in persons in this age group. That report, however, tacitly assumed that a reduction in the frequency of pneumococcal bacteremia among vaccinees translated into a decrease in the frequency of pneumonia — a premise that the current study and the previous randomized trials refute. The appropriate comparison in terms of cost analysis is that between the expenses associated with hospitalization for pneumococcal bacteremia among unimmunized persons and the combined expenses associated with immunization plus an equivalent number of hospitalizations among vaccinees for pneumonias without pneumococcal bacteremia. A reexamination of the estimates demonstrates that such a calculation would not justify vaccination for the prevention of pneumococcal bacteremia.
3 ReferencesJan V. Hirschmann, M.D.
Puget Sound Veterans Affairs Medical Center, Seattle, WA 98108
pepsi@u.washington. edu 1
Jackson LA ,Neuzil KM ,Yu O , et al. Effectiveness of pneumococcal polysaccharide vaccine in older adults. N Engl J Med 2003;348:1747-1755
Full Text | Web of Science | Medline2
Hirschmann JV . Use of the pneumococcal polysaccharide vaccine is unwarranted in the US. ASM News 2000;66:326-3273
Sisk JE ,Moskowitz AJ ,Whang W , et al. Cost-effectiveness of vaccination against pneumococcal bacteremia among elderly people. JAMA 1997;278:1333-1339[Erratum, JAMA 2000;283:341.]
CrossRef | Web of Science | Medline
To the Editor:
In addition to bias from misclassification of the outcome of pneumonia, a residual bias due to confounding in the observational study by Jackson et al. may explain the increased risk of hospitalization for pneumonia among persons who received pneumococcal vaccination.1,2 Patients at higher risk will be more strongly advised to receive the vaccine than those at lower risk. This difference is clearly demonstrated in Table 1 of the article, which shows that the frequency of many risk factors was significantly higher among persons who had been vaccinated than among those who had not. This “confounding by indication” will reduce the estimate of a beneficial effect. The conventional regression analysis performed by the authors should have been complemented with other powerful methods such as propensity scores to examine the potential for residual bias.3 However, misclassification of cases of pneumonia cannot be controlled for, and there might be residual confounding. Therefore, a sufficiently powered prospective, randomized, controlled trial among the elderly remains the best approach to providing a valid estimate of the effect of pneumococcal vaccination on the incidence of pneumonia.4
4 ReferencesEelko Hak, Ph.D.
Marc J.M. Bonten, M.D., Ph.D.
Arno W. Hoes, M.D., Ph.D.
University Medical Center Utrecht, 3508 AB Utrecht, the Netherlands
e.hak@med. uu. nl 1
Hak E ,Verheij TJM ,Grobbee DE ,Nichol KL ,Hoes AW . Confounding by indication in non-experimental evaluation of vaccine effectiveness: the example of prevention of influenza complications. J Epidemiol Community Health 2002;56:951-955
CrossRef | Web of Science | Medline2
Hak E ,Nordin J ,Wei F , et al. Influence of high risk medical conditions on the effectiveness of influenza vaccination among elderly members of 3 large managed-care organizations. Clin Infect Dis 2002;35:370-377
CrossRef | Web of Science | Medline3
Joffe MM ,Rosenbaum PR . Propensity scores. Am J Epidemiol 1999;150:327-333
Web of Science | Medline4
Mangtani P ,Cutts F ,Hall AJ . Efficacy of polysaccharide pneumococcal vaccine in adults in more developed countries: the state of the evidence. Lancet Infect Dis 2003;3:71-78
CrossRef | Web of Science | Medline
To the Editor:
In the carefully performed cohort study by Jackson et al., the finding of a significant reduction in the rate of the most specific end point, pneumococcal bacteremia, is both plausible and useful. However, the discordant finding of an increased risk of hospitalization for community-acquired pneumonia is confusing. If, in contrast to the study's findings, the true rate of hospitalization due to community-acquired pneumonia was actually lower among vaccinated subjects, the observed result might be attributable to either the misclassification of admissions that were actually not associated with community-acquired pneumonia (inadequate specificity) or the misclassification of admissions that actually were associated with community-acquired pneumonia (inadequate sensitivity). Given the highly specific definition that was used for hospitalization associated with community-acquired pneumonia, I wonder whether admissions that were actually associated with community-acquired pneumonia could have been misclassified. This theory would be supported by a finding that previous vaccination was associated with a decreased risk of hospitalization for all causes. Did the analysis include testing for an association between the rate of hospital admission for all causes and vaccination status?
Theodore K. Marras, M.D.
University of Toronto, Toronto, ON M5G 2C4, Canada
ted.marras@utoronto. ca To the Editor:
In the report by Jackson et al., it is difficult to understand how an intervention that increased the rate of hospitalization for pneumonia could also offer such extraordinary protection against death from any cause. Among immunocompetent persons, for example, vaccination reduced the risk of bacteremia by 54 percent (39 events); it also reduced the risk of death from any cause by 12 percent (3613 events). Vaccination should prevent death only by preventing pneumococcal bacteremia itself or reducing its severity, yet the number of deaths from any cause exceeded the number of bacteremic episodes by a factor of more than 90. How were such apparently contradictory effects on these two important outcomes achieved?
David S. Fedson, M.D.
57 Chemin du Lavoir, 01630 Sergy Haut, France
dfedson@wanadoo.fr Author/Editor Response
Our study, along with others, provides strong evidence that pneumococcal polysaccharide vaccine protects against bacteremic infection. Although such infection is a relatively uncommon outcome, the prevention of these serious events saves money.1 Hirschmann speculates that vaccination is of little clinical benefit if bacteremic pneumonias are replaced by nonbacteremic pneumonias. In clinical trials involving young adults in South Africa, where bacteremia commonly accompanied pneumococcal pneumonia, receipt of pneumococcal polysaccharide vaccine was associated with a reduction in the risk of pneumococcal bacteremia and with a reduction of more than 50 percent in the risk of any radiographically confirmed case of pneumonia.2 Thus, there was no evidence of replacement by other pneumonias in that population. In our population, bacteremic pneumococcal pneumonia accounted for only approximately 2 percent of all cases of pneumonia. Hence, a 50 percent reduction in the frequency of bacteremic pneumonia could, at most, result in a 1 percent reduction in the frequency of pneumonia of any cause, which would be difficult to detect even in a large clinical trial. Clearly, new vaccines that could prevent a larger proportion of cases of pneumonia as well as invasive disease are desirable, but the evidence supports the current recommendations for the use of pneumococcal polysaccharide vaccine in elderly persons.
Hak et al. and Marras suggest that the inaccurate ascertainment of outcomes of pneumonia may have influenced our estimates of the effectiveness of the vaccine. We think this is unlikely. Episodes of pneumonia were presumptively identified according to broad criteria (any diagnostic code for pneumonia). To ensure specificity, presumptively identified hospitalizations for pneumonia were validated with the use of chart review. Our rates of hospitalization for community-acquired pneumonia are very similar to those reported in a population-based active-surveillance study in Ohio.3 Since our report was published, we have also validated all episodes of pneumonia in outpatients with the use of radiographic criteria. The association between pneumococcal vaccination and this outcome is consistent with the previously reported association between pneumococcal vaccination and pneumonia in outpatients as defined on the basis of administrative data.
We included death as a secondary outcome and did not find an association between vaccination and the risk of death in the analysis including all subjects, but as Fedson observes, there was a significant reduction in the risk of death in some subgroup analyses. As we mentioned, it is likely that factors such as the severity of illness are more important confounders of the association of vaccination with death than of the associations with the more specific outcomes of pneumonia and pneumococcal bacteremia. Accurate estimation of the relation between vaccination and mortality from any cause may require the identification of other covariates for adequate adjustment.
3 ReferencesLisa A. Jackson, M.D., M.P.H.
Group Health Cooperative, Seattle, WA 98101
jackson.l@ghc. org Kathleen M. Neuzil, M.D., M.P.H.
Veterans Affairs Puget Sound Health Care System, Seattle, WA 98108William W. Thompson, Ph.D.
Centers for Disease Control and Prevention, Atlanta, GA 303331
Sisk JE ,Moskowitz AJ ,Whang W , et al. Cost-effectiveness of vaccination against pneumococcal bacteremia among elderly people. JAMA 1997;278:1333-1339[Erratum, JAMA 2000;283:341.]
CrossRef | Web of Science | Medline2
Austrian R ,Douglas RM ,Schiffman G , et al. Prevention of pneumococcal pneumonia by vaccination. Trans Assoc Am Physicians 1976;89:184-194
Medline3
Marston BJ ,Plouffe JF ,File TM Jr , et al. Incidence of community-acquired pneumonia requiring hospitalization: results of a population-based active surveillance study in Ohio. Arch Intern Med 1997;157:1709-1718
CrossRef | Web of Science | Medline
- Citing Articles (1)
Citing Articles
1
Ali A. El Solh, Thomas Brewer, Mifue Okada, Omar Bashir, Michael Gough. (2004) Indicators of Recurrent Hospitalization for Pneumonia in the Elderly. Journal of the American Geriatrics Society 52:12, 2010-2015
CrossRef
