Skip navigation

Pneumococcal vaccine update


Clinical bottom line

There is no evidence from systematic reviews of randomised trials that polysaccharide pneumococcal vaccines have any effect in patients in the high-risk or elderly whom it is used.

Systematic reviews can be a bit like buses - you wait for ages and then four come along together! Bandolier first visited this topic in issue 74 , but three more systematic reviews have now been published [1-4]. The main features of the four reviews are in Table 1.

Table 1: systematic reviews of pneumococcal vaccine efficacy

Systematic review, origin, sponsorship Search and dates Inclusion and exclusion criteria Trials found Sensitivity analysis Conclusion
Hutchison et al. Can Fam Physician 1999 45 2381-2393
McMaster, Canada
Sponsorship not mentioned
MEDLINE, EMBASE and SCISEARCH to November 1996, plus reference lists, and writing to authors and organisations. Adults, pneumococcal vaccine, one clinical outcome, randomised controlled trial, or prospective cohort or case-control study. 20 trials, including 13 randomised or quasi-randomised, 11 properly randomised Unblinded studies omitted.
Pooled results for different outcomes
Some attempt to examine different clinical situations
Statistical significance only for systemic pneumococcal infection in six trials, including one large unblinded study from 1947.
Moore et al. BMC Family Practice 2000 1:1.
Oxford, England
Sponsorship not mentioned
Cochrane Library, MEDLINE, PubMED, EMBASE to August 2000, reference lists and previous reviews Prospective, randomised polysaccharide pneumococcal vaccine (any valency) with placebo or no treatment in adults. Quasi-randomised design not included 13 RCTs in 12 reports Analysis as to whether vaccine recipients likely to have healthy or impaired immune response No significant benefit for any outcome in elderly or high risk patients.
Cornu et al, Vaccine 2001 19: 4780-4790
Lyons, France
Commercial sponsorship
MEDLINE, EMBASE to 1999, reference lists, manufacturers Randomised trials of polysaccharide pneumococcal vaccine in immunocompetent adults. 14 RCTs in 11 reports Age over 55 and high-risk patients No significant benefit in high risk patients
Watson et al, Vaccine 2002 20: 2166-2173 MEDLINE to May 2000, plus Cochrane Library, CINAHL, Inter-Dec, ASSIA, Gray literature, SIGLE plus Internet sources, references, and manufacturers and experts Adults, in randomised trials with no treatment or placebo control (not including conjugate vaccine) 16 RCTs in 12 reports Quality score, high risk, and over 65 years No significant difference for any outcome in high risk patients, industrialised countries, or high quality studies

The reviews

The four reviews have more similarity than differences. They are all looking at the use of pneumococcal vaccines in adults, using clinical outcomes, and largely confined to randomised trials. The later three exclude some of the case-control and quasi or non-random studies conducted in the 1940s, and one large more recent quasi-randomised study from Finland. The early study was positive, the later was not.

Sensitivity analyses of various sorts are defined and performed. If we excuse the inclusion of an open study from 1947 in the Hutchison review, they all come to the conclusion that there is no statistical benefit for any clinical outcome in patients like those to whom we give it - over the age of 65 years or with chronic organ dysfunction or immunosuppression.

They are all pretty good reviews of essentially the same trials. They use different search strategies, but review essentially the same trials. They make slightly different decisions about studies included or omitted, and slightly different decisons about numbers extracted from the papers, for reasons that are not always obvious. But these are quibbles. Four different groups look at the same trials and come up with the same answer.

The problem

The problem we have is whether or not we believe that trials done 30 years ago in South African gold miners or in New Guinea highlanders are representative of patients like ours. In these groups, pneumococcal vaccines were effective, and, because the trials were large, they have the potential to influence our thinking. Look, for instance, at Figure 1, a cumulative meta-analysis showing the relative risk of pneumonia-related death from 1977 to the present time. It looks like the next trial might just tip the confidence interval below 1, so giving statistical significance.

Figure 1: Relative risk of pneumonia-related death from 1977 to 1998 (bars are 95% confidence interval).

Now look at the same information plotted from the present back in time to 1976 in Figure 2. Now it looks more like confidence intervals where each additional trial just gives more confirmation of a lack of effect.

Figure 2: Relative risk of pneumonia-related death from 1998 to 1977 (bars are 95% confidence interval).

What is true is that the amount of information in patients like ours is deficient in numbers. It probably is insufficient to definitely say that there is evidence that pneumococcal vaccines do not work in our patients. We are left with a lack of evidence that they do work in our patients.

Event rate, or something else?

There are two possibilities. First is whether this is just an event-rate phenomenon. If event rates were higher in South African gold miners or New Guinea highlanders 30 years ago, perhaps that would account for an effect there and then, but no here and now. Maybe or maybe not. Table 2 shows a simple calculation of the incidence of different outcomes in studies before 1980 and those after 1980, effectively separating patients with ours from those we are unlikely to see. There is not much difference in overall event rates.

Table 2: Event rates with control for different outcomes before and after 1980


Incidence (%)

1980 or before

1985 or later

Lower respiratory tract infection









Pneumococcal pneumonia



Pneumonia-related death



That leaves us with the possibility, even probability, that something else may be different. For gold miners working at incredible depths and heat, it may be the combination of working and living conditions 30 years ago. For New Guinea highlanders, it may have been communal living in smoke-filled atmospheres of their longhouses. Whatever, there is no obvious cause of the difference, other than the obvious one that they are not like the patients we treat.

Is invasive streptococcus pneumonia infection a problem?

An epidemiological study from the United States shows the size of the burden in industrialised countries [5]. This was a population-based survey in nine US states with active surveillance to ensure as complete collection of data possible.

In the years 1995-1998 there were 15,860 cases of invasive pneumococcal disease, in persons whose ages ranged from 0 to 104 years. The overall incidence was 23 per 100,000, with highest incidence in children under two years and adults over 65 years. About one in 10 cases resulted in death for all ages, and rose steadily from low rates in children to high rates in the over-80s. Higher case-fatality rates were also seen in persons with congestive heart failure, cancer, or who abused alcohol or had cirrhosis.

For the United States in 1998, the estimated number of cases was 62,800, and deaths 6,080.

The future

New conjugate pneumococcal vaccines are on their way to replace the polysaccharide vaccines predominating now. Should we use them without good evidence that they reduce clinically important events? It is an interesting question worth returning to when evidence becomes available.


  1. BG Hutchison et al. Clinical effectiveness of pneumococcal vaccine. Can Fam Physician 1999 45: 2381-2393.
  2. RA Moore et al. Are the pneumococcal polysaccharide vaccines effective? Meta-analysis of the prospective trials. BMC Family Practice 2000 1:1 ( ).
  3. C Cornu et al. Efficacy of pneumococcal polysaccharide vaccine in immunocompetent adults: a meta-analysis of randomized trials. Vaccine 2001 19: 4780-4790.
  4. L Watson et al. Pneumococcal polysaccharide vaccine: a systematic review of clinical effectiveness in adults. Vaccine 2002 20: 2166-2173.
  5. KA Robinson et al. Epidemiology of invasive Streptococcus pneumoniae infections in the United States, 1995-1996. Opportunities for prevention in the conjugate vaccine era. JAMA 2001 285: 1729-1735.