Article Text

Download PDFPDF

Systematic review
Available evidence points to low effectiveness of influenza vaccines for older people
  1. Lone Simonsen
  1. Department of Global Health, School of Public Health and Health Services, George Washington University, Washington, District of Columbia, USA
  1. Correspondence to Lone Simonsen
    Department of Global Health, School of Public Health and Health Services, George Washington University, 2175 K Street, NW, Suite 200, Washington, DC 20037, USA; lone{at}

Statistics from

Commentary on:

Each winter, most influenza deaths occur among seniors (although they were spared during the 2009 pandemic period). For nearly a decade, there has been a controversy brewing about the benefits of vaccinating seniors against influenza. The policy of vaccinating seniors was put in place in the 1960s in USA and other countries without much in the way of clinical trials demonstrating benefits, and the question was already raised at that time whether seniors could actually mount a protective response.1 More recently national analysis of influenza mortality time trends in several countries fueled this concern, as these studies saw no decline in flu-related deaths as vaccine coverage rose fourfold in seniors.2

This observation seemed to be in disagreement with the large body of evidence from observational studies that consistently reported a halving in total winter deaths (all cause mortality) in vaccinated seniors.3 4 But the paradox was there – how could there be no reduction in mortality from national studies while the observational studies showed vaccination had astonishing mortality benefits.5

In this issue, the Cochrane collaboration led by Tom Jefferson takes on this question systematically by conducting a structured meta-analysis of all experimental and non-experimental studies of flu vaccine efficacy and effectiveness studies with a clinical end point. This update follows two similar Cochrane reports on influenza vaccine benefits in seniors in recent years. Jefferson and colleagues conclude that most of the evidence base is observational studies with few placebo-controlled trials, that the role of vaccines to prevent flu is not clear and that investment in more immunogenic vaccines is needed. They call for a large multi-year placebo-controlled trial to resolve the uncertainties.

Because of the controversy, the systematic and careful reviews provided by the Cochrane group are necessary and important. But the rigidity of the meta-analysis methodology strategy may have been counterproductive in this case, in that key observational studies published since 2005 that really helped to resolve the paradox and point a way forward did not meet the inclusion criteria. These studies demonstrated that most observational studies were severely biased and greatly overestimated the vaccine benefits in seniors by showing that the observed differences in mortality in vaccinated and unvaccinated seniors were most pronounced in the autumn, months before flu ever circulated!6 These authors also showed that the standard approach for adjusting the bias is actually counterproductive.7 Furthermore, my colleagues and I used a modelling approach to demonstrate that the use of non-specific outcomes (all-cause mortality) combined with frailty selection bias (lower vaccination rates in frail seniors at higher risk dying for any reason) could comfortably explain the observed patterns of a halving in all winter deaths seen in observational studies without assuming any vaccine benefits in the model at all.5 8

And so I would like to think the field has moved forward recently to a mode of fixing the problem of inadequate protection of seniors rather than just pointing it out. Indeed, there is a budding literature of observational studies that carefully control for selection bias in updated study designs using specific clinical end points and adjusting for effects during the preflu periods. These studies systematically find low or no vaccine benefits in seniors.9,,11 Unfortunately the ‘new generation’ bias-free observational studies of vaccine effectiveness9,,11 are not included in any tables or calculations in the Cochrane review. Meanwhile vaccine dosing studies have pointed to solutions, showing that the antibody response is greatly enhanced with higher antigen dosing12,,15 or with adjuvanted formulations.16 So a technical solution is at hand.

In the meantime, the vaccine industry has also been taking on the challenge. A new senior vaccine formulation will be on the US market in the 2010–2011 season featuring a fourfold higher antigen content.17 In Europe, an adjuvanted vaccine formulation designed for seniors has been in use for some time. Finally, vaccine policy has been rapidly changing in the USA and Canada; most recently in USA to universal recommendation of influenza vaccine which may lead to a general reduction in flu transmission and thereby indirect (herd) the protection of seniors.18,,21

So some countries have embraced the new insights and are beginning to implement new strategies for control of influenza. It will be interesting to see whether the new strategies make a difference in mortality over the years to come, and whether other countries will follow.


View Abstract


  • Competing interests LS has received support from Pfizer for research related to pneumococcal vaccine and consulting fees from SDI, a health data analytics business.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.