March 2, 2020

expert reaction to study looking at the flu vaccine and hospitalisation and mortality in older people

Research, published in Annals of Internal Medicine, reports that flu vaccines aimed at the elderly may not be as effective as previously thought. 

Prof Jonathan Ball, Professor of Molecular Virology, University of Nottingham, said:

“We know that the previous vaccine used to prevent influenza virus in the elderly sometimes did not induce good levels of protection in this vulnerable group and this varied from year to year.  For this reason, the adjuvanted vaccine – which produces much higher levels of immunity in the over-65s – was introduced.  The period that the current study focusses on does not include any data collected after this new vaccine was introduced, and therefore all it is doing, in a very powerful way, is confirming that the previous vaccine did not work as well as we would want, and underlines the reason why the new vaccine was introduced.

“It will be important to continue to monitor vaccine effectiveness to ensure that the new adjuvanted vaccine provides the protection that we hope for in this specific risk group.”

Dr Michael Skinner, Reader in Virology, Imperial College London, said:

“The paper reports an observational study which cleverly takes advantage of the availability of free influenza vaccination in the UK from age 65 and, by applying a ‘Regression Discontinuity Design’, becomes analogous to a ‘randomized trial with imperfect adherence’.

“The paper will make a useful, specific contribution to the question of how to protect the elderly from seasonal influenza but its conclusions are overstated in the title (“The Effect of Influenza Vaccination for the Elderly on Hospitalization and Mortality”), which extrapolates their results for those around 65 to “the elderly”, which is not necessarily a justified extrapolation.  The title should have adhered to the conclusion in the first line of their Discussion: “Our results showed a sharp increase in influenza vaccination rates at age 65 years with no matching decrease in hospitalization or mortality rates”; I would suggest “A sharp increase in influenza vaccination rates is not matched by decreases in hospitalization or mortality rates at age 65”.  The press release goes even further in exaggerating the scope of the study with the title “Flu vaccine may not decrease hospitalization or mortality among elderly persons” – again data about those age 65-70 cannot necessarily just be extrapolated to ‘the elderly’.

“There are some specific issues with the paper.  I should point out that I am not expert in such detailed statistical analysis (and am not familiar with all of the literature in the field to which they compare their approaches and results) but neither are the authors expert in immunology or infectious diseases.

“The USA-based authors used UK NHS hospital and GP databases.  Ideally, one would use randomised trials to assess the efficacy of influenza vaccination by age across the “elderly” spectrum, but these data are not available.  However, in the UK, the availability of free influenza vaccination in the UK from age 65 leads to a sharp increase in vaccination rates (the increase is not absolute because those with a range of conditions that predispose them to serious influenza infections receive free vaccination before 65, as do many in certain occupations, such as the NHS or teaching).  Essentially, this allowed the authors to use 64-year-olds as controls for the effects of vaccination on hospitalisation and mortality of 65-year-olds.  In practice, they needed higher numbers than that comparison offered so they extended the age ranges down to 60 for the controls and up to 70 for the ‘vaccinated’.

“Of course, vaccination at 65 is not the end but only the start of the process to protect ‘the elderly’.  Moreover, it is not the 65-74 year-olds who are the primary target of this process: 75-84 year-olds are considerably more susceptible to hospitalisation and mortality, and those over 85 are vastly more susceptible¹.  It is a pity the authors of the new study did not demonstrate this fact in their Fig. 2.

“Unfortunately it becomes harder to stimulate immune responses to vaccines as we get much older, so introduction at 65 is viewed as a compromise that is close enough to the target ages but early enough to be reasonably effective.  It also has the advantage of encouraging the habit of annual vaccination.

“Ideally we need to know whether the increased vaccination at 65 leads to reduced hospitalisations and mortality in 75-84 and >85 year-olds but the required control data are not available.

“There are also issues around influenza that need to be considered.  Seasonal influenza varies in its severity from year to year (and vaccination varies in its effectiveness)¹, but the authors report only across a broad range of years (2000-2011 for hospitalisation, 2000 to 2014 for deaths).  We really needed to see year by year details.  This is particularly so for 2009-2011 which corresponds to the H1N1 pandemic, against which seasonal flu vaccine was not effective.

“The authors want to contribute to the debate about the relative effectiveness of vaccinating the elderly with ‘killed’ flu vaccine versus vaccinating children with ‘live’ flu vaccine – with the aim of reducing morbidity and mortality in the elderly.  However, they have not taken into account the progressive introduction of the live vaccine in children in the UK during the years since the 2009 pandemic, which might be a confounder.”

1 Czaja et al. 2019; https://academic.oup.com/ofid/article/6/7/ofz225/5510081

‘The effect of influenza vaccination for the elderly on hospitalization and mortality’ by Michael L. Anderson et al. was published in the Annals of Internal Medicine at 22:00 UK time on Monday 2 March 2020.

DOI: 10.7326/M19-3075

Declared interests

Prof Jonathan Ball: “No CoIs.”

Dr Michael Skinner: “Dr Mike Skinner is a (semi-retired) Reader in Virology at Imperial College London.  His introduction to virology was with mouse coronaviruses at the hands of Stuart Siddell working in Germany long before SARS emerged.  He’s since worked on poliovirus vaccines, HIV and avian viruses (including viruses that have emerged to threaten global poultry production).  He takes avid interest in emerging viruses and vaccines.  He still works on projects led by others and funded by the BBSRC.  He no longer has any direct or indirect current interest in influenza vaccines.”