Article Text

Download PDFPDF

Should all children be immunised against influenza?
  1. Valtyr Thors1,2,
  2. Calum Smith1,
  3. Adam Finn1,2
  1. 1Schools of Clinical Sciences and Cellular and Molecular Medicine, University of Bristol, Bristol, UK
  2. 2University Hospital Bristol NHS Foundation Trust, Bristol, UK
  1. Correspondence to Dr Valtyr Thors, Level 6 Education Centre, Upper Maudlin Street, Bristol BS2 8AE, UK; valtyr.thors{at}

Statistics from

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.

Introduction—flu vaccines in children and indirect effects

Influenza is an important cause of morbidity and mortality, especially in combination with secondary bacterial infections.1–3 Annual influenza vaccination is recommended for everyone at risk by the WHO.4 In recent years, a number of countries have recommended influenza vaccination for all children older than 6 months although the uptake has been variable. The effectiveness of inactivated influenza vaccines in children has been questioned.5 Numerous studies have been published on the subject but outcome measures used vary with some studies using influenza-like-illness while others use culture or PCR-proven influenza, making comparison and meta-analyses difficult. In several randomised clinical trials, live attenuated influenza vaccine (LAIV) has been found safe, effective compared to placebo and consistently more effective than trivalent inactivated vaccine (TIV) in children. In 2012, UK authorities announced plans to offer annual LAIV to all children aged 2–17 years and in July 2013 that a single dose of the vaccine will be offered to all 2-year-olds and 3-year-olds from September 2013. Although the evidence base supporting this decision is robust, some important questions remain unanswered. Such a campaign, if carried out successfully, could significantly reduce the burden of disease in children and, since children are thought to be important in the propagation of infection within the population and thus development of influenza epidemics, this initiative may well impact on disease in other age groups through indirect protection. However, few studies quantifying such effects have been done to date. In this paper, aspects of the implementation of universal childhood influenza vaccination are discussed.

Disease burden, epidemiology and interactions with bacteria

Although it is widely believed to be a relatively insignificant infection in childhood, between 10% and 30% of children are infected with influenza during every annual epidemic,6–8 an attack rate higher than any other age group with a rate of complications that is especially high in the youngest. In the UK, it has been estimated that a typical epidemic results in approximately one million general practice consultations of all ages, 25 000 hospital admissions and 20 000 excess deaths.9 The youngest children have comparable influenza-related admission rates to the elderly, and most of the former have no recognised risk factor. Although paediatric deaths due to influenza are very rare, among them, over half occur in previously healthy children who are not eligible for vaccine in programmes targeted only at high risk groups.10 ,11 Estimating the exact burden of influenza is imprecise as only a small proportion of infected people seek healthcare and a small fraction of these are virologically tested. In addition, time off work because of influenza-related illness or caring for sick children is hard to measure as such data are not collected systematically. As a consequence, estimates of excess morbidity and infectious disease models are often used as tools to infer burden of disease and are useful as far as they go in the absence of better surveillance data.9 Since children are ill and excrete the virus for longer and at higher titres than adults, it is often stated that young children are the engines of every influenza epidemic.12 ,13 This hypothesis is supported by a reported association between transiently lower rates of excess deaths in all age groups attributed to influenza and pneumonia in Japan during a period when school children were required to receive annual flu vaccination.14 Similarly, during the 2009 influenza pandemic, after two Japanese districts closed all schools, the attack rates of influenza declined15 only to increase again later that year when schools reopened.16

However, influenza does not occur in isolation. Common complications in young children include acute otitis media and pneumonia, driving prescription of antibiotics and thus emergence of antimicrobial resistance. Secondary bacterial infections do undoubtedly occur and viral–bacterial interactions are turning out to be complex. There is evidence that respiratory viral infections including influenza alter bacterial colonisation patterns in the nasopharynx in children, potentially setting the stage for bacterial disease.17 ,18 Accordingly, prevention of influenza in children could impact on incidence of bacterial disease as well. Vu et al reported a case–control study performed in Vietnam including 274 children with radiologically confirmed pneumonia, 276 with other lower respiratory tract infections and 350 controls from the community. It showed very significantly higher colonisation rates and densities of pneumococci in the pneumonia cases. Among those coinfected with a virus, there was a 15-fold higher density.19 Several studies have also reported coinfection with viruses in bacterial otitis media.20 ,21 While pneumococci, Haemophilus influenzae and Moraxella catarrhalis are most often cultured from the middle ear, the human nasopharynx is also colonised with many other bacterial species that are not detected by standard microbiological techniques as has been demonstrated by Bogaert et al22 using next-generation sequencing methodologies. Further intricacies of nasopharyngeal ecology and the effects that flu has upon it remain to be elucidated.

Former strategies and new developments

Until recently, the available primary prevention tools have been inactivated vaccines made from influenza A (H1N1 and H3N2) and B viruses mass produced in hens’ eggs and given by intramuscular injection. Vaccines have been reformulated each year based on springtime predictions of the strains considered to be most likely to be circulating the following autumn. Vaccines are thought to rely predominantly upon haemagglutinin and to lesser extent neuraminidase antigens to induce specific protection primarily as antibody responses. Since these antigens are subject to antigenic drift, their effectiveness depends heavily on how well matched the predicted strains are to those which actually circulate and this can vary widely from 1 year to another. Much of the controversy around the usefulness or otherwise of flu vaccines stems from this problem simply because, not infrequently, the vaccine in use or in trials has the wrong antigens in it.

Several new approaches are now being taken towards the manufacture of inactivated flu vaccines. First is the use of cell culture to propagate virus instead of eggs, thus circumventing the potential limitation of egg supply and the issue of egg allergy. Second, the use of adjuvants to enhance immune response and reduce individual dose requirements, potentially allowing limited supply of vaccine to reach more people. However, adjuvanted flu vaccines have now twice been associated with severe adverse events,23 ,24 experiences which may inhibit their future development and use. Third, recognition that the single B strain hitherto included in trivalent vaccines has resulted in frequent B mismatch in the context of emergence of two distinct B lineage viruses worldwide over several years has resulted in the development of inactivated and live quadrivalent vaccines containing two A and two B strains. Such vaccines will be available in the USA in 2013 and should reach Europe in 2014.25 Finally, efforts are in progress to identify and develop effective vaccines using antigens or epitopes not subject to continuous genetic change, which is such a problem with current vaccines.26 These last vaccines remain some way away from the clinic at present.

In the context of these progressive developments in the existing portfolio of vaccines, the standard strategic approach to vaccine use, as recommended by the WHO, has been to target those considered to be at enhanced risk of developing severe disease when infected, the largest group of whom are the elderly. Aside from the generic difficulty in all settings of consistent successful implementation of targeted immunisation programmes, this approach also suffers from the paradox that as vulnerability to flu increases with advancing age or other forms of incapacity, capacity to make useful immune responses to immunisation is also diminished.27

Alternative approaches have been taken including the two-decade-long universal school programme in Japan alluded to above and a progressive broadening of recommendations in the USA during the first decade of this century, until they became universal for all ages above 6 months. More recently, there is evidence that this approach has led to somewhat higher coverage so that evidence of impact on flu incidence and mortality is eagerly awaited. In Europe, Finland recommended universal flu immunisation of children up to 3 years old in 2007 following detailed cost–benefit evaluation.28 However, uptake there has been severely hit in the aftermath of the recognition of an association between the adjuvanted monovalent H1N1 pandemic flu vaccine Pandemrix and cases of narcolepsy. Even though this adverse event has not been described in association with the standard unadjuvanted trivalent seasonal vaccines in current use, general public acceptance of flu vaccines in Finland has been badly dented.

The new arrival on the European scene is LAIV. Although it has been used in a relatively small proportion of the growing number of vaccine recipients in the USA since 2003 and is licensed there for between the ages of 2 and 49 years, the European licence, for ages 2–17 years, only arrived in 2011 and supplies of vaccine in 2012. The viruses for this vaccine are also cultured in hens’ eggs, but induction of allergic reactions has not been observed and is not expected among the many children who have degrees of egg allergy and a safety study which may confirm this is in set-up in the UK for the 2013 season. This vaccine is also formulated every year following WHO strain recommendations. The attenuated vaccine strains are cold-adapted and temperature sensitive so that they can remain viable and replicate efficiently but only in the relatively cold environment of the upper respiratory tract, where they induce transient and sometimes mildly symptomatic infection and protective immune responses.29 The modified gene segments responsible for this phenotype derive from master donor virus strains while the dominant haemagglutinin and neuraminidase antigens derive from relevant wild-type circulating viruses selected for the year of manufacture.

LAIV has been found to be significantly more effective in preventing virologically proven influenza than TIV in head to head trials,30–33 a difference which is most evident in younger children. Side effect profiles of the two vaccines are distinct but overall rates are similar. LAIV has an absolute efficacy of about 80% (CI 69 to 91) based on six randomised placebo controlled trials over several seasons.34–40 A potentially important limitation to this vaccine is the lack of licensure for children younger than 2 years old. Although shown to be efficacious in this age group, one study showed higher hospitalisation rates over 6-month postvaccination for diverse causes in 6–11-month-old vaccine recipients, while 6–23-month-olds had more episodes of wheezing.30 These safety signals were not seen in other trials.34–40

Universal childhood immunisation in the UK

Recent and current innovations in approaches to immunisation in the UK include a new schools-based programme (against human papillomavirus) and consideration of predicted indirect effects into the cost–benefit studies and transmission models that drive policy (on control of pneumococcal and meningococcal disease). Indeed, the best way to construct cost–benefit calculations for an intervention which will drive savings over a period of many years or even decades is a topic of hot debate. Taken together with the historical experience that a mucosal vaccine (oral polio vaccine) is highly acceptable to parents, and despite remaining uncertainties regarding the exact size of its power to induce indirect protection in the wider population, the prospect of a flu vaccine that can be widely and easily administered in a national programme is highly attractive. One recent paper predicts that immunising all children 2–17 years old would be highly cost-effective when added to existing programmes for risk groups and the elderly and even when indirect effects are ignored and that 50% coverage would lead to up to 95% reduction in hospitalisations and deaths due to influenza and influenza-related disease in all age groups.41 Personal experience of using the vaccine in nurseries and a primary school in Bristol during the 2012–2013 season has certainly suggested to us that most parents can be expected to embrace the vaccine for their children with enthusiasm.

In this context, a series of published minutes and then public announcements have progressively provided certainty and details about a new UK programme which will gain momentum over a period of 2–3 years. In 2013, a single dose of LAIV will be offered to all 2-year-olds and 3-year-olds and a series of local pilot roll-out programmes will be established for different age groups in different regions.42 If the programme extends to cover the entire 2–17-year-old population, as currently envisaged, the logistics of implementation will clearly be very challenging indeed given the large numbers of recipients and short time scale for delivery in the autumn. It will be interesting to see how this is tackled but, given that most of this group are school age, a schools-based programme involving school staff seems highly probable. It is also safe to assume that the uptake rates of vaccine and national and local epidemiology of whatever flu epidemic emerges during the 2013–2014 winter season will be followed very closely and with great interest by the UK authorities and by flu watchers everywhere. It would be no exaggeration to state that, if successful in terms of public acceptance and strain match, this programme could fundamentally change strategy towards flu control worldwide, particularly if, in addition to reduced illness among children, predicted indirect protection of the elderly and reduced associated morbidity and mortality are also seen.

Indirect effects

Although the trodden pathway to vaccine licensure is based on studies which measure direct protection of vaccine recipients in randomised controlled trials and serological correlates of direct protection (antibody titres) intended either to predict efficacy or to demonstrate equivalence to vaccines already in use, the truth is that the way many vaccines actually prevent disease once deployed in universal schedules is primarily by blocking transmission within the population so that it largely or entirely disappears even though a substantial minority remain non-immune. Despite near-universal unawareness of the fact, when you have your child immunised you are doing a lot more than just protecting your child.

In the past, these effects have usually been ignored entirely in formulating policy, as they usually are during vaccine development and are recognised later as a kind of bonus extra. Sometimes, they are predicted rather approximately either using epidemiological data from other countries already using the vaccine under consideration or with epidemiological modelling techniques of rapidly increasing mathematical sophistication but which characteristically lack good quality real epidemiology to feed them, relying often on very approximate estimates of incidence of infection and disease.

The path towards LAIV introduction in the UK has been little different in this respect, although it is clear that R0—the index of infectiousness which determines the level of vaccine coverage necessary to interrupt transmission—is perilously (for the virus) close to one for influenza.43 Accordingly, there are very good reasons to expect potent indirect effects of high coverage immunisation with an efficacious vaccine. This has been demonstrated previously for TIV in a high-quality cluster randomised study conducted in Canada.44 Studies of this type have the potential to measure the capacity of a vaccine to induce indirect effects with great accuracy by observing whether immunising some individuals in defined communities (in this case 80% of children) reduces the disease in the rest (in this case by 60% in adults) when compared to similar communities where the vaccine is not deployed. Although two non-randomised observational studies of this kind using LAIV have been published45 ,46 both suggesting an effect, to date a prospective trial of this kind using LAIV remains to be done. The results could be enormously valuable in informing policy makers in the UK and elsewhere what coverage they really need to be aiming for and how broad an age range actually needs to be offered vaccine in order to achieve disease control.

Conclusions—the future

Given the limited understanding, we have of viral–bacterial interactions in the respiratory tract and specifically what role flu and other respiratory viruses play in the pathogenesis of bacterial respiratory disease, and given the uncertainties that exist around the impact, if any, of previous influenza vaccination programmes, it is possible to speculate that universal childhood LAIV, given good strain match and high coverage, may do a lot more than just prevent influenza. However, for the present, this is not possible to predict with accuracy. Studies and surveillance to evaluate the effects of the vaccine on bacterial colonisation and disease, as well as its impact on flu, would make very good sense. Ultimately, it could conceivably prove that viral vaccines are as useful in the prevention of bacterial otitis, pneumonia, sepsis and even meningitis as bacterial vaccines are—but if we do not look, we will not find out.

The fact that the relatively obvious superiority of LAIV over TIV in younger children is less evident as they get older could simply reflect a better capacity of the live vaccine to stimulate the relatively less mature immune system of the young child. Alternatively or additionally, it could indicate that previous exposure to flu, as is present in older children, results in a degree of cross-strain immunity which might inhibit replication of the live attenuated vaccine viruses and thus its capacity to induce strain-specific protection. Such hypothetical possibilities, added to the existing uncertainty about the size of capacity of the vaccine to induce indirect protection in the population at large, will make it very important to devise innovative studies to inform future modifications to the programme as a highly immune, annually immunised population of children emerges over coming years.

Finally, we all know very well that flu and other respiratory viral infections are (usually) seasonal but, despite no shortage of hypotheses, we do not actually know why. If we are going to deliberately infect a large proportion of the population with this and, potentially in the future, other live attenuated respiratory viruses, we should not ignore the opportunities this offers to increase our understanding of the mechanisms which explain how these infections happen.


View Abstract


  • Contributors VST conceived the content of the manuscript, co-wrote and finalised the manuscript. CS co-wrote the manuscript and contributed to the final edit. AF conceived the content of the article, co-wrote the manuscript and performed the final edit.

  • Competing interests VT is an ESPID Research Fellow and is currently undertaking research partially funded by a grant to the University of Bristol from Astra-Zeneca. AF undertakes consultancy and clinical research for all the main vaccine companies including most of those manufacturing flu vaccines. All funding related to these activities is paid to his employers, the University of Bristol and University Hospitals Bristol NHS Foundation Trust. He receives no personal remuneration related to these activities and has no other financial interest in these companies or any related intellectual property.

  • Provenance and peer review Commissioned; internally peer reviewed.

Linked Articles

  • Atoms
    R Mark Beattie