Varicella vaccine

Editor, – Despite the risks, the article 'Frequently asked questions about varicella vaccine' (Aust Prescr 2005;28:2-5) recommends widescale immunisation. There are three arguments against this strategy. Firstly, vaccine immunity may wane over time leaving susceptible adults. Secondly, immunising part of the population may shift the disease burden to those who are not vaccinated and because they will be less likely to acquire chickenpox in childhood they risk more severe disease in adulthood. Thirdly, the effect of vaccination on the incidence of herpes zoster is unknown.

The data so far show that chickenpox in immunised individuals is less severe. However, it is too early to know how this will change as immunised infants reach adulthood.

In 2000 mathematical modelling showed that immunising 90% of infants would produce an initial 'honeymoon' period of low incidence, one or more post-honeymoon epidemics in adolescents and young adults 10-20 years later, and an equilibrium reached after 20-40 years in which the incidence in adults is similar to that in the pre-vaccine years.¹ The evidence from the USA on reduced incidence in all age groups covers only five years of experience, which is within the honeymoon period predicted by the modelling. This is insufficient time for epidemics in adults to occur through the build-up of susceptible people, as partial population immunity increases the interepidemic interval.

The impact of varicella vaccine on herpes zoster is complex. There is reasonable evidence that adults exposed to children, or exposed to chickenpox, have less chance of developing zoster, through presumed immunologic boosting by exposure to varicella zoster virus.² Modelling shows that immunisation causes an increase in herpes zoster for up to 50 years until immunised infants reach old age.

Due to the infectivity of reactivated herpes zoster it is not possible to eliminate varicella zoster virus in the way measles or polio could be eliminated completely. The aim of immunisation is therefore to reduce the burden of varicella disease rather than disease elimination. Since the burden of serious disease, particularly mortality, is in adults, and the modelling shows that in the long term the incidence in adults will not be affected by even high levels of vaccination coverage, the logic of universal vaccination has to be questioned.

Vaccination undoubtedly reduces childhood disease and saves the costs of medical care, childcare costs and lost income for parents while they look after sick children. Health decisions, however, should be primarily based on health considerations rather than economics.

The current low burden of disease from varicella means that it would take only a small rise in varicella in adults for us to be worse off than we were without the vaccine.

Ben Ewald
General practitioner and Lecturer in epidemiology
Centre for Clinical Epidemiology and Biostatistics
University of Newcastle
Newcastle, NSW

Associate Professor Jonathan Carapetis, one of the authors of the article, comments:

Dr Ewald raises the question of how mathematical modelling should be used in determining public health policy. Should we refrain from using a vaccine that can bring immediate reductions in morbidity and mortality because of predictions that there might be ill effects in the future? Some reassurance comes from US data, which have failed to show any change in rates of zoster up to seven years after introduction of varicella zoster virus immunisation.³ It is still early days, and this study may have taken place during the initial 'honeymoon' period. Even if the models prove correct and we begin to see early increases in adult zoster followed by later increases of varicella in adolescents and young adults, there is an obvious solution: booster doses. We already give booster doses of pertussis, diphtheria and tetanus vaccine in adolescence, and regular influenza and pneumococcal immunisation is recommended for high-risk adults.

A recent study of more than 38 000 elderly people in the USA found that a live attenuated varicella zoster virus vaccine reduced the incidence of zoster by 51%.⁴ This provides reassurance that vaccination of adolescents or adults will be an effective countermeasure to the model predictions, if they eventuate.

The models of post-varicella zoster virus vaccine disease patterns are important in highlighting the need for better surveillance of varicella and zoster, the longer-term questions relating to duration of immunity, and the importance of a flexible immunisation policy that can react quickly to changes in the epidemiology of vaccine preventable diseases. The uncertainty surrounding predictions from models means that they should not be used as a reason to withhold an intervention that can prevent illness and death and save health dollars at the same time.