The other pox

February 2003

---Philip A. Brunell, MD

Many of you probably have had calls and much discussion about the article in The New England Journal of Medicine (2002;347: 1909) concerning the poor performance of varicella vaccine (Varivax, Merck) in a day care center. I am certain this has been viewed as validating the beliefs of parents who elected not to give their children the vaccine.

The article describes an outbreak in which the efficacy of varicella vaccine was found to be only 44%. This is poorer than reported in earlier studies (New Engl J Med. 2001;344:955, Pediatr Infect Dis J. 1999;18:1047, Clin Infect Dis. 2002;35: 104, J Infect Dis. 2002;106: 102). It has been suggested that the efficacy figures from the “research” trials were inflated, as those who were vaccinated were monitored serologically — and any who did not become seropositive were reimmunized. Thus, the “no takes” were eliminated. The follow-up of clinical efficacy eliminates these vaccine failures in their analysis.

However, there have been studies of vaccine efficacy as it is used routinely in the community and efficacy rates this low have not been observed (N Engl J Med. 2001;344:955, Pediatr Infect Dis J. 1999;18:1047, Clin Infect Dis. 2002;35:104, J Infect Dis. 2002;106:102).

One of the most disturbing findings was that the risk of vaccine breakthroughs has increased with the passage of time since vaccination. One previous study has reported a similar phenomenon, although the data were not robust (Pediatrics. 1993;91:17).

The duration of vaccine-induced immunity and concern about whether the risk of zoster in vaccinated individuals might increase were the two major considerations about vaccine licensure. Many cited the experience with existing vaccines, eg, measles and polio, which appeared to produce durable protection. Others cautioned that this was a herpesvirus and that we had no prior experience with a human herpesvirus vaccine. It was argued that varicella-zoster virus (VZV) produces a latent infection, which appears to be activated when immunity wanes. Indeed, there now is a trial to try to prevent zoster by boosting immunity.

If the observation that cases increase with the passage of time is real, this does not portend well. Conversion of chickenpox from a childhood disease to one of adults is the last thing we want. The number of deaths from varicella are as great in adults older than 20 as in childhood yet only about 2% of reported cases occur in adults. Giving booster doses to adults would not be a viable strategy based on our experience with the influenza vaccine, because adults are difficult population to reach. Influenza kills tens of thousands of people (mainly adults) each year, yet, many still choose not to be vaccinated.

Although it has been argued that immunization of a large proportion of children would eliminate the risk of exposure to varicella in adult life, zoster will continue to be a source of contagion in the foreseeable future. In fact, the index case in the day care outbreak contracted varicella from a sibling with zoster.

The authors carefully examined factors that are suspected of being associated with poor vaccine efficacy — eg, administration within a month following MMR (MMWR. 2002;47:1058), asthma (JAMA. 1997;278:1495), faulty storage, immunization under 14 months of age (Clin Infect Dis. 2002;35:104, J Infect Dis. 2002;106:102), bad lots of vaccine, problems in a particular office — but none seem to explain their findings.

Decline in cases

Varicella vaccine has been very successful as judged by the decline in cases of varicella and the decrease in hospitalization due to varicella. These effects have been directly related to the number of children who are immunized (JAMA. 2002;287:606). As has been found with many other vaccines, there also has been a decline in cases in those not immunized with varicella vaccine, in this case, those under age 1 (JAMA. 2002;287:606). Ordinarily, we would welcome the development of herd immunity in the population but this may not be an effect that is desirable for VZV. If unimmunized children grow to adulthood without getting varicella because of decreased opportunity for exposure to natural infection, pools of susceptible adults will increase.

There may be yet another effect of the lack of boosting due to decreased opportunity for exposure, which is an accelerated loss of immunity. Much has been made about the role of boosting in the maintenance of immunity to VZV for the prevention of zoster (Lancet. 2002;360:678, Proc Roy Soc Med. 1964;58:200). Perhaps, it may also be required for the maintenance of immunity in vaccinated individuals. That boosting of antibodies that occurs in vaccinated individuals following exposure has been known for some time (Pediatrics. 1988;81:779).

In fact household contacts of vaccinated individuals, whether or not they develop varicella, have almost a log increase in the relatively low VZV antibody titers found post immunization. That this is not simply an anamnestic response can be gleaned from the experience in immunized children with leukemia. A second dose of vaccine given to those who had lost antibody did not produce an appreciable response. In fact, some of these children again became seronegative.

In contrast, those who had household exposure had very high and sustained levels of VZV antibody. This raised the possibility that durable protection against varicella requires exposure to natural infection. Whatever figure one uses, exposure to varicella in the prevaccine era was common (Nat Med. 2000;6:451).

Moreover, it is clear that natural infection produces better protection than vaccine. Those with a history of having had varicella in the day care study were protected against infection (N Engl J Med. 2002;347:1909).

The answer to the need for boosting by natural exposure will come with time. If this experience is an aberration, and we all hope that it is, nothing more need be done. If however, there is increasing evidence of loss of protection with increasing duration since vaccination, a new strategy will need to be considered.

The most obvious approach would be to boost immunity by an additional dose or doses of varicella vaccine. This would be facilitated by the licensure of MMR-V (measles-mumps-rubella-varicella), which would be given two times. MMR-V as a vaccine, has spent a long time in development.

Our first studies on MMR-V were published almost 15 years ago (Pediatrics. 1988;81:779). If a two-dose strategy were adopted, one would then need to consider whether the best time for a booster is at school entry or during early adolescence. You might recall that there was a similar issue with the second dose of measles vaccine. It was thought that a dose during adolescence might boost waning immunity. Logistically, it was easier to give it at school entry. This, for measles vaccine, was designed to protect those who did not respond to initial vaccination in order to be protected. In the case of varicella vaccine, it also would be used to boost immunity. However, there is no good evidence that a second dose of varicella vaccine would increase the duration of protection, although those who did not respond to the first dose probably would no longer be susceptible. Another approach would be the development of a better vaccine and this too is in the works.

As indicated the vaccine has had a very impressive effect on decreasing varicella morbidity. In the original equation, parental loss of time from work accounted for about 90% of the savings that the vaccine would produce (JAMA. 1994;271:375). Children who had breakthrough varicella could return to school sooner than those who had unprotected infection. However, the change in recommendation for when children could return to school, may have contributed in some part to the savings accrued. Previously, it had been seven or five days following onset and not when “all lesions are crusted” (Red Book 2002). The cost-benefit analysis of the program would have to be recalculated if additional doses of vaccine are recommended at the current price, not $35 per dose, which was used in the original estimate.

What then can you contribute to the discussion with parents and colleagues? I believe that the risk of an unimmunized child growing to adulthood susceptible to infection is real. The risk of loss of protection over time will have to be determined. If you are more uncomfortable than you were previous to this report of poor efficacy, welcome to the club.