Analysis and implications of:
by Fraser et al. and The WHO Rapid Pandemic Assessment Collaboration. Science. May 11, 2009.
This paper is the first detailed analysis of the epidemiology of the new H1N1 virus (swine flu). It contains a mixture of empirical data and modeling. The goal of the paper is to make estimates of the case fatality rate (CFR) and clinical attack rate to guide both policy and further studies.
There are a number of significant findings/estimates, but here are the two likely to generate the most interest:
1. The CFR of the new H1N1 is likely between .3% and 1.5%.
Seasonal flu has a CFR of less than .1%. So, the authors are estimating that this flu is 3 to 15 times more lethal than seasonal flu.
2. The clinical attack rate of the new H1N1 is 61% among children 15 and under.
This means that most children will be infected in an unmitigated pandemic.
We know from other reports that deaths are focused in younger age groups and that children are not spared. So, let us consider what the findings above mean for a school district with 10,000 students. We can expect that 6,100 of the children will be infected if the schools are left open. Assuming a .4% CFR, 24 children will die of their infections.
So, when a school district decides to remain open despite clear evidence that the virus is spreading within schools, they have made the deliberate decision to sacrifice the lives of some of their students.
SouthernBlueNeck at PFI_Forum further points out that there are 74,431,511 children in the United States under the age of 18. Thus, allowing the pandemic to continue to spread without restraint in the US, which is current US government policy, may result in the deaths of 136,000 - 681,000 children.
Applicability of the current study to the US and other countries
Although there have been only a few deaths have been reported in the US thus far, there is insufficient data to draw any conclusions about the CFR within the US. It is not clear if the CDC is making any attempt to find out how many deaths have actually occurred within the US. They are not releasing what information they do have in a timely fashion. Finally, flu deaths usually lag infections as people typically die after they have been on a ventilator for some period of time (often weeks).
A highly transmissible virus that targets children and young people for both infection and death, preferentially, at the rates described in this paper is chilling enough. However, there are reasons to believe that higher rates of death may be observed in the future. The estimates for the likely CFR are based on incomplete information, which the authors acknowledge. There are two important variables are not included in their analysis: weather and Tamiflu administration. Weather in Mexico City when many of the infections occurred was in the 80 degree (F) range. Influenza is generally thought to spread more efficiently in cooler weather. Tamiflu was administered to many, but not all of the Mexican cases. Anecdotal evidence suggests that this drug is effective in reducing the severity of symptoms with the new H1N1 if given soon after infection (AP). There have been reports that some of the people who died of the new H1N1 in Mexico had not gone to the hospital until they were quite ill. Thus, the CFR may be quite different for people who received Tamiflu as opposed to those who did not. The data in the current paper included an uncertain mixture of both types of patients.
In countries with more conducive environmental conditions, transmissibility of the new H1N1 virus may be greater than observed in this study. Although R0 is often portrayed as a constant, intrinsic property of the viral genome, it is in fact an average of the number of people infected by individuals with a certain virus under certain conditions. Both cold, dry weather and insufficient Tamiflu stockpiles may reasonably be expected to result in a higher apparent R0 than was reported in this study.
Most countries have few or no Tamiflu stockpiles. Although some Tamiflu may be provided by international donors, available supplies are only a tiny fraction of what will be needed. Further, even in rich countries like the US, there is insufficient Tamiflu to provide treatment for all who will likely be infected. Thus, almost all people in poor countries and most people in rich countries will face this virus without the only known effective remedy. It is therefore not unreasonable to expect that the CFR may be higher than that proposed in this study.
Future studies should attempt to take into account the effects of weather and Tamiflu administration on transmissibility and lethality. Given the wide disparity in temperatures in April in North America, it should be possible to test the hypothesis that the new H1N1 spreads more rapidly in cold, dry areas. Retrospective analysis of Mexican cases should focus on whether, and when Tamiflu, was administered. For all new cases of infection, it is critical that Tamiflu administration be recorded and be considered as a critical variable in future analysis.
The paper by Fraser et al. does not provide a final answer on either transmissibility or lethality of the new H1N1 virus. However, it is a welcome first step in data collection and analysis of the epidemiology of the first new pandemic virus of the 21st century.
Monday, May 11, 2009
