We have developed a strategy for generating mortality burden estimates comparable across influenza epidemic and pandemic seasons using US mortality surveillance data. Here we presented an estimated range of Years of Life Lost (YLL) for the 2009 pandemic up to December 31, 2009. Because of the unusually low mean age of the H1N1 pandemic deaths (37 years), using the YLL metric changes the perception of relatively low impact, both when compared to seasonal influenza (where more than 90% of deaths are in persons aged over 65 years) and to past pandemics. The resulting range of YLL estimates for the 2009 H1N1 pandemic included in its lower end the burden of an average seasonal epidemic caused by A/H3N2, the most virulent seasonal virus subtype ciruclating in the last 3 decades. The upper range of YLL estimates for 2009 exceeded the burden of the 1968 pandemic adjusted to the year 2000 population, a pandemic with a relatively high mean age of death at 62 years. We believe our YLL approach more accurately reflects the qualitative differences between pandemic and inter-pandemic influenza mortality, using methods frequently used in health outcome evaluations.
We derived the lower end of the YLL range using the conservative assumption of counting only deaths coded as pneumonia and influenza. This was most similar to the approach used to quantify the burden of the 1918 pandemic, the only other pandemic we are aware of that had a low median mortality age (27 years). Interestingly in 1918, 95-100% of the influenza burden was captured by deaths coded as pneumonia, influenza, and bronchitis, based on data from NYC and Copenhagen
. By contrast, the higher end of our YLL range estimate for the 2009 pandemic derives from methods used to estimate burden of the 1957 and 1968 pandemics and seasonal epidemics. This method accounts for influenza-related deaths in part attributable to underlying high risk conditions, but occurring in excess of a normal baseline mortality level during the influenza activity period. Given the unseasonal nature of this pandemic, it is difficult to determine what best describes the 2009 pandemic mortality burden at this time – but final estimates will resolve this once national vital statistics data are available.
We believe that the use of qualitative designations such as mild, moderate and severe in describing the health impact of seasonal and pandemic influenza is insufficient, and possibly inappropriate. This terminology describes a pandemic from a single outcome measure; namely an estimate of all direct and contributing deaths. However, the substantial difference in the age distribution of pandemic-related deaths, the number of patients requiring intensive care, and the loss to society in work productivity when a younger age population is disproportionately impacted, supports the use of alternative measures to describe the burden of influenza pandemics and compare with typical influenza seasons. While we have not identified a single outcome measure that can simultaneously account for all the above variables, we believe that using the YLL metrics is the best single measure at this time. From that perspective, the 2009 pandemic had a significant impact on the world’s health from April through December 2009..
Based on US mortality surveillance data, we conclude that the YLL burden of the 2009 pandemic may in fact be as high as for the 1968 pandemic – but that at this time the assessment is still tentative. More H1N1 pandemic waves are likely to occur over the next seasons, increasing the cumulative burden of the emerging virus, and severity may also change over time. For example, the 1918 1st
wave was relatively mild compared to the devastating autumn 1918 wave that killed 2% of the global population 
. In Europe and Asia, the majority of deaths associated with the 1968 pandemic occurred a full year after the pandemic virus first began circulating 
. Only several years after the emergence of the H1N1-pandemic, when national vital statistics become available, will we know the full extent of the severity of this pandemic in the US. Finally, the mortality burden of the 2009 pandemic remains poorly studied in other countries and potential geographical differences in excess mortality rates remain unclear. An early assessment using Mexican data suggested the severity may have been similar to the 1957 pandemic, a pandemic believed to have caused 2 million deaths worldwide. However, data from Europe and the US, including the present study, suggest a milder impact there similar to what occurred in the first year of the 1968 pandemic. Geographical variability in mortality impact was also seen during the 1918 pandemic, where for example Scandinavian cities and NYC experienced a mild 1st
wave, followed by the catastrophic second wave in the autumn 
while cities like Geneva 
and Madrid 
experienced their severe impact in the first wave.
In conclusion, based on US data for May-December 2009, we estimate that the H1N1 pandemic was associated with an impact ranging from that of a virulent influenza A/H3N2 season to that of the 1968 pandemic, when one takes into account the markedly young age distribution of influenza-related deaths in occurring in 2009. The 2009 H1N1 virus may replace one or more A type influenza viruses and cause additional waves either late this season or the following winter, potentially with additional mortality impact to come. Given the historic record of changing disease severity geographically and over time, it is critical to keep up vigilance with this emerging pandemic virus, and continue vaccination efforts and other preparations to protect the population. The recent recommendation to include the pandemic H1N1 virus antigen in the seasonal vaccine formulation for the next winter is a good step towards this goal.