Study raises questions about age-group impact of H1N1

Jun 14, 2010 (CIDRAP News) – A study by French researchers suggests that the H1N1 influenza pandemic may not have differed from seasonal influenza epidemics in its effects on different age-groups quite so much as has been supposed.

The general understanding is that H1N1 has disproportionately affected children and young adults, leaving elderly people relatively unscathed. The new study agrees that the younger groups have borne a much larger share of pandemic deaths than they do in seasonal flu, but it suggests that the age distribution of total cases is much the same as in past seasonal flu epidemics.

"The age distribution of influenza-like illness was similar between the 2009 H1N1 pandemic and seasonal epidemics whereas the proportion of under-60s among influenza deaths was markedly higher during the 2009 pandemic (peak<20 years) than during the seasonal epidemics," says the report by Magall Lamaitre and Fabrice Carrat. It was published online last week by BMC Infectious Diseases.

The US Centers for Disease Control and Prevention (CDC) has estimated that about 90% of all pandemic H1N1 cases were in people younger than 65 and that 32% were in children (under 18). The BMC study implies that this disproportionate burden of illness in young people is not much different from what happens in seasonal flu.

The French researchers compared the age distribution of cases and deaths during two different seasonal flu epidemics in the United States and France with the age distribution of pandemic H1N1 cases and deaths in the two countries.

For the seasonal epidemics, the researchers chose the H1N1 epidemics of 1978-79 in the United States and 1998-89 in France and the H3N2 epidemics of 1989-90 in both countries.

They used influenza-like illness (ILI) as the indicator of cases (morbidity), gathering data from published studies, the CDC, and France's Sentinel flu surveillance system. The age distributions of flu-related deaths were obtained from national death registries.

To assess the relative burden of illness and deaths on different age-groups, the investigators calculated a relative illness ratio (RIR) and a relative mortality ratio (RMR). The RIR is the percentage of sick people in a given age-group divided by the percentage of the total population belonging to that age-group. Similarly, the RMR is the percentage of total flu deaths in a given age-group divided by the percentage of total all-cause deaths in that age-group. Any ratio greater than 1 signals a disproportionate burden.

Overall, the age profile of the 2009 pandemic resembled that of the seasonal epidemics, the researchers found. The highest illness burdens (RIR values) in the pandemic were in 5- to 9-year-olds in France (RIR, 2.97) and in 0- to 4-year-olds in the United States (RIR, 3.49).

The 5- to 9-year-old group had the highest illness burdens in both seasonal epidemics in France (H1N1, 2.15; H3N2, 1.77). In the US seasonal epidemics, the highest illness ratio in the 1978-79 H1N1 outbreak fell on 10- to 14-year-olds (2.99) and, in the 1989-90 H3N2 epidemic, on 5- to 24-year-olds (1.84). All the RIRs for those 35 and older were lower than 1.

In contrast to the morbidity pattern, the age pattern for deaths differed sharply between the pandemic and the seasonal epidemics, the researchers found. In the pandemic, RMRs for those under 60 were all higher than 1, with peak ratios in those under age 20. In the United States the hardest hit age-group was 5- to 24-year-olds (RMR, 38.66), while 5- to 9-year-olds were hardest hit in France (RMR, 37.39).

In three of the four seasonal epidemics, the highest RMRs were in those 65 and older, the report says. The exception was the 1978-79 H1N1 epidemic in the United States, in which the 5- to-24-year-old age-group had the highest relative mortality (RMR, 3.91). As the authors note, the H1N1 virus re-emerged in 1977 after a 20-year absence, leaving young people at the time relatively unprotected.

The researchers say their morbidity findings suggest that "the age distribution of risk of infection did not differ" between the pandemic and the seasonal epidemics. They add that previous studies have shown that school-age children are most susceptible to contracting seasonal flu and that studies of the 1918, 1957, and 1968 pandemics showed an age distribution of cases similar to that of seasonal flu epidemics.

As for mortality findings, the authors write that while mortality was highest in children and younger adults, it decreased with age between 20 and 60. "Prior exposure to seasonal influenza viruses thus seems to protect against the 2009 H1N1 virus," they add. They further comment, as have others, that the relatively low mortality in elderly people suggests they have immunity related to their exposure to H1N1 viruses that circulated before 1957.

Cecile Viboud, PhD, a staff scientist in the division of international epidemiology and population studies at the National Institutes of Health's Fogarty International Center, observed that the finding of a shift in the mortality burden to younger people in the H1N1 pandemic has been detected in several other studies and also has been described in the previous three pandemics.

However, "the authors do not find evidence of an age shift in the distribution of influenza cases" in the 2009 pandemic, "which is perhaps a little surprising," Viboud told CIDRAP News by e-mail.

She said the study may lack the statistical power to detect changes in the age distribution of cases in pandemics. This is "mostly because they only have two seasonal epidemics to compare with, and also because the morbidity data come from several different studies with likely large variations in sampling and testing practices."

Viboud added that at least two previous studies "have noted a change in the age distribution of cases in [past] pandemics, with proportionally more children being sick than expected. Also, we know that seasonal A/H1N1 epidemics tend to cause illness in younger individuals than H3N2 epidemics, which is not obvious from their [Lamaitre and Carrat's] data, and suggests that the study may be underpowered."

She said another possible problem with the study is its use of data on deaths specifically coded as influenza, which underestimate the true flu burden and may introduce age-related biases in coding. She added that she looks forward to further studies on the age distribution of cases and deaths as more data on the 2009 pandemic become available.

Source: http://www.cidrap.umn.edu/cidrap/content/influenza/swineflu/news/jun1410age.html