Jun 27, 2012 (CIDRAP News) – Working with admittedly sparse data, a research team led by the US Centers for Disease Control and Prevention (CDC) has estimated the global death toll from the 2009 H1N1 influenza pandemic at more than 284,000, about 15 times the number of laboratory-confirmed cases.
The World Health Organization (WHO) has put the number of deaths from confirmed 2009 H1N1 flu at a minimum of 18,449, but that number is regarded as well below the true total, mainly because many people who die of flu-related causes are not tested for the disease.
The CDC-led team, which included researchers from several other countries, based its estimates on H1N1 case data from 12 countries and case-fatality ratios (CFRs) reported from five countries. Their report was published online yesterday in the Lancet Infectious Diseases.
The researchers estimate that the pandemic virus caused 201,200 respiratory deaths and another 83,300 deaths from cardiovascular disease associated with H1N1 infections. They also calculate that Africa and Southeast Asia, which have 38% of the world's population, accounted for a disproportionate 51% of the deaths.
The team estimated that 80% of those who died were younger than 65, which is in accord with previous observations that the pandemic H1N1 burden fell heavily on younger people, unlike the pattern for seasonal flu.
The team, with the CDC's Fatimah S. Dawood, MD, as first author, set out to estimate H1N1 deaths in the period April 2009 through August 2010, focusing on the first 12 months of H1N1 circulation in each country. Their general approach was to estimate the cumulative symptomatic-case attack rate (sAR) with data from high-income, middle-income, and low-income countries and multiply it by the estimated symptomatic case-fatality ratio (sCFR) derived from data from certain high-income countries.
The authors stratified their estimates of sAR and sCFR into three age-groups: 0 to 17 years, 18 to 64, and over 64. They obtained estimates of the attack rates in those groups from 12 countries: Bangladesh, Denmark, Germany, India, Kenya, the Netherlands, New Zealand, Nicaragua, Peru, the United Kingdom, the United States, and Vietnam.
The sAR estimates ranged from 4% to 33% for the pediatric group, 0 to 22% for 18- to 64-year-olds, and 0 to 4% for those over 65.
Next the team gleaned age-based sCFR estimates from six studies conducted in Denmark, the Netherlands, New Zealand, the United Kingdom, and the United States. These estimates ranged from 0.002% to 0.013% for children, 0.018% to 0.159% for those aged 18 to 64, and 0.090 to 0.308% for seniors.
The authors devised a way to adjust for differences in the risk of respiratory-disease death between high- and low-income countries. They used the WHO's country-specific mortality rates for lower respiratory tract infections and the agency's classification of countries into five mortality strata to come up with a "respiratory risk multiplier."
As part of steps to minimize distortions related to the use of data of uneven quality, they assigned countries to one of three mortality risk groups: African countries, non-African countries with high child and adult mortality, and all other countries. The estimation process yielded thousands of mortality estimates for each age-group and risk group.
The team also devised a way to estimate deaths due to cardiovascular complications of H1N1 infections. This involved using estimates of excess circulatory and respiratory deaths attributed to the virus in Argentina, Brazil, Chile, Mexico, and the United States and the base respiratory mortality rate in each country.
The total of median estimates of country-specific respiratory deaths was 201,200, with a range of 105,700 to 395,600, calculated on the basis of the 25th- and 75th-percentile estimates for each age-group and country. Another 83,300 deaths (range, 46,000 to 179,900) were attributed to cardiovascular complications, for a total of 284,400 (range, 151,700 to 575,400). (Because of rounding of regional estimates, the totals don't add up precisely.)
Twenty-nine percent of the respiratory deaths were ascribed to African countries, where the estimated mortality rate was about two to four times as high as in countries elsewhere. The addition of cardiovascular deaths reduced regional disparities in deaths, but African mortality still was two to three times higher than elsewhere, the report says.
Overall, only 20% of the deaths were in people older than 64. In contrast, about 90% of seasonal flu deaths are in seniors.
The authors also estimated the years of life lost to the pandemic by multiplying the estimated number of deaths in each age-group by the number of years of life expected for each age-group by country. They came up with a figure of 9,707,000 years.
They project that this number is 3.4 times higher than it would've been if the age distribution of deaths had been similar to what happens in seasonal flu epidemics in developed countries.
The authors say their estimate of 151,700 to 575,400 deaths represents 0.001% to 0.011% of the world population. For comparison, they observe, the WHO estimates that 250,000 to 500,000 people (0.004% to 0.008% of the population) die of seasonal flu annually, and flu deaths in past pandemics ranged from 0.03% of the population in 1968 to 1% to 3% in 1918.
However, the team says its numbers are not directly comparable to the WHO seasonal flu estimate and the past pandemic numbers for various methodologic reasons. For example, the estimates for the earlier pandemics include data from several years of virus circulation, not just the first year.
The estimation approach used by the authors received a general endorsement from Marc Lipsitch, PhD, an epidemiologist and director of the Center for Communicable Disease Dynamics at the Harvard School of Public Health.
Lipsitch told CIDRAP News via e-mail that he was not completely independent from the study because he served on a WHO working group that gave the authors some advice, and he cautioned that he had not yet read the full report. But he said, "I know the broad outlines of their method and think it is a plausible approach to a very difficult problem of making estimates with extremely limited data, where even the available data are really very uncertain.
"I don't know if the number is right, but certainly any reasonable estimate would be far above the WHO number of confirmed deaths," he added. "Confirming flu deaths is not routine (and much less than 100% sensitive) in the richest countries, and would be exceptional rather than likely in much of the world. Thus the disparity from the WHO confirmed number is a good sign, not a red flag."
In a commentary accompanying the study, Cecile Viboud, PhD, and Lone Simonsen, PhD, write, "Future research should focus on obtaining additional country-specific estimates of the burden of seasonal and pandemic influenza in understudied locations through traditional excess mortality approaches or innovative surveys of mortality, and address the lack of estimates of case fatality rates from middle-income and low-income regions."
Viboud works at the National Institutes of Health's Fogarty International Center, and Simonsen is with George Washington University's School of Public Health. The two were part of a team that in March 2010 estimated years of life lost because of the 2009 H1N1 pandemic in the United States.
The authors of the study say their findings point up the need to expand delivery of flu vaccines to Africa and Southeast Asia, since those regions may have borne a disproportionate share of the pandemic mortality burden.
Dawood FS, Iuliano AD, Reed C, et al. Estimated global mortality associated with the first 12 months of 2009 pandemic influenza A H1N1 virus circulation: a modelling study. Lancet Infect Dis 2012 Jun 26 [Abstract]
Viboud C, Simonsen L. Global mortality of 2009 pandemic influenza A H1N1. (Commentary) Lancet Infect Dis 2012 Jun 26 [Extract]
August 6, 2010, WHO update listing global H1N1 death toll of at least 18,449
Mar 23, 2010, CIDRAP News story "Study: In life-years lost, H1N1 pandemic had sizable impact"