If a worldwide flu pandemic like the one in 1918-20 were to strike today it could kill between 51 and 81 million people, with 96 per cent of deaths being in poorer countries. This was the conclusion of a US study led by Professor Chris Murray of the Harvard Initiative for Global Health.

The study is published in The Lancet.

The researchers used high-quality vital registration data of deaths occuring at the time of the 1918-20 pandemic. From these figures they estimated what the likely globaly mortality would be from a flu pandemic if it happened today.

In a podcast interview, Professor Murray says that the impetus for the research came from listening to colleagues talk at meetings about what the death toll from a flu pandemic might be. Estimates ranged from 50 to 100 million to as many as 1 billion deaths worldwide, but nobody had yet made a systematic attempt to quantify it.

They decided to have a go, and started by examining the data for the three major flu pandemics of the 20th century: the 1918-20, the ’57-58 and the ’68-70. The figures for the last two were not large enough to give a reliable starting point for their analysis, so they focused on the 1918-20 pandemic.

However, one of the first problems they encountered is establishing a reliable estimate of the number and distribution of deaths that occured as a result of the 1918-20 pandemic.

The only casualty count they could rely on was what they call in the paper “vital registration data”. In the more developed, higher income countries, tight legislation governs the reporting of births and deaths. This is not the case in the less developed world. 250 years ago Sweden was the first country to bring in such legislation, and gradually others followed suit.

By 1918-20 most of what we now call the developed world, Western Europe, North America, Australia, New Zealand, had reliable vital statistics registration. This also applied to the colonies such as the Philippines, Taiwan, Sri Lanka, and especially India. This gave the researchers confidence to go ahead and use 1918-20 data based on registered deaths.

The deaths were recorded by date, age and sex. However, Prof Murray and his team found that the reasons given for death were not a reliable method for working out the number of deaths due to flu. So instead they looked at the overall pattern of mortality in each region for the time before, during and after the pandemic and calculated in a consistent way, the “excess mortality”. On this basis they estimated the number of deaths resulting from the pandemic for each region affected. And they came up with a very wide range which they went on to try and explain.

The death toll of the 1918-20 pandemic ranged from a low 0.2 per cent of the population in Denmark to an astonishing 4.4 per cent of the population in India. Even within India there was huge variation by province, for instance for the Central Provinces and Berar the figure was over 7 per cent.

If we put that into context in today’s terms, about 1 per cent on average of the world’s population dies every year. So for India, during the pandemic of 1918-20, the annual mortality went up fourfold.

To project from the 1918-20 data to produce an estimate for today, the scientists looked for explanations of these large variations. One very simple one is the socioeconomic differences between the countries. They found a strong correlation between the deaths due to pandemic and the income per head of the countries affected. They went ahead and did the extrapolation, taking into account socioeconomic status (based on income per head), and the age structure of the countries, since the highest mortality appeared to be in young adults.

Prof Murray accepts that basing the projections on socioeconomic status and age profile could be debatable, but it does give a figure to work with, and it is based on a systematic, quantitative analysis using reliable vital registration statistics.

Their final estimate ranges from 51 to 81 million deaths worldwide (with a median, or mid-range value of 62 million). And they estimate that 96 per cent of these would be the poorer countries. The reason they estimated this last figure as so high is because of two things. Since 1918-20 the more developed countries have become much richer, thus increasing their socioeconomic status in the “model”. And the second reason is the younger age profile of the less developed world.

Professor Murray pointed out that this 62 million should be viewed as an upper limit since they did not take into account factors like improved symptom management, treatment of secondary pneumonia (which develops after the flu and was a major killer in the 1918-20 pandemic), vaccination, and so on.

He also says that while the world has paid a lot of attention to the risk of a flu pandemic, and there has been a “remarkable concerted response among high income countries”, not much attention has been paid to what is practical for the low and low to middle income countries, which according to this study would carry the major burden of deaths.

Prof Murray says the public health message is “Don’t let the perfect be the enemy of the good. We need to think of strategies that might work in these poorest settings. It doesn’t seem likely that vaccination strategies are ever going to reach the poorest communities in a time frame that would be beneficial.” He goes on to say that “we have to look at this problem from where most of the burden and the harm is going to be and think of strategies that will benefit those populations”.

“Estimation of potential global pandemic influenza mortality on the basis of vital registry data from the 1918�”20 pandemic: a quantitative analysis.”
Prof Christopher JL Murray DPhil email address a Corresponding Author Information, Prof Alan D Lopez PhD b, Brian Chin ScB a, Dennis Feehan AB a and Prof Kenneth H Hill PhD.
The Lancet 2006; 368:2211-2218

Click here for full text of article (free registration).

Click here for podcast (mp3) interview with Professor Murray (The Lancet).

Written by: Catharine Paddock
Writer: Medical News Today