Historical Study Epidemiological isolation causing variable mortality in Pacific island populations during the 1918-1920 influenza pandemicArchived

Shanks GD, Hussel T Brundage TF. Epidemiological isolation causing variable mortality in Island populations during the 1918–1920 influenza pandemic Influenza and other respiratory viruses 6: 417-23. DOI: 10.1111/j.1750-2659.2011.00332

During the 1918-1920 influenza pandemic all populations experienced an increase in influenza related deaths. However that mortality varied considerably between different locations and demographic subgroups.(1,2) Nowhere were these differences greater than in isolated island populations. This publication from an Australian-American research group experienced in studies of the history of influenza explored and analysed the relevant mortality data from the Pacific after an extensive literature search. Their hypothesis was that the relationship between prior exposure to a variety of respiratory agents could explain the differences and mortality during the influenza pandemic on islands and amongst the varied sub-populations.(2) The researchers hypothesised that exposure to a variety of respiratory infectious agents before the 1918-1920 pandemic caused cross protective immunity from influenza virus strains and that on the whole the relative isolation of island communities would make them more vulnerable than communities that mixed more internationally. At the same time they attempt to exclude the possibility of the differing mortality among various island populations being due to differences in medical care and/or viral mutations..

The data sources used by the authors for this study included both civil and military influenza mortality records and the acquired results were presented in population based percentages, allowing direct comparison between both the islands and their demographic subgroups. Groups of respiratory causes of death were combined such as influenza and pneumonia. Since they were dealing with events of nearly a century ago what they were not able to do was to obtain evidence of prior infection and what those infections were. The researchers considered both the experience in the islands and what happened when islanders went abroad, usually to serve towards the end of the 1914-1918 war in Europe The results demonstrate that during the 1918-1920 pandemic, the mortality variation on the Pacific Islands, was greater than 50 fold, ranging widely between 0.4% in Australia through 0.7% in New Zealand, 5% in Fiji, 16% in French Polynesia to 22% in Samoa. A further observation included the differences between demographic subgroups. For example on Fiji the mortality rate was 5.7% among native Fijians but 1.4% among Europeans. In the Commonwealth of the Northern Mariana Islands, Saipan, the mortality among the Chamorros ethnic group was 12% while among the more recently arrived Caroline Islanders it was only 0.4%.

Comparison of those going out of the islands for military related service found that certain groups had worse experience. Among the Maori ethnic group from New Zealand mortality was four times as high as for other New Zealanders but very similar to what was experienced by the Maori groups in New Zealand itself. (2) Conversely though the mortality experienced by French Polynesians soldiers in France (0.13%) was much less than what was experienced among civilians remaining in French Polynesia (16.3%). The authors argue for the hypothesis that the island residents who experienced exposure to varied, antigenically diverse respiratory pathogens before infection with the pandemic influenza strain from 1918-1920 experience lower mortality. This argument has been made before but in addition they go on to show how that prior exposure might work with not only immunological mechanisms but also airway remodelling, modification of epithelial cells and alteration or distribution of fluid channels. They also go on to suggest that having had prior exposure to a single influenza type may have perverse effects if there is later exposure to a similar type resulting in a cytokine storm.(2)

ECDC Comment (25 November 2012):

The case made by this group for the importance of prior exposure having important protective effects is a credible and important one, though it is by no means original. A recurrent theme in pandemic studies is of older people having some protection against a novel strain even if they are the most susceptible age-group for severe disease if they do become infected.(3) The authors down-play the importance of differences in mortality being due to genetic susceptibility or access to care. However the 2009 pandemic still found that Maori and Pacific Islanders in New Zealand were at higher risk of death from influenza than the white population.(4) Also in the only studies that looked at ethnic group experience in Europe (in the UK) a recurring theme was of higher risk of severe disease and death in South Asian ethnic groups compared to white population.(5,6)

There are other difficulties with these studies. Firstly the numbers are small in some groups, for example estimates of mortality in Maori soldiers in Europe is based on a sample of only 800 persons. Also the cytokine storm theory is not very scientific in this context as there is no way to prove or disprove it. Finally there are anomalous findings like that the much lower mortality in French Polynesian soldiers in France compare to those living in Polynesia. Another anomaly is the thirty fold difference in mortality in Saipan among the Chamorros ethnic group and the more recently arrived Caroline Islanders. The authors have a plausible explanation, for the latter that the Caroline Islanders had been exposed to an earlier pandemic before they came to Saipan but there is a post hoc nature to this. Reading historical articles like this is important not just for scientific purposes but also to remind us just how severe influenza can be. A pandemic that kill one in five in a population as in Samoa (then population 38,000) are a very serious affair. However a positive closing note is that globalisation by increasing the likelihood of early exposure to respiratory pathogens makes this less likely to be repeated in the future.