Abstract
There were multiple waves of influenza-like illness in 1918, the last of which resulted in a highly lethal pandemic killing 50 million people. It is difficult to study the initial waves of influenza-like illness in early 1918 because few deaths resulted and few morbidity records exist. Using extant military mortality records, we constructed mortality maps based on location of burial in France and Belgium in the British Army, and on home town in Vermont and New York in the USA Army. Differences between early and more lethal later waves in late 1918 were consistent with historical descriptions in France. The maps of Vermont and New York support the hypothesis that previous exposure may have conferred a degree of protection against subsequent infections; soldiers from rural areas, which were likely to have experienced less mixing than soldiers from urban areas, were at higher risk of mortality. Differences between combat and disease mortality in 1918 were consistent with limited influenza virus circulation during the early 1918 wave. We suggest that it is likely that more than one influenza virus was circulating in 1918, which might help explain the higher mortality rates in those unlikely to have been infected in early 1918.