The Spanish flu was the 1918-1920 influenza pandemic caused by an H1N1 influenza A virus. It infected about 500 million people worldwide and killed an estimated 17 to 50 million, with its heaviest toll falling, unusually, on healthy adults between 20 and 40 years old [3][4].
Published: 2026-06-05. Last reviewed: 2026-06-05.
What the Spanish Flu Actually Was
The Spanish flu was a 1918-1920 pandemic of H1N1 influenza A that infected roughly 500 million people, about a third of humanity at the time, and killed somewhere between 17 and 50 million of them within roughly two years [3][4]. It remains the most lethal short-duration disease event in recorded history, and for the curious reader it sits beside the small set of outbreaks that the science of its own day could neither name nor stop. The work of Dr. Felix Chen tracks several of these in our archive.
An H1N1 Virus, Not a Spanish One
The name is a historical accident. The 1918 pandemic spread during the final year of World War I, when wartime censors in the United States, Britain, France, and Germany suppressed reports of mass illness to protect military morale [15]. Spain stayed neutral, its press uncensored, and Spanish newspapers reported the epidemic freely, including the illness of King Alfonso XIII. The free flow of bad news out of Madrid created the false impression that the disease started there, and the label “Spanish flu” stuck even though Spain was neither the origin nor the worst-hit nation [15][4]. The technical referent is influenza A subtype H1N1, the same lineage whose descendants still circulate as seasonal flu more than a century later [1][2].
The Scale: 500 Million Infected
On the number: for decades the standard death estimate was 50 million, sometimes quoted as high as 100 million, figures that trace back to a 2002 reassessment by Niall Johnson and Juergen Mueller. A 2018 study by Peter Spreeuwenberg and colleagues at the Netherlands Institute for Health Services Research re-ran the arithmetic using the Human Mortality Database and Indian colonial records and arrived at a best estimate near 17.4 million, with a central figure around 15 million [3]. The two numbers are not contradictions so much as different bookkeeping: the lower figure is better anchored in primary mortality data, the higher one extrapolates across regions where records were thin. The United States alone lost about 675,000 people, more than its combined military deaths in the First and Second World Wars [4].
Where the 1918 Pandemic Began
The geographic origin of the 1918 pandemic remains contested, with the leading candidates being Haskell County, Kansas in January 1918, the Camp Funston army cantonment in March 1918, and the British troop depot at Étaples, France [5][7]. No site has been confirmed by physical evidence; the debate runs on epidemiology, troop-movement records, and molecular phylogenetics rather than a recovered patient zero.
The Kansas Hypothesis: Haskell County and Camp Funston
Historian John M. Barry, in his 2004 book The Great Influenza, traced the most-cited origin story to rural Kansas. Local physician Loring Miner reported an unusually severe influenza in Haskell County in January and February 1918, severe enough that he warned the U.S. Public Health Service [5]. Young men from Haskell traveled to nearby Camp Funston at Fort Riley, a crowded mobilization camp, and on March 11, 1918, an army cook named Albert Gitchell reported sick. Within a week the camp infirmary held more than 500 cases [5][4]. The camp funneled soldiers to Europe, and the timing fits the first wave’s appearance in France that April. The Kansas account is a strong circumstantial chain, not a proven one.
The Étaples and Molecular-Clock Alternatives
Where the consensus and the evidence diverge: virologist John Oxford of Queen Mary University of London argued that the pandemic seeded earlier, around 1916, at the massive British army camp and hospital complex at Étaples in northern France, where a respiratory illness with the right clinical signature appeared among densely packed troops, pigs, and poultry [7]. A separate line of work led by Michael Worobey at the University of Arizona used a host-specific molecular clock to read the virus’s own genome as a calendar; the 2014 analysis concluded that the 1918 H1N1 arose from a human H1 virus already circulating since roughly 1907, which then picked up avian neuraminidase and internal genes shortly before 1918 [6]. Worobey’s later phylogenetic work found the North American origin plausible but not decisive [8]. Three hypotheses, no autopsy of the question: the trigger event is still open.
Why It Killed the Young and Healthy
The 1918 influenza produced a W-shaped mortality curve, killing infants, the elderly, and an unusually large number of healthy adults aged 20 to 40, a peak that ordinary seasonal flu never produces [16][6]. Normal influenza traces a U: the very young and very old die, the middle stays largely spared. The 1918 strain added a third hump in the prime of adult life, and explaining that middle spike has driven a century of immunology.
The Cytokine-Storm and Prior-Exposure Hypotheses
The leading mechanistic idea is the cytokine storm: in a young adult with a vigorous immune system, the 1918 virus could provoke an inflammatory overreaction so violent that immune cells flooded and destroyed the lung tissue they were meant to defend, producing fatal acute respiratory failure within days [16]. A complementary hypothesis is darker and more specific. People born around 1889-1890 were children during the so-called Russian flu pandemic, probably an H3 virus, and that early imprinting may have left their immune memory mistuned for an H1 challenge, dysregulating their response precisely in the 20-to-40 cohort that died hardest in 1918 [16][6]. Neither hypothesis is closed, and they are not mutually exclusive; both point to the host’s own immune status, not just the virus, as the lethal variable.
The Bacterial Second Strike
Most 1918 deaths were not caused directly by the virus at all. David Morens, Jeffery Taubenberger, and Anthony Fauci reviewed 118 published autopsy series covering 8,398 individual autopsies from 15 countries and found that the overwhelming majority of fatalities showed the histology of secondary bacterial pneumonia, ordinary upper-respiratory bacteria like pneumococcus and streptococcus invading lungs the virus had already stripped bare [12]. In an antibiotic-free world, that second infection was the actual killer. As Fauci put it, the virus landed the first blow and the bacteria delivered the knockout [12][16].
Three Waves and the Cities That Diverged
The pandemic arrived in three waves between March 1918 and the spring of 1919, and the second wave, from August to December 1918, was by a wide margin the deadliest [9]. The spring first wave was widespread but comparatively mild; somewhere over that summer the virus changed character, and the autumn wave that followed killed with a speed that overwhelmed hospitals, undertakers, and city governments across the Northern Hemisphere.
On the math: two American cities became the textbook natural experiment in how response timing changed outcomes. Philadelphia held a Fourth Liberty Loan parade on September 28, 1918, drawing roughly 200,000 spectators along a 23-block route while influenza was already circulating; within two weeks the city’s hospitals were full, and at the epidemic’s peak about 1,700 Philadelphians died in a single day [10][11]. St. Louis, by contrast, closed schools, theaters, and gatherings within days of its first cases. Studies by Richard Hatchett and by Howard Markel, both published in 2007, quantified the gap: cities that acted earlier and kept measures in place longer saw lower peak mortality, and St. Louis’s excess death rate ran far below Philadelphia’s [10][11]. The reopening lesson was equally sharp, since cities that lifted controls too soon saw cases rebound.
| City response (1918) | Key action | Outcome vs peers |
|---|---|---|
| Philadelphia | Held Liberty Loan parade Sept 28; delayed closures | ~1,700 deaths in one day at peak; highest-class excess mortality [10][11] |
| St. Louis | Closed schools and gatherings within days of first cases | Flatter curve; excess death rate roughly half Philadelphia’s [10][11] |
| Cities lifting controls early | Reopened before transmission ended | Documented case rebound (second local spike) [10] |
Resurrecting the Virus: Hultin, Taubenberger, and the 2005 Reconstruction
In 1997, retired pathologist Johan Hultin recovered frozen lung tissue from a 1918 flu victim buried in the permafrost at Brevig Mission, Alaska, and shipped it to Jeffery Taubenberger’s laboratory for sequencing [14]. The recovery closed a search that had run, for Hultin personally, since 1951, and it gave molecular virology the intact viral RNA it had never had.
The Frozen Lungs of Brevig Mission
Brevig Mission lost 72 of its 80 residents to influenza in five days in November 1918, and the dead were buried in a mass grave that the Alaskan permafrost then preserved [14]. Hultin had first opened that grave as a graduate student in 1951 and failed to culture any live virus. In 1997 he read Taubenberger’s first partial-genome paper, written from formalin-fixed army autopsy slivers, and offered to return. At 72 years old, funding the trip from his own savings, he reopened the grave and removed the frozen lungs of a woman the team came to call Lucy. Her tissue carried recoverable 1918 viral RNA, and it supplied the gene segments the slide archive could not [14][1].
What the Reconstructed Genome Revealed
Working from Hultin’s samples and the archival slides, Taubenberger and Ann Reid completed all eight gene segments, and in 2005 Terrence Tumpey rebuilt the live virus by reverse genetics under high containment at the Centers for Disease Control and Prevention [2]. The reconstructed 1918 virus killed mice faster than any modern seasonal strain and replicated to extraordinary titers in human lung cells [2]. No single gene carried the lethality. The hemagglutinin let the virus bind deep in the lower airway, the polymerase complex drove rapid replication, the small protein PB1-F2 worsened lung damage and bacterial co-infection, and the NS1 protein blunted the interferon alarm the body uses to fight back [2][13]. Virulence was a committee decision among genes, not a single villain.
What the 1918 Pandemic Still Teaches
The reconstructed 1918 virus, rebuilt gene by gene and characterized in 2005, proved that no single mutation made it lethal; its danger came from a combination of viral genes acting together with secondary bacteria [2][12]. That finding reframed pandemic preparedness. If the worst outcomes of 1918 came from bacterial pneumonia, then antibiotics, oxygen, and intensive care, none of which existed then, change the lethal arithmetic now, which is part of why no later H1N1 wave, including the 2009 swine-flu pandemic descended from the same lineage, approached 1918’s toll [2][16].
The harder lesson is the one the cities taught. The contrast between Philadelphia and St. Louis showed that the timing of ordinary public-health measures, closing gatherings, staggering exposure, acting before the curve steepens, can move death rates by a factor of two with no drug involved at all [10][11]. The 1918 pandemic is not a sealed historical episode. Its genome was read backward out of Alaskan ice, its mortality curve still anchors immunology, and its city-by-city record still informs how governments respond. For the broader catalog of unresolved natural-world puzzles, our science and natural anomalies pillar holds the collection, and the mysterious diseases and epidemics sub-niche gathers the parallel cases. The data did the work; the wonder is that a virus dead for a century could still be made to confess.
