In April 1979, anthrax spores escaped from a Soviet military facility in Sverdlovsk, now Yekaterinburg, and people who had been downwind began to die of inhalation anthrax. Soviet authorities blamed contaminated meat. That explanation was false, and officials maintained it for thirteen years. A team led by the Harvard biologist Matthew Meselson settled the cause in 1994 by mapping where each victim had been on the day of the release. This guide covers the outbreak, the denial and how the field work found the source. It also explains why the death toll is a floor and what investigators can learn from a case where the source was deliberately hidden.

Compound 19, a closed city and an agent that does not spread between people

Sverdlovsk was a large industrial city in the Ural Mountains, and it was closed to most foreign visitors during the Soviet period. It was renamed Yekaterinburg in 1991. Military Compound 19 belonged to the Soviet Ministry of Defense, and the published literature describes it as a military microbiology installation. It stood in the southern part of the city, with factories and residential blocks close around it. As far as I am aware, the Russian government has never published a full account of what the compound was doing in 1979. I describe it only as far as the established record goes: it handled Bacillus anthracis, and the investigation published in 1994 placed the source of the outbreak there.

The location mattered internationally because the Soviet Union was a party to the Biological Weapons Convention. The convention opened for signature in 1972 and entered into force in 1975. It prohibits developing, producing or stockpiling biological agents in types and quantities that have no justification for protective or other peaceful purposes. An anthrax outbreak traced to a military facility would have been evidence about Soviet compliance, which gave the government in Moscow a strong reason to explain the deaths some other way.

The organism shapes everything that follows. Bacillus anthracis forms spores that survive for long periods in soil and on animal products. The classic forms of the disease depend on how the spores enter the body:

• through broken skin (cutaneous anthrax)
• through eating meat from infected animals (gastrointestinal anthrax)
• through breathing spores deep into the lungs (inhalation anthrax)

Inhalation anthrax does not spread from one person to another. In an inhalation outbreak, every human case therefore marks a place and a time where that person breathed spores, and the form of disease a patient had is direct evidence about the route of exposure. That property made the Sverdlovsk outbreak solvable once someone was allowed to ask where the victims had been.

The contaminated meat explanation and the response built around it

Patients began arriving at city hospitals in the first week of April 1979 with high fever and difficulty breathing. Many deteriorated rapidly and died within days of falling ill. Accounts of the outbreak describe City Hospital No. 40 as the main hospital receiving cases and the place where autopsies were performed. Physicians recognized anthrax early, so the open question from the start concerned the route of exposure rather than the diagnosis.

The official explanation was false. It held that the victims had eaten meat from animals that died of anthrax, sold privately outside state inspection. The public health response was organized around that story. Accounts describe several measures:

• confiscation and destruction of privately sold meat
• killing of stray dogs
• washing of buildings and trees in the affected district
• a vaccination campaign for adults in the area

Observers later pointed out that washing trees and building exteriors makes little sense for a food source and considerable sense for spores deposited from the air. The public message nonetheless stayed on meat.

The records did not survive intact either. Both the pathologists and the later investigators describe state security officers removing hospital records and case files during or soon after the outbreak. That is why the 1990s investigation had to rebuild the case list from other sources, and why every figure from it carries the limits of that reconstruction.

The meat explanation outlived 1979. In 1988, during the glasnost period, a delegation of Soviet medical officials that included Pyotr Burgasov, a former deputy minister of health, visited the United States. They presented the outbreak at scientific meetings as gastrointestinal anthrax from contaminated meat, with clinical and epidemiological detail. As accounts of the visit describe it, several American scientists who heard the presentations found them coherent. A cover story prepared by medical professionals can survive expert questioning when the questioners cannot see the underlying data.

Early Western reports, an explosion story and a plausible-looking denial

Word of the outbreak reached Western readers months after the deaths, first through emigre publications and then through European and American newspapers. Some of those early reports described an explosion at the facility and death tolls far larger than any later documented count. None of those details was confirmed at the time. The later field work found a pattern consistent with an aerosol release from the compound, and it did not need an explosion to explain the cases. I treat the explosion accounts as unconfirmed reports that the evidence never supported.

The United States government raised the outbreak with the Soviet Union in 1980 as a possible violation of the Biological Weapons Convention. The Soviet reply repeated the meat explanation. Western analysts had no access to the city, the patients or the records, so they could only weigh intelligence reporting of uncertain quality against an official account backed by medical claims they could not check.

Meselson was among the scientists who examined the question during the 1980s. He had worked on chemical and biological weapons policy since the 1960s and is widely credited with influencing the American decision in 1969 to renounce biological weapons. As accounts of his involvement describe it, he judged the meat explanation plausible on the information then available, and he was criticized for that view. His earlier position belongs in this account because it shows how a careful expert reasons when the evidence comes from the party concealing the source. It also matters that he later went to look for himself.

For an outbreak investigator, this phase shows two opposite errors happening at once. Unsourced Western reports overstated the event with explosions and very large tolls. The official account understated it by assigning a local, ordinary cause that implicated nobody in uniform. Neither version could be tested until investigators could stand in the neighborhoods, interview families and examine the tissue.

Yeltsin in 1992 and the pathologists who kept their notes

Boris Yeltsin had been the Communist Party first secretary of Sverdlovsk region in 1979. In May 1992, as president of the Russian Federation, he said in a newspaper interview that the KGB had acknowledged military developments as the cause of the outbreak. Because the statement came from the man who had run the region at the time, it ended official denial at the political level. It came without documents, a description of how the release happened or a case list, and no full official Russian account has followed that I know of.

The medical evidence had been preserved by people who were expected to give it up. Faina Abramova and Lev Grinberg were pathologists who performed autopsies on victims in 1979. They kept their notes and tissue specimens after official records were removed. What they had documented was the characteristic pathology of inhalation anthrax: hemorrhagic inflammation of the mediastinum and of the lymph nodes in the chest, where inhaled spores are carried and germinate. Primary gastrointestinal anthrax centers its damage in the gut and its lymph nodes, so the chest findings pointed to the lungs as the portal of entry.

Abramova and Grinberg published their findings in 1993 in the Proceedings of the National Academy of Sciences, with Olga Yampolskaya and the American pathologist David Walker. The paper described the pathology in 42 cases. As it is commonly summarized, it also reported lesions in other organs, including the intestine in many cases, which is consistent with spread through the bloodstream in inhalation disease. That secondary involvement may help explain how a gastrointestinal story could be sustained by people inclined to sustain it. Readers who need the detail should work from the paper itself rather than from my summary of it.

The 1992 and 1993 field work: names, families and daytime locations

Meselson organized a team that traveled to Yekaterinburg in June 1992 and returned in 1993. Its members included:

• Jeanne Guillemin, a medical anthropologist who conducted many of the household interviews
• Martin Hugh-Jones, a veterinary epidemiologist
• Alexander Langmuir, who had founded the Epidemic Intelligence Service at the Centers for Disease Control
• Russian colleagues, among them Olga Yampolskaya and Ilona Popova

The hospital records had been taken, so the first task was to establish who had fallen ill.

Accounts of the field work describe a case list assembled from several sources:

• names compiled locally
• grave markers in the section of a city cemetery where victims had been buried
• what families, neighbors and coworkers could recall

Each name then had to be supported as an anthrax case by whatever evidence survived. Guillemin’s 1999 book, Anthrax: The Investigation of a Deadly Outbreak, describes the interviewing in detail and is the most useful single source on how the list was built.

The central question in every interview was where the person had been during the days just before the first cases, especially during working hours. People spend the working day away from home, and in an industrial district that place may be a factory floor a few hundred meters from a fence. A home address is the easiest location to collect and the least useful one for an airborne point source. The team recorded both home and daytime locations so they could be plotted and compared.

The team also collected evidence that did not depend on human memory. It obtained weather records for the period and veterinary records of animal anthrax in villages south of the city. Veterinary services had recorded livestock cases for their own purposes at the time, so those cases formed an independent data set that the human interviews could not have influenced. If animals had died along the same line as the people, the airborne explanation would have support that no family’s recollection could have produced.

A narrow zone, the April 2 wind and the livestock deaths farther out

The team published its results in Science in November 1994. It identified 77 patients, of whom 66 died. Most of them had lived or worked within a narrow zone that ran from Compound 19 to the southern limit of the city. Plotting by home address alone would have blurred that pattern, because some victims were inside the zone only during working hours, and the daytime locations made the shape clear.

Weather records showed a northerly wind on April 2, and the zone of human cases lay along the path that wind would have carried an aerosol from the compound. Livestock had also died of anthrax farther south, in villages along the extended axis of the same zone. Common summaries of the findings put those villages as far as roughly 50 kilometers from the compound. The team concluded that an aerosol of anthrax spores had escaped from the military facility and caused the outbreak, and it placed the release on April 2. Anyone citing the distances or the timing should take them from the paper.

The demographics pointed the same way. The cases were adults, and no children were reported among them, although a contaminated household food supply would be expected to sicken children as well. Accounts of the investigation describe a cluster of cases among workers at a ceramics factory close to the compound, people who had been at work during the day. Case onsets were spread over roughly six weeks, from early April into May. Inhalation anthrax can have a long and variable incubation period, so that spread is consistent with a single release and does not require repeated exposure.

The onset data have been used well beyond this case. Sverdlovsk is one of very few documented inhalation anthrax outbreaks in a human population, so researchers modeling incubation periods have used its onset dates. Discussion of how long antibiotics should continue after a suspected inhalation exposure has drawn on those data alongside animal studies. Responders should take the current post-exposure regimen from CDC guidance rather than from any historical account, including this one.

66 deaths as a documented floor, and how much was released

The death toll is commonly given as at least 66. That figure comes from the 1994 Science paper, which documented 66 deaths among 77 cases. The authors built their list without the original hospital records, from surviving names, cemetery evidence and interviews. Their count is therefore the number of deaths they could document, not a census. A victim who left no trace in any of those sources would be missing, and the true figure may be higher.

Higher estimates appear in some later accounts and press retellings. I have not been able to tie them to a documented case list I trust, so I give the range this way: at least 66 deaths documented by the Meselson team, with higher estimates in other accounts that rest on weaker sourcing. The very large tolls in the early Western reports were never confirmed and should not be cited. Soviet officials gave their own figures under the meat explanation. Because those figures came attached to a false account of the cause, I would not use them as an independent check without the original documents behind them.

The amount of material released is the other figure that gets inflated. The 1994 paper estimated the release from the size and shape of the affected zone and the number of cases. As its estimate is commonly summarized, the amount could have been as little as a few milligrams of spores or as much as about a gram, and the authors presented that as a rough estimate with wide uncertainty. Later modeling by other researchers has revisited the dose question under different assumptions. Retellings that give a precise spore count, or a quantity far above the published range, are not drawing on the field data, and I would not repeat them.

The mechanism of the release is the least settled part of the record. Ken Alibek, a former senior official in the Soviet biological weapons program who defected to the United States in 1992, wrote in his 1999 book Biohazard that a clogged filter in an exhaust system at the facility had been removed and not replaced. He gave this as an account he had learned within the program rather than as something he witnessed. It has not been confirmed by any released Russian document I know of, and the 1994 epidemiology neither depends on it nor tests it. I treat it as an unconfirmed account of how the release happened. That fact about the cause rests on the field data and stands whether the filter story is true or not.

Outbreak epidemiology when someone is concealing the source

The Sverdlovsk investigation used no technique that was unavailable in 1979. Its tools were all present locally at the time:

• household interviews
• a city map
• weather records
• veterinary reports
• autopsy pathology

What was missing in 1979 was the freedom to put those pieces together. For public health responders, the practical point is that ordinary field epidemiology can locate a hidden point source if the data are collected properly and kept long enough for someone to analyze them.

Several specific practices follow from the case:

• Daytime location. Case investigation forms need a field for where each person was during the plausible exposure window, with times, separate from the residence, because an airborne release is defined by where people were standing.
• Testing each hypothesis. Every proposed source should be checked against the demographic pattern it predicts. A household food source predicts illness across family members, including children. An outdoor or workplace airborne source predicts adults clustered by where they spent the day.
• Animal data. Animal health records belong in the same analysis, which means the health department needs a working relationship with the state veterinarian before an event rather than during one.

Records need protection from more than state security. Sverdlovsk could be reconstructed because two pathologists kept material they were expected to surrender, families remembered, and gravestones recorded dates. A comparable event in the United States is unlikely to involve seized files, but evidence can still be lost:

• records purged under a routine retention schedule
• a medical examiner’s office that did not keep specimens
• a dispatch system that overwrote location and time data after a set period

Retention rules written for ordinary administration can destroy the material an epidemiologist will want a year later.

The explanation that arrives first deserves the hardest look. In Sverdlovsk, the meat story came from authorities connected to the facility, and it was delivered with medical detail. Investigators should notice when response measures, such as washing trees, do not fit the stated cause. They should also record case data in a form that would let someone else test the official explanation later. Local responders will rarely be the people who settle a concealed-source question, but they create and keep the data that whoever settles it will need.

What to do at your agency

• The local health department epidemiologist, or the EMS medical director acting as liaison, pulls the current reportable disease case investigation template and confirms it records where each case was during the exposure window, with times, separately from the home address. If it does not, the epidemiologist adds that field to the template this month.
• The emergency management coordinator adds the state veterinarian’s office to the contact list in the biological incident annex and confirms the after-hours number with a call, so that livestock reports can reach the health department during an outbreak.
• The communications center manager asks the CAD administrator for the current retention period on incident location and time data. The manager then runs a test export of one month of medical calls by geographic area and sends it to the health department epidemiologist as a sample of what dispatch data can show.
• The emergency management coordinator asks the county or state medical examiner for the written policy on retaining tissue and autopsy records in deaths from suspected infectious causes, and files a copy with the biological incident annex.
• The emergency management coordinator writes one paragraph into the biological incident annex naming the source of archived wind data for the county, such as the nearest National Weather Service office or airport weather station, so that a plume question can be answered with recorded wind rather than recollection.
• The public information officer writes one paragraph into the public information annex requiring that every public statement about an outbreak’s source be labeled as confirmed or under investigation, with the agency that made the finding and the date.

Takeaways

• In April 1979, an aerosol of anthrax spores escaped from Military Compound 19, a Soviet Ministry of Defense facility in Sverdlovsk, according to the investigation the Meselson team published in Science in November 1994.
• Soviet authorities attributed the deaths to contaminated meat, a false explanation that Soviet medical officials were still presenting to American scientists in 1988.
• Boris Yeltsin, who led the Sverdlovsk region in 1979, said in May 1992 that the KGB had acknowledged military developments as the cause of the outbreak.
• Pathologists Faina Abramova and Lev Grinberg preserved autopsy notes and tissue showing inhalation anthrax, and their findings on 42 cases were published in 1993.
• The Meselson team documented 77 cases and 66 deaths. That count is a minimum because the original records had been removed, and higher estimates in other accounts have not been tied to a documented case list.
• Most victims lived or worked in a narrow zone running from the compound along the April 2 wind, and livestock died of anthrax along the same axis farther from the city.
• The released quantity, as the 1994 estimate is commonly summarized, ranged from a few milligrams to about a gram of spores, and the clogged-filter account of the mechanism comes from Ken Alibek and remains unconfirmed.
• Sverdlovsk was solved with ordinary field epidemiology, which works only when case investigations record where people were and the records survive long enough to be analyzed.