SARS originated in horseshoe bats, likely crossed into humans through infected animals sold at live-animal markets, and first appeared in Foshan, a city in southern China’s Guangdong Province, in November 2002. The earliest known patient fell ill on November 16 of that year. Within months, the virus had spread across continents, ultimately infecting more than 8,000 people and killing nearly 800 before it was contained in mid-2003.
The First Cases in Guangdong
The earliest known SARS patient was a 45-year-old village leader and administrator in Foshan city. He became ill on November 16, 2002, and was only identified later through retrospective case searching. Investigators noted that he had prepared food including chicken, domestic cat, and snake, all common in the regional cuisine of southern China.
Among the first wave of cases in Guangdong, a striking pattern emerged: nine of the early patients worked in the food trade. Seven were restaurant chefs at township restaurants where live animals were slaughtered on-site. One was a market produce buyer for a restaurant, and one sold snakes at a produce market that also offered other live animals. Six of these food handlers lived or worked in Shunde, an urban district within Foshan with about 1.7 million people. None of them could be directly linked to each other through contact, which suggested they were each catching the virus independently from animals rather than passing it person to person in a single chain.
By late April 2003, Guangdong Province had reported 1,454 SARS cases and 55 deaths. The overall fatality rate was 3.8%, but it jumped to 12.7% in people over 65.
Bats as the Natural Reservoir
Horseshoe bats are the natural reservoir for the family of viruses that includes SARS. These bats carry a range of closely related coronaviruses, collectively called sarbecoviruses, without getting sick themselves. The SARS virus belongs to this subgroup within the broader category of betacoronaviruses, the same viral genus responsible for MERS and COVID-19.
Researchers have found SARS-like coronaviruses in several horseshoe bat species sampled in caves across China’s Yunnan Province. Bat viruses don’t typically infect humans directly, though. They usually need to pass through another animal species first, adapting along the way until they can bind to human cells efficiently. That’s where the live-animal markets came in.
The Role of Civets and Markets
Masked palm civets, small cat-like mammals sold as food in Guangdong, were initially suspected of carrying SARS to humans. In January 2004, authorities culled more than a thousand civets in the province based on this concern. But the picture turned out to be more complicated than a simple civet-to-human story.
Testing revealed that civets on farms were largely free of SARS infection. Yet roughly 80% of civets sampled from one animal market in Guangzhou carried significant antibodies to the virus. This gap pointed to the markets themselves as the problem. Overcrowded conditions and the mixing of multiple animal species in close quarters created a perfect environment for a bat virus to jump between species and mutate. Civets appeared to be getting infected at the markets, not in the wild or on farms. Many farms had even obtained their original breeding stock from markets, further blurring the chain of transmission.
Another telling detail: people who got sick early in the epidemic were more likely to live near a produce market than near a farm. The virus was circulating in the market ecosystem, not in the countryside.
How SARS Went Global
SARS smoldered in Guangdong for months before the rest of the world noticed. The turning point came in February 2003, when a doctor from Guangdong Province traveled to Hong Kong and checked into a hotel. He was already infected. Other hotel guests caught the virus from him, and they carried it to Vietnam, Singapore, Canada, and elsewhere. One of his contacts was hospitalized in Hong Kong on February 17 with severe pneumonia; by February 22, healthcare workers who had treated him were falling ill.
On February 28, a hospital in Hanoi, Vietnam, contacted the World Health Organization about a patient with unusual pneumonia. Dr. Carlo Urbani, an Italian infectious disease specialist working for the WHO, visited the hospital on March 3 and recognized he was looking at something new: severe, fast-moving, and highly contagious. He learned the patient had recently stayed at the same Hong Kong hotel where other guests had been infected, connecting the dots of international spread.
Urbani pushed Vietnamese authorities and the WHO to act immediately. Infected patients were isolated, infection control measures went into effect, and the hospital was cordoned off by security guards. On March 9, Urbani and a WHO representative called for urgent action, and Vietnam’s Vice Minister of Health assigned a local team that same day. This rapid response in Vietnam became a model. By identifying and isolating patients before they could infect hospital staff, the outbreak was prevented from spiraling further. Urbani himself contracted SARS and died on March 29, 2003.
After the Epidemic Ended
The WHO declared the SARS outbreak contained in July 2003. No natural human cases have occurred since. But the virus did resurface briefly through laboratory accidents. In Singapore, a graduate student working at the Environmental Health Institute became infected in September 2003 after handling a sample that had been cross-contaminated with SARS. He had worked with the sample on August 23, three days before his symptoms began. Similar lab-acquired infections occurred in other countries in 2004, underscoring how dangerous the virus remained even in controlled settings.
SARS-like viruses continue to circulate in horseshoe bat populations across Asia and beyond. The 2002 outbreak was not a one-time fluke but a demonstration of what happens when bat coronaviruses find a bridge species in the right conditions. The same mechanism, a sarbecovirus jumping from bats through an intermediate host to humans, is the leading explanation for how SARS-CoV-2 emerged to cause COVID-19 nearly two decades later.