Subscribe to Fortune’s Outbreak newsletter for a daily roundup of stories on the coronavirus outbreak and its impact on global business.
The number of COVID-19 coronavirus cases worldwide reached over 98,600 on Friday, including 233 in the U.S., where 12 people have died. Germany’s health minister called the outbreak, which has spread to at least 84 countries, a “global pandemic.”
Along with the global rise in infections and deaths is an increase in recoveries—more than 55,000 people have recovered from the disease as of Friday. While those recoveries are welcome news for a world on edge, public health officials in China and Japan have reported cases in which patients see their symptoms subside and are discharged from hospitals—only to test positive for the coronavirus again.
On March 2, a man in Wuhan, China, with COVID-19 died after being readmitted to the hospital two days after he was discharged, according to a Chinese news site.
The exact number of these cases is unclear—Japan has had at least one case and China has had several, including the Wuhan man, one man in Sichuan province, and 13 discharged patients from one hospital in Guangdong province.
Being reinfected with COVID-19 is possible, said Sharon Lewin, director of the Peter Doherty Institute for Infection and Immunity in Melbourne, but such an instance would be “surprising.” It’s possible that patients are not actually being reinfected, but that other factors—misdiagnosis, human error, or faulty tests—are giving that appearance.
Testing done with animal models for SARS, a similar coronavirus, indicates that patients who recover from the disease will have immunity from it. “From what we know of other coronaviruses, you recover, you make an antibody response, and you clear the virus, and then if you rechallenge the animal with the same virus they’re protected,” Lewin said.
Lewin put stocks in explanations other than reinfection, but she emphasized that the virus still has too many unknowns and there are too few reported twice-positive cases to conduct a study or draw a definitive conclusion.
Human error and faulty tests
The mysterious double positives could simply be the result of human error. Hospitals have been testing for the presence of the virus using swab samples from a patient’s nose, throat, and sometimes lungs. Swabs can yield different quantities of the virus depending on where the clinician swabs and how they do it, Lewin said. If the swab sample yields too little of the virus, it could lead to a false negative for a patient who is still infected.
Problems with the testing kits themselves can also lead to false negatives or false positives—like the batch of faulty kits the U.S. Centers for Disease Control and Prevention rushed out and then ordered back, a blunder that delayed testing in the U.S. by weeks and may have contributed to the virus’s spread there.
The Wuhan man who died after he was discharged and readmitted had twice tested negative for the virus before being released, though a pre-discharge CT scan indicated a remaining infection in his lungs.
Virus ‘reservoir’
Another possible culprit for double positives is residue virus in a patient’s system following their discharge from the hospital, said Vijay Dhanasekaran, an associate professor of microbiology at Monash University who studies viral pathogens and infectious disease outbreaks.
Since hospitals largely discharge patients based on the reduction of their symptoms, Dhanasekaran said, the virus could be lingering in patients with milder COVID-19 cases who are discharged. (The World Health Organization said 80% of cases are “mild to moderate.”)
One hospital in Wuhan said it would start conducting antibody tests before releasing patients, after multiple discharged patients fell ill again and returned to the hospital, suggesting that a symptom check or swab test is insufficient to determine whether a patient has recovered. (Antibodies are produced by the body’s immune system in response to infection.)
The coronavirus could also be lingering in a “reservoir” within the patient’s body. Ebola virus and HIV are known to sequester in viral reservoirs, though the trait is “not typical of coronaviruses that we know of,” Lewin said. Nevertheless, the reservoir theory is “the sort of thing that would be on my mind in trying to understand [the COVID-19 reinfection reports],” she said.
Real reinfection
Scientists are not ruling out the prospect of reinfection.
Dhanasekaran said reinfections are “possible,” and pointed to influenza and respiratory syncytial virus, two viruses—neither coronaviruses—that patients can get more than once.
“[I]n short, it is possible, but I would suggest in a minority of cases,” said Brett Lidbury, an epidemiologist at Australian National University. “All we can do is speculate at the moment, based on previous experience and prior data.”
“If this phenomenon of being positive-negative-positive is very rare, you need a very large study to capture those people,” Lewin said. What’s needed going forward, Lewin said, is more research into the “natural histories of the disease—so knowing exactly how long people are excreting virus for, how that relates to system improvement or deteriorating, and how that relates to infectiousness.”
“All my ideas around being positive-negative-positive, they’re all theories,” Lewin added. “We don’t have the answer yet.”
More must-read stories from Fortune:
—Testing for coronavirus should be free, but it’s not always that simple
—Coronavirus is mutating: Chinese scientists find second strain
—Coronavirus is giving China cover to expand its surveillance. What happens next?
—The coronavirus is officially claiming its first corporate casualties
—Why the U.S. is so far behind other countries in coronavirus testing
—Travel insurance is booming, even though it doesn’t help flight changes and cancellations
Subscribe to Fortune’s Outbreak newsletter for a daily roundup of stories on the coronavirus outbreak and its impact on global business.
