How many COVID-19 infections are asymptomatic? Research shows it may be a lot more than thought

Genetic tests are the gold standard for detecting if someone’s sick with COVID-19, but recent research argues large numbers of the infected don’t test positive for the disease and never develop symptoms. That might mean serological tests that reveal antibodies to the disease are essential to know the true scope of infection.
Written by Tiernan Ray, Senior Contributing Writer

Two-thirds or more of people infected with COVID-19 who infect others may do so before ever showing symptoms themselves, a prospect that is prompting intense activity to understand the so-called asymptomatic nature of some carriers.


The protein array used by Lei and team, referred to as a proteome array, holds 23 different virus proteins. Reactions by a person's blood to the proteins show up under a laser scanner as different colored dots in different degrees of intensity. Depending on the color and intensity, people can be positive for one type of antibody, negative for another, or positive or negative in combination, in ways that show specificity for each protein. 

Lei et al.

Many, perhaps most asymptomatic patients will never actually develop symptoms, some researchers believe, making them a silent population of the infected, a covert cohort that is undercounted by typical COVID-19 genetic tests.

One study out of China last week argues that so many asymptomatic people go undetected that it may be time to have a new definition of infectious people -- one that relies not only on genetic tests but possibly on serological tests that find antibodies in the blood even when a person tests negative for the disease. 

Also: With unprecedented amounts of genetic data, researchers are tracking how COVID-19 mutates around the world

Qing Lei and colleagues from China's Tongji Medical College screened 177 people in the city of Wuhan, where the virus started, from February to April, they report this week in a paper posted on the medRxiv pre-print server. The study matched the 63 asymptomatic people with 51 people who showed mild symptoms of COVID-19, and 63 people who were healthy control subjects. All of the participants were followed for 65 days.  

Among the asymptomatic, 12 people who had no symptoms tested positive for COVID-19 according to the widely used RT-qPCR genetic test, the so-called asymptomatic carriers. Another 51 asymptomatic people came up negative on the genetic test but were found to have antibodies to COVID-19 in their blood, meaning they were infected. Antibodies are particles that form in the blood as the body's response to the virus. 

In other words, 81% of people with the disease in this sample who had no symptoms appeared negative on typical tests -- all false negatives.

"Asymptomatic infection has clear infectivity," one of the scientists, Xionglin Fan, told ZDNet in an email, citing past studies by the Chinese Health Commission. For that reason, said Fan, "early identification of asymptomatic infection is urgently needed."

Researchers from many institutions have pointed out for months now that people who show no symptoms of COVID-19 -- no fever, no cough, no shortness of breath -- can nevertheless infect others. The Tongji work is part of a broad trend in COVID-19 research to more sharply delineate just how big the asymptomatic population is. 

Also: How a smartphone coupled with machine learning may become a simple, efficient test for COVID-19

If most asymptomatic people go undetected, it could be many more than generally believed, and if many of those undetected infect others, then a whole part of the transmission of the disease might be obscured.

Fan and colleagues propose that the definition of people who are infected with COVID-19 be extended beyond the definition of merely those who have tested positive via the genetic test, a so-called nucleic acid test, or NAT for short. 

Instead, "asymptomatic infection should be defined as a person [who] has positive NAT or/and antibodies without clinical symptoms," write Lei and colleagues in the paper, titled, Antibody dynamics to SARS-CoV-2 in Asymptomatic and Mild COVID-19 patients.

The paper has not yet been peer-reviewed, and so its findings should be taken with a degree of caution. 

Asked about the definition, Fan told ZDNet: "We think this definition more reasonable and practical."

"If only NAT detection [is used] for identification, 81% of asymptomatic infection can not be found out," wrote Fan. That would mean "the proportion of asymptomatic infections might be much higher than the current rate reported in China."

Fan cites support for the importance of asymptomatic cases in research published in Nature magazine last week by fellow researchers at Tongji, led by Xingjie Hao. That study looked back at 32,583 cases in Wuhan and found that 87% of infections prior to March 8 were cases that were "unascertained, potentially including asymptomatic and mild-symptomatic cases," a phenomenon the authors call a "high covertness."

The scope of asymptomatic infection and transmission is widely debated. Asymptomatic cases of COVID-19 have been notoriously difficult to identify and count, leading to controversy about the term.

For example, a review of the literature posted on June 4 on medRxiv by Oyungerel Byambasuren and colleagues at Bond University in Australia looked at nine different primary studies of COVID-19. The authors concluded that asymptomatic patients made up on average 15% of total cases of COVID-19, "lower than many highly publicized studies."

Lei and colleagues at Tongji say the data show otherwise. 

"I don't agree with the perspective that the truly 'asymptomatic' may be rare," Fan told ZDNet via email. 

Fan directed ZDNet to a statement by China's National Health Commission on April 15 that reported 6,764 cases of asymptomatic infection in China up until that time. Only 1,279 of those asymptomatic cases were ultimately confirmed via nucleic acid test. "These 1,297 cases might be thought of as pre-symptomatic," Fan told ZDNet. "The others really remain asymptomatic."


Patterns of generation of immunoglobulin G, or IgG, antibodies in response to different proteins of the COVID-19 virus, by healthy, mildly sick, and asymptomatic individuals. 

Lei et al.

If asymptomatic people are truly a group that never develop symptoms and don't even show up on a genetic test, and therefore are not confirmed cases, then that means the transmission by people without symptoms may be much higher than believed. 

For example, a study by Tapiwa Ganyani and colleagues at I-BioStat, part of Belgium's Hasselt University, concluded that upwards of 62% of transmission may be pre-symptomatic. However, that study only looked at symptomatic and pre-symptomatic transmission and explicitly excluded asymptomatic transmission. 

Besides pointing out the scope of asymptomatic transmission, it's conceivable serological studies of the kind conducted by Lei and team may lead to a more nuanced understanding of infection. Because serological tests can measure different kinds of reactions, they can classify people into sub-groups. 

The serological test used by Lei et al. to detect the asymptomatic patients' antibodies is a marvel of modern technology known as bio-printing. 

Proteins of the virus are loaded in solution into a kind of inkjet printer from British firm Arrayjet. The printer emits those proteins through its nozzles onto the surface of a thin strip of film. The film with the dots of protein, known as a protein microarray, is then exposed to the blood sera of individuals. If an individual's blood forms a reaction to any of the proteins, that means that the individual's blood may contain antibodies that have formed in response to exposure to the virus. The reaction can be seen based on the hue and intensity of a color change in the film when it's run through a laser scanner. 

There are 23 different virus proteins printed on the assay used by Lei and the team. Depending on the kind and number of dots that show up, people can be positive for one type or antibody, negative for another, in ways that show specificity for each protein. 

Those variations are so specific they can reveal where some people have been misdiagnosed, Fan told ZDNet

"Our proteome array really demonstrated there are several misclassifications [of] healthy contacts, asymptomatic infections, and mild patients in clinic," said Fan. 

Also: 'We are doing in a few months what would normally take a drug development process years to do': DoE's Argonne Labs battles COVID-19 with AI

One big potential change of that sub-typing of different groups, if the work is ultimately supported by further findings, is that the definition of those who've recovered versus those who are asymptomatic may also change. 

The conventional wisdom is that recovering from the disease is marked by the build-up of antibodies. But then how does one distinguish people who've recovered and people who are infectious if both exhibit antibodies? 

When ZDNet put that question to Fan, the researcher replied that serological tests are blurring the important distinctions found by first contact tracing and only then testing individuals. The truly asymptomatic, in view of the team, are marked by the response to the so-called S1 protein (actually a subunit of the S protein, containing the receptor binding domain, the part that attaches to human cells). That protein, the spike glycoprotein, as it's known, is the characteristic bud you see sticking out of the body of the virus in all the pictures, forming the "corona" of the coronavirus. Those who have recovered are marked by the response to a different protein, the N protein.

"Our result found that S1 specific IgM [immunoglobulin type M, one of the antibodies] response can be used for early identification of asymptomatic infection, and should be performed alone, in contrast to that used in current commercial serological kit targeting both S and N proteins," Fan told ZDNet in email. The study showed that IgM responses showed up as early as 7 days after the initial estimated time of infection. 

In contrast, an antibody called immunoglobulin G type, or IgG, was more specific to the N protein than S1, and so should be more an indicator of people who've recovered, notes Fan.

"IgG response to S or N [proteins] means that they recovered from infection," Fan added. "S1 or S or RBD specific IgG might indicated recovered and immune."

The results argue for being more specific in how various antibody tests are handled, Fan told ZDNet.

"So the serological test should be performed by the kit only targeting these proteins, not N protein or combination to improve the sensitivity of detection," advised Fan, "which also is an important finding of our paper."

Somewhat ominously, the scientists also observed the possibility that the neutralizing antibodies in a person's blood, the kind that fight the disease, seem to go away over time. Lei and team found that antibodies began to erode 25 days after the infection began. Within a couple of months, they were generally gone from individuals' bloodstreams. 

That echoes another study last week of serological tests from Jeffrey Seow and colleagues at the Department of Infectious Diseases, School of Immunology and Microbial Sciences at King's College in London. 

If antibodies disappear within months, then immunity to COVID-19 after recovery may not last long. 

As a consequence, "Both papers do not think immunity passport is okay," Fan told ZDNet, rejecting the notion of a special exemption for someone to avoid lockdown and masks after having recovered. 

Editorial standards