Epstein's body is subjected to an autopsy



COVID-19 is an infectious disease caused by the SARS-CoV-2 virus that is easily transmitted between people in a similar way to the common cold or the flu. Most people with COVID-19 have mild to moderate respiratory disease and some have no symptoms (asymptomatic infection). Others have severe symptoms and need specialist treatment and intensive care.

The immune system of people with COVID-19 responds to this infection by making proteins in the blood that can fight off the virus (antibodies). Tests that detect antibodies in people's blood could show whether people currently have COVID-19 or have already had it.

Why are accurate tests important?

Through careful testing, it is possible to identify those who may need treatment or who should isolate themselves to contain the spread of the infection. Failure to detect COVID-19 in people even though they are infected (false negative result) could delay treatment and put the infection at risk of spreading to other people. If COVID-19 is incorrectly diagnosed even though the disease is not present (false positive result) it could lead to unnecessary further testing, treatment and isolation of the person and their close contacts. Properly identifying people who have had COVID-19 before is important in order to measure the spread of the disease, to study the success of public health interventions (such as isolation), and possibly to identify people with immunity (if detected for the antibodies in the future that they indicate immunity).

To identify false negative and false positive test results, antibody test results are compared in people known to have COVID-19 and people known not to be infected. Study participants will be classified as to whether or not they have COVID-19 based on a criterion called "reference standard". Many studies use nose and throat swabs to identify people with COVID-19. The samples go through a test called the reverse transcriptase polymerase chain reaction (RT-PCR). With this test procedure, infection can sometimes be missed (false negative result), but additional tests can identify COVID-19 infection in people with negative RT-PCR results. These include measuring clinical symptoms, such as a cough or high temperature, or imaging tests, such as chest x-rays. People known to not have COVID-19 are sometimes identified through stored blood samples taken prior to the advent of COVID-19, or through patients with respiratory symptoms related to other medical conditions.

What did the review examine?

The studies looked at three types of antibodies, IgA, IgG and IgM. Most tests measure both IgG and IgM, but others measure only one of these three antibodies or a combination of all three antibodies.

Antibody levels rise and fall at different times after infection. IgG is the last to rise but lasts the longest. Antibody levels are usually highest a few weeks after infection.

Some antibody tests can only be performed with special laboratory equipment. Other tests are single-use tests, similar to pregnancy tests. These tests can be performed in laboratories or wherever the patient is located (point-of-care), in the hospital or at home.

We wanted to find out if antibody tests:

- are sufficiently accurate to diagnose infection in people with or without COVID-19 symptoms, and

- Can be used to find out if someone has already had COVID-19.

How did we go about it?

We looked for studies that measured the accuracy of antibody tests against criteria of the reference standard in detecting current or past COVID-19 infection. The studies were able to compare any antibody test with any reference standard. The subjects could be tested in the hospital or outside clinical facilities. The studies were able to test people who are known to have had COVID-19 or not, or who were suspected of having an infection.

Study characteristics

We found 54 relevant studies. They were carried out in Asia (38), Europe (15) and in the USA and China (1).

Forty-six studies (46) only included those who were hospitalized with suspected or confirmed COVID-19 infection. Twenty-nine studies (29) compared test results from people with COVID-19 with test results from healthy people or people with other medical conditions.

Not all studies provided data on the age and gender of the participants. Often times, we were unable to determine whether the studies looked at current or past infections, as few reported whether participants were in recovery. We did not find any studies that only tested asymptomatic people.

Main results

Our findings are largely based on 38 studies that provided results based on the time since symptoms were first noticed.

The antibody tests, done a week after the first symptoms, only detected 30% of people who had COVID-19. The accuracy increased in the second week, 70% was detected and was highest in the third week (with more than 90% detected). Little evidence was available for after the third week. The tests gave false positive results in 2% of those without COVID-19.

Results from IgG / IgM testing three weeks after the onset of symptoms suggest that if 1,000 people do antibody tests and 50 (5%) of them actually had COVID-19 (as we might expect in a national survey):

- 58 people would test positive for COVID-19. Of these, 12 people (21%) would not have a COVID-19 (false positive result).

- 942 people would test negative for COVID-19. Of these, 4 people (0.4%) would actually be infected with COVID-19 (false negative result).

If we tested 1,000 healthcare professionals (a high risk setting) with symptoms, and 500 (50%) would actually have COVID-19:

- 464 people would test positive for COVID-19. Of these, 7 people (2%) would not have a COVID-19 (false positive result).

- 537 people would test negative for COVID-19. Of these, 43 (8%) actually had COVID-19 (false negative result).

We did not find convincing differences in accuracy for different types of antibody tests.

How reliable were the results of the studies in this review?

Our confidence in the evidence is limited for several reasons. In general, the studies were small, did not use the most reliable methods, and incompletely reported their results. Often times, patients with COVID-19 who may have had a false negative result with PCR were not included, and the studies determined data on people who did not have COVID-19 based on records of tests performed before COVID-19 emerged. This could have affected the test accuracy, but it is impossible to determine by how much.

Who do the results of this review apply to?

Most participants were hospitalized with COVID-19 and therefore likely had more severe disease than people with mild symptoms who were not hospitalized. This means we don't know how accurate antibody tests are in people who have milder disease or who have no symptoms.

More than half of the studies looked at tests they had developed themselves, most of which are not available for purchase. Many studies were quickly published online as “preprints”. Preprints do not undergo the usual rigorous reviews of published studies. Hence, we are not sure how reliable they are.

Since most of the studies were done in Asia, we don't know if the test results would be similar in other parts of the world.

What are the conclusions of this review?

The review shows that antibody tests could play a useful role in detecting COVID-19 infection, but that when the tests are done is important. Antibody tests could help confirm COVID-19 infections in people who have symptoms for more than two weeks and no RT-PCR test or who have negative RT-PCR test results. The tests are better at detecting COVID-19 in people if it has been two weeks or more since symptoms first appeared. However, we don't know how well they work five weeks after the onset of symptoms. We don't know how well the tests work in people who experience milder disease or no symptoms, as the majority of the studies in this review were conducted on people in hospitals. Over time, we will learn whether people with a previous COVID-19 infection have immunity to future infection.

More research is needed on the use of antibody tests in people recovering from COVID-19 infection and in people who experience mild or no symptoms.

How current is this review?

This review includes evidence that was published as of April 27, 2020. As a lot of new research is published on this topic, we will update this review regularly.