COVID-19 Antibody Testing
What Do the Results Mean and Will It Be Helpful?
By Stuart M. Caplen, M.D.
As society tries to reopen during this COVID-19 pandemic, one of the tools most commonly cited as being important in that process is the antibody blood test. How accurate is that test and what does it measure?
After a person gets infected, an immune response is mounted. Initially the antibody response is with immunoglobulin M (IgM), whose presence indicates an acute infection. Levels of IgM may be present a few days after symptoms start, but then drop over a few weeks and immunoglobulin G (IgG) starts to rise. IgG typically starts to rise 10 to 14 days after infection,[1] but may start later. In certain diseases, like measles, IgG antibodies give lifelong immunity if a person is exposed to measles again. In COVID-19, no one yet knows how long the protection may last or even if having IgG antibodies will confer immunity to prevent reinfection, but as the pandemic continues, those questions will eventually be answered. For the SARS coronavirus, one study found elevated IgG antibody levels last an average of two years in recovered patients, although about 10% of patients did not have detectable levels at that time. At three years after SARS infection, 50% of the patents still had significant IgG levels.[2]
Some antibody tests measure both IgG and IgM, but more commonly the current antibody testing is just for IgG to see if the person has had the disease in the past. There are several types of antibody tests. An ELISA (enzyme-linked immunosorbent assay), which is run in a laboratory and can take a few hours. Another immunoassay uses chemiluminescence and produces light that can be measured if antibodies are detected.[3] A third test is the lateral flow assay, which can be done in minutes in a doctor’s office. A rapid pregnancy test is an example of a lateral flow assay test.[4] Finally, neutralization assays assess the ability of antibodies in serum to neutralize the virus. It takes several days to perform and is generally done only in public health and research facilities.[5] The lateral flow assays are thought to be the least accurate assay.[4] Adding to testing confusion is that the FDA in April emergently allowed over 70 companies to produce antibody tests, and they are only now trying to ensure that those tests are actually accurate.[3]
In medical testing one needs to know the sensitivity and specificity of the test. Sensitivity is the ability of a test to correctly identify all the true positives in blood samples, in this case the presence of COVID-19 IgG antibodies. The sensitivity of a test is derived by the formula:
true positives / (true positives + false negatives).
A false negative is when the test result was negative, but the tested patient actually had antibodies to COVID-19. A sensitivity of 100% or 1 is perfect and means there were no false negatives.
Specificity is the ability of the test to identify all the true negatives in blood samples, in this case the absence of COVID-19 IgG antibodies. The specificity of a test is derived from the formula:
true negatives / (true negatives + false positives).
A false positive is when the test result was positive, but the patient never had COVID-19. A specificity of 100% or 1 means there were no false positives.
To obtain the sensitivity and specificity percentages for an IgG COVID-19 antibody test, results must be compared against another test with known results. The nasopharyngeal PCR (polymerase chain reaction) test used to diagnose acute cases of COVID-19 infection would be the comparison test typically used. It is very unlikely that a medical test will be perfect, having both 100% sensitivity and 100% specificity. In COVID-19 antibody testing there may be cross reactivity with other coronavirus antibodies, which may give some erroneous false positive results.[5] Another potential source of statistical error is that manufacturers’ claims of test sensitivity and specificity are typically based on a selected sample group of patients. When the test is in general use and many more people have been tested, it may turn out that the initial sensitivity and specificity estimates were inaccurate.
There are many COVID-19 antibody tests currently on the market with different sensitivities and specificities. Each test will have a different number of false positives and negatives. In a prepublication manuscript, 12 different COVID-19 antibody tests were analyzed, and the range of specificities for those tests was 84% to 100%.[6] A problem with some of the antibody tests is they have low specificities which result in a higher rate of false positives. The false positive rate of a testing method is important information, as people who get a positive result may assume they have had the disease and have antibodies to it, when in fact they do not. This is not an insignificant issue, as a test with a specificity of 90% and a sensitivity of 100% performed in an area of low disease prevalence, such as upstate New York, will have an estimated 73% of positive results being false positives.[7]
While an estimated percentage may or may not be accurate compared to real-world results, how can a test have a statistical probability that there will be more false positives than true positives? To correctly calculate the false positive rate for a test you also need to consider the prevalence of the disease in the population tested. The positive predictive value (PPV) calculation is a more accurate measure of the rates of true and false positives. The formula is complicated, but it uses the incidence of the disease in the population along with the test’s sensitivity and specificity percentages. There is an inverse relationship between prevalence and the false positive rate. The higher the prevalence of disease in a population, the lower the false positive rate. The lower the prevalence, the higher the false positive rate. If the disease has affected 3% of the population, the percentage of false positives will be higher than if the disease has affected 50% of the population.
In New York State 3,000 people were randomly tested for COVID-19 IgG antibody levels. The findings, reported on April 23rd, were a positive IgG antibody rate of 21.2% in New York City and 3.6% in upstate New York.[8] As this blog was being written, Roche announced they had developed a COVID-19 antibody test that was 100% sensitive and 99.81% specific when testing was done at least 2 weeks after infection.[9] Using the New York antibody prevalence statistics and the Roche test parameters results in an estimated 0.7% false positive rate in New York City antibody testing, and an estimated 4.8% false positive rate in upstate New York.[7] It is clear that using a test with a 99.81% specificity results in a marked decrease in the estimated false positive rate, compared to a test with a 90% specificity.
Trying to apply population statistics, such as those in the previous paragraph, to diagnose or treat an individual patient can be problematic. If someone has a positive result in upstate NY, how can you tell if that person is in the 4.8% false positive group or the 95.2% true positive group? In New York City, you may not be able to tell if a specific individual is part of the 0.7% false positive group, or the 99.3% true positive group. A clinical history of the patient having symptoms consistent with COVID-19, or having a close contact or work exposure to someone who was infected, will make it more likely a positive test is a true positive. For the asymptomatic patient, there may be no way to know for sure which group they are in. In New York City, not correctly diagnosing a false positive result in seven out of a thousand people with positive test results is a lower error rate than many of the tests currently used for diagnosing other diseases.
Key Points:
COVID-19 IgG antibody testing has usefulness as an epidemiologic tool to determine actual disease prevalence.
With newer tests being developed with better specificity, COVID-19 antibody testing is becoming a significant clinical tool. At the present time its most important clinical function is to test for antibodies in people who have recovered from COVID-19 who wish to donate antibody-rich plasma to help treat critically ill patients. It may also be possible that some patients who received these antibody-rich infusions won’t produce their own antibodies and testing will help determine if that is the case or not.[3]
Antibody testing may have value in people who were symptomatic, or have had possible exposure, such as health care workers to confirm they have had the disease. Many symptomatic individuals were not tested for COVID-19 or could have had a false negative PCR nasopharyngeal test. The PCR test can be falsely negative in up to 30% of people infected with the virus.[10] A strong past clinical history of disease symptoms or possible exposure to the virus will increase the likelihood that a positive result is a true positive.
In the asymptomatic, or minimally symptomatic patient group, who might want to know if they were infected, a high specificity IgG antibody test may possibly be helpful for them, and would be especially helpful epidemiologically. Patients getting this testing should be informed that a small percentage of positive results may be false positives.
If the test is done early in the course of the disease, before IgG antibody is being produced, there may be false negative test results.
The sensitivity and specificity of the test used in measuring the antibody are important factors in interpreting results. Disease prevalence in an area may also be a consideration.
Whether having COVID-19 IgG antibodies confers immunity to repeat COVID-19 infection is still under investigation.
Finally, the true and false positive percentages in this article were calculated using statistical equations with data from a manufacturer’s claim about the test validation and results from limited epidemiological testing data. As is the case in much of statistics, the calculated percentages may or may not be totally accurate in reflecting the real-world situation. However, the results and concepts can be used as a tool to better understand the issues involved in interpreting COVID-19 antibody testing.
References
Accelerated Emergency Use Authorization (Eua) Summary Covid-19 Elisa Igg Antibody Test (Mount Sinai Laboratory). Retrieved from: https://www.fda.gov/media/137029/download
Wu LP, Wang NC, Chang YH, et al. Duration of antibody responses after severe acute respiratory syndrome. Emerg Infect Dis. 2007;13(10):1562‐1564. doi:10.3201/eid1310.070576
Abbasi J. The Promise and Peril of Antibody Testing for COVID-19. JAMA. Published online April 17, 2020. doi:10.1001/jama.2020.6170 https://jamanetwork.com/journals/jama/fullarticle/2764954
Dance, A, Covid-19 antibody testing: Tougher than true/false, Knowable Magazine from Annual Reviews, 5/1/2020, Retrieved from: https://www.knowablemagazine.org/article/health-disease/2020/coronavirus-antibody-testing
SARS-CoV-2 Serology (COVID-19) Antibody (IgG), Immunoassay, Clinical Education Center, Quest diagnostics, 04/24/2020. Retrieved from: http://education.questdiagnostics.com/faq/FAQ219
Whitman, J et al. Test performance evaluation of SARS-CoV-2 serological assays, Retrieved 5/3/20, https://www.dropbox.com/s/cd1628cau09288a/SARS-CoV-2_Serology_Manuscript.pdf?dl=0
Lowry, R, Clinical Calculator 2, Predictive Values and Likelihood Ratios. Retrieved from: http://vassarstats.net/
Higgins-Dunn N et al., New York antibody study estimates 13.9% of residents have had the coronavirus, Gov. Cuomo says, CNBC Health and Science, April 23 2020. Retrieved from: https://www.cnbc.com/2020/04/23/new-york-antibody-study-estimates-13point9percent-of-residents-have-had-the-coronavirus-cuomo-says.html
Roche’s COVID-19 antibody test receives FDA Emergency Use Authorization and is available in markets accepting the CE mark. Retrieved from: https://www.roche.com/media/releases/med-cor-2020-05-03.htm
Yang Y, et al., Evaluating the accuracy of different respiratory specimens in the laboratory diagnosis and monitoring the viral shedding of 2019-nCoV infections, medRxiv, 2020. Retrieved from: medrxiv.org PDF
The author has no direct financial connection to Roche. Although unknown to the author, it is possible a mutual fund or ETF he owns may include some Roche stock. The Roche antibody test was chosen as an example because it has a high specificity and could be used to demonstrate the effect of an actual test’s specificity level on results.
ABOUT THE AUTHOR
Stuart M. Caplen, MD, FACEP, MSM
Dr. Caplen is a former emergency physician and emergency department medical director, now retired from clinical practice. His current interests include how quality is produced and maintained in health care, and he recently achieved greenbelt certification in lean/six sigma.
Try our medical search Compendium or Annotate your EMR, PDFs, and docs on our FibonacciMD.app.
DISCLAIMER
IMIT takes pride in its work, and the information published on the IMIT Platform is believed to be accurate and reliable. The IMIT Platform is provided strictly for informational purposes, and IMIT recommends that any medical, diagnostic, or other advice be obtained from a medical professional. Read full disclaimer.