A recent article by Dr E Vardas and colleagues published in Daily Maverick on 8 July 2020 purports that the use of serology assay to diagnose Covid-19 could redress challenges being faced with diagnosing Covid-19 cases and could assist in better managing the outbreak being experienced.
The article correctly points out that, “Laboratory tests are essential for the control of this pandemic, contributing to case identification, isolation, contact tracing, and rationalisation of infection control measures”. Notably, for testing to play a role in addressing any of these objectives requires the testing to identify current and not past infections. For laboratory tests to be useful, it needs to be designed for purpose.
While antibody-based serology assays may have a role in identifying past infections, it has zero role in identifying individuals that are infectious (current infection).
The evidence is clear that the antibody responses (IgG and IgM) following onset of symptoms is positive in 80% of individuals only approximately eight days after symptom onset in those with Covid-19. However, people with Covid-19 are most infectious in the pre-symptomatic phase (two to three days before the onset of symptoms) and for approximately 5-7 days after the onset of symptoms. Beyond eight days, very few cases with Covid-19 shed live virus, even if the molecular detection test (PCR) remains positive.
Consequently, alluding that there may be a role of an antibody-based serology assay to identify infectious cases, which is required for determining who needs to be isolated and which contacts need to be traced, or to decide on management of suspected Covid-19 cases, is scientifically unsound.
Also, serology assays will not redress inherent limitations of the sensitivity of a PCR (polymerase chain reaction) assay, as it is not a substitute (and has no role) for identifying infectious cases. Furthermore, deploying serology tests will not address the challenges faced with turnaround time of doing PCR testing laboratories, as the purpose of undertaking serology testing differs from the reasons why PCR testing is done. Also, a serology assay in a person with active infection would have a high “false-negativity” rate at a time when it is critical to identify whether a person is infectious or not. This is critical for triaging and management of patients hospitalised for Covid-19, as well as when advising on isolation of infectious cases and tracing their contacts to be quarantined.
Antibody-based serology assays may, however, have a possible role in characterisation of the population prevalence of past infections and possibly assist in geospatial mapping of hotspots that existed at least two weeks previously.
While serology may have some utility in identifying past infections, the question is whether this is of any value at an individual level. Serological evidence of past infection may allude to underlying immunity; however, at this stage the evidence for this is tenuous. In fact, it’s uncertain whether natural infection by SARS-CoV-2 induces long-lasting immunity at all (which coincidentally the serology assays proposed for use do not measure in any case); and more so in the majority of infected individuals that are asymptomatic or who only develop mild illness.
Also, there is rapid waning of antibodies following natural infection over a period of two to three months after developing Covid-19, and probably even sooner (within a month) in those infected with SARS-CoV-2 who were asymptomatic or had mild illness. Consequently, the sensitivity of serology tests in identifying even past infections is likely to be compromised, if infections occurred more than three months ago.
As such, an antibody-based serology test, even if positive, is unlikely to provide any information that would guide the management of the majority of people hospitalised with Covid-19; nor does it have any value in defining whether a person should be in isolation, or if their contacts (which would have occurred sometime back than when the assay is positive) should be in quarantine. At an individual level, it provides little to no value other than making a hole in the pocket of already financially distressed citizens.
Antibody-based serology assays may, however, have a possible role in characterisation of the population prevalence of past infections and possibly assist in geospatial mapping of hotspots that existed at least two weeks previously. However, then too, it is limited in light of the rapid natural waning of antibodies that would be expected to result in lower sensitivity to diagnosing infections that may have occurred a few months ago. Such sero-epidemiological studies need to be done systematically at a population level as part of coordinated surveillance.
In summary, diagnosing SARS-CoV-2 active infection does not fall within the same paradigm of doing an antibody test for HIV, as the implications of a positive result are vastly different. For SARS-CoV-2, all that it tells you, with limited certainty, is that you previously knowingly or unknowingly had been infected by the virus, but in all likelihood you are no longer infectious (and consequently it is very much of academic interest). On the other hand, a positive serology test for HIV tells you that you have a virus in the body and need to get treatment. Obfuscating the role of Covid-19 serology assays in relation to identification of current or past infection is dangerously misleading to the public. DM/MC
Shabir Madhi is Professor of Vaccinology, University of the Witwatersrand and Director of the South African Medical Research Council: Vaccines and Infectious Diseases Analytical Research Unit.