March 27, 2020

expert comment on the reliability of swab testing

There have been question from journalists around the reliability of swab testing.

Dr Andrew Preston, Reader in Microbial Pathogenesis at the University of Bath, said:

“The ‘swab test’ for COVID-19 detects the presence of the viral RNA (its genome) in a sample, and thus the presence of the virus. The test uses a version of PCR, RT-PCR. This technique is highly specific, i.e. it is excellent for detecting very specifically COVID-19 virus, so will produce very few false positives.

“However, it is expected that in terms of missing a positive specimen (a false negative), these tests will be very similar to the performance of very similar tests for diagnosing other viruses (e.g. SARS, influenzae). For these, a 10% false negative level is widely accepted.

“False negatives are most likely to arise due to a few factors.

“Sampling. Several different body sites can be used for sampling, including a nasal swab, an oropharyngeal swab and a lower respiratory tract site. For the nasal swab, it is recommended that the swab is inserted quite far, not simply into the entrance to the nose. Likewise, of the oropharyngeal swab it is recommended to sample the posterior pharynx, i.e. the back of the throat, not just around the inside of the cheeks. Lower respiratory tract samples are really limited to hospitalised patients. If virus is not picked up on the swab, the result will be negative. Thus, how effectively the swab is taken, and the amount of virus present at the sampling sites will determine whether the virus is detected from an infected person. There is little clear understanding of differences in the amount of virus among infected people, and what any differences might mean clinically. It is likely that those showing symptoms have a higher viral load than those who might be carrying the virus, but are symptom free. This has consequences for using the test to determine the level of infection among the population, and certainly in determining whether a healthcare worker poses a risk to others.

“Another important consideration is that the swab test is a snap-shot of the status of the person at the time the sample is taken. To use the virus test to determine whether people are safe in terms of transmitting the virus to others, they would need to have regular tests, as being negative on one day does not mean they will be negative several days later, particularly if working in an environment where the risk of transmission is high.

“There has been a huge debate over the role of mass testing for controlling the spread of the virus. Clearly it has a place in control measures. However, it is essential that the test kits used are validated in terms of their performance and that all those performing the tests are doing so properly. Recently, the Spanish government returned a batch of thousands of test kits as they were found to be ‘faulty’. Using the results from poorly performing test kits to decide whether someone should isolate or not could have severe consequences. Thus, it is not easy to quickly and massively ramp up testing capacity when it requires both high quality kits and properly trained staff to run the tests. Even then, the limitations of these tests need to be recognised, and the need to run regular tests if we want assurance that someone is truly negative, and that they remain so over a period of time.”

Dr Al Edwards, School of Pharmacy, University of Reading, said:

“The technique is one factor that can affect accuracy. However, it’s also important that the amount of virus will vary depending on the patient and the stage of infection. So even with perfect technique, you can have an infection but not test positive- for example very early during the incubation phase, or right at the end when you are fighting the virus off and most of the virus has gone.

“The actual lab test (RT-PCR) is incredibly sensitive and can typically detect 1 virus particle.

“There is an increasing amount of data published discussing accuracy, but in the real world there is always some degree of uncertainty about any diagnostic test result. This is why diagnostic test results must be used with care by healthcare professionals, and we must not over-interpret test results.”

All our previous output on this subject can be seen at this weblink: www.sciencemediacentre.org/covid-19/

Declared interests

None received.