How do swab tests work?

As of early morning 2nd July, nearly 8 million tests have been processed in the UK. The vast majority of those were molecular tests, so-called swab tests, which can reveal if you’re currently infected with coronavirus. A swab is taken from the nose or throat, and the test looks for specific genetic material that indicates the virus is there. Ravi Gupta is a physician at Cambridge University and he’s been working on speeding up tests like this. He spoke to Katie Haylor about how they work...

Ravi - It's based around trying to pick up or detect small amounts of the genetic material of the virus, which of course is present during an infection. The method relies really on making large numbers of copies of what's already there because of course what's already there can be very small and we need to make the signal very large in order to detect it. And so that's called nucleic acid amplification, and that's exactly how most tests are developed and function. So these can be fairly complex or they can be more simple. And we've been working on a simplified version that instead of taking three to five hours to run, we'll be able to give your results in about 75 minutes. This particular platform is called Samba two. The reason it's very successful in terms of the technology and how quickly we've been able to implement it is because this is a machine that was really developed for monitoring HIV in Africa, and to give a robust, quick result in a very sort of sturdy platform. And it's been tested and tried over many, many years. And so what the inventors did is to adapt it for SARS coronavirus two, which is the virus that causes COVID-19. And so the new chemistry was developed very rapidly and we had the opportunity to test it at Addenbrooke's hospital in around 200 patients in a trial that showed that the test was equally as good as the existing test in the lab.

Katie - And what is this new chemistry that enables it to be so much faster?

Ravi - So the new chemistry is based around the fact that traditionally we use something called polymerase chain reaction, and that basically needs temperatures to go up and down in a very precise way. And that takes time and we have these things called cycles involved. And the new chemistry really does something somewhat similar except performs things at the same temperature. And it's what we call isothermal. And that makes it much quicker and simpler to run.

Katie - How good are these tests? How worried do we need to be about false results, false negatives, false positives, stuff like that?

Ravi - In the trial that we ran, where we put the Samba Two rapid test against the hospital laboratory test developed by Public Health England, our test was 98 to 99% sensitive. In other words, picked up 99 out of every 100 positive tests using the standard laboratory assay. And it didn't get anything wrong the other way in that it didn't call anything positive that wasn't actually positive. And that's very important because you don't want to be telling people to have COVID when they do not in fact have it.

Katie - So 98 to 99%, is that good enough? Can it be made better? Can it be made faster?

Ravi - Well, there's always a trade off here. The reason it takes 75 minutes is because you still have to amplify the stuff that's there in the first place. And that takes time. So if you're going to develop a highly sensitive test, you need to have time for amplification to occur. If you just use a standard antigen test, which is just looking for protein that cannot be amplified, your tests are much less sensitive in general. And so there is a, there is a trade off here and a cost to having accuracy. And that is time. And I think an hour is actually pretty good in terms of getting a test result. The biggest delays sometimes actually are getting the swab from the nose and throat and getting the sample to the machine. If we can cut that down, for example, using saliva, I think things could get even faster.

Katie - Why is speed so important when it comes to testing?

Ravi - Well, speed is of the essence because when a patient comes into hospital, we want to prevent any transmission within the hospital. And so we have to pick out those people with symptoms who may or may not have COVID. Of course, the problem is that the symptoms are very vague, ranging from runny nose to cough, to headaches, to problems with your sense of smell. So it's very difficult to tease out who has COVID and who doesn't. And therefore we need to have a rapid test that can give us that answer quickly because the clinical diagnosis is not particularly accurate and we need to isolate infected patients rapidly to prevent spread.

Katie - Ravi, it's July at the minute. But looking ahead to winter of 2020, there's all kinds of seasonal coughs, colds, bugs that come along with that time of year. How does this relate to testing for COVID? Does winter pose a threat?

Ravi - Yes, we're all somewhat concerned about winter here in the UK, because as we can see COVID has not gone away and we are relaxing certain measures, and therefore we expect the number of cases to rise. And this rise may indeed end up coinciding with the seasonal influenza epidemic. And it's probably going to be a smaller one this year because of social distancing worldwide, but it will still probably arrive. And of course, other cold viruses that circulate in winter. And the key thing here is that we need to design tests or have tests available that can test for multiple different viruses at the same time in order to differentiate patients and to treat them appropriately in terms of isolation. Of course, the most important virus of those three or four is going to be coronavirus even going into winter.

Katie - But we don't have those tests right now is that, is that right?

Ravi - So a number of individuals and companies are looking at things called multiplex tests, which are rapid tests often that can detect a number of different viruses at the same time. We know that the Samba machine is being adapted to do a number of different pathogens at the same time, but will require validation as we did for the single test in order to make sure that it's as good as it needs to be. And the same goes for all the commercial tests that are being developed that go on to much larger machines and can run much larger numbers of samples. Those will also need to be tested. And this is a bit of a challenge because of course, we're going into a period where we have declining numbers of cases and actually evaluating the tests in real patient samples becomes more difficult as the number of cases in hospitals decreases.