‘Any breed could do it’: dogs might be a Covid tester’s best friend

It is simple and pain-free, could be used to test for coronavirus in care homes, airports and schools, and might just be more realistic than the UK government’s £100bn “Operation Moonshoot” mass screening plan. Its name? Fido.

Around the world – from the UK to Finland, Spain, Brazil, Lebanon and Australia – teams of researchers are training dogs to sniff out Covid-19. And some say the idea of training hundreds of thousands of canine noses to check for coronavirus is not as far-fetched as it may sound.

How do dogs do it? At Finland’s Helsinki airport, where four Covid-19 sniffer dogs have begun work in a state-funded pilot scheme, passengers dab their skin with a wipe, which is placed in a beaker next to others containing control scents. If the dog detects the virus – shown by yelping, pawing, or lying down – the passenger takes a free swab test to verify its verdict.

Speaking to the Guardian, scientists said any breed could in theory be trained – a process that takes between two and 10 weeks – raising the prospect of pet canines joining an army of Covid sniffers.

Prof Dominique Grandjean, of the national veterinary school of Alfort in France, who is leading a research team using bomb detection, cancer detection and search and rescue dogs, said the canines were not sniffing the virus itself but rather tell-tale volatile chemicals produced when the virus infects cells, and released by the body.

The chemicals should be produced whether or not an infected person has symptoms, and only if the virus is active – suggesting that unlike current lab techniques, dogs are unlikely to pick up “dead” virus, Grandjean said.

Results from Grandjean and his colleagues, which are yet to be peer-reviewed, show sweat samples from Covid patients were correctly identified by eight dogs at least 83% of the time, with some making a correct identification in 100% of the trials they underwent. The team say they have since validated their approach in three separate trials, although the results have yet to be published.

‘Close to 100% accuracy’: Helsinki airport uses sniffer dogs to detect Covid

Read more

Grandjean thinks the approach has potential to become widespread. “We can have one dog per retirement house that is trained and this dog would be able every single morning to check everybody, just by walking by,” he said. His team plans to work with a French organisation to provide Covid-sniffing dogs to care homes.

“Pet owners could have their dog trained in order to search for Covid, but not only for them,” he added. “If we had 10,000 dogs able to sniff for Covid, well, that means that every dog should be able to sniff 200-300 samples a day, so that means 2-3 million samples a day.”

He said it would be better to use samples from individuals rather than let dogs wander among crowds sniffing for Covid.

Another research project is under way in Germany, using saliva rather than sweat samples. In a pilot study using eight dogs and 1,012 samples, the animals correctly spotted Covid-positive samples 83% of the time on average, and correctly identified Covid-negative samples 96% of the time.

The lead author of the research, Prof Holger Volk of the University of Veterinary Medicine in Hanover, said the current “have you got it” Covid lab test correctly identified the virus was present about 75% of the time, and correctly ruled it out almost 100% of the time.

It took Volk’s team just two weeks to train their dogs – he said hunting breeds were best suited to the work – but the French team said any breed, including mongrels, could potentially be trained. Grandjean said it could take eight to 10 weeks to train a dog with no prior experience of scent detection.

Rowland Kao, a professor of veterinary epidemiology and data science at the University of Edinburgh, who is not involved in the work, said larger studies would be needed but the approach appeared to be simple, non-intrusive and “a very good addition to the surveillance ‘armoury’”.

Prof Ian Jones, a virologist at the University of Reading, was less optimistic, saying such efforts detract from the real challenges of mass testing. “All that dogs can detect is an odour difference,” he said. “For explosives and drugs and even chronic disease like MS, that is fine, but many viruses infect the same cells as Covid and lead to similar changes in metabolism – so the gas you exhale is the same.”

Volk said his team was working on whether the dogs can distinguish between different viruses, and Grandjean was upbeat. “Different types of virus have different volatile organic compounds coming from cell cultures, meaning [these] compounds are specific to each virus,” he said, although he noted this had yet to be proved for Sars-Cov-2.

Dr David Strain, a senior clinical lecturer at the University of Exeter medical school, said the canine approach was likely to be hugely beneficial.

Winning by a nose: the dogs being trained to detect signs of Covid-19

Read more

“If dogs can be appropriately trained, there is a high likelihood that they will have a higher success rate than the current screening strategies, given that they will be able to pick up the scent from wherever it emanates not just for those who have Covid in their upper airways,” he said. “Real time” screening could be particularly useful when infection levels fall.

“They could work in ports, harbours and airports to limit the risk of travellers returning with the infection,” he said. “Importantly to the economy, all of this can be performed at a fraction of the cost of the ‘Moonshot’ program, and are likely to be with us much sooner.”