An early-stage study suggests dogs could one day be used to sniff out the coronavirus in a patient’s sweat.
Detection dogs were trained for one to three weeks to identify the infection in underarm swabs.
Scientists from the University of Paris-Est then exposed the animals to samples from 177 individuals, 95 of whom tested positive for the coronavirus and endured symptoms.
The remaining 82 participants swabbed negative for the infection.
Results reveal each dog correctly identified a coronavirus-positive sample between 76% and 100% of the time.
Although unclear, the coronavirus’ cellular action or replication in a patient’s sweat glands may result in the break down of compounds that dogs can sniff out.
Early in the coronavirus pandemic, the World Health Organization’s director-general Dr Tedros Adhanom Ghebreyesus urged countries to “test, test, test”.
Quick, reliable and widespread testing is a key measure to control the infectious outbreak, with some nations managing to stem the spread by identifying patients and having them isolate.
Research has long suggested dogs may be able to detect a specific odour given out by people with infections like malaria or even cancerous tumours.
In countries with a “lack of diagnostic tests to perform mass detection”, dogs may therefore offer a “rapid, reliable and cheap ‘tool’ to pre-screen people or perform rapid [coronavirus] checking”, the Paris scientists wrote in the journal PLOS One.
Watch: Can you catch coronavirus twice?
To learn more, the French scientists trained six detection dogs at testing sites in Paris and Beirut.
The 177 participants were recruited from five hospitals, with some receiving care for a non-coronavirus related condition.
One underarm sweat sample was collected per patient.
After training, each dog had to detect a coronavirus-positive sample that had been randomly placed behind one of several cones.
The remaining cones contained a coronavirus-negative or “mock” sample.
The dog’s handler was also unaware which cone housed which sample.
The results reveal each dog correctly identified a coronavirus-positive sample between 76% and 100% of the time.
This was calculated according to the number of correct identifications divided by the total experiments.
Filming each session enabled the scientists to better understand why some were unsuccessful, with the dogs reportedly being thrown off by a horse walking past the testing room and too much noise.
Overall, the scientists believe their results “provide some evidence detection dogs may be able to discriminate between sweat samples from symptomatic COVID-19 [the disease caused by the coronavirus] individuals and those from asymptomatic COVID-19 negative individuals”.
They stressed, however, this was just a “proof-of-concept” study, with further research being required.
The trial had some limitations, including certain coronavirus samples being used more than once.
This raises the risk a dog may memorise other odours coming off a positive swab, with research suggesting the animals can remember at least 10 scents.
The samples’ odours may have also varied according to the patient’s age or underlying health.
Dogs should therefore never be seen as a “perfect diagnostic method”, but rather a “complementary tool”, added the scientists.
This comes after a Spanish study suggested dog walkers were 78% more likely to catch the coronavirus during the first lockdown.
Watch: What is long COVID?