Dogs can detect COVID-19 with 97% accuracy, study finds

pharmafile | May 21, 2021 | News story | Medical Communications  

A French trial has found has found that dogs are able to detect COVID-19 with 97% accuracy in humans, which means they are more effective than lateral flow tests.

The trial was conducted in March and April by France’s national veterinary school and the clinical research unit at the Necker-Cochin hospital in Paris.

This could mean a rise in the use of canines in diagnosing people with coronavirus, particularly for mass screening at airports, train stations, and live events.

The study, which was the first of its kind, also found that the dogs were 91% correct in identifying negative samples.

Lateral flow tests, meanwhile, correctly identify on average 72% of people infected with the virus who have symptoms, and 58% who do not, a recent review found.

The Paris hospital board said: “These results are scientific confirmation of dogs’ capacity to detect the olfactory signature of Covid-19.”

Professor Jean-Marc Tréluyer told Agence-France Presse said: “These are excellent results, comparable with those of a PCR test.”

Though these canine tests won’t replace polymerase chain reaction (PCR) tests, which are significantly more reliable than LFT tests, Tréluyer said: “They could help identify those people who should undergo a full viral test and – because the dogs’ response is so quick – facilitate mass testing.”

In the study, samples were collected through cotton pads pressed for two minutes under participants’ armpits from 335 people aged between 6 and 76, who presented themselves for PCR tests in Parisian testing centres.

The pads were then sealed in jars and given to at least two of the nine dogs used in the trial – none of whom came into contact with the volunteers – to be sniffed. The dogs’ handlers did not know in advance which samples were positive.

The dogs detected 97% of the 109 people whose PCR test subsequently proved positive, and 91% of those whose PCR test was negative.

Lilly Subbotin

Related Content

No items found

Latest content