How Scientists Are Training Dogs To Sniff Out COVID-19 Infection

Three security detection dogs at the airport.

yacobchuk / Getty Images

Key Takeaways

  • Scientists have had considerable success at training dogs to identify people with COVID-19.
  • The dogs can distinguish between bodily fluid samples taken from people with and without COVID-19.
  • People with COVID-19 smell different than people without COVID-19 due to changes in body odor caused by the activity of virus molecules.

Using nothing more than their powers of smell, dogs can find explosives and illegal drugs and even diagnose cancer. For some months now, they have been in the process of adding another skill to their repertoire: COVID-19 detection.

Dogs around the world have already demonstrated an impressive ability to detect COVID-19 with nothing more than a single whiff of a sample of bodily fluid. In a recent study, Dominique Grandjean, DVM, professor at the University Paris-Est’s National Veterinary School of Alfort in France, and colleagues found that the six participating dogs were able to detect COVID-19 with an accuracy rate higher than that of some widely available diagnostic tests. The December study was published in the journal PLOS One.

What This Means For You

Depending on the success of dog detection research, it’s possible you may be sniffed by a canine in the future as a form of COVID-19 detection. An airport in Finland is already implementing this method.

Body Odor Points To COVID-19 Infection

The dogs—experienced sniffers named Guess, Maika, Gun, Bella, Jacky, and Oslo—received one to three weeks of training prior to the start of testing. With the exception of Jacky, a Jack Russell terrier, they were all Belgian Malinoises, a common French working breed similar in appearance and personality to German shepherds. 

The testing sessions evaluated the dogs’ ability to distinguish between sweat samples collected from 95 symptomatic COVID-19-positive individuals and 82 asymptomatic COVID-19-negative individuals at sites in Paris, France, and Beirut, Lebanon.

Sweat, as Grandjean and colleagues wrote in the study, contains high levels of volatile organic compounds (VOCs), chemical signatures that can “convey important information about metabolic processes.” When SARS-CoV-2, the virus that causes COVID-19, enters the bloodstream, it induces the production of specific catabolites, chemical byproducts whose presence manifests as changes in body odor. We can’t pick up on these slight changes, but dogs—some of which have 300 million olfactory receptors in comparison to our six million—can.

In addition to sweat, other bodily fluids like breath, tears, urine, feces, saliva, and tracheobronchial secretions such as phlegm, contain VOCs, making them viable candidates for experimentation as well. 

At the University of Pennsylvania’s School of Veterinary Medicine, Cynthia M. Otto, DVM, professor of working dog sciences and sports medicine, is testing dogs with urine samples. At the University of Veterinary Medicine Hannover in Germany, Holger Andreas Volk, DVM, head of the department of small animal medicine and surgery, is utilizing saliva samples. These studies are currently still being conducted.

Grandjean says that he does not expect COVID-19 detection accuracy rates to vary between the type of fluid being used to detect the illness.

“If we consider what's been done by some other countries, the answer is probably not, but [for] now we do not know and [continue to] work on it," he says. "Elimination routes of specific catabolites might not be the same. For example, we excrete urea in urine but not in sweat."

Evaluating the Dogs

During the testing sessions, the dogs, accompanied by their handlers, were presented with three or four axillary (armpit) sweat samples, each stationed behind a cone. They were then tasked with correctly identifying the sample that had been sourced from a symptomatic COVID-19-positive individual.

Every lineup included one COVID-19 positive sample and one or more COVID-19 negative samples, according to the original study. The dogs were successful between 76% and 100% of the time. Bella and Jacky led the pack with 100% success, while Maika brought up the rear at 76%. 

The initial round of testing did not include sweat samples taken from asymptomatic COVID-19-positive individuals, who constitute a significant percentage of the disease population. However, Grandjean says, since publishing this study, subsequent rounds of testing have found that a lack of visible symptoms does not fool dogs.

“So, yes, it works on asymptomatic [people],” he says.

How Does Dog Detection Compare To Other Tests?

COVID-19 tests—which include saliva tests, antigen tests, and nasal, throat, and nasopharyngeal swab tests—are not created equal in terms of their precision or price.

RT-PCR tests, for example, are highly accurate but are expensive and time-consuming to run. Temperature checks, on the other hand, provide instant results but cannot distinguish between fevers caused by COVID-19 and those caused by other infectious agents. They are also unable to detect COVID-19 cases that don't cause fevers.

So how does dog detection stack up against these more orthodox methods of surveying for the virus? Pretty well, according to Grandjean.

“The results obtained by the numerous teams working now on the subject in different countries are quite similar, with sensitivities ranging between 85 to 99 p[er] 100 and specificities around 95,” he says. “So it is close to PCR but way over Ag [antigen] and saliva tests.” 

Dog detection as a method of diagnosing COVID-19 has several major benefits. According to Grandjean, dog detection is:

  • Easy and non-invasive
  • Provides an immediate answer
  • Inexpensive

However, it also has several major drawbacks. You “need to train [the] dog first, so [you] need sweat samples from positive people,” Grandjean says. Finally, of course, live animals are harder to care for than machines. RT-PCR testing equipment, after all, does not get bored, need bathroom breaks, or eats and drinks. 

But at least one major city has decided that the pros outweigh the cons: Helsinki, Finland. Upon arriving at the airport, international travelers are asked to pat their skin with a wipe, which is then presented to one of four waiting dogs: ET, Kossi, Miina, or Valo. If they bark, paw, or lie down, signaling that they have detected COVID-19, the traveler is asked to take a free PCR test to confirm the assessment. If early laboratory results are any indication, the dogs' accuracy rate hovers around 100%.

The information in this article is current as of the date listed, which means newer information may be available when you read this. For the most recent updates on COVID-19, visit our coronavirus news page.

5 Sources
Verywell Health uses only high-quality sources, including peer-reviewed studies, to support the facts within our articles. Read our editorial process to learn more about how we fact-check and keep our content accurate, reliable, and trustworthy.
  1. Grandjean D, Sarkis R, Lecoq-Julien C, Benard A, Roger V, et al. (2020) Can the detection dog alert on COVID-19 positive persons by sniffing axillary sweat samples? A proof-of-concept study. PLOS ONE 15(12): e0243122. doi:10.1371/journal.pone.0243122

  2. Phoenix Veterinary Center. How powerful is a dog’s nose?

  3. Else H. Can dogs smell COVID? Here’s what the science says. Nature.

  4. Jendrny P, Schulz C, Twele F, et al. Scent dog identification of samples from COVID-19 patients – a pilot studyBMC Infect Dis. 2020;20(1):536. doi:10.1186/s12879-020-05281-3

  5. Food and Drug Administration. Coronavirus disease 2019 testing basics.

By Caroline Tien
Caroline Tien is a journalist with degrees in English and biology. She has previously written for publications including Insider and Cancer Health.