COVID-19: Study Suggests that Sniffer Dogs May Detect SARS-CoV-2 in Humans

by John M. Simpson.

Researchers at the National Veterinary School in Alfort, France, recently made available an un-refereed pre-print of a proof-of-concept study that suggests that the axillary perspiration (i.e., underarm sweat) of humans infected with SARS-CoV-2 emits an odor that detection dogs can be trained to detect.  The work has not yet been peer-reviewed, but it offers some interesting results.

The researchers began with the premise that, due to their particularly acute sense of smell, dogs can be effectively trained to recognize certain substances, including diseases.  Thus, there are many examples wherein canine olfaction has been used to detect explosives, narcotics and currency.  In the medical field, there likewise are multiple reports of dogs being effective at detecting certain cancers, certain diabetic and epileptic events, certain disease parasites and certain bacteria.

The researchers collected axillary sweat samples from patients who had been shown positive by testing for COVID-19 and showed symptoms of the disease and from patients who were shown to be negative by testing for COVID-19 and who had no symptoms of the disease.  The dogs used in the study had all been trained in some form of detection and came from three groups:  explosives detection; disaster and rescue (which also is based on the sweat of the persons being searched for); and colon cancer detection.  Most of the dogs were Belgian Malinois Shepherds.

The dogs were trained to recognize the odor of the COVID-19 positive sweat which was determined to be different than the axillary sweat from negative individuals.  Some of the animals could not adapt to this information and were excluded, with the study narrowed to eight dogs.  Once the dogs were able to differentiate COVID-19 positive, from negative, axillary sweat, each of these dogs was asked to find the COVID-19-positive sweat sample which was placed in a box randomly located in a series of boxes containing negative samples or mocks.  Each dog was tasked with finding the positive sample over a number of individual trials, ranging from 15 to 68 trials.

The dogs’ success in recognizing the COVID-19 positive sample ranged from 85 to 100 percent correct identification.  The researchers calculated that the prospect that the dogs’ success rate was due to chance alone ranged from 14 to 33 percent.  The authors concluded:

The results of this first proof of concept study demonstrate that COVID-19 positive people produce axillary sweat that has a different odour, for the detection dog, than COVID-19 negative persons.  It strongly suggests the hypothesis according to which dogs can be trained to detect SARS-CoV-2 contaminating people, contained in axillary sweat which is easy and safe to collect.

One particularly interesting aspect of the study occurred when the dogs marked two of the samples positive that previously had been classified as negative.  The health institutions for those patients were notified and the patients were retested and actually found to be positive which indicated that the dogs were right.

If the proposition examined in this proof-of-concept study pans out, it could play a role in the detection of COVID-19.  Whether sniffer dogs would actually become part of any business’ return-to-work protocol remains to be seen.  However, as the authors noted, canine detection could be important in scenarios in which equipment and resources are not available for serological tests.