Accuracy of three COVID-sniffing dogs exceeds 85%

A team of scientists from Italy and Brazil has recently conducted a training program for sniffer dogs to detect coronavirus disease 2019 (COVID-19). They have trained the dogs to accurately sense COVID-19 specific volatile organic compounds in sweat samples collected from COVID-19 patients. The study is currently available on the medRxiv* preprint server.

Background

Because of extraordinarily high olfactory sensation, dogs are often used by law enforcement to detect narcotics, explosives, dead bodies, etc. In addition, sniffer dogs are used in medical fields to diagnose diseases. The olfactory system of dogs exhibits some unique characteristics, including continuous regeneration of olfactory sensory neurons and the presence of about 200 million olfactory receptors. These make the olfactory system of dogs highly sensitive to odors.   

Regarding diagnostic precision, studies have shown that sniffer dogs are able to diagnose breast and lung cancers with up to 99% accuracy. Regarding viral and bacterial infections, sniffer dogs exhibit a detection range of 77% to 93%.

Several volatile organic compounds are produced in the human body in response to inflammation, infection, or neoplastic transformation.  With adequate training, sniffer dogs can sense these volatile organic compounds that are released from the body via expiration, perspiration, urination, or salivation.

In the current study, the scientists have trained sniffer dogs to detect volatile organic compounds that are released from COVID-19 patients via perspiration.

Study design

One male dog and two female dogs from different breeds, including Black German Shepherd, German Shepherd and Dutch Shepherd, were used. Their training was conducted in two steps. In the first step, specific conditioning to COVID-related volatile organic compounds was performed using sweat samples from hospitalized COVID-19 patients. In the second step, the dogs were trained for “olfactory discrimination” so that they can differentiate between COVID-specific and non-specific volatile organic compoundss.

For the training, sweat samples of COVID-positive and COVID-negative individuals were collected on sterile cotton gauze, which was placed in glass jars with a metal top. After removing the metal top, the glass jars with sweat samples were placed in metal boxes, which were then randomly positioned in front of the dogs for discrimination training. Each training session lasted for about 2 min 30 sec.

Important observations

Overall, the entire training program was conducted for four weeks, including 227 sessions and 700 tests with 92 different sweat samples. The specific aim of the training was to identify the time frame (“switch” moment) when the dogs became trained to identify COVID-specific volatile organic compounds accurately. The accuracy level was fixed at 80% to make it comparable to the gold standard COVID-19 diagnostic methods, such as RT-PCR and rapid antigen test.

For each training session, the scientists assigned a coefficient of difficulty directly proportional to the number of boxes in front of the dogs.

With further analysis, the scientists noticed that three trained dogs identified the sweat samples collected from COVID-19 positive patients with 85%, 87%, and 88% accuracy, respectively. Moreover, they observed a statistically significant difference between the percentage of correct and incorrect identifications. This indicates that each dog reached the “switch” moment after adequate training.  

   

Study significance

The study provides a promising method for large-scale COVID-19 diagnosis using trained sniffer dogs. The sniffer dogs trained in the study exhibit more than 80% accuracy in detecting sweat samples collected from COVID-19 patients.

The scientists believe that the trained dogs can be used for mass detection of COVID-19, particularly at airports, stadiums, or any large-scale events.

Future research could include the detection of volatile compounds emanating from human skin, comparing sniffer dogs against the RT-PCR gold standard test for COVID-19, and using SARS-CoV-2 proteins to train dogs to directly target viral particles instead of volatile organic compounds.

*Important Notice

medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.

Written by

Dr. Sanchari Sinha Dutta

Dr. Sanchari Sinha Dutta is a science communicator who believes in spreading the power of science in every corner of the world. She has a Bachelor of Science (B.Sc.) degree and a Master's of Science (M.Sc.) in biology and human physiology. Following her Master's degree, Sanchari went on to study a Ph.D. in human physiology. She has authored more than 10 original research articles, all of which have been published in world renowned international journals.