Background Sniffer dogs are able to detect certain chemical particles and are suggest to be capable of helping diagnose some medical conditions and complications, such as colorectal cancer, melanoma, bladder cancer, and even critical states such as hypoglycemia in diabetic patients. With the global spread of COVID-19 throughout the world and the need to have a real-time screening of the population, especially in crowded places, this study aimed to investigate the applicability of sniffer dogs to carry out such a task. Methods Firstly, three male and female dogs from German shepherd (Saray), German black (Kuzhi) and Labrador (Marco) breeds had been intensively trained throughout the classical conditioning method for 7 weeks. They were introduced to human specimens obtained from the throat and pharyngeal secretions of participants who were already reported positive or negative for SARS-COV-2 infection be RT-PCR. Each dog underwent the conditioning process for almost 1000 times. In the meantime another similar condition process was conducted on clothes and masks of COVID-19 patient using another three male and female dogs from Labrador (Lexi), Border gypsy (Sami), and Golden retriever (Zhico) breeds. In verification test for the first three dogs, 80 pharyngeal secretion samples consisting of 26 positive and 54 negative samples from different medical centers who underwent RT-PCR test were in a single-blind method. In the second verification test for the other three dogs, masks and clothes of 50 RT-PCR positive and 70 RT-PCR negative cases from different medical center were used. Results In verification test using pharyngeal secretion, the sniffer dogs’ detection capability was associated with a 65% of sensitivity and 89% of specificity and they amanged to identify 17 out of the 26 positive and 48 out of the 54 true negative samples. In the next verification test using patients’ face masks and clothes, 43 out of the 50 positive samples were correctly identified by the dogs. Moreover, out of the 70 negative samples, 65 samples were correctly found to be negative. The sensitivity of this test was as high as 86% and its specificity was 92.9%. In addition, the positive and negative predictive values were 89.6 and 90.3%, respectively. Conclusion Dogs are capable of being trained to identify COVID-19 cases by sniffing their odour, so they can be used as a reliable tool in limited screening.

Background: Sniffer dogs have been shown to detect certain chemical particles and to help diagnose many diseases and complications, such as colorectal cancer, melanoma, bladder cancer, and even dangerous states such as hypoglycemia in DM-1 patients. With the spread of COVID-19 throughout the world and the need to have a real-time screening of the population, especially in crowded places, this study aimed to investigate the applicability of these sniffer dogs. Methods: This study was done in two phases. In the rst step, three dogs (including one German shepherd, one German black, one Labrador) were intensively trained by the classical conditioning method for seven weeks. Human specimens were obtained from the throat culture and pharyngeal secretions of both de nitely positive and negative COVID-19 participants. During the rst seven weeks, each dog underwent the conditioning process for averagely 1000 times. In the veri cation process, 80 pharyngeal secretion samples consisting of 26 positive samples of hospitalized patients and 54 negative samples of hospitalized patients for other medical reasons were provided to the training team in a single-blind manner. The veri cation test was done using three dogs (Lexi, Sami, and Kozhi). Another similar conditioning process was done using COVID-19 patients' clothes and masks and the veri cation test was done using 50 positive and 70 negative samples by three other trained dogs (one Labrador, one Border gypsy, and one Golden retriever). Results: In the pharyngeal secretions veri cation test, the sniffer dogs showed 65% sensitivity and 89% speci city. In this process, they could identify 17 out of the 26 true positive samples and 48 out of the 54 true negative samples. According to the results, the positive predictive value and negative predictive value for this experiment were 74% and 84%, respectively. In the next veri cation test for the patients' face masks and clothes, 43 out of the 50 positive samples were correctly identi ed by the dogs. Moreover, out of the 70 negative samples, 65 samples were correctly found to be negative by the canines. The sensitivity of this method was as high as 86% and its speci city was 92.9%. In addition, the positive and negative predictive values were 89.6% and 90.3%, respectively. Conclusion: Dogs are capable of being trained as the identi er of people with COVID-19 by detecting their odor and can be used as a reliable tool in limited screening. body result in a set of certain odors that may be detectable for animals, such as dogs, as they have an extreme olfactory sensation and a very complex and unique analytical mechanism for detecting odors at the molecular level (25). In such a scenario, COVID-19 may not be an exception. At present, the diagnosis of this disease is based on sophisticated biochemical and genetic tests and CT-scan imaging. These modalities are relatively expensive and require accurate medical equipment and interpretation, which are usually positive during the initiation of symptoms. Thus, the golden time may be lost to ...