BMJ Global Health reports that dogs can reliably detect COVID-19 infection from skin smears.
Prompt and accurate identification and isolation of patients with COVID-19 infection is of great importance. Currently, the diagnosis of coronavirus infection is based on an accurate and sensitive PCR test. However, PCR tests, among other reasons, are not suitable for the examination of large masses due to their long delivery time and high cost.
A recent study by the University of Helsinki and Helsinki University Hospital (DOI 10.1136 / bmjgh-2021-008024) confirmed that odor-detecting dogs can be trained to recognize people infected with the SARS-CoV-2 coronavirus with 92% accuracy. leather sticks ..
Researchers at the University of Helsinki and the Department of Veterinary Medicine at Helsinki University Hospital have developed a joint triple-blind, randomized, controlled study to test the accuracy of dogs trained to detect the characteristic odor of infection.
Neither the dog handler nor the researcher knew which of the smelly skin sticks was positive or negative. The study also looked at factors that could potentially interfere with a dog’s ability to recognize a positive pattern.
In the first phase of the study, dogs were trained to distinguish skin rods from voluntary samples from coronavirus patients who tested negative. After several weeks of training, the dogs moved from the training center to Helsinki-Vantaa Airport, where the next experiments were conducted.
In the second phase of the study, four trained dogs underwent a validation test to prove their ability to differentiate between samples. During the experiment, each dog was presented with a series of 420 samples over seven days. As several parallel samples were collected from each donor, each dog sniffed the same 114 sets of samples from coronavirus patients and 306 controls. The presence or absence of coronavirus in all donor samples was confirmed by PCR.
Sensitivity to coronavirus detection was 92% and specificity was 91%. Only small differences in accuracy were observed between the four dogs.
Coronavirus variants have greatly contributed to the misidentification of samples by dogs.
The study confirms previous reports that dogs can use their sense of smell to identify people infected with the coronavirus.
The third phase of the study included a real-time inspection of passengers and staff at Helsinki-Vantaa Airport. Dogs correctly identified 98.7 percent. negative examples. In fact, the small number of coronavirus-infected specimens in the experiments made it impossible to accurately assess the results in dogs with positive specimens. However, in this part of the study, the results of correctly identified positive samples based on positive “work motivation patterns” given to dogs on a regular basis were also assessed at 98.7%. Motivation samples were originally collected from PCR-positive patients but had not previously been sniffed by dogs. In situations and environments where the percentage of positive samples is very low, they are given to dogs on a regular basis to maintain interest in the target odor.
“Our research structure was highly scientific. The sample sizes were large enough and all the dogs sniffed the same set of samples, which allows us to compare their results. Dogs also had to successfully demonstrate sets of samples containing only negative samples – that’s important. Helsinki Anna Hielm-Björkman, head of the DogRisk research team at the university and an associate professor, says collecting samples from outpatients is more important than inpatients. In addition, the tests were performed in real conditions, not in a laboratory.
“I was particularly impressed by the fact that dogs performed less well with samples from patients suffering from coronavirus disease. The explanation is simple: dogs were initially trained with a wild-type virus and therefore did not always report the disease. “It shows the ability to vote,” said Anu Kantele, a professor of infectious diseases and chief physician at the University of Helsinki and Helsinki University Hospital.