key: cord-0847561-s9jzcohd authors: Jendrny, Paula; Twele, Friederike; Meller, Sebastian; Schulz, Claudia; von Köckritz-Blickwede, Maren; Osterhaus, Ab; Ebbers, Hans; Ebbers, Janek; Pilchová, Veronika; Pink, Isabell; Welte, Tobias; Manns, Michael Peter; Fathi, Anahita; Addo, Marylyn Martina; Ernst, Christiane; Schäfer, Wencke; Engels, Michael; Petrov, Anja; Marquart, Katharina; Schotte, Ulrich; Schalke, Esther; Volk, Holger Andreas title: Scent dog identification of SARS-CoV-2 infections, similar across different body fluids date: 2021-03-05 journal: bioRxiv DOI: 10.1101/2021.03.05.434038 sha: 46a7e5f435124467fd874d7f306edb3e8bde6e9a doc_id: 847561 cord_uid: s9jzcohd Background The main strategy to contain the current SARS-CoV-2 pandemic remains to implement a comprehensive testing, tracing and quarantining strategy until vaccination of the population is adequate. Methods Ten dogs were trained to detect SARS-CoV-2 infections in beta-propiolactone inactivated saliva samples. The subsequent cognitive transfer performance for the recognition of non-inactivated samples were tested on saliva, urine, and sweat in a randomised, double-blind controlled study. Results Dogs were tested on a total of 5242 randomised sample presentations. Dogs detected non-inactivated saliva samples with a diagnostic sensitivity of 84% and specificity of 95%. In a subsequent experiment to compare the scent recognition between the three non-inactivated body fluids, diagnostic sensitivity and specificity were 95% and 98% for urine, 91% and 94% for sweat, 82%, and 96% for saliva respectively. Conclusions The scent cognitive transfer performance between inactivated and non-inactivated samples as well as between different sample materials indicates that global, specific SARS-CoV-2-associated volatile compounds are released across different body secretions, independently from the patient’s symptoms. Funding The project was funded as a special research project of the German Armed Forces. The funding source DZIF-Fasttrack 1.921 provided us with means for biosampling. The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which causes COVID-68 19, infects the upper respiratory tract and in more serious cases may also cause severe 69 pneumonia and acute respiratory distress syndrome. The clinical presentation of SARS-CoV-2 70 infection is heterogeneous, ranging from asymptomatic infection to typical symptoms such as 71 fever, cough, fatigue, ageusia and anosmia, but may also present atypically and lead to 72 multiorgan dysfunction and death 1, 3, 4 . Containing this global pandemic requires a high rate of 73 testing, as an effective tool to contain viral spread. Viral loads can be detected by reverse 74 transcription polymerase chain reaction (RT-PCR) assays and with slightly less sensitive and 75 usually more rapid antigen detection tests in nasal or pharyngeal swabs 2,4 and saliva 5,6,7 with a 76 peak at days three to ten after infection. The peak of infectiousness is around symptom onset 8 . 77 It remains unclear if sweat or urine are also sources of virus transmission 9,10 . Odour detection 79 Different infectious diseases may cause specific odours by emanating volatile organic 80 compounds (VOCs). These are metabolic products, primarily produced by cell metabolism and 81 Page 5 of 20 released through breath, saliva, sweat, urine, faeces, skin emanations and blood 11 . The VOC-82 pattern reflects different metabolic states of an organism, so it could be used for medical 83 diagnosis by odour detection and disease outbreak containment 12 . 84 Canines are renowned for their extraordinary olfactory sense, being deployed as a reliable tool 85 for real-time, mobile detection of, e.g., explosives, drugs and may identify certain pathogen-86 and disease-specific VOCs produced by infected body cells. The limit of detection for canines 87 is at concentrations of one part per trillion, which is three orders of magnitude more sensitive 88 than currently available instruments 12 In contrast to our first study 24 , which only included hospitalised COVID-19 patients suffering 120 from severe courses of disease, we now additionally included non-hospitalised asymptomatic approved the study (ethic consent number 9042_BO_K_2020 and PV7298, respectively). To ensure safety for presentation of non-inactivated samples, glasses specially designed for Kynoscience UG, Hörstel, Germany) was utilised, which provided automated and randomised 144 sample presentations for the dogs as well as automatic rewards as described previously 24 . The 145 recorded results were verified by manual video analysis. Training procedure 147 The training procedure was exclusively based on positive reinforcement. Dogs were 148 familiarised to the DDTS for six days using a replacement odour, followed by specific training 149 for 8 days to condition them for the scent of SARS-CoV-2 infections in twelve inactivated 150 positive saliva samples and negative control samples from healthy individuals, respectively. The final study was conducted on four days (four hours a day) and included non-inactivated 152 saliva samples as well as urine and sweat samples. All of the samples used in the final study 153 had not been presented to the dogs before. Study design of the double-blinded study 155 The study was conducted in compliance with safety and hygiene regulations according to the 156 recommendations of the Robert Koch Institute (Berlin, Germany), and approved by local 157 authorities (regional health department and state inspectorate's office; Hannover, Germany). Page 8 of 20 All samples were handled by the same person with personal protective equipment including 159 powder-free nitrile gloves to prevent odour contamination which may irritate the dogs. In the 160 first session non-inactivated saliva samples were used to assess whether dogs were able to 161 transfer their trained sniffing performance from inactivated to non-inactivated saliva samples. In the following sessions, the detection performances for non-inactivated sweat, urine, and, CI: 66.67-100%) and 98% (95% CI: 94.87-100%) respectively (Fig. 1, suppl. table 4) . pivotal not only for limiting the spread of the current pandemic, but also for providing a tool to 217 limit the impact on public health and the ecomony. Data from the current scent dog detection 218 study confirm our former pilot study (sensitivity 84% versus 83% and specifity 95% versus 219 96%, respectively). In the current study, dogs were after only eight days of training not only 220 able to immediately transfer their scent detection abilities from inactivated to non-inactivated 221 saliva samples, but also to sweat and urine, with urine having the highest sensitivity of 95% and 222 specificity of 98%. These results suggest a general, non-cell specific, robust VOC-pattern 223 generation in SARS-CoV-2 infected individuals and provide further evidence that detection 224 dogs could provide a reliable screening method providing immediate results. In the former pilot study from our group 24 , only BPL-inactivated samples of COVID-19 patients 226 and controls were used. The first step in the current trial was therefore to evaluate if dogs can for SARS-CoV-2 infections. They require a diagnostic sensitivity of above 80% and specificity 291 above 97% 34 . The scent dog method would meet these criteria. The COVID-19 epidemic Viral Load in Upper Respiratory Specimens of Infected Patients Canine detection of the volatilome: A 396 review of implications for pathogen and disease detection Biomedical Scent Detection Dogs: Would They Pass as a Health 399 Technology? Diagnostic accuracy 401 of canine scent detection in early-and late-stage lung and breast cancers Trained dogs identify 404 people with malaria parasites by their odour Using dog scent 407 detection as a point-of-care tool to identify toxigenic clostridium difficile in stool Detecting Staphylococcus aureus in 410 milk from dairy cows using sniffer dogs How effective are trained dogs at alerting 413 their owners to changes in blood glycaemic levels?: Variations in performance of glycaemia 414 alert dogs Cellular scent of 416 influenza virus infection Can the detection 418 dog alert on COVID-19 positive persons by sniffing axillary sweat samples? A proof-of-419 concept study Dog Savior: Immediate Scent-422 Detection of SARS-COV-2 by Trained Dogs Protocol: Real-time RT-PCR assays for the detection of SARS-CoV-2. 425 pcr-assays-for-the-detection-of-sars-cov-2-institut-pasteur-paris.pdf?sfvrsn=3662fcb6_2 Accessed 31 A universal heterologous internal control system 429 for duplex real-time RT-PCR assays used in a detection system for pestiviruses Scent dog identification of samples from COVID-19 patients -A pilot study Values: Foundations, Pliabilities, and 435 Pitfalls in Research and Practice Interval estimation for the difference between independent proportions: 437 Comparison of eleven methods COVID-19 by analysis of breath with gas chromatography-ion mobility spectrometry -a 441 feasibility study Rapid detection of SARS-CoV-2 443 infection by multicapillary column coupled ion mobility spectrometry (MCC-IMS) of breath. A 444 proof of concept study False Negative Tests for SARS-CoV-2 Infection -Challenges and 447 False positives in reverse transcription PCR testing for SARS-CoV-2 323 The authors declare that they have no competing interests.