Canine olfaction as a disease detection technology: A systematic review
Introduction
The ability of dogs to detect health conditions via olfaction has been empirically investigated for the past three decades. The first case report of a dog alerting its owner of cancer, published in the Lancet in 1989, described a dog that persistently sniffed at a lesion on its owner’s leg, which was later diagnosed as melanoma (Williams and Pembroke, 1989). A second case report more than a decade later described a similar phenomenon (Church and Williams, 2001). Since then, numerous studies have examined dogs’ ability to detect disease. These have included cancers (reviewed in Edwards et al., 2017), bacterial infections (Bomers et al., 2012, Maurer et al., 2016), seizures (Catala et al., 2019), and humans infected with COVID-19 (Grandjean et al., 2020, Jendrny et al., 2020). Canine olfaction has also been demonstrated to be effective in identifying conditions or diseases affecting non-human animals, cattle and dogs, (Angle et al., 2016a, Dorman et al., 2017) and plants, avocado and citrus trees (Mendel et al., 2018, Gottwald et al., 2020).
Several reviews have previously summarized portions of this field but are relatively narrow in scope. Jezierski et al. (2015) and (Pomerantz et al., 2015) published reviews focused on the utilization of canine olfaction to detect cancer in humans. These works provided valuable insight into the methodological characteristics of detection studies, but excluded those in which dogs were trained to detect diseases other than cancer, or to detect disease in organisms other than humans. Another systematic review of olfactory disease detection by animals included detector animals other than dogs, but excluded patient species other than humans, and excluded some health conditions, such as seizures (Edwards et al., 2017). Hackner & Pleil (2017) focused primarily on offering recommendations for potential applications of canine olfactory disease detection. Other articles have reviewed selected studies of canine olfactory detection of disease, yet these reviews focused on developing technological alternatives to canine olfaction in detection of disease using volatile organic compounds (Pleil and Giese, 2017, Marzorati et al., 2019).
In the context of human medicine, the risk of canine misbehavior or inconsistency and human fear of, or allergies to, dogs may outweigh the benefits of canine olfaction as a rapid, mobile, low-tech screening technology. However, in both animal and plant agriculture, the capacity for individualized laboratory diagnosis of disease is lower than in human medicine (Gottwald et al., 2019, Larson et al., 2005). Dogs may be highly valuable for screening large numbers of animals and plants for conditions that impact animal welfare and can have economic implications, including bovine viral diarrhea virus (BVDV) in cattle and Liberibacter asiaticus in citrus (Angle et al., 2016a, Gottwald et al., 2019). These studies have been excluded from previous reviews of disease detection by canine olfaction. This review is therefore necessary both to include the several studies that have been published since 2017, and to offer a comprehensive collection of all studies pertaining to the olfactory detection of disease by canines.
In addition to varying in subject matter, prior studies have yielded variable results. Dogs’ ability to detect lung and breast cancers has been reported to have near-perfect sensitivity and specificity (McCulloch et al., 2006), with less accurate performance in others (Amundsen et al., 2014, Gordon et al., 2008), sometimes no better than chance. Because few studies include trials of disease detection dogs outside the laboratory (Bomers et al., 2012, Gottwald et al., 2020), there is a lack of information about the potential for dogs in real-world disease screening applications. The circumstances under which dogs can detect any given health condition by scent remain unclear, and the limits to what conditions dogs are able to detect diseases remain unknown.
In analyzing all available published studies of canine olfactory detection of health conditions, the objectives of this review were to characterize the subjects and methodology of such studies and to conduct a meta-analysis that evaluated to what extent dogs can serve as reliable detectors of disease in humans, other animals, and plants. Comparators were aspects of study design such as sample size, test apparatus, and number of samples presented during testing. Outcomes measured included detection accuracy, sensitivity, and specificity. This will provide a clearer view of the canine olfactory disease detection landscape, a strong foundation for future research in the field, and analysis of what training methods and testing paradigms result in the greatest rates of success in canine disease detection.
Section snippets
Methods
This systematic literature review was conducted via a broad, structured search of scientific databases.
Results
Searches of three databases yielded 3731 search results (Fig. 1). After 1618 duplicates were automatically removed and four additional duplicates were merged, 2109 studies moved to title and abstract screening. Of those, 159 studies moved on to full-text review, and 58 were included for analysis (Table 2). For some measures, because a single study might report testing dogs on multiple diseases, or under multiple sets of conditions, these studies were divided into 105 sub-studies.
Discussion
Most of the studies of canine olfactory detection of disease included in this review reported high levels of success, whether reported as accuracy or as sensitivity and specificity. While there are methodological concerns related to blinding and sample re-use in some studies, only blinding had an impact on detection success. Sub-studies without double-blind tests had greater mean specificity, which suggests that dogs may take cues from experimenters in tests that are not double-blind.
Conclusions
Canine olfaction is a promising method for screening individuals for diseases and health conditions. However, training and testing methods, results, and reporting strategies are highly variable across studies, limiting comparability. Further research assessing detection accuracy under realistic conditions is needed before canines can be widely deployed to screen for disease.
Funding
This work was supported by the Universities Federation for Animal Welfare Small Project, United Kingdom Grant #26–20/21 and the Texas A&M University USA Triads for Transformation Round 4.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
We would like to thank the student workers who contributed to this project: Gabriela Diaz, Nicole Fernandez, Lily Grajeda, Hana Henderson, Kenzy Hoffman, and Brooke Vandenbergh.
References (73)
- et al.
A detection dog to identify patients with Clostridium difficile infection during a hospital outbreak
J. Infect.
(2014) - et al.
Identifying environmental reservoirs of Clostridium difficile with a scent detection dog: preliminary evaluation
J. Hosp. Infect.
(2017) - et al.
Detecting Clostridioides (Clostridium) difficile using canine teams: what does the nose know?
Infect. Prev. Pract.
(2019) - et al.
Another sniffer dog for the clinic?
Lancet
(2001) - et al.
Olfactory detection of prostate cancer by dogs sniffing urine: a step forward in early diagnosis
Eur. Urol.
(2011) - et al.
Animal olfactory detection of human diseases: guidelines and systematic review
J. Vet. Behav.
(2017) - et al.
Detecting Staphylococcus aureus in milk from dairy cows using sniffer dogs
J. Dairy Sci.
(2018) - et al.
Trained dogs can identify malignant solitary pulmonary nodules in exhaled gas
Lung Cancer
(2019) - et al.
Canine scent detection—Fact or fiction?
Appl. Anim. Behav. Sci.
(2013) - et al.
Canine scent detection as a tool to distinguish meticillin-resistant Staphylococcus aureus
J. Hosp. Infect.
(2017)