John Murphy, MDLinx | December 19, 2017
Dogs have been known to identify cancer, diabetes, and even epileptic seizures. But are these stories scientifically valid? Can man’s best friend truly detect disease? And if they can, how do they do it?
It’s not entirely clear, but it’s apparently the dog’s nose that knows how to sniff out disease—which leads to yet another question: What does cancer (or hyperglycemia or an epileptic seizure) smell like?
To understand that, one must understand the dog’s remarkable ability to smell. Dogs have up to 300 million olfactory receptors (compared to just 6 million in humans), and can smell 10,000 to 100,000 times more acutely than humans.
“We’ve always known that dogs have an incredible sense of smell. I like to say that they ‘smell in color,’” said Cindy Otto, DVM, PhD, associate professor, University of Pennsylvania School of Veterinary Medicine, and executive director of the Penn Vet Working Dog Center, Philadelphia, PA. “They really see the world through their noses.”
Dr. Otto and her team are studying how dogs use their highly developed sense of smell to detect ovarian cancer. “We know that there is an odor signature—a unique chemical ‘fingerprint,’ if you will—of malignant ovarian cancer, so we are all working backwards to determine what that is.”
Once the researchers determine the signature smell of malignant ovarian cancer, they hope to put the dogs out of work and replace them with more reliable cancer-detecting machines. “What we’re trying to do with this study is harness that remarkable ability to inform and create an electronic system for detecting early-stage ovarian cancer,” Dr. Otto explained.
In her study, dogs are rewarded for correctly identifying blood samples from ovarian cancer patients. A well-publicized trial in 2006 tested dogs’ ability to correctly distinguish breath samples of lung or breast cancer patients from those of healthy subjects. The dogs were able to sniff out lung cancer with 99% sensitivity and 99% specificity and breast cancer with 88% sensitivity and 98% specificity.
But when dogs were put to the test in a screening situation, not just a find-the-smell lab experiment, their cancer-sniffing reliability dropped drastically. As shown in a 2016 study, dogs’ ability to detect lung cancer from breath samples had a positive predictive value of only 30.9% and a negative predictive value of 84%.
“This disparity is not likely to be a detection issue; dogs have been shown to have extremely sensitive noses as proven by their use in tracking, bomb detection, and search and rescue,” said the study’s first author Dr. Klaus Hackner, Krems University Hospital, Krems, Austria. “However, in contrast to analytical instruments, dogs are subject to boredom, limited attention span, fatigue, hunger, and external distraction.”
Some researchers worry that these limitations may prevent dogs from detecting disease in real-world situations. The canine’s limitations are a particular concern for patients with epilepsy, a condition in which there’s ample anecdotal evidence of dogs detecting seizures, but nearly no solid scientific studies.
“There may be much hand-wringing over the lack of scientific evidence that dogs can alert to epileptic seizures; but this does not necessarily mean that they can’t. It could simply mean we haven’t done the right experiments,” wrote Adrian Burton in a 2017 editorial in The Lancet Neurology.
But until those experiments take place, the evidence for dogs’ seizure-alerting ability must be considered weak, the editorial noted.
Dogs are also used as four-footed continuous blood glucose monitors (CGM) to warn patients with type 1 diabetes of impending hypo- or hyperglycemic attacks. An upcoming study found that diabetes-alert dogs detected extremes in blood glucose more often than not, with 57% total sensitivity. Dogs’ sensitivity to detecting low blood glucose was higher (59.2%) than their sensitivity for high blood sugar (56.1%). However, abilities differed widely among dogs.
Indeed, electronic CGMs perform better and more reliably than the furry variety. But even if dogs’ ability to detect hypo- or hyperglycemia is imperfect, it’s still pretty amazing. How do dogs do it?
They smell the chemical “fingerprint” that Dr. Otto described. More precisely, they smell volatile organic compounds (VOCs) in humans’ exhaled breath.
In a recent study, researchers sought to identify VOCs characteristic of hypoglycemia by using not a Lab, but a mass spectrometer in a lab. The researchers artificially lowered the blood glucose of subjects with type 1 diabetes, and then compared their “before” and “after” breath samples. One chemical stood out: isoprene.
“Humans aren’t sensitive to the presence of isoprene, but dogs with their incredible sense of smell find it easy to identify and can be trained to alert their owners about dangerously low blood sugar levels,” said the study’s corresponding author Mark L. Evans, MD, FRCP, Institute of Metabolic Science, University of Cambridge, Cambridge, England.
Dr. Evans and colleagues hope that their research will do more than replace dogs as the detectors of diabetes. “It’s our vision that a new breath test could at least partly—but ideally completely—replace the current finger-prick test, which is inconvenient and painful for patients, and relatively expensive to administer.”
Diagnostic devices may detect disease better than dogs, but there’s one thing that machines can’t do: Make patients feel better.
In all three categories—cancer, epilepsy, and diabetes—dogs led to improvements in patients’ conditions. People with epilepsy who have seizure-detecting dogs experience up to 40% fewer seizures. In cancer patients, dogs improve feelings of social and emotional (though not physical) well-being. Also, patients with diabetes report fewer diabetic ketoacidosis hospitalizations, severe hypoglycemic episodes, and driving accidents when they have diabetes-alert dogs.
“These potential benefits should not be readily discounted and deserve more research,” the diabetes investigators stated.