Request from Kelleigh Nelson, Tennessee USA.

Gibber! Gibber!

Chugley

Please post and spread far and wide.

Six Diseases Your Dog Can Smell

Inside the Canine Mind, LLC

Sep 29, 2026

How six kinds of medical dogs changed the way we detect, manage, and live with illness.

In 1989, a woman in London went to a dermatology clinic because of her dog. For months, he had sniffed one mole on the back of her thigh and ignored every other mole on her body. One day, while she wore shorts, he tried to bite it off. The doctors removed the mole and tested it. It was malignant melanoma, the deadliest form of skin cancer.

Nobody had taught him what cancer was. He had no training. He noticed that one spot on his owner smelled different. He kept coming back to it until he got her attention. Two doctors wrote about the case in a short letter to The Lancet, one of the world’s oldest medical journals. That letter led scientists to ask a new question. Can a dog smell disease?

The answer is yes. A dog can detect lung cancer in a person’s breath and Parkinson’s disease in the oil on a person’s skin. He can register a drop in blood sugar, the odor of a seizure, and the chemistry of human stress. He needs no lab, no machine, and no blood. All he needs is his nose and training.

Disease changes your body’s chemistry. As your cells change, they release different compounds. Those compounds leave your body through your breath, sweat, skin oil, and urine. Your dog has up to 300 million scent receptors, compared to your 6 million. He can detect some odors at concentrations as low as a few parts per trillion. He can smell changes in your body that almost no person can detect.

I have spent my career training dogs to find one missing person by that person’s scent. I know what a trained nose can do, and these medical dogs still amaze me. Today, dogs work in six areas of medical detection: cancer, Parkinson’s disease, diabetes, epilepsy, PTSD, and narcolepsy. With cancer, diabetes, and epilepsy, dogs reacted to changes in their owners before scientists knew these diseases had an odor. Researchers then set out to prove what the dogs were detecting.

These dogs do two very different kinds of work. Some work in a lab. They sniff rows of samples, such as breath in a tube or a skin swab, and signal the one that carries the disease. Others live with their owner and go everywhere together, ready to alert and jump into action the moment their owner’s body changes.

Every medical dog is measured the same way: how often he is right. The lab dogs score the highest. Lung cancer dogs have been right up to 99 percent of the time. But look at what those dogs are working with. A sample in a tube is not a person. In a way, it is stagnant. It does not eat, sweat, sleep, or get stressed. It smells the same from one minute to the next. A living person changes all day long. A dog at home has to catch one small change in a scent that never stops changing. When researchers checked diabetic alert dogs against glucose monitors, the dogs caught about a third of low blood sugar events. In another study, the best dogs caught more than 90 percent. Much of the difference was not the dogs.

Start with the lab dogs. The London dog inspired the first cancer studies. One of the best-known studies came in 2006. Researchers in California trained five ordinary household dogs to sniff breath samples in tubes. The dogs correctly identified 99 percent of the lung cancer samples. Since then, dogs in labs around the world have detected ovarian, prostate, and bladder cancer. Their work proved that cancer has an odor. That proof sent scientists looking for the exact compounds the dogs detect. Engineers now use those findings to develop breath tests for cancer.

A woman in Scotland, not a dog, brought the odor of Parkinson’s disease to scientists. Joy Milne, a retired nurse, noticed a new musky odor on her husband, Les. She first noticed it about 12 years before doctors diagnosed him with Parkinson’s. Researchers tested her by giving her T-shirts worn by people with and without the disease. She correctly identified every Parkinson’s shirt. She also flagged one shirt from the healthy group, and the researchers counted it as her only mistake. Within a year, the man who wore that shirt learned he had Parkinson’s. Scientists traced the odor to sebum, the natural oil on human skin.

In 2025, researchers in the United Kingdom trained dogs on skin swabs to find that same odor. The dogs picked out most of the Parkinson’s samples. This matters because doctors diagnose Parkinson’s mostly by its symptoms, such as tremors and stiffness. By the time those symptoms appear, the disease has already damaged the brain. An odor test could help diagnose patients years earlier.

Narcolepsy is the newest question for the lab dogs. In 2013, researchers in Spain trained a dog to identify people with narcolepsy from their body odor alone. This single study suggests the disease also has an odor.

The lab dogs show scientists which diseases carry an odor. Scientists now use those odors to develop breath and skin tests that could find these diseases earlier. Dogs who live beside their people do a harder job. They work in real time. In mantrailing, I ask a dog to pick one person’s scent out of every other scent on the trail. A medical alert dog does a related job with one person. He knows his person’s normal scent, and he detects the moment that scent changes.

Diabetic alert dogs started with a man who understood both the science and the dogs. Mark Ruefenacht had type 1 diabetes. He worked as a forensic metrologist, a scientist who tests measuring instruments for accuracy. His specialty was breathalyzers. He also volunteered to raise puppies for Guide Dogs for the Blind. Around 2000, on a trip to New York, he took extra insulin and went to bed without checking his blood sugar. It dropped dangerously low during the night. Benton, the one-year-old black Labrador puppy he was raising, pawed and barked until Ruefenacht woke up. With his background in breath testing and dog training, Ruefenacht understood what his puppy had detected.

He spent four years testing the idea with Jeannie Hickey, a diabetes specialist who also has diabetes. He trained another Labrador from Guide Dogs for the Blind, named Armstrong. For one test, Hickey gave Ruefenacht a paper towel with her sweat on it. Armstrong picked her sample out from the others in about 30 seconds. That test showed the idea could work for other people with diabetes, not only for Ruefenacht. In 2004, he founded Dogs4Diabetics, one of the first programs to train diabetic alert dogs. Researchers later found that a compound called isoprene rises in the breath when blood sugar drops. The researchers believe isoprene is one of the compounds a diabetic alert dog detects on his person’s breath.

A diabetic alert dog will never replace a glucose monitor. Research shows the monitor usually catches a blood sugar change first and sounds an alarm. The dog detects the same change by scent. That makes the dog the second alert. People sleep through monitor alarms or silence them. A dog is much harder to ignore. He gets up, finds his person, and keeps alerting until the person takes action. Owners of alert dogs report fewer ambulance calls, fewer episodes of losing consciousness, and more independence.

So why did some diabetic alert dogs score so much higher than others? Much of it came down to follow-through. Researchers who studied diabetic alert dogs found that the most accurate dogs belonged to owners who tested their blood after every alert. Those owners rewarded the dog only when the test confirmed a low. Some owners ignored alerts. The dog did his job and got nothing for it. Other owners rewarded the dog before they tested, so he sometimes got paid for being wrong. Either way, the dogs became less accurate, and their accuracy kept slipping over the years. This is the same science I wrote about in “The Reward That Never Came.” When the reward stops coming, the behavior fades. The dog’s nose did not fail. The follow-through did.

Seizure dogs have one of the most unlikely beginnings of all. In 1981, the Prison Pet Partnership began at a women’s prison in Washington state. Inmates trained rescue dogs to become service dogs. In the mid-1980s, a German Shepherd named Sheba, trained in the program, began to sense seizures coming in a person with epilepsy. Sheba would ignore commands and force her person to stop all activity. As inmate trainer Sue Miller put it, the girl’s job was to learn to listen to Sheba. After her release, Miller placed nearly a dozen seizure dogs with people who had epilepsy.

In the 1990s, Val Strong and neurologist Stephen Brown built a formal training program for seizure alert dogs in England. Some of their dogs alerted 15 to 45 minutes before a seizure. For years, nobody could prove that a seizure had an odor a dog could detect. In 2019, a research team in France answered that question. Their trained dogs picked out sweat and breath samples taken during a seizure. That study proved that a seizure has its own odor.

A seizure alert dog is meant to warn his person before a seizure begins. Researchers are still testing whether dogs can detect the seizure odor that early. A seizure response dog does a different job. He stays with his person during a seizure, gets help, or presses an alarm button. With his dog beside him, a person with epilepsy does not face a seizure alone. Owners report more confidence going to work, to school, and out in public.

The bond shows up in epilepsy research too. Owners whose dogs alerted on their own reported a stronger bond than owners whose dogs did not alert. An alert dog and his person work as a team, and the dog’s accuracy depends on both of them. In mantrailing, I teach handlers the same lesson. The dog does the detecting, but the handler has to trust the alert and reward it. It is the same bond I wrote about in “Your Dog’s Hidden Compass.” The drive that brings a lost dog home to his person is the same drive that keeps an alert dog working beside the person he loves.

One of the best-known PTSD service dog programs started with a mother who wanted her son back. Brett Simon was a civilian K-9 police officer before he deployed to Iraq twice. There, he worked with explosive detection dogs in Baghdad and Mosul. He came home with PTSD and a traumatic brain injury. His mother, Shari Duval, said, “He was somebody different, he was not my son that went over there.” She searched for anything that might help and found early reports that service dogs could help veterans with PTSD. Almost no one was placing those dogs with veterans. So in 2010, she started K9s For Warriors. The program trains mostly rescue dogs for up to a year each and has placed dogs with hundreds of veterans. Brett went on to oversee the program’s dogs.

Dogs like these have changed the daily lives of many military veterans. They wake a veteran from a nightmare. They stand between their handler and a crowd to give him space. They lean their weight against their handler when his anxiety rises. Studies of veterans found that those with service dogs had fewer PTSD symptoms and less depression than veterans on a waiting list. Researchers have also shown that dogs can detect the odor of stress in human breath and sweat. We still do not know how much of a service dog’s work comes from scent and how much comes from watching his handler. Either way, the veterans in these studies report a better quality of life.

An accuracy score cannot measure what these dogs do in daily life. It cannot measure Benton waking Mark Ruefenacht the night his blood sugar dropped dangerously low. It cannot measure Sheba warning her person before a seizure hit. It cannot measure a PTSD alert dog interrupting a veteran’s nightmare so he can finally get a good night’s rest. None of these dogs replaces a doctor. But each one gives a person something no doctor can: a partner who is there every hour of every day.

Now think about your dog. He smells you every day. He knows what you normally smell like, and he knows when you are stressed or excited. He is not a doctor, and he cannot diagnose you. But if he keeps returning to one spot on your body, I would pay attention.

Remember the dog in London. He had no training and no idea what melanoma was. He knew only that his owner smelled alarmingly different. He kept returning to that one spot until she went to a doctor. Cancer, diabetic, and seizure detection all started that same way. A dog noticed a change, and a person chose to take him seriously.

The comprehensive research behind each condition, with every study and source, is posted in the Chronicles in the Field Notes Library.

© 2026 Lauren Duncan, PhD, and Inside the Canine Mind, LLC. All rights reserved.

Inside the Canine Mind is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Subscribe