Cynthia Hatch, a mother in Brookfield, Wisconsin, is convinced that her family’s diabetes alert dog Sunny, a golden retriever/Irish setter mix, has saved her son Nathan’s life multiple times. Nathan has all the latest diabetes gear: a CGM tied into a pump that automatically shuts off when his glucose goes below a certain level. But he also has a rare combination of type 1 diabetes and Addison’s disease, another autoimmune condition where hormonal imbalances can cause severe and precipitous drops in blood sugar. Sunny is an additional tool in Nathan’s arsenal, and is one of a growing number of dogs in the U.S. being used to alert their owners of oncoming lows or highs in blood sugar.
Many different organizations now train dogs for this purpose; some are for profit and may charge tens of thousands of dollars for a dog trained from birth until over two years old. Others, like Dogs for Diabetics, are non-profits that charge people no more than a small application fee, but are very selective in terms of the clients they will take on. Dogs for Diabetics’ founder, Mark Ruefenacht—a forensic scientist with type 1 diabetes—is considered a pioneer in the field, the person who first set out to test this ability in dogs in a systematic way.
Back in the 1990s, he was traveling for work, accompanied by a seeing eye dog that he was helping to train. (On his mother’s side, several people suffer from advanced macular degeneration; on his father’s a relative has vision impairment from diabetes, so Ruefenacht thought that a seeing eye dog might be a part of his own future some day.) Alone at his hotel, Ruefenacht had a severe episode of hypoglycemia, and his dog worked to rouse him until he could treat himself. That chance event got Ruefenacht thinking that maybe identifying lows was something that dogs could be trained to do. He talked to his doctor, who also had type 1 diabetes, and she suggested that with his scientific and dog training background, he was the perfect person to test out this prospective ability in dogs. For the next several years, he worked on training his dog Armstrong and then his dog Danielle, trying to answer two questions: could he train the dogs to reliably detect his own lows? And if so, could he train them to detect the lows of others? He tested them using saliva and sweat samples collected when his blood glucose was low. He knew he was onto something when his doctor hid a brown paper towel with a sample on it among a bunch of other towels in a 3,000 square foot space. It took Armstrong 30 seconds to find and bring back the right towel.
Once he had trained four alert dogs, he started Dogs for Diabetics, a non-profit which takes dogs that have already been trained as service dogs and trains them for up to a year more until they can alert their owners at a rate of 80% or better of lows or precipitous drops in blood sugar. Whenever a dog alerts its owner of a low correctly, it’s given a treat to reinforce the behavior. High glucose levels, which even people can detect by smell, are much easier for a dog to identify (and less dangerous in the short term). So dogs are generally only rewarded for identifying lows. (Ruefenacht says that his own dog alerts on him about twice a day, and has an accuracy rate of 95-98%.) In his experience, the dogs are more accurate than constant glucose monitors, which have a fair amount of measurement uncertainty, and which lag behind both conventional meters and dogs.
A few different groups of researchers have set out to test this ability in dogs. An ongoing study by researchers at the University of Virginia is using “scientifically rigorous research methods similar to those required to demonstrate accuracy in blood glucose meters” to collect data over a four-week period for each dog studied. Results will be published next year. In a completed study, published last year in the open access peer-reviewed journal PLOS One, researchers examined the abilities of ten dogs trained in hypoglycemic scent detection and paired with owners with type 1 diabetes. The study found that “Most dogs showed a significantly higher proportion of alerts when their owners’ blood sugars were out of target range than within target range, indicating that alerting behaviour is unlikely to be random. In the case of the best performing dog, the odds of an alert being when bloods were out of range were 10,000 times higher than that of routine tests.” The authors pointed out that there is “marked variation” in dogs’ abilities, which could spring from their natural aptitude for the task, their original training, or the training abilities of their owners. This, combined with the high cost of many diabetes alert dogs, means that before committing to a dog it’s extremely important to research the organization and its training methods carefully; Ruefenacht has served as an expert witness in several law suits over dogs whose abilities and training were not as promised. One place to start is to see which organizations are accredited by the standards-based organization Assistance Dogs International.
Ruefenacht explains the science behind dogs’ ability this way: “Dogs, such as Labrador Retrievers, have over 200,000,000 sensors that can smell individual elements in parts per trillion, versus current technology’s ability to identify items in parts per million. Rapid changes in the blood sugar levels cause chemical changes in the body that are expressed through a person’s breath and skin, and include unique chemical elements that the dog can smell. Our experience indicates that the identifiable changes in a diabetic’s chemistry derived from his breath or sweat precedes the measurable change in blood sugar currently measured by glucose meters by 15 to 30 minutes. The dog can be trained to identify the onset of these changes and react to his handler when it is smelled.”
Dogs for Diabetics looks for people who will use their dogs to give them the confidence to achieve tighter glucose control. If they know their dogs will catch their lows, then they will be more willing to try to hover in that 85-100 area that would result in an A1C similar to that of someone without diabetes. All of the dogs he uses come from Guide Dogs for the Blind. Ruefenacht points out that 30-40 percent of that organization’s dogs go to people who are blind from diabetes. “If we can use their dogs to prevent blindness, we will be much further ahead.” The non-profit will only work with people who live within a three hour commute of its training center in California, and with people who are older than 12. Clients must complete 100 hours of classroom and field training with their dog, as well as monthly refresher courses.
The National Institute for Diabetic Alert Dogs, on the other hand, is a business that will deliver and train dogs throughout the U.S. It was founded by Lily Grace, a retired nurse who has been training dogs for thirty years. The dogs she uses are scent imprinted from birth—given saliva samples from a person with hypoglycemia even before the puppies can see or hear. The base price for a diabetic alert dog from NIDAD is about $10,000, but clients may ask Grace to do more training, up to teaching the dog how to call 911 with a special dialing device on its owner’s phone. Just as with seeing eye dogs, diabetes alert dogs can go everywhere with their owners—to school, on airplanes, into restaurants. Grace emphasizes to her clients that the dog is just another tool for managing diabetes, to be used in conjunction with a meter and perhaps a CGM. And while they are lovable companions, these dogs are working dogs, not family dogs. Their constant presence both marks their owner as someone with a medical condition, and requires their owner to work with and care for the dog throughout the day. “It can be a great answer for some people, but it’s not an end all for everybody. Dogs aren’t going to cure your diabetes,” said Grace.
Grace is the one who placed Sunny with Nathan Hatch, whose combination of Addison’s and type 1 diabetes makes his blood sugar especially volatile and difficult to manage. When Sunny senses a dramatic change in Nathan’s blood sugar (either up or down), he gets antsy and starts pawing at Nathan’s leg. If he doesn’t get a response, he runs to find someone else. Cynthia Hatch described one night when Nathan was having an Addison’s crisis, and they couldn’t get his blood sugar up. (They use glucagon only as a last resort, since the hormone worsens his Addison’s.) Cynthia had been in his room for four hours, feeding him juices and gels, giving him everything she could to get his blood sugar up. Finally, after giving him 320 carbs worth of sugar, she checked his blood sugar and it was around 140. She figured that he was all set for a while. “He was exhausted, I was exhausted, Sunny had been up all night.” She got Nathan situated and went to sit down in the living room. “Next thing I know Sunny is alerting like crazy. Nathan was 52 and dropping like crazy. Without Sunny, I’m not sure when I would have gone back in there, since I thought it was safe to let him sleep.”
Hatch tells of another time, when Nathan was swimming in a crowded local pool, over on the other side from his mother. His blood sugar had been 320 when he went in, and because his CGM sensor was underwater, his CGM wasn’t working. After about fifteen minutes, Sunny started to pace and then alerted Hatch. She went over to test Nathan and he was 48 and dropping. “I think without Sunny the first time we would have realized would have been when he had a seizure in the water,” she said.
Hatch seconds Grace’s point that having a diabetes service dog is a lot of work: whenever Sunny alerts, Nathan needs to test himself, and reward Sunny if he has correctly identified a low. At times he gets frustrated. But Sunny has become not just an important part of Nathan’s diabetes care but something that brings comfort and support to Nathan wherever he goes. “Before we had Sunny, one of the worst feelings was to walk in and have Nathan in a seizure, and not have any idea how long he had been in it. With Sunny we have had seizures, but we’ve prevented many of them, and we’re on top of them fast. He has brought a safety factor into our lives. When Sunny is sleeping next to Nathan he feels safer, even emotionally. He is happy that wherever he goes has a buddy next to him,” said Hatch.
Dennis McGrath-Wagner, who got his black Labrador Yancey from Dogs for Diabetics in March 2014, likewise sees his experience with a diabetic alert dog as both life changing and lifesaving. McGrath-Wagner decided to get a dog as a tool to lower his blood sugar. He would consistently run his blood sugar high as a way to guard against going low. He hoped that having a dog would help him to keep his blood sugar nearer its target. During the past six months, he feels that Yancey has saved his life more than once. In one instance, Yancey alerted McGrath-Wagner, and when he tested he saw that his glucose level was 70. About five or ten minutes later, Yancey started alerting frantically, and McGrath-Wagner had dropped to 30. He ate a lot of glucose tablets and gel, and got himself up to 80, and got in the car to drive home. Yancey alerted again, McGrath-Wagner checked again, and saw that he’d dropped down to 50. “I would have been on a highway in less than 5 minutes. Yancey saved my life, and maybe someone else’s that day, too.” He mentions other times when Yancey has refused to get into the car with him until he’s checked his blood sugar—only to find that he’s low. Yancey’s success rate for alerting stands at 95% (with extensive initial training, and constant follow-up training).
But Yancey, and Dogs for Diabetics, have brought more to McGrath-Wagner than a new sense of safety. Dogs for Diabetics, which has an extensive training and volunteer program, has become an invaluable network for McGrath-Wagner as he navigates life with type 1 diabetes. And Yancey’s skill and companionship have helped his owner lower his A1C from a high of 10 to 6.9, a change that McGrath-Wagner says is due not only to his confidence not to run his blood sugars high, but to the calming effect Yancey has. McGrath-Wagner has had pets before, but his relationship with Yancey is on a different plane. “When a dog is there staring you in the eyes, and you know that he is probably saving your life, it is a totally different bond. It’s an incredible connection we have.” McGrath-Wagner hopes it’s a connection that will be available and accessible to more and more people with diabetes as this field grows.
My poodle, Trixie, alerted me many times while I was asleep, that my bg was very low. This was before I had ever heard of any dogs being trained. I felt foolish telling my dr that my dog had saved my life. Prove positive. She did. Absolutely astounding… There was NO training involved.