The Sniff Test – Benefits and limitations of diabetes alert dogs

Dogs serve an important role in enhancing human health and well-being, from guiding people with disabilities and calming people on the autism spectrum to sensing allergens and other potential medical emergencies—such as hypoglycaemia. The use of diabetes alert dogs (DADs) has been steadily gaining in popularity, and research is yielding new insights into their benefits and limitations. In this article, an expert on hypoglycaemia detection strategies provides a balanced overview of DADs and evidence-based guidance on their use.

Dogs help humans  in many ways: they provide companionship, comfort, and motivation to be more physically active. Service dogs go a step further and help people stay safe. It’s hardly surprising, then, that so many people with diabetes have taken an interest in diabetes alert dogs (DADs)—dogs trained to smell chemical changes in the human body when blood glucose falls outside the normal range.

I believe that DADs have value. At the same time, they should not be regarded as a panacea or substitute for vigilant glucose monitoring with reliable technologies. In order to provide informed advice to their patients, health providers need to understand how DADs have fared in scientific studies and to balance this information with the well-established contribution of dogs to human well-being.

The role of DADs in hypoglycaemia

Early detection of hypoglycemia is critical for the prevention of more severe and potentially dangerous episodes. Patients and their families, especially parents of children with type 1 diabetes, are understandably eager to find better tools to avoid severe hypoglycaemia. While technologies such as continuous glucose monitoring (CGM) can help in this regard, a growing number of people with type 1 diabetes are turning to DADs, in some cases as a substitute for CGM.

DADs are especially popular with people who have impaired awareness of hypoglycaemia (IAH), defined as a compromised ability to perceive or experience hypoglycemic symptoms. DADs can alert these individuals to check their blood glucose and treat hypoglycaemia before it becomes serious. While DADs cannot and should not replace glucose monitoring devices, they provide an extra layer of protection—at least in theory.

In practice, it can take time and (in some cases) a lot of money to find a well-trained DAD. As with all purchases, caveat emptor applies here, as there are no standards regulating the training and performance of DADs. Ideally, the dog should be no younger than 1.5 years, adequately trained to perform distinct “alert behaviours”, and obtained from an organization willing to provide references from past clients.1

The evidence

There’s a further reason for caution: while patient surveys and anecdotal reports almost uniformly praise DADs for their ability to  detect lows and improve diabetes control,2,3 clinical-trial evidence for the accuracy of DADs remains scant and inconclusive. Here’s what we know so far.

Only a few studies have tested the  accuracy of DADs under experimental conditions. In one study, trained dogs were largely unable to identify a “hypoglycaemic scent” in skin swabs obtained from subjects with type 1 diabetes (who were not the owners of the dogs).However, a  later, similar study  found that DADs’  ability to recognize hypoglycaemia in human perspiration samples was statistically significant.5

In a different type of study, 18 DAD owners completed diaries of DAD alerts during the first year after they obtained their dogs.6 Diary readings included daily blood glucose readings and DAD alerts (in response to perceived low or high glucose). Results showed that DAD alerts had a sensitivity of just 57%, with a slightly better sensitivity to low than to high glucose, and a specificity just under 50%.6 In addition, high variability was observed across DADs, suggesting that dog’s detection abilities may vary widely.

Because CGM readings are a reliable method of detecting hypoglycaemia, it makes sense to study how DADs compare to CGM. At least two studies have done just that. In the first, participants reported a high level of satisfaction with their DADs, but the dogs provided timely alerts in just 36% cases of hypoglycaemia (sensitivity), and showed a high rate of false positives.7 Time to detection was also significantly longer with DADs than with CGM (median difference of 22 minutes).7 The second study also found  low sensitivity and specificity (with high inter-dog variability), and rate-of-change analyses indicated that the dogs were responding to absolute blood glucose levels rather than rapid changes in glucose.8


Most diabetes health providers already have or will soon have patients interested in using DADs and need to be prepared to discuss the science behind DAD performance in a balanced way. While the evidence to date suggests that DADs are not as accurate as tools such as CGM, an appraisal of DADs must also consider their demonstrated benefits which include improvements in quality of life and diabetes control. . We cannot discount these positive reports and more research is needed to understand how DADs help people to manage their diabetes more effectively.


  1. Cattet J, Hardin DS. Diabetes alert dogs: buyer beware. Diabetes forecast 2014. Accessed at
  2. Petry NM et al. Perceptions about professionally and non-professionally trained hypoglycemia detection dogs. Diabetes Res Clin Pract 2015;109:389-96.
  3. Gonder-Frederick LA et al. Diabetic alert dogs: a preliminary survey of current users. Diabetes Care 2013:e47.
  4. Dehlinger K et al. Can trained dogs detect a hypoglycemic scent in patients with type 1 diabetes? Diabetes Care 2013;36:e98-9.
  5. Hardin DS et al. Dogs can be successfully trained to alert to hypoglycemia samples from patients with type 1 diabetes. Diabetes Ther 2015;6:509-17.
  6. Gonder-Frederick LA et al. Variability of Diabetes Alert Dog Accuracy in a Real-World Setting. J Diabetes Sci Technol 2017. doi:1932296816685580.
  7. Los EA et al. Reliability of Trained Dogs to Alert to Hypoglycemia in Patients With Type 1 Diabetes. J Diabetes Sci Technol 2016. doi:1932296816666537.
  8. Gonder-Frederick LA et al. Diabetes alert dogs (DADs): an assessment of accuracy and implications. Diabet Res Clin Pract 2017; 134:121-30.