VIDEO: Using dogs to detect hypoglycemia remains an imperfect science

NEW ORLEANS – In the first known controlled study of its kind, researchers demonstrated the reliability of trained dogs to alert hypoglycemia in human companions with type 1 diabetes, but at the cost of a high false-positive rate, compared with continuous glucose monitoring.

“The use of trained dogs to alert to changes in blood glucose in human companions with diabetes is increasing yet we don’t know how effective they are,” lead author Dr. Evan Los said in an interview in advance of the annual scientific sessions of the American Diabetes Association. “Enthusiastic anecdotal reports and limited in vitro studies are promising but rigorous studies have not been done. Clinicians are unable to provide informed advice when patients ask, ‘What do you think about diabetes alert dogs?’”

Dr. Los, a pediatric endocrinology fellow at Oregon Health and Science University, Portland, and his associates set out to conduct a controlled study of the reliability of diabetes alert dogs to detect hypoglycemia in their human companions with type 1 diabetes under real-life conditions. Capillary blood glucose (CBG) and continuous glucose monitoring (CGM) were used as comparators, and the study participants consisted of eight dog-human pairs. The human subjects ranged in age from 4 to 48 years old. Hypoglycemia was defined as CBG and/or CGM below 70 mg/dL and the human subjects answered survey questions regarding their impressions of the dog’s reliability and reasons for obtaining a diabetes alert dog. A timely alert was defined as occurring within 10 minutes before to 30 minutes after onset of hypoglycemia.

On a 10-point Likert scale, the dog users reported being “very satisfied” and “largely confident” in their dog’s ability to detect hypoglycemia (scores of 8.9 and 7.9, respectively), and they universally cited detection of hypoglycemia as the chief reason for obtaining a trained service dog. The researchers found that spontaneous dog alerts occurred 3.2 times more often than the rate of alerts during euglycemia, which was defined as 70-180 mg/dL, but the positive predictive value of dog alerts for hypoglycemia was only 12%. When Dr. Los and his associates reviewed event diaries and blinded CGM data, they observed that trained dogs provided a timely alert in 36% of all 45 hypoglycemia events. In the 30 events when both the dog-alerted and the blinded-CGM reached the hypoglycemia threshold, CGM would have alerted prior to the dog in 73% of events, which translated into a median difference of 22 minutes sooner.

“Trained dogs often alert a human companion to otherwise unknown hypoglycemia,” Dr. Los said. “However, due to high false-positive rates, a dog alert alone is unlikely to be helpful in differentiating hypo-/hyper-/euglycemia. CGM often detects hypoglycemia before a trained dog by a clinically significant margin.”

He acknowledged the study’s small sample size as a limitation. “Dog breeds and training methods were not universal, so there are likely to be variation in the reliability between dogs,” Dr. Los added. “This is not the final word on whether trained dogs might be helpful for patients with diabetes and there may be other benefits not assessed by this study such as having a positive partner in the daily management of a chronic disease.” He reported having no financial disclosures.

[email protected]

NEW ORLEANS – In the first known controlled study of its kind, researchers demonstrated the reliability of trained dogs to alert hypoglycemia in human companions with type 1 diabetes, but at the cost of a high false-positive rate, compared with continuous glucose monitoring.

“The use of trained dogs to alert to changes in blood glucose in human companions with diabetes is increasing yet we don’t know how effective they are,” lead author Dr. Evan Los said in an interview in advance of the annual scientific sessions of the American Diabetes Association. “Enthusiastic anecdotal reports and limited in vitro studies are promising but rigorous studies have not been done. Clinicians are unable to provide informed advice when patients ask, ‘What do you think about diabetes alert dogs?’”

Dr. Los, a pediatric endocrinology fellow at Oregon Health and Science University, Portland, and his associates set out to conduct a controlled study of the reliability of diabetes alert dogs to detect hypoglycemia in their human companions with type 1 diabetes under real-life conditions. Capillary blood glucose (CBG) and continuous glucose monitoring (CGM) were used as comparators, and the study participants consisted of eight dog-human pairs. The human subjects ranged in age from 4 to 48 years old. Hypoglycemia was defined as CBG and/or CGM below 70 mg/dL and the human subjects answered survey questions regarding their impressions of the dog’s reliability and reasons for obtaining a diabetes alert dog. A timely alert was defined as occurring within 10 minutes before to 30 minutes after onset of hypoglycemia.

On a 10-point Likert scale, the dog users reported being “very satisfied” and “largely confident” in their dog’s ability to detect hypoglycemia (scores of 8.9 and 7.9, respectively), and they universally cited detection of hypoglycemia as the chief reason for obtaining a trained service dog. The researchers found that spontaneous dog alerts occurred 3.2 times more often than the rate of alerts during euglycemia, which was defined as 70-180 mg/dL, but the positive predictive value of dog alerts for hypoglycemia was only 12%. When Dr. Los and his associates reviewed event diaries and blinded CGM data, they observed that trained dogs provided a timely alert in 36% of all 45 hypoglycemia events. In the 30 events when both the dog-alerted and the blinded-CGM reached the hypoglycemia threshold, CGM would have alerted prior to the dog in 73% of events, which translated into a median difference of 22 minutes sooner.

“Trained dogs often alert a human companion to otherwise unknown hypoglycemia,” Dr. Los said. “However, due to high false-positive rates, a dog alert alone is unlikely to be helpful in differentiating hypo-/hyper-/euglycemia. CGM often detects hypoglycemia before a trained dog by a clinically significant margin.”

He acknowledged the study’s small sample size as a limitation. “Dog breeds and training methods were not universal, so there are likely to be variation in the reliability between dogs,” Dr. Los added. “This is not the final word on whether trained dogs might be helpful for patients with diabetes and there may be other benefits not assessed by this study such as having a positive partner in the daily management of a chronic disease.” He reported having no financial disclosures.

[email protected]

NEW ORLEANS – In the first known controlled study of its kind, researchers demonstrated the reliability of trained dogs to alert hypoglycemia in human companions with type 1 diabetes, but at the cost of a high false-positive rate, compared with continuous glucose monitoring.

“The use of trained dogs to alert to changes in blood glucose in human companions with diabetes is increasing yet we don’t know how effective they are,” lead author Dr. Evan Los said in an interview in advance of the annual scientific sessions of the American Diabetes Association. “Enthusiastic anecdotal reports and limited in vitro studies are promising but rigorous studies have not been done. Clinicians are unable to provide informed advice when patients ask, ‘What do you think about diabetes alert dogs?’”

Dr. Los, a pediatric endocrinology fellow at Oregon Health and Science University, Portland, and his associates set out to conduct a controlled study of the reliability of diabetes alert dogs to detect hypoglycemia in their human companions with type 1 diabetes under real-life conditions. Capillary blood glucose (CBG) and continuous glucose monitoring (CGM) were used as comparators, and the study participants consisted of eight dog-human pairs. The human subjects ranged in age from 4 to 48 years old. Hypoglycemia was defined as CBG and/or CGM below 70 mg/dL and the human subjects answered survey questions regarding their impressions of the dog’s reliability and reasons for obtaining a diabetes alert dog. A timely alert was defined as occurring within 10 minutes before to 30 minutes after onset of hypoglycemia.

On a 10-point Likert scale, the dog users reported being “very satisfied” and “largely confident” in their dog’s ability to detect hypoglycemia (scores of 8.9 and 7.9, respectively), and they universally cited detection of hypoglycemia as the chief reason for obtaining a trained service dog. The researchers found that spontaneous dog alerts occurred 3.2 times more often than the rate of alerts during euglycemia, which was defined as 70-180 mg/dL, but the positive predictive value of dog alerts for hypoglycemia was only 12%. When Dr. Los and his associates reviewed event diaries and blinded CGM data, they observed that trained dogs provided a timely alert in 36% of all 45 hypoglycemia events. In the 30 events when both the dog-alerted and the blinded-CGM reached the hypoglycemia threshold, CGM would have alerted prior to the dog in 73% of events, which translated into a median difference of 22 minutes sooner.

“Trained dogs often alert a human companion to otherwise unknown hypoglycemia,” Dr. Los said. “However, due to high false-positive rates, a dog alert alone is unlikely to be helpful in differentiating hypo-/hyper-/euglycemia. CGM often detects hypoglycemia before a trained dog by a clinically significant margin.”

He acknowledged the study’s small sample size as a limitation. “Dog breeds and training methods were not universal, so there are likely to be variation in the reliability between dogs,” Dr. Los added. “This is not the final word on whether trained dogs might be helpful for patients with diabetes and there may be other benefits not assessed by this study such as having a positive partner in the daily management of a chronic disease.” He reported having no financial disclosures.

[email protected]