Dogs can be trained to identify individuals with Parkinson’s disease (PD) by sniffing out a disease-specific odor in skin swabs with up to 98% specificity and 80% sensitivity, a recent study suggested.
The findings may help inform key biomarkers for PD before symptoms appear. There is currently no single diagnostic test for PD, and reported estimates show roughly a quarter of patients are misdiagnosed.
“Sensitivity levels of 70% and 80% are well above chance, and I believe that dogs could help us to develop a quick noninvasive, and cost-effective method to identify patients with Parkinson’s disease,” lead investigator Nicola Rooney, associate professor at Bristol Veterinary School at the University of Bristol, Bristol, England, said in a news release.
The study was published online on July 14 in the Journal of Parkinson’s Disease.
The Nose Knows
Previous studies have shown that dogs can detect disease-specific odors in human breath or urine in patients with cancer, diabetes, and even posttraumatic stress disorder. This study adds to the growing evidence they can also identify PD.
Seborrheic dermatitis and increased sebum production are premotor signs of PD. Previous research published last year by a nonprofit Parkinson’s organization based in Washington showed that various breeds of dogs can be trained to recognize sebum unique to PD patients. This led the UK-based researchers to suspect sebum could be used as an indicator of PD.
For this double-blind, randomized controlled trial, investigators collected dry skin swabs from 130 patients with PD and 175 individuals without the condition. Of these, 205 swabs were used for training dogs from the organization Medical Detection Dogs. For double-blind testing, 40 samples were from patients who didn’t use drugs for PD (42.5% women; mean age, 69.6 years), and 60 samples were from individuals without PD.
After 38-53 weeks of training, a Golden Retriever named Bumper and a Black Labrador named Peanut were shown samples on four stands in a line. In each line, there was either one positive sample or all controls. Both the trainer and experimenter were blind to the location of positive or negative samples. Computer software confirmed whether the dog’s response was correct.
If the dog was unsure, samples were retested in new lines until a clear decision was made. Dogs were rewarded for correctly identifying a positive or negative sample.
The dogs showed high specificity by correctly ignoring 90% and 98.3% of swabs from the control group and high sensitivity by correctly identifying 70% and 80% of the swabs from patients with PD (P < .0001). Of note, the Labrador had 15 weeks more training than the Golden Retriever and achieved the higher of the two results.
“While we do not envisage dogs being diagnostic, they could potentially, with refinements, help to validate methods of clinical utility…and aid rapid screening and diagnosis,” the investigators wrote.
The authors declared having no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.