Super Noses: Dogs Have Them!

NEW ORLEANS — Dogs’ keen odor-detection capabilities have been put to use (at least for investigational purposes) in a variety of cancer indications and diabetes, and now a researcher speaking here said they may be useful in diagnosing malaria — particularly in regions where elimination of the disease has been accomplished or is in sight.

In a proof-of-principle study, two dogs trained to recognize malaria-associated human odors were modestly successful in identifying socks worn by malaria-infected schoolchildren in Gambia, said Steve W. Lindsay, PhD, of Durham University in England.

Speaking at the American Society of Tropical Medicine and Hygiene’s (ASTMH) annual meeting, Lindsay reported correct detection rates of 70.0% and 73.3% for each of the two dogs for the 30 children in the study known to be infected. False positive rates among 145 malaria-negative participants were 9.0% and 9.7% for the two dogs. Accuracy was somewhat poorer in cases involving malaria parasites in the sexual stage of their life cycle.

The two dogs responded identically to 86% of samples. Lindsay noted that both dogs indicated malaria detection in six socks worn by children known to be uninfected; he speculated that these were not really false positives, but rather reflected bedsharing by study participants with infected siblings — a common practice in Gambia — whose odors were picked up in the participants’ socks.

Lindsay said the study’s impetus came largely from knowledge that mosquitoes can pick up odors from asymptomatic malaria-infected individuals, allowing them to feed preferentially on them. “If mosquitoes can do it, what about dogs?” Lindsay said, noting that canines have been trained to identify people with various forms of cancer as well those with diabetes who are on the verge of hypoglycemia.

For the study, Lindsay and colleagues provided asymptomatic children ages 5 to 13 with nylon socks to wear overnight. Socks were then frozen and shipped to England. The dogs were separately trained for 19 weeks in one instance and 24 weeks in the other. The test on which the current results were based involved placing the socks on stands, which the dogs sniffed. Handlers were blinded to which socks came from infected children, and as a second precaution, stayed behind a screen so as not to provide visual cues while the dogs sniffed the samples.

Lindsay said dog-based malaria detection was not intended for clinical diagnosis, but rather for screening in situations where moderately accurate but instantaneous results would be helpful. For example, he said, “once a country is malaria free or nearing elimination, it is critical to identify individuals with asymptomatic infections.” He noted that six countries that once had endemic malaria are now considered free of the disease, and another 12 are reporting no home-grown cases.

In consequence, detection dogs could be deployed at ports of entry, or for detecting cases in areas where prevalence is believed to be low — less than 1 case per 1,000, perhaps — which would make more conventional methods that require blood testing impractical.

Indeed, other presentations preceding and following Lindsay’s at ASTMH highlighted the limitations of current testing methods, all of which use blood samples and which carry substantial tradeoffs between accuracy and cost.

Lindsay cited several limitations to the dog study, however, chiefly the small numbers of samples used and the restriction to one country in Africa. “Since parasites strains differ around the globe, we need to replicate this with people from different parts of the world,” he said.

LAST UPDATED 10.29.2018