Smart, docile, and hard-working, African giant pouched rats may one day play a role in diagnosing TB. NIAID grantee Bart Weetjens, M.Sc., and colleagues in Tanzania, Africa, are training Cricetomys gambianus rats to react to Mycobacterium tuberculosis (Mtb), the bacteria that cause TB. In early tests, trained rats correctly identified Mtb-containing samples at rates comparable to standard diagnostic techniques. The rodents also worked quickly; a rat can sniff its way through 140 samples in just 15 minutes, the researchers found. This compares well with the standard diagnostic method in which trained technicians use a microscope to examine sputum—typically, a technician can assess about 40 samples a day. (Sputum is any material from the lungs, trachea, or bronchial tubes expelled through the mouth by coughing.)
Using rats to detect TB is an outgrowth of an earlier project involving Mr. Weetjens and a team of Belgian and Tanzanian researchers. In the previous work, rats were trained to detect vapors emitted by landmines. This was because, although they tip the scale at about 3 pounds, African giant pouched rats are too light to trigger a landmine explosion. The success of the previous project led the team to modify the training regimen—instead of landmine vapors, rats were trained to react to volatile organic compounds produced by TB bacteria.
In their pilot study, the team first trained 20 rats to sniff along a row of 10 lidded containers that were sequentially uncovered by the experimenters. Rats received a food reward when they learned to pause for 5 seconds at any container holding Mtb. The best-performing rats were given additional training and learned to react to TB-positive human sputum samples that had been collected (and rendered non-infectious through sterilization) from several TB treatment centers in Tanzania.
After 6 months of training, the rats could reliably distinguish between TB-infected sputum and uninfected sputum. In a period of 71 days, 3 trained rats sniffed nearly 10,000 sputum samples. The sensitivity (the rate at which the rats correctly identified Mtb-positive samples) was 86 percent, while the specificity (the rate at which the rats correctly excluded Mtb-negative samples) was 89 percent.
The researchers also found that rats could detect TB bacteria even in samples that had earlier been judged by human evaluators as negative for Mtb. This suggests, say the researchers, that trained rats may be capable of detecting pulmonary TB at a very early stage of the disease. This has important public health implications, they add, because some 55 percent of TB cases in Tanzania are undetected through standard surveillance techniques and early detection and treatment could minimize the spread of TB to other people.
If these initial results are confirmed through additional research, the scientists envision teams of trainers and “doctor rats” travelling to locations throughout Tanzania to perform fast and accurate TB case detection.
back to top
Last Updated July 14, 2009