Tuberculosis is a serious public health challenge in the developing world, where the infection claims roughly two million lives each year, according to the World Health Organization (WHO). Yet the disease, which is a leading killer of patients with HIV/AIDS, is cumbersome to detect, resulting in delayed or inappropriate treatment, greater spread of the infection and preventable deaths.

So, researchers in Colorado are developing a portable, rapid TB sensor that could help reduce the death toll and make treatment more efficient.

This "field friendly" device relies on readily available and relatively low-cost components and can find the lethal pathogen in blood in just 20 minutes, says Diego Krapf, an applied physicist and assistant professor of engineering at Colorado State University in Fort Collins, who designed the new device. That is far faster than conventional methods such as sputum tests, which examine secretions from the lungs and bronchi and can take days—sometimes weeks—to return results. The new detector can identify active infection as well as dormant, or latent, microbes that might one day flare up into full-blown illness, the developers say. It also can distinguish between patients who carry TB and those who once were infected but have successfully fought off the illness.

"Our detector senses proteins, not an immune response against them," Krapf says. "The benefit of doing this is that while proteins mark the presence of disease, antibodies remain in the organism after the disease is long gone." So the new approach greatly reduces the chances of false positives.

A portable detector would greatly aid efforts to fight the infection in developing countries, particularly parts of Asia and Africa where as much as 40 percent of the population carries the microbe, says Robert Belknap, a physician and TB expert at the Denver Health Medical Center in Colorado. "If it works, it's truly portable and doesn't require special conditions, it would be a tremendous advance," he adds.

The device could overcome a few key failings of conventional TB screening. Long delays, in some cases up to six weeks, between test and result often force public health workers to waste time and resources on people suspected of having the infection but who, in fact, do not. Potential patients are frequently isolated—from their children, colleagues and others—for fear that they may spread the disease.

Those suspected of having TB may also be given potent antibiotics as a precaution—drugs that can cause serious side effects, including vomiting, headaches, rash and drug-induced hepatitis (liver inflammation). For these drugs to be effective, and to prevent the body's development of resistant bacteria, patients must take the medications on a strict schedule and complete the regime. The consequences of straying from the regimen are so high that public health authorities try to monitor TB patients to make sure they take their pills, an approach called directly observed therapy. Such programs require substantial money and manpower, and often fail for lack of both.

"An accurate, rapid, point-of-care test would help public health workers better focus resources quickly on the individuals who need it—including providing directly observed therapy—and avoid unnecessary medications and isolation in those who don't," Belknap says.

Krapf and his colleagues built a shoe box–size prototype with off-the-shelf parts including diode lasers from CD and DVD players, photodiode detectors, and other low-cost electronics. A handheld version could easily be made from somewhat more expensive components, he says.

The detector works by a capture-illuminate process: It first analyzes a blood sample in search of two TB proteins—antigen 85b and alpha-crystallin. If either is present, it will bind to a specific molecular brush linked to immune molecules, called antibodies, which affix to antigen 85b and alpha-crystallin but reject all other proteins. Then a bath of additional TB-sensitive antibodies, modified with a gene that produces fluorescence, is added. The result is in effect a glowing antibody–TB–antibody protein sandwich.

Eventually, signature proteins from drug-resistant strains of the bacteria could be incorporated into the device, Krapf says.  According to WHO, an estimated 511,000 cases of multidrug-resistant TB were reported in 2007. Most of these occurred in China, India and the Russian Federation.

The detector research, which will be presented at a meeting this month of the American Physical Society, is of primary benefit for people in the developing world, yet it could help in the U.S., too. Although the disease is largely under control in humans there, it still poses a threat to the cattle industry—as well as those who depend on it for food.

Kathy Orloski (pdf), a veterinary epidemiologist at the U.S. Department of Agriculture's National Tuberculosis Eradication Program in Fort Collins, Colo., says the country's cattle industry would welcome a rapid test for bovine TB. Cows currently are screened for the disease with skin tests like those used in people, a process that requires animals to be handled twice in three days, driving up costs for ranchers. Dairy or beef cattle infected with TB are euthanized, she says.

Government inspectors at slaughterhouses would likewise benefit from an immediate on-site detection method, helping them identify carcasses of infected cows faster than current laboratory testing, which takes up to 48 hours to process tissue samples. "A test used in this capacity would have to be very accurate," Orloski says, "because the United States slaughters 33 million cattle each year."

The government also would like to see a speedier way of testing the million-odd cows imported each year from Mexico, which has had a problem with bovine TB in its beef and dairy stock for several years, Orloski says. Whereas Mexico and Canada are the two leading exporters of cattle to the U.S., Canadian cattle have for the most part been TB-free (pdf) for the past decade, with the exception of certain areas of Manitoba (pdf).

*Correction (3/24/10): This article originally stated that the researchers were from the University of Colorado.