Medical Examiner

Sniffing Out Swine Flu

Researchers hope to create a better way to diagnose swine flu and other ailments.

Swine flu detection

Crude methods of detecting swine flu have so far provoked hand-wringing and no small amount of ridicule. Planeloads of travelers to China have had laser beams aimed at their foreheads, landing some under quarantine (and spurring a YouTube minifest of airport videos). This summer, Slate reported on a camp that tried to prescreen kids for flu by checking campers for fevers—and failed to detect a sick child whose physician parent brought his temperature down with Tylenol, fueling an outbreak. Meanwhile, people infected with the virus can pass it on before they develop symptoms; others never develop fever at all.

With swine flu cases surging again in much of the United States, a portable and cheap detector that could ID the infected—even before they fall sick—would be a boon to public health. Despite some evidence that it gets deeper into the lungs than other seasonal viruses, most cases of swine flu are described as mild. But for people with asthma, diabetes, and heart disease; young children; and pregnant women, H1N1 can be grim. Even worse, some parents irrationally plan to avoid vaccination for their kids, and fears remain that a more virulent form may yet emerge—both of which could fuel trouble. Better methods of detection could make a big difference, especially in day care centers, hospitals, and schools, by pinpointing who should stay home and who might benefit from early treatment.

Happily, researchers are gaining ground with a novel strategy to detect a wide range of infections and assess their severity in individuals. Traditional detection efforts have focused on fishing for pathogens like viruses or bacteria, and this remains a critical approach. But a growing number of scientists are also zeroing in on the other side of the battle—the individual’s immune response. The molecular signatures of that response may offer clues as to what ailment, exactly, a person has, how sick he may become, and what treatments could help him. It’s like analyzing Napoleon’s invasion of Russia by focusing on the Russians’ scorched-earth retreat. It’s a big idea, with possibilities extending well beyond flu season.

Scientists at Duke, for instance, are mapping broad biological changes, including immune-related changes, among students who are exposed to or come down with H1N1. The hope is that these signals will appear early enough and prove adequately specific to be used for early screening. The researchers first made strides looking at a range of respiratory infections. In September, they reported on combinations of genes that were expressed in response to cold virus, the respiratory syncytial virus, and seasonal influenza A: They exposed healthy (and willing) subjects to the virus and tracked them with repeated blood samples as some developed symptoms. Then they analyzed molecular changes in those blood samples, homing in on signatures that predicted who got sick. Using these signatures, they could distinguish viral infections from one another—as well as from bacterial infections—with signals appearing relatively early on. Now the group has turned to swine flu, which is on the rise at Duke, with several hundred cases so far this year.Students enrolled in the study are providing blood samples, e-mail check-ins, and, if they get sick, information on their close contacts, with whom researchers will also try to follow-up.

The goal is to develop a molecular signature for H1N1 and other viruses that could be used for early detection. The team also hopes to build a portable device that would search for disease with just a small amount of blood—or simply a nasal swab or saliva sample. Such an approach might help ID people who would benefit from fast antiviral treatment, those who should be isolated, and those who need not be—perhaps saving planeloads of irritation down the road. The Duke work is funded by the Defense Advanced Research Projects Agency, the central research and development office for the Department of Defense, which would also like to identify infected soldiers before they spread disease or are sent on assignments they might become too sick to perform. DARPA apparently sees parallels between students and soldiers, who live in crowded quarters, get little sleep, and endure intensive (if different) kinds of stress, lead researcher Geoff Ginsburg told me.

The question of how, exactly, the immune system responds to various kinds of flu is already a hot topic. In 2007, researchers took a sample of the deadly 1918 virus and infected a group of monkeys in order to study their responses and ensuing (ugly) deaths. This work, which made even some scientists queasy, contributed to the theory that the 1918 flu caused an aberrant and unusually strong immune response—and that this was what killed a lot of people, especially healthy young ones. A similarly overwrought response, characterized by massive inflammation, seems to occur in cells infected with the avian flu H5N1, in contrast to less pathogenic strains. (H1N1 appears to have some similarities to the 1918 strain but is nowhere near as dangerous.) Other researchers are untangling why and how the immune system goes haywire in response to certain viruses, with the hope of finding treatments that might calm it down, says Ben Greenbaum of the Institute for Advanced Study in Princeton (who is also a good friend).

Other researchers are running with the signatures-of-response approach, too. When parents bring babies with fevers to the emergency room, for instance, doctors often don’t have a fast, reliable way to tell which infections are bacterial as opposed to viral and which are truly serious. As a result, babies tend to be admitted to the hospital for two to three days of observation, says Octavio Ramilo, chief of infectious diseases at Nationwide Children’s Hospital in Columbus, Ohio. His group is studying whether tests of molecular changes in the babies’ blood might spare unnecessary hospitalizations—or allow for better, faster care of those who require antibiotics. Ramilo’s team is also working on immune signatures of a condition called Kawasaki disease, which causes fevers and is sometimes confused with other kinds of infection. Again, the goal is to understand how, exactly, different diseases provoke the body and which patients are most likely to become very sick. Another aim is to develop portable technologies to give clinicians and patients fast access to this information.

Meanwhile, researchers are also homing in on signatures related to cancer, heart disease, and other complex conditions. Geoff Ginsburg and colleagues have shown that profiles of the gene expression from tumor tissues in blood may predict how well cancer patients respond to specific chemotherapeutic drugs. Similar methods can reveal the extent of coronary artery disease in patients undergoing an imaging procedure called angiography. Others have reported on potential applications related to aortic aneurisms, Type 2 diabetes and autoimmune diseases. By the looks of it, researchers are just starting their engines.

To be sure, none of this will replace traditional methods of hunting down pathogens directly, especially in the midst of new and weird outbreaks. Consider the success of pathogen-focused work in untangling West Nile virus and a hemorrhagic virus called LuJo, among others. But for some diseases, looking at molecular responses in infected people may prove easier and faster. It may also work better for screening, assessing how bad a sickness is, and figuring out the best response. This is the new game in town—and someday it might save us.