Diabetes Forecast

A Breath Test for Blood Glucose

Making strides in finding an alternative to finger sticks

By Andrew Curry ,

Researcher:Pietro Galassetti, MD, PhD
Occupation: Director, Metabolism and Bionutrition Core, Institute for Clinical and Translational Science, University of California–Irvine
Research Funding: ADA Clinical/Translational Research Award

Imagine if instead of a finger prick, you could measure your blood glucose with a puff of breath. The potential of such technology is enormous, going far beyond what would clearly be an improvement in the lives of people with diabetes. But because of its complexity, breath analysis of something as complicated as blood glucose has remained out of reach.

With help from an ADA research grant, University of California­–Irvine scientist Pietro Galassetti, MD, PhD, is hoping to change that. He's looking for ways to detect changes in blood glucose levels in the breath using technology originally developed to sense chemicals in the atmosphere. If he can prove the technique works, he hopes private industry can refine it and make breath-based glucose monitors a reality in the next 20 years. "Gases have for decades been courted by researchers," Galassetti says. "If we could find them, it would be the holy grail—it's very easy to take a breath sample."

The idea is simple. Think of the body as a car. As it burns fuel—fats and sugars—it creates exhaust. Tweak the fuel mixture, and the hundreds of different gases change, too. When it comes to diabetes, the fuel-mixture metaphor is particularly apt. "The body usually has a balance of energy sources," Galassetti says, but diabetes cuts the potential fuel sources in half: Without insulin, the body can't burn sugar (glucose) and must rely on fat alone for fuel.

In fact, before insulin was discovered in 1921, one of the characteristic signs of end-stage diabetes was the smell of acetone—the same chemical as nail-polish remover—on patients' breath as the body used up its stores of fat. "I hoped it would be something as simple as acetone," Galassetti says. But acetone is just one of hundreds of chemicals the body produces while generating energy, and by the time it becomes easily detected, the body is already at the point of no return.

Translating theory to practice is complex. "There's a lot of potential, but so far very little concrete has come out of it," Galassetti says. There have been two main barriers: finding machines sensitive enough to detect all the different gases that make up human breath, and finding algorithms that would enable computers to make sense of the resulting information. The first barrier began to crumble in the 1970s, thanks to a UC–Irvine chemist named F. Sherwood Rowland. Rowland won the Nobel Prize in chemistry in 1995 for his work detecting tiny amounts of gases in the atmosphere. Though he has since retired, his Irvine lab—run since Rowland's retirement by Galassetti's collaborator and UC–Irvine chemistry department chair Donald Blake—is still a leader in the field. "Their main thing is being extremely good at picking up tiny concentrations of gas in extremely large gas mixtures," Galassetti says.

The technology Rowland pioneered is now sensitive enough to detect gases in concentrations as low as 10 parts per quadrillion— Galassetti says that's like covering the western United States in white golf balls, then picking 10 red ones out—and was originally deployed to measure the depletion of the Earth's ozone layer. More recently, it saw service during the 2008 Beijing Olympics, measuring pollution levels in that famously smoggy city. Working with Blake, Galassetti has harnessed it to pick apart the gases in exhaled air.

But that's only part of the battle. In the complex stew of hundreds of chemicals that make up breath, the challenge is figuring out which ingredients are relevant. "When glucose changes over time, not one but 20 or 30 different gases in the body change," Galassetti says. "We're looking for a mathematical algorithm that could correlate with enough accuracy that we could just use the breath [for blood glucose and other tests]." To do that, Galassetti and his team take hundreds of samples and run them through the computer to look for what they have in

common. To train the machines, Galassetti's team conducts tightly controlled experiments. Volunteers are brought to a normal glucose level using intravenous insulin. Then, taking breath and blood samples at five-minute intervals, the team gives them glucose infusions, bringing them up to the point of hyperglycemia over the course of an hour or so, then back down again to normal.

Over the past year, they've tested people in four groups: people who don't have diabetes to create a baseline, and then people with pre-diabetes and types 1 and 2. The data are plugged into a computer, which analyzes and compares more than 100 different variables in the hunt for patterns. So far the researchers have amassed over 100 data sets. "We have very strong data that allow us to predict glucose and insulin levels in healthy subjects," Galassetti says. "With the help of ADA, we're making quite [a lot of] headway." Perfecting the technology could mean an end to painful finger-prick blood tests, making it easier for people with diabetes to monitor their blood glucose levels; at some point, it could conceivably make screenings for pre-diabetes economical on a mass scale.

But don't expect to trade in your glucose meter for a Breathalyzer-style monitor any time soon. Galassetti freely admits there's a long way to go before the technology makes it out of the lab. "The way we do it is extremely complex and extremely expensive," he says. "But this proves the point that it can be done."

To sponsor an ADA research project at the Research Foundation's Pinnacle Society level of $50,000 or more, call Elly Brtva, MPH, managing director of Individual Giving, at (703) 253-4377, or e-mail her at ebrtva@diabetes.org.



Take the Type 2
Diabetes Risk Test