What Clinical Niche Will Breath Tests Fill?
By Genna Rollins
After a flurry of Food and Drug Administration (FDA) clearances in the early 2000s, breath-based tests have faded from the radar screen, seemingly consigned as an afterthought in the in vitro diagnostics industry. While it’s true that breath tests haven’t surged into widespread clinical practice, their story is far from over, according to experts. Research in the field is robust, and investigators have been quietly building the science base, edging a variety of new breath tests closer to everyday use in areas such as personalized medicine, infectious diseases, and cancer risk prediction.
“This is the frontier of medical testing. Clearly, blood, urine, and radiography tests in some cases will be displaced, and in others, complemented by breath testing, because there are so many advantages of breath testing that are not true for other tests,” said Raed Dweik, MD, director of the pulmonary vascular program in the department of pulmonary and critical care medicine at Cleveland Clinic and associate professor of medicine at Case Western Reserve University Lerner College of Medicine. “Right now we’re kind of scratching the surface. But the field’s grown tremendously in recent years and with technological advances in sampling, sensor design, standardization, and analytics, it has the potential to clinically benefit patients on a global scale.”
What’s Old is New
The notion of breath as a bellwether of disease dates to Hippocrates, who commented on fetor oris and fetor hepaticus in a discourse on breath aroma and disease. The modern science of breath testing took off in the early 1970s when Nobel Prize recipient Linus Pauling demonstrated that more than 250 volatile organic compounds (VOCs) were present in exhaled breath. That milestone focused considerable research on quantifying and understanding the role of various endogenous and exogenous gasses in the breath, which primarily is comprised of nitrogen, oxygen, carbon dioxide, water vapor, and inert gasses.
Investigators now have identified in excess of 3,000 different molecules in the breath, although less than 400 have been well-characterized, according to Dweik. “It’s almost like the Human Genome Project. We need to profile and analyze all those compounds so that we have a breathprint. The next step will be to understand how all the different breath components measure and monitor wellness and disease,” he said. Dweik is researching the role of nitric oxide (NO) in lung physiology, along with other exhaled markers of lung disease. He also heads a public-private consortium in Ohio seeking to develop sensors and point-of-care (POC) devices for breath analysis.
Of the thousands of trace VOCs with concentrations ranging from parts per million to parts per trillion, most make their way into breath via passive diffusion across the pulmonary-alveolar membrane. As byproducts of core metabolic processes, many, such as ethane, pentane, acetone, methanol, isoprene, and alkanes, have been linked to oxidative stress.
The Sampling Challenge
Breath also consists of exogenous VOCs inhaled from the environment, an issue that complicated progress in the field for some time because these compounds can interfere with breath sampling and analysis. A spate of studies performed in the 1980s with inconsistent and nonreproducible findings made clear that collecting breath was not as simple as phlebotomy or urine sampling and needed some work. “Many of those studies were not looking at how breath was collected, and we now know that sampling plays a huge role in breath. If you don’t collect samples properly, it’s garbage in–garbage out,” explained Terence Risby, PhD, professor emeritus of environmental health sciences at the Johns Hopkins University Bloomberg School of Public Health in Baltimore. “Breath is a composite of many compounds. It reflects your physiology as well as your microenvironment.”
Sample collection problems for some time stymied research involving NO, which probably is the most investigated breath analyte. Recommendations for standardized NO measurement published in 1999 by the American Thoracic Society and European Respiratory Society picked up the pace of research considerably, and established a model for collecting other breath VOCs.
Why Test Breath?
Given the technical challenges associated with breath collection and all the information still to be gleaned about the components of breath, why bother with breath testing? Why not just stay with tried-and-true blood, urine, cerebrospinal fluid—and to a lesser extent, saliva and sweat—biomarkers?
Breath researchers emphasize that although most people can endure a fingerprick or blood draw without too much inconvenience, the fact remains that these modalities are invasive and tolerated better by some than others. “Who wants to have a needlestick for a blood sample? We take it because we have to, but a lot of people would rather not,” contended Michael Phillips, MD, founder and CEO of Menssana Research in Newark, N.J. “On the other hand, breath testing is immensely simple for patients, and it’s noninvasive, painless, and safe.”
Experts also point out that breath testing offers flexibility not always possible or easily accomplished with blood or urine testing, including continuous sampling. “You can continuously collect and analyze breath real-time during sleep or exercise and analyze patterns and any changes while the activity is going on,” explained Anton Amann, PhD, professor of chemistry at Innsbruck Medical University in Austria. Risby added that properly collected breath samples can be stable, thereby not subject to transport and storage issues associated with other types of specimens.
Breath testing also offers the potential of detecting endogenous compounds not measurable today in blood or urine, because they’re at such low concentrations. “Theoretically, if you had super-sensitive technology, you’d detect these compounds in blood and urine, because they’re partitioning out into the breath,” said Phillips. “But at least in our sample collection method, you’re collecting breath over a period of two minutes, during which time the entire body content of blood has passed through the lungs. So we’re really doing our sampling from an enormous quantity of blood, and that achieves a sensitivity that’s not achievable with a small quantity of blood.”
Tests in Use
To most people, the term ‘breath test’ is synonymous with breath alcohol tests—better known as breathalyzers—and for sure, these tests have been adopted far-and-wide. However, capnography, used in operating rooms every day, is the most widely implemented breath test. Recommended in the mid-1980s as a standard for patients receiving general anesthesia, anesthesiologists use this test as both a direct and indirect measure of carbon dioxide (CO2) concentration. FDA has cleared numerous breath alcohol and capnography tests (see Table, below).
Types of FDA-Cleared Breath Tests
A variety of different breath-based tests have been cleared by the Food and Drug Administration. Although one test dates to the 1970s, the majority of clearances came within the past decade.
Asthma, airway inflammation
Grade 3 heart transplant rejection
CO poisoning, carboxy-hemoglobin
CO2, O2, N2O, N2, anesthetic agents and gases
Gas isotope ratio MS
IR gas analysis
IR absorption spectrometry
Gas chromatography MS
MS=Mass spectrometry; IR=infrared; 1IR, paramagnetic, miniCO2 IR measuring, colorimetric CO2, narrow band IR; 2electrochemical gas; 3electrochemical, sol-gel-heme protein; 4semiconductor oxide, semi-conductive alcohol, fuel cell, fuel cell electrochemical; 5electrochemical gas
Source: adapted from F1000 Medicine Reports 2010, 2:56
Aside from these specialized applications, breath tests for NO have gained the most acceptance in general clinical practice, with at least three cleared by the FDA for use in monitoring asthma and airway inflammation. Scientists suspected early on that NO had particular potential as a breath test because it is released in the airway as a result of airway inflammation rather than through the blood-lung interface. This supported the hypothesis that NO might be useful in detecting or monitoring reactive airway and inflammatory diseases such as asthma and chronic obstructive pulmonary disease. Standardized collection procedures lead to validated studies confirming that asthmatics and allergy patients have higher levels of NO than healthy controls.
At this point, NO has a solid, if somewhat limited, place in the asthma management armamentarium. A recent review article found only equivocal benefits from adding NO to usual clinical guideline-recommended care, but suggested that it may serve as an important adjunct for diagnosis and management of certain asthmatics (Chest 2010;138:682–92).
“NO is clearly abnormal in a lot of patients with asthma, so it has some clinical value in diagnosis and monitoring. Although it’s not as useful as originally believed because not all patients have increased NO, for some, it can be very useful,” explained Peter Barnes, FRS, professor of thoracic medicine and head of airway disease at the National Heart and Lung Institute and of respiratory medicine at Imperial College London. “We find it very useful for monitoring whether patients are taking their inhaled steroids, which is very difficult to monitor because you can’t measure the blood levels like you would with a tablet. If we find a patient has very high NO levels, despite being given steroids, it’s almost always due to the fact they’re not taking inhaled steroids.”
Another well-established area of breathing testing involves using a 13C-labelled urea substrate to detect Helicobacter pylori. When H. pylori is present it interacts with the substrate, producing 13CO2, which is detected in breath. Although these tests have been in available in the U.S. for close to a decade, they are used more in Europe and Japan, possibly due to reimbursement issues, according to observers.
Reimbursement challenges also have played a role in adoption of Menssana’s Heartsbreath test, which combines a specially developed collection system, gas chromatography mass spectrometry (MS), and a proprietary algorithm to detect alkanes and monomethylalkanes indicative of Grade 3 heart transplant rejection. The test was FDA-cleared in 2004, but without a coverage determination from the Centers for Medicare and Medicaid Services (CMS), it has had limited clinical implementation, according to Phillips. The company has received NIH funding for a large validation study that he hopes will convince CMS to cover the test.
On the Horizon
Based on Menssana’s experience with Heartsbreath, the company has three other breath tests in various stages of development and validation having to do with lung cancer, breast cancer, and pulmonary tuberculosis. The lung cancer test, which analyses a cluster of alkanes and methylalkanes associated with oxidative stress, is the furthest along in the pipeline. “We’re nearing the final phases of an NIH-funded blinded, multicenter study, and if the results come out as we hope, we plan to apply for FDA clearance by the end of this year or beginning of next,” said Phillips.
He sees the role for these tests as early risk predictors rather than definitive diagnostics. “That’s an important point to emphasize. These are not either/or tests. The idea is to start off at the bottom rung of the ladder with the least invasive, least expensive test that has a high negative predictive value. If the test is positive, you move up to the next rung of the ladder,” Phillips explained. In the case of the lung cancer test, “if it’s positive, you wouldn’t say the patient had lung cancer. It would be an indication that he or she was at high risk and would move up to the next level, with a chest CT or perhaps a bronchoscopy or biopsy.”
Whereas Menssana’s technology measures endogenous VOCs that are byproducts of oxidative stress, other promising breath tests close to clinical use work off the same principals as the H. pylori tests, i.e. measuring breath metabolites after administration of an isotope-labeled substrate. Observers contend that this approach could be one of the best uses of breath testing in the future. “This is an area that’s elegant,” said Risby. “It’s an example of breath testing providing something that’s either not available or not easily available by other means.”
The reason this newer generation of xenobiotics have the potential to go farther commercially than H. pylori breath tests is that they involve personalized medicine. With most FDA-approved drugs metabolized in the liver by the CYP family of enzymes, there is a need from both efficacy and safety standpoints to evaluate how well these medications are cleared.
Enter isotope-labeled breath tests, which can provide crucial phenotype information faster and easier than existing blood or urine tests. “There are phenotype tests available which are considered the gold standard, but they require either four-hour urine pooling or blood sampling every 10 to 15 minutes for three hours,” explained Anil Modak, PhD, associate director of medical products research and development at Cambridge Isotope Laboratories in Andover, Mass.
Modak is chief investigator for several breath tests in various stages of clinical evaluation that eventually could become companion diagnostic tests. These include a dextromethorphan-13C labeled test to detect CYP2D6 enzyme activity for selection of tamoxifen or aromatase inhibitors, a 13C-pantoprazole labeled test to detect CYP2C19 enzyme activity for identifying responders and non-responders to antiplatelet therapy with clopidogrel, and a uracil-2-13C labeled test to evaluate dihydropyrimidine dehydrogenase (DPD) activity responsible for catabolism of the chemotherapy agent, 5-fluorouracil. Each of these tests require only two breath samples, one prior to ingestion of the 13C-labeled substrate and one 20–50 minutes after ingestion, depending on the test.
In the case of the uracil-2-13C labeled test, “people with the enzyme will have 13CO2 in their breath after taking the substrate, and people without the enzyme will have no13CO2 in their breath because there’s no enzyme available and they won’t be able to metabolize the 13C uracil ingested,” explained Modak. “Patients who don’t have the enzyme shouldn’t take 5-fluorouracil because it’s a highly toxic drug. It’s best for physicians to know their patients status prior to initiating therapy.”
The Long Term Outlook
Numerous other breath test technologies are further away from clinical implementation but of intense interest to researchers. One, a focus of Barnes’s lab, involves exploring inflammatory mediators in exhaled breath condensate. “We’ve found it reliable for picking up certain substances like lipid mediators, and we use it to detect oxidative stress in the lung,” he explained. “However, we’ve encountered some problems with variability which relate to dilution because of the water vapor in breath, and we’ve been trying to find a way to correct for that dilution.”
Another hot area of breath research—an interest of Dweik’s—involves electronic noses. “These are a compilation of sensor arrays—16, 32, 64 or even more sensors—that react to exhaled breath,” said Dweik. “They can be used to identify a pattern, as in lung cancer, without knowing exactly what compounds are in that pattern.” Proof-of-concept studies have reported using this technology to detect acetones, glucose, melanoma, and lung cancer.
Dissemination of MS has promoted breath research, but many in the field, Risby included, believe breath testing will burgeon as MS and other spectrometry methods go the way of the microchip and become smaller and nearer to the patient, if not POC, allowing for real-time results and even at-home self-monitoring. “Very often big mass spectrometers have been used for breath analysis, but I don’t see that as the future of breath analysis. I think it’s going to be where we take the instrument to the patient with small hand-held devices,” he predicted. Risby currently is investigating quantum cascade lasers and mid-infrared sensors that one day might enable the type of miniaturization he envisions.
Dweik believes the field will take off in earnest when broader teams of scientists—including medical laboratorians—begin collaborating on specific projects, an idea he’s been promoting. “We have chemists and physicists working on these great sensors but they don’t have a great concept of how they would work in practice, and at the same time, physicians and other practitioners are looking for tests and devices to help them better monitor patients,” he observed. “They haven’t been collaborating too much, but if we get them together upfront we’ll see some real progress. And that’s starting to happen.”
For Further Information:
Cao W, Duan Y. Breath Analysis: Potential for Clinical Diagnosis and Exposure Assessment. Clin Chem 2006;52:800–11.
Mashir A, Dweik RA. Exhaled breath analysis: The new interface between medicine and engineering. Adv Powder Technol 2009; 20:42–25.
Modak A. Breath biomarkers for personalized medicine. Personalized Medicine 2010;7:643–53.
Paschke KM, Mashir A, Dweik RA. Clinical applications of breath testing. F1000 Medicine Reports 2010;2:56.
Solga SF, Risby TH. What is Normal Breath? Challenge and Opportunity. IEEE Sensors Journal 2010;10:7–9.