Critical issues for breath analysis

Abstract

The current resurgence of interest in breath analysis follows directly from the ingenuity and foresight of the researchers in the area of nitric oxide, the gas that was named molecule of the year in 1992 [1]. Pioneering studies have defined the physiologic role of nitric oxide including its involvement in reactive airways. Understanding this role in relation to the mechanism of asthma has led to the establishment of international task forces that developed recommended procedures to quantify nitric oxide in vivo [2–4]. These recommended protocols have spurred interest in breath nitric oxide and there were almost 1000 publications during the year 2007 that involved measurement of this gaseous radical. Despite terrific promise of breath analysis and the fact that there have been literally thousands of articles published during the last hundred years, only a handful of breath tests are used clinically (table 1) with a few others used for research purposes.Table 1. Clinical breath tests. Breath carbon dioxide test or capnography Breath carbon monoxide test for neonatal jaundice Breath hydrogen test to detect disaccharidase deficiency, gastrointestinal transit time, bacterial overgrowth, intestinal statis Breath nitric oxide test for asthma therapy Breath test for detection of heart transplant rejectionMost of these tests require that breath be collected under carefully controlled conditions that include careful monitoring of the subject's breathing. Capnography—real-time measurement of carbon dioxide—is the most widely used clinical breath test and provides important information on cellular metabolism, and pulmonary ventilation, critical to monitoring the 'well-being' of a patient. The remaining tests are less widely used. A common feature to the tests listed in table 1 is that most are based upon real-time measurement of the breath biomarker. The subsequent discussion presents some of the steps that researchers in the field of breath analysis must take to ensure that breath analysis is generally accepted by clinicians and researchers. Breath analysis is not normally used clinically for a number of reasons. Most breath analyses have been performed using typical analytical chemistry instrumentation that was not designed for breath analysis. This has hindered both research and, ultimately, enthusiasm for clinical applications. Custom-designed breath monitors that are portable and are easy to use would be a significant advance. There are also multiple potential confounders unique to breath analysis, including contamination from ambient air, interference from other molecules, a subject's cardiopulmonary status, tobacco use, and so on. Without careful attention to these factors, breath analysis research can fall victim to the cliché 'garbage in/garbage out'.Ideally, standard protocols should be generated for the collection and analysis of single breath samples, for the collection and analysis of end-tidal breath samples, and for the analysis of breath samples collected during constant tidal breathing. Also guidelines should be generated for methods of breath collection that involve breath holding. The development of these guidelines will allow breath samples collected and analyzed in different laboratories to be compared and contrasted. In this same vein a standard commercial gas mixture should be developed that includes all the major species that are found in exhaled human breath. Analysis of this gas mixture should be included whenever a novel or improved method of breath analysis is submitted for publication. At this time it is reasonable to propose that the method for the collection of a single breath could be the same as the standard method for on-line breath nitric oxide analysis [2–4]. If the method does not involve the use of a real-time monitor for the biomarker then an end-tidal concentration of the carbon dioxide should also be monitored. In general when breath is collected under any sampling protocol, mouth pressure and carbon dioxide should be monitored continuously. The profile of carbon dioxide will define the quality of the breath sample, and the variation of mouth pressure with breathing cycle will demonstrate that the patient is maintaining a tight seal at the mouth and is exhaling through the mouthpiece. In addition, volume flow should be monitored continuously during tidal breathing.The composition of the inspiratory air will contribute significantly to breath analysis. Many molecules that are important for clinical diagnosis can also be present in the ambient environment. Currently, there is no consensus for a standard method to allow the background levels to be subtracted. At least part of the reason for this deficiency is the fact that there are no data that define how long it takes for a subject to reach steady state with his or her ambient environment. It has been suggested the lung can be washed out in approximately four minutes if a subject breathes pure air [5]. However, the washout of the entire body may take days or weeks depending upon the identity of the molecule. Similarly, the body may take a significant time to reach steady state with the composition of inspiratory air. At this time there are few toxicokinetic models that can predict the disposition of molecules present in inspiratory air [6, 7]. Until robust models are generated, a sample of inspiratory air should be collected and if the concentrations of biomarker molecules in the inspiratory air are greater than 25% of the concentrations in breath, then the analytical data should be treated with caution. This limitation is proposed since the study subject may not be in steady state with his or her environment and the resulting analysis will have a significant error. Guidelines should include a definition for the way the results of breath analysis are expressed. These guidelines will allow intrasubject and intersubject breath analyses to be compared and contrasted. Breath analysis for single breath samples could be expressed in terms of concentration units that are dimensionless (i.e., parts per million, etc) or in terms of moles per unit volume (pmol/l). Alternatively, single breath analyses could be normalized to a physiological based parameter such as carbon dioxide (i.e., pmol/ml of CO2). Normalization to carbon dioxide allows breath analysis data for subjects with widely different body masses to be compared. This latter method of data expression should definitely be used for reporting analysis of breath collected after breath holding [8]. Collection of breath during tidal breathing presents additional problems, since when human subjects are asked to breathe normally they tend to hyperventilate. Hyperventilation will change the distributions of molecules across the alveolar–capillary junction with time. Hyperventilation during breath collection can be prevented by requiring the subject to breathe at a constant defined rate (10 breath/min) and at a constant tidal volume based upon height and body weight [9]. Breath collected during tidal breathing will provide the average composition of all the breaths sampled. The resulting breath analyses can be normalized to minute ventilation per body mass (pmol/kg min), normalized to minute ventilation per body surface area (pmol/m2 min) or normalized to carbon dioxide production (pmol/ml of CO2). The latter method of normalization is preferred when cross-sectional breath data are compared for subjects with different body mass indices. This method of normalization assumes that there is no ventilation/perfusion mismatch. Finally, when standard techniques for breath collection and procedures for background correction have been adopted, then it should be possible to generate normal concentration ranges for diagnostic breath biomarkers as a function of gender, age, ethnicity, body mass index, pulmonary function, etc. These ranges will allow limits to be set that identify abnormal concentrations of breath biomarkers. Similarly, it will be possible to set limits for concentrations of breath biomarkers not normally present in breath so that breath biomarkers can be used to diagnose abnormal physiology, tissue injury or disease. Once this basic information on the molecules that have already been identified in breath has been obtained, then pioneering studies can be performed to identify new breath biomarkers of normal and abnormal physiologies. Clinical breath analysis remains in its infancy, despite the fact that its potential has been recognized since antiquity. Recent advances in instrumentation may enable more of this potential to be realized. In particular, the wider availability of real-time, portable monitors would be a breakthrough. Progress will require teamwork amongst device makers, experts in breath analysis and clinicians. The International Association for Breath Research (IABR) must play a critical role in this task and coordinate its efforts with the European Respiratory Society and the American Thoracic Society.