How accurately should we estimate the anatomical source of exhaled nitric oxide?

FOR MORE THAN A DECADE it has been recognized that nitric oxide (NO) appears in the exhaled breath and the level is altered in numerous pulmonary diseases in which inflammation plays an integral role (e.g., asthma) (1, 8). Thus the exhaled NO signal has the potential to uniquely delineate the contribution of inflammatory processes to lung disease in a noninvasive manner complementing more traditional measurements of lung function, namely, lung volumes and expiratory airflow that focus solely on structural properties of the respiratory system. This potential, combined with the relative ease with which it can be detected and the serious and ongoing epidemic of asthma and other chronic lung diseases, has imbued the measurement of exhaled NO with the promise of becoming a useful clinical tool. However, this promise has not yet been fulfilled. Skepticism regarding the measurement of exhaled NO persists due to both its variability among clinically similar subjects and the inconsistent correlations with other indexes of lung function and symptoms. This skepticism is appropriate and stems both from the relatively crude techniques currently employed to characterize exhaled NO and from our still incomplete understanding of the fundamental biological mechanisms that determine the appearance of NO in the exhaled breath. Since the initial observation that NO appears in the exhaled breath, several research groups have made seminal contributions toward our understanding of the unique features of NO exchange in the lungs. In particular, exhaled NO has sources from both the airway and alveolar regions, which has been determined from a combined approach implementing experimental observations (5, 7, 15, 19) and “two-compartment” mathematical models (9, 14, 20, 21). The present clinical approach for exhaled NO measures the concentration during a vital capacity maneuver while holding expiratory flow and pressure constant (2). The recommended exhalation flow is low enough (50 ml/s) to cause the concentration to be predominantly of airway origin and is thus ineffective at describing the lower alveolar concentration of NO, ignoring this potentially important signal. Although asthma is traditionally thought to be an inflammatory disease of the airways, several groups have employed the two-compartment model of NO exchange and reported an elevated alveolar concentration of NO during periods of enhanced symptoms (11, 13), or in patients who are refractory to inhaled corticosteroids and bronchodilators (3, 6, 12). The observations of increased alveolar NO are particularly relevant as these patients have proven to be difficult to manage, are hospitalized more frequently, and could well benefit from early detection of disease exacerbation and alternate therapeutic regimens. Since the alveolar concentration cannot be directly measured, estimating the alveolar concentration requires a model of NO exchange in the lungs that, when combined with experimental measurements, can partition the exhaled NO signal into proximal and peripheral contributions. This feature is frequently neglected when the “alveolar concentration” or “airway flux” of NO is reported, and we are led to believe that these numbers are direct experimental measurements. This concept is not new to the physiology community. Other examples include the Fick method to determine cardiac output, and the measurement of lung diffusing capacity. Buried in these measurements are mathematical models approximating the physiology. In fact, the accuracy of these estimates depends not only on the accuracy of the model but also on the experimental protocol (i.e., algorithm) and instrumentation. It is therefore pertinent in our quest to interpret exhaled NO to consider the question: how accurately can the anatomic source of NO be estimated, and at what cost? In general, as both the computational complexity and accuracy of an algorithm and model increase, the ease of clinical translation decreases. The initial two-compartment models were extremely simple in structure, essentially describing NO gas exchange using a single expansile balloon (alveolar region) connected to a rigid tube (airways). The algorithms were equally simple involving linear fits of experimental measurements in which the slope and intercept reflected region-specific (i.e., alveolar) NO parameters (9, 14, 20, 22). While these early models were elegant in their simplicity and ability to explain the strong flow dependence of exhaled NO, they neglected potentially important physical and physiological phenomena such as axial (or longitudinal) gas phase diffusion, the trumpet shape of the airway cross-sectional area, and spatial heterogeneity in flow. Recently, more advanced models have been developed (18, 23) and validated with new experimental measurements (16, 17) demonstrating the importance of axial diffusion of NO. In particular, the airway source of NO is large enough to create an axial gradient in NO concentration that leads to diffusion of NO from the airway tree into the alveolar region (i.e., “backdiffusion”). In other words, the alveolar region can serve as a sink for airway NO; and conversely, NO from the airway tree can contaminate the alveolar region, leading to a falsely elevated estimate of the alveolar concentration. We recently quantified this potential effect and presented a simple method to account for axial diffusion of NO on the estimation of the alveolar concentration (4); however, we only tested the model in healthy subjects. Two new studies (10, 24) are presented in the Journal of Applied Physiology, both of which elegantly combine experimental measurements and a mathematical model of NO exchange that advance our knowledge and understanding of NO gas exchange and thus our interpretation of the exhaled NO signal. Kerckx and colleagues (10) have independently developed a similar method to account for axial diffusion of NO into the Address for reprint requests and other correspondence: S. C. George, Dept. of Biomedical Engineering, Univ. of California-Irvine, Irvine, CA 92595-2715 (e-mail: [email protected]). J Appl Physiol 104: 909–911, 2008; doi:10.1152/japplphysiol.00111.2008.