Sputum eosinophils have shown some promise in this regard. They serve as an indirect metric of the extent of eosinophilic airway inflammation that, according to early studies, tracks with indexes of asthma control and maybe even predicts exacerbations in children with asthma.^3,4 Unfortunately, samples for sputum eosinophils are relatively difficult to collect, especially in children, because the hypertonic saline solution induction protocols are not well tolerated; moreover, the cytologic examination required is both relatively expensive and time-consuming.⁵ An alternative and perhaps better tolerated noninvasive diagnostic test for monitoring airway inflammation is exhaled nitric oxide (NO). Some studies have demonstrated a correlation between exhaled NO levels and eosinophilic airway inflammation. NO is produced abundantly in asthmatic airways as a result of the up-regulation of inducible NO synthase by proinflammatory mediators.⁷ However, exhaled NO, too, has its limitations, including an unclear relationship to risk for asthma exacerbation.^8,9

Beyond NO, the inflammatory milieu of asthmatic airways gives rise to a diverse array of oxidant species.¹⁰ Although these oxidants are generally too labile to measure directly, the molecular footprints they leave behind can serve as quantitative indexes of the precise inflammatory processes at play. For example, we have shown that eosinophils generate potent brominating oxidants, such as hypobromous acid, when recruited to asthmatic airways and activated by proinflammatory mediators.^11,12 On activation, eosinophils generate a respiratory burst and secrete granule proteins, including eosinophil peroxidase. We have shown that eosinophil peroxidase uses respiratory burst–generated hydrogen peroxide to form reactive brominating species such as hypobromous acid, a highly potent antimicrobial oxidant that can preferentially brominate protein tyrosine residues to form 3-bromotyrosine (BrTyr), a stable post-translational modification of proteins.^11,12 BrTyr can be measured noninvasively in urine as a relatively specific molecular marker of eosinophil activation (respiratory burst and degranulation). After oxidative modification (eg, bromination) of proteins, as airway remodeling occurs and the modified proteins are digested/degraded, the oxidized amino acids are efficiently removed by excretion in urine. In fact, in a recent clinical study of patients with asthma and healthy control subjects, we showed that urinary BrTyr serves as a noninvasive marker that is higher in patients with asthma and associated with multiple spirometric parameters of airway obstruction.¹³

In this study, we sought to investigate the potential clinical utility of urinary BrTyr as a sensitive noninvasive marker for both assessing the degree of asthma control and the risk of subsequent acute decompensation of chronic asthma control in pediatric participants. In parallel, the ability of exhaled NO, a distinct noninvasive marker of airway inflammation, along with traditional blood and spirometric measures of asthma control, were examined for their ability to predict asthma exacerbations over a near term (6-week) follow-up period.

Participants were recruited consecutively from the Sections of Pediatric Pulmonology and Pediatric Allergy and Immunology at the Cleveland Clinic if they were aged 5 to 21 years and diagnosed with asthma by a pediatric pulmonologist or allergist. All subjects enrolled were considered to be at baseline status and were being seen as part of a regularly scheduled clinic visit. Subjects considered to be having an exacerbation at clinic visit were not included in this study. Additional exclusion criteria included comorbid lung disease or inflammatory conditions, cystic fibrosis, diabetes mellitus, any smoking history, and current immunosuppressive therapy other than corticosteroids for asthma. Only three of the participants were receiving oral corticosteroids at their initial baseline visit. Recruitment occurred between July and September 2006, and follow-up data were collected between August and November 2006.

Asthma control was assessed at the initial study encounter with the six-item Asthma Control Questionnaire (ACQ6), developed by Juniper et al,¹⁴ which is a validated metric of asthma control that has been administered to children.^15,16 Six weeks after the initial study encounter, participants (with parent/guardians) were readministered the ACQ6 by telephone and asked about any asthma exacerbations since the initial study encounter, defined a priori as an acute de-compensation of asthma control resulting in a visit to the emergency department, unexpected doctor’s appointment or phone call, oral or parenteral corticosteroid burst, or an acute onset of asthma-related respiratory symptoms per self-report. All clinicians involved with the care of study participants were blinded to ACQ6 results, urinary BrTyr levels, and exhaled NO measurements.

Spirometry was performed as part of routine clinical care during the initial study encounter with an automated spirometer, as described elsewhere,¹⁷ consistent with American Thoracic Society standards.¹⁸ When clinically indicated, spirometry was repeated after the administration of two puffs (180 μg) of albuterol. Exhaled NO was measured during the initial study encounter by an online method at a constant flow rate of 50 mL/s, consistent with American Thoracic Society standards.^19,20

Spot clean-catch urine samples were collected at the initial study encounter, from which levels of free BrTyr were determined by stable isotope dilution high-performance liquid chromatography with online electrospray ionization tandem mass spectrometry using [¹³C₆]-BrTyr as an internal standard, as described elsewhere.¹³ To adjust for variations in urinary dilution, BrTyr concentrations are reported as a ratio with urinary creatinine concentrations, which was determined by standard laboratory techniques.

Charts were reviewed for relevant medical history, current medications, and laboratory results on record, such as complete blood counts, serum IgE levels, and allergy skin tests that were performed as part of routine clinical care. Allergy skin test results were considered “positive” if the participant reacted to at least two of 16 allergens tested.

Urinary BrTyr and exhaled NO levels did not differ between males and females, who had median (IQR) levels of BrTyr of 0.07 (0.00–0.21) ng/mg and 0.25 (0.00–0.53) ng/mg creatinine, respectively, P = .12, and median (IQR) levels of NO of 8.4 (5.4–16.4) ppb and 11.0 (6.7–31.9) ppb, respectively, P = .27. NO, but not BrTyr, correlated directly with age (R = 0.38, P = .007); neither correlated with BMI percentile. Urinary BrTyr levels demonstrated no significant correlation with spirometric parameters; NO correlated inversely with FEV₁/FVC (R = −0.32, P = .04), but did not correlate with FEV₁ percent predicted or the percent change in FEV₁ after bronchodilator. Neither BrTyr nor NO correlated with any of the complete blood count parameters, including percent eosinophils, or IgE levels. BrTyr and NO did not correlate with one another (Table II).

NO, but not BrTyr, was higher in participants with positive allergy skin test results (atopic) than those with negative allergy skin test results (nonatopic). The median (IQR) NO level among atopic participants was 11.1 (6.1–25.9) ppb versus 6.3 (3.9–11.5) ppb among nonatopic participants (P = .03. The median (IQR) BrTyr level among atopic participants was 0.14 (0.00–0.47) ng/mg creatinine versus 0.00 (0.00–0.47) ng/mg creatinine among nonatopics, P = .39.

Urinary BrTyr levels demonstrated a striking and consistent direct correlation with ACQ6 scores at baseline (R = 0.38, P = .004) and follow-up (R = 0.39, P = .008) (Table I). In contrast, NO did not (R = 0.20, P = .17 at baseline; R = 0.27, P = .10 at follow-up). BrTyr further correlated directly with each of the six individual items on the ACQ6 at baseline and all but one (β₂-agonist use) of the items on the ACQ6 at follow-up. NO, on the other hand, did not correlate significantly with any of the individual items on the ACQ6 at baseline; at follow-up, NO correlated directly with only one item (activity limitation, R = 0.41, P = .01). Interestingly, eosinophil counts, in contrast to BrTyr levels, did not correlate with either baseline (R = 0.03, P = .88) or follow-up (R = 0.27, P = .26) ACQ6 scores.

We next examined whether baseline levels of either urinary BrTyr or exhaled breath NO were elevated among the subset of subjects who experienced an asthma exacerbation over the ensuing 6 weeks of follow-up versus those who did not experience an exacerbation. BrTyr levels were significantly higher among those destined to experience a near-term exacerbation (BrTyr median [IQR] 0.00 [0.00–0.18] vs 0.23 [0.02–0.68]; P = .02). In contrast, baseline levels of exhaled breath NO were comparable among those who experienced an exacerbation and those who did not (median [IQR] NO levels of 12.0 [5.7–16.9] vs 9.1 [6.7–22.9]; P = .71). To further assess the capacity of urinary BrTyr and exhaled NO to predict asthma exacerbations and indexes of asthma control, BrTyr and NO were dichotomized into high and low levels. High levels were defined as greater than their respective median levels; low levels were defined as less than or equal to their respective median levels. FEV₁ percent predicted, FEV₁/FVC, and percent change in FEV₁ after bronchodilator were also dichotomized by their respective medians. ACQ6 scores at baseline and follow-up were dichotomized as ≥1.5 versus <1.5, a cut point that previous research has deemed optimal to detect inadequately controlled asthma (positive predictive value of 0.87).²¹ Individual ACQ6 item scores were dichotomized as at least one versus zero, indicating that the participant, in the week preceding questionnaire administration, has woken up at night because of asthma, experienced asthma symptoms on waking in the morning, was limited in his/her activities because of asthma, experienced shortness of breath, wheezed, or used a short-acting β₂-agonist. Finally, the presence of an asthma exacerbation between baseline and follow-up was assessed as previously defined.

A high level of BrTyr at baseline conferred 18.1-fold the odds of inadequately controlled asthma at baseline, as defined by an ACQ6 score ≥1.5, than a low level of BrTyr (95% CI of 2.1–153.1, P = .0004). Moreover, participants with high levels of BrTyr were more likely to experience morning symptoms (OR, 95% CI of 3.0, 1.0–9.1; P = .04), activity limitation (4.7, 1.5–14.5; P = .005), shortness of breath (3.2, 1.0–9.9; P = .04), and wheezing (3.3, 1.0–10.7; P = .04) at baseline, as well as use a short-acting β₂-agonist (4.2, 1.3–13.0; P = .01) at baseline. High levels of BrTyr showed a trend toward experiencing nocturnal waking at baseline, but this failed to reach statistical significance (2.7, 0.8–9.2; P = .11) (Figure 1, A). Unlike BrTyr, a high level of NO at baseline did not confer significantly greater odds of inadequately controlled asthma at baseline, neither with respect to the composite ACQ6 score nor to any of the individual asthma symptom items (Figure 1, B).

Similarly, a high level of BrTyr at baseline conferred 5.0-fold the odds of inadequately controlled asthma at follow-up than a low level of BrTyr (95% CI 1.1–22.2, P = .02). Participants with high levels of BrTyr were also more likely to experience morning symptoms (OR, 95% CI of 3.8, 1.1–13.5; P = .03), activity limitation (4.6, 1.3–16.3; P = .02), and wheezing (4.3, 1.1–17.1; P = .03) at follow-up. High levels of baseline BrTyr showed a trend toward experiencing nocturnal waking (5.3, 0.9–29.1; P = .06) and shortness of breath (3.0, 0.9–10.4; P = .07). They were not, however, more likely to use a short-acting β₂-agonist at follow-up (Figure 2, A). Again, unlike BrTyr, participants with high levels of NO at baseline were not significantly more likely to have inadequately controlled asthma at follow-up, neither with respect to the composite ACQ6 score nor to any of the individual items (Figure 2, B).

Urinary BrTyr, exhaled NO, asthma control as monitored by ACQ6, and traditional asthma laboratory diagnostics were next examined for their ability to predict asthma exacerbation over the ensuing 6-week interval. Remarkably, of all blood, urine, and spirometric measures examined, only a high level of BrTyr predicted the presence of an asthma exacerbation by follow-up; participants with elevated BrTyr levels demonstrated a 4.0-fold increased likelihood of experiencing an exacerbation by follow-up (95% CI 1.1–14.7, P = .03) compared with subjects with low BrTyr levels (Figure 3).

Asthma is one of the most common chronic diseases in children, accounting for substantial morbidity rates and cost in terms of medical treatment and missed school days.²⁴ Asthma exacerbations are particularly problematic because they herald declining lung function independent of other risk factors.^25,26 Monitoring of asthma control is largely guided by clinical history, physical examination, and spirometry, not by direct assessment of underlying inflammatory processes. This poses an especially formidable problem in children with asthma, in which case the clinical history is often provided or supplemented by the caregiver, not the patient, and spirometric parameters frequently fall within normal limits.^27,28 This latter observation is corroborated in our cohort, because FEV₁ percent predicted, FEV₁/FVC, and the percent change in FEV₁ post-bronchodilator fell within normal limits in most of the participants in our study. Many children, especially younger ones, are unable to perform spirometry with adequate technique, substantially limiting the utility of spirometry in pediatric pulmonary practice.²⁹

The direct correlation that we observed between NO levels and age, as well as the higher NO levels present in children with atopic versus nonatopic asthma (as determined by allergy skin testing) corroborates the findings of others.^30,31 BrTyr levels, on the other hand, appeared robust to variations in age and degree of atopy, along with spirometric parameters, suggesting that its relationship to asthma control and capacity to predict exacerbations are independent of these factors. Furthermore, BrTyr and NO did not track with BMI percentile, which was higher, on average, in this cohort than in the general pediatric population. Adiposity has been associated with asthma in previous studies and has even been shown to track with asthma control (but not with exhaled NO levels) in select cohorts of children with asthma.^32,33

Urinary BrTyr levels correlated directly with ACQ6 scores at baseline and follow-up, along with all six individual item scores at baseline and five of six individual item scores at follow-up. Consequently, high BrTyr levels were seen to confer significantly greater odds of inadequately controlled asthma, including increased symptoms, activity limitation, and short-acting bronchodilator use at baseline, plus significantly greater odds of inadequately controlled asthma, including all of the above (save short-acting bronchodilator use) at follow-up. We did not observe such relationships with respect to NO levels and traditional spirometry- and blood- (IgE, eosinophil count) based metrics of asthma control.

High levels of BrTyr conferred 4.0-fold the odds of having an asthma exacerbation in the subsequent 6 weeks. Alongside spirometric parameters, ACQ6 scores, and NO levels, BrTyr levels were the only variable to demonstrate this predictive capacity. These findings strongly support the potential clinical utility of BrTyr as both an aide in objectively monitoring asthma control and predicting risk of decompensation and subsequent exacerbation in a noninvasive manner. They thus suggest that urinary BrTyr levels may potentially serve as a quantitative index with which to guide asthma therapy.

Our findings that exhaled NO, contrary to BrTyr levels, neither track with asthma control nor predict asthma exacerbations are not unique to this study. Previous studies in children that have investigated putative relationships among NO levels, asthma control, and subsequent exacerbations, have reported weak associations, at best.^8,34 Several studies looking at such relationships in adults, however, have found that NO does, indeed, track with asthma control and the risk of exacerbations.^35–37 Taken together, these findings suggest that (1) the inflammatory milieu may be somewhat different in adults versus children with asthma; (2) for purposes of monitoring asthma control and predicting exacerbations, exhaled NO is more effective in adults than children; and (3) BrTyr serves as a potentially more useful noninvasive marker than NO for monitoring asthma control and predicting exacerbations, especially in pediatric populations.

Even though we did not look specifically at sputum eosinophils in this study, previous studies have demonstrated that sputum eosinophils also track with asthma control and possibly even the risk of exacerbations in children.^3,4 Unfortunately, the potential utility of sputum eosinophils in everyday pediatric pulmonary practice is largely thwarted by cost and technical difficulties in both obtaining sufficient quantities of sputum in children and processing samples in the laboratory.⁵ Urinary BrTyr, on the other hand, can be obtained easily by spot (random) urine collection and quantified by a well-validated stable isotope dilution high performance liquid chromatography/tandem mass spectrometry assay. Furthermore, urinary BrTyr tracks with an even more specific inflammatory process than sputum eosinophils, that is, eosinophil activation per se. BrTyr production by eosinophils requires that the leukocytes are activated. This emphasizes the important distinction between eosinophil levels, as determined by blood or sputum eosinophil counts, and eosinophil activation, as determined by BrTyr levels. Detection of elevated urinary BrTyr requires that eosinophil activation with both respiratory burst and eosinophil degranulation has occurred, enabling significant protein oxidative injury within an inflamed asthmatic airway.^11–13 Perhaps it is not surprising, then, that BrTyr levels appear more closely aligned with asthma symptoms and risk of exacerbation than eosinophil count. Indeed, BrTyr levels do not correlate with circulating eosinophil levels in this study, a finding also seen in our previous study in adults with severe asthma.¹³

Although our findings present compelling support for the potential utility of urinary BrTyr to assess asthma control and risk of subsequent exacerbations, several limitations should be noted. The study lacked the power necessary to examine how the reported findings may vary depending on particular asthma phenotypes, such as severe versus nonsevere and atopic versus nonatopic asthma.³⁸ Most of the participants were being treated with corticosteroids, which may have attenuated the strength of the relationships observed. Exhaled NO levels, for example, are known to decrease with corticosteroid treatment.³⁹ Moreover, complete blood counts with differentials, serum IgE levels, and allergy skin testing results were accrued from chart review only on participants for whom these data were available and not necessarily at the time of the initial study encounter. This may, at least in part, explain why exhaled NO did not correlate with serum IgE despite previous findings to the contrary.¹³ We did not confirm asthma diagnoses with provocative challenge testing in all subjects; however, all of the participants had been diagnosed with asthma by physicians specializing in pediatric pulmonology or allergy and immunology by clinical history, physical examination, and spirometry, and all diagnoses were made before exhaled and urinary marker measurements. Finally, because participants for the study were recruited from specialty pediatric pulmonary clinics and follow-up data collected during allergy season, participants may have been at higher risk of having an asthma exacerbation than the pediatric asthmatic population at large. The findings suggest that this high-risk population may benefit from a predictor of asthma exacerbations such as BrTyr; however, a broader sample of participants recruited from primary care clinics or at varying times of the year is necessary to validate that findings are generalizable to the pediatric asthmatic population at large and not exclusively to patients at high risk.