Abstract
Exhaled nitric oxide fraction (FeNO), which is a reliable marker of eosinophilic airway inflammation, is partially suppressed by tobacco smoking. Consequently, its potential as a biomarker in asthma management has never been evaluated in smoking patients. In the present study, the authors tested the validity of FeNO to predict asthma control in this population.
FeNO and the Asthma Control Questionnaire (ACQ) were recorded at least once in 411 nonsmoking (345 with at least two visits) and 59 smoking (51 with at least two visits) asthma patients.
Despite similar mean ACQ scores (1.5 versus 1.7), FeNO was reduced in smoking asthmatics (18.1 ppb versus 33.7 ppb). A decrease in FeNO of <20% precludes asthma control improvement in nonsmoking (negative predictive value (NPV) 78%) and in smoking patients (NPV 72%). An increase in FeNO <30% is unlikely to be associated with deterioration in asthma control in both groups of patients (NPV = 86% and 84% in nonsmoking and smoking patients, respectively).
It is concluded that, even in smokers, sequential changes in FeNO have a relationship with asthma control. The present study is the first to indicate that cigarette smoking does not obviate the clinical value of measuring FeNO in asthma among smokers.
Although the debate is not over, it is generally accepted that the exhaled nitric oxide fraction (FeNO) has the potential to be useful in the management of asthma 1–6. However, several factors confounding FeNO measurement have been recognised 7. Among them, tobacco smoking has been consistently shown to reduce FeNO levels 6, 8–15, by a factor varying from 0.63 to 0.80 according to the multivariate analyses that have compared FeNO in smokers and nonsmokers 6, 14, 15. The mechanism by which smoking causes FeNO reduction is not fully understood, but may include reduction in nitric oxide (NO) synthesis due to feedback inhibition induced by high concentrations of NO contained in cigarette smoke 9. NO oxidation or interaction with other molecules present in tobacco smoke might also occur 16. However, regardless of the mechanism of FeNO reduction reported in smokers, it is generally assumed that FeNO should not be assessed in asthmatic patients who smoke. Perhaps, consequently, this population (∼25% of adult asthma patients 17) has been excluded from clinical trials that have explored the potential of FeNO as a biomarker in asthma management. Even in the present authors’ own studies, where it has recently been shown that FeNO is a reliable marker of asthma control over time in unselected patients, smoking patients were not enrolled 18. Interestingly, the present authors’ previous study 18 strongly suggested that it is the change in FeNO values, rather than absolute cut-off points (i.e. individualised FeNO profiles), that may be meaningful for the longitudinal assessment of asthma control in daily practise. Therefore, in the present study, the authors investigated whether, despite the FeNO reduction reported in smoking asthma patients 6, 8–15, changes in FeNO might also be significantly related to changes in asthma control in this population.
To do this, FeNO was monitored on several occasions in smoking and nonsmoking patients attending a tertiary asthma clinic. Its ability to reflect improvement or worsening of asthma control over time was compared in both groups, using the Asthma Control Questionnaire (ACQ) as a gold standard for the assessment of asthma control 19.
METHODS
Subjects
Between January 1, 2004, and July 30, 2008, 411 adult nonsmokers and 59 adult smokers attending the Allergy and Asthma clinic in the Chest Dept of the Erasme University Hospital, Brussels, Belgium, for treatment of persistent asthma diagnosed according to standard criteria 20 were enrolled in the study.
The present study was approved by the local ethics committee, and patients signed to give informed consent.
Study procedures and design
Study design
The present study is a post hoc analysis of an existing database that is continuously updated. A significant part of the database was reported in the present authors’ previously published study, which documented a relationship between asthma control and FeNO in nonsmoking asthma patients 18. The present analysis focuses on whether current smoking annuls the validity of FeNO measurements to predict asthma control. ACQ scores and FeNO were recorded independently on one or more occasions for each patient, including smokers who were excluded from the initial analysis for reasons mentioned earlier. At each visit, asthma treatment was adjusted according to the Global Initiative for Asthma (GINA) guidelines recommendations 20, regardless of ACQ score or FeNO value, which were recorded separately.
Since optimal asthma control appears more difficult to achieve in smoking patients 21, the 1.5 optimum cut-off point identifying poorly controlled asthma 22 was selected as the reference ACQ score in the receiver operating characteristic (ROC) curve analysis. For clarity's sake, it was considered that an ACQ score of <1.5 identified asthma that is controlled (i.e. partly or well controlled), whereas an ACQ score of ≥1.5 identified uncontrolled asthma. Using the ROC curve analysis, the present authors assessed the ability of FeNO to: 1) reflect asthma control cross-sectionally using an ACQ threshold of 1.5; 2) detect a significant improvement or worsening of asthma control that resulted in a change from uncontrolled (ACQ ≥1.5) to controlled (ACQ <1.5) asthma, either respectively or vice versa; and 3) detect a significant improvement or worsening of asthma control defined as a decrease or increase in ACQ of ≥0.5, even though it was not large enough to result in a change of the asthma control status.
Patients treated with low (≤500 µg beclomethasone diproprionate (BDP) eq·day−1) and high-to-moderate (>500 µg BDP eq·day−1) inhaled corticosteroid (ICS) doses were considered separately. Indeed, in the present authors’ previous study 18, it was found that the overall ability of FeNO to reflect asthma control was reduced in patients using high ICS doses.
Study procedures
Asthma Control Questionnaire
Asthma control was assessed using a French translation of the short version of the ACQ of Juniper et al. 23. This version does not include FEV1 rating. Patients subjectively evaluate the degree of impairment caused by their asthma during the preceding 7 days by responding to six questions using a seven-point scale: a score of 0 indicates no impairment and a score of 6 indicates maximal impairment. The total ACQ score is the mean of the six responses, thus varying between 0 (totally controlled asthma) and 6 (totally uncontrolled asthma). A score of >1.5 is used to identify poorly controlled asthma 22. A change of 0.5 in the ACQ score is considered to be the minimum change that is clinically relevant 22.
FeNO
FeNO was measured before any forced expiratory manoeuvres using a daily calibrated LR 2000 chemoluminescence analyser (Logan Research Ltd, Rochester, UK) with online measurement of a single exhalation at a flow rate of 50 mL·s−1 (American Thoracic Society (ATS)/European Rspiratory Society (ERS) standard) 24. FeNO levels were read at the plateau corresponding to 70–80% of the carbon dioxide curve. Absolute FeNO values are expressed in ppb, and changes in FeNO are expressed as a percentage of the initial value (Δ%).
Statistical methods
ROC curve analysis was performed in the whole population, as well as in the two different subgroups, i.e. patients treated by low and high-to-moderate doses. The area under the ROC curve (AUC) was computed, and its difference from 0.5 was statistically evaluated. For a given type of assessment, the optimal cut-off value was determined for the whole population by maximising the Youden's index 25, i.e. the true-positive rate (sensitivity) minus the false-positive rate (1-specificity; see online supplementary material). Geometrically, this index is the vertical distance between the ROC curve and the first bisector. The cut-off value corresponding to the maximum value of Youden's index was then used to derive sensitivity (Se), specificity (Sp), positive predictive values (PPV) and negative predictive values (NPV), and accuracy in the whole population and in the subgroups of patients. In the online supplementary material, Se, Sp, PPV, NPV and accuracy may be found for other cut-off values, as well as the amounts of true-positive, true-negative, false-positive and false-negative cases (contingency tables).
Unpaired t-tests were used when considering forced expiratory volume in 1 s (FEV1) and log-transformed FeNO values, and Mann–Whitney U-tests when considering ICS doses and ACQ scores. Proportions were compared using a Chi-squared test. The limit of significance was a p-value of 0.05.
RESULTS
Of the 411 nonsmoking patients and 59 smoking patients, 345 and 51, respectively, were seen at least twice, representing 646 and 92 pairs of successive visits for nonsmoking and smoking patients, respectively (the median time between two visits for nonsmoking patients was 88 days, range 10–1,255 days, interquartile interval 42–189; and for smoking patients it was 93 days, range 7–525 days, interquartile interval 49–182). Table 1⇓ presents demographic data for the nonsmoking and smoking groups, as well as FeNO, FEV1, ACQ score and ICS dose values at study onset for the total studied population (included in the cross-sectional analysis) and for the subgroup of patients who were seen at least twice (included in the longitudinal analysis).
Demographic data and indices values at study onset
Tables 2⇓–⇓⇓5⇓ display the cut-off values (resulting from Youden's index maximisation), the number of positive and total cases, and therefore the prevalence, the Se, the Sp, the PPV and NPV, the accuracy and the p-value allowing rejection (or not) of the null hypothesis AUC = 0.5.
Cross-sectional assessment of asthma control
Assessment of a change from uncontrolled(Asthma Control Questionnaire (ACQ) score of ≥1.5) to controlled (ACQ score <1.5) asthma
Assessment of a change from controlled(Asthma Control Questionnaire (ACQ) score <1.5) to uncontrolled (ACQ score ≥1.5) asthma
Assessment of asthma control improvement(change in Asthma Control Questionnaire (ΔACQ) <-0.5)
Cross-sectional assessment of asthma control
Asthma control was cross-sectionally assessed at study onset for 411 nonsmoking and 59 smoking asthma patients. Controlled asthma (ACQ score of <1.5) was considered as a positive event. Table 2⇑ shows that, in smoking asthma patients, FeNO is unable to cross-sectionally assess asthma control.
Assessment of change in asthma control between pairs of visits
Change from uncontrolled (ACQ score ≥ 1.5) to controlled (ACQ score <1.5) asthma
In nonsmoking and smoking patients, asthma was uncontrolled at visit 1 in 283 pairs (out of 646) and 52 pairs (out of 92), respectively. A change to controlled asthma (spontaneous as well as treatment-induced) at visit 2 was considered as a positive event. In nonsmoking and smoking patients, this was observed on 133 and 17 occasions, respectively.
Table 3⇑ shows that FeNO exhibits high operating characteristics in both nonsmoking and smoking groups. The cut-off values for decreases in FeNO which had the highest NPVs for establishing control were 30% in nonsmokers and 20% in smokers.
Change from controlled (ACQ score <1.5) to uncontrolled (ACQ score ≥1.5) asthma
In nonsmoking and smoking patients, asthma was controlled at visit 1 in 360 pairs (out of 643) and 40 pairs (out of 92), respectively. A change to uncontrolled asthma at visit 2 was considered as a positive event. In nonsmoking and smoking patients, this was observed on 65 and 10 occasions, respectively.
Table 4⇑ shows that FeNO exhibits high operating characteristics in both nonsmoking and smoking groups. The cut-off values for an increase in FeNO, which had the highest NPVs for a change to uncontrolled asthma was 50% in both nonsmokers and smokers.
Improvement of asthma control (ΔACQ <-0.5)
A significant improvement in asthma control between two consecutive visits was considered to be a positive event. As a whole, in nonsmoking and smoking patients, this occurred on 257 and 40 occasions, respectively.
Table 5⇑ shows that, in the entire population, FeNO exhibited similar operating characteristics in nonsmoking and smoking patients. Figure 1⇓ illustrates this feature.
Receiver operated characteristic curve characterising the ability of exhaled nitric oxide fraction (FeNO) to assess an improvement of asthma control defined as a significant Asthma Control Questionnaire (ACQ) score decrease (ΔACQ score of >0.5) between two consecutive visits. FeNO exhibits similar operating characteristics in both populations. ––––: nonsmoking patients; ------: smoking patients.
When considering the subgroup of smoking patients treated with >500 μg eqivalents BDP·day−1, FeNO was no longer significant in assessing an improvement of asthma control.
Worsening of asthma control (ΔACQ >0.5)
A significant worsening of asthma control between two consecutive visits was considered a positive event. As a whole, in nonsmoking and smoking patients, this occurred on 161 and 26 occasions, respectively.
Table 6⇓ shows that, as for improvement assessment, FeNO exhibited analogous operating characteristics in nonsmoking and smoking patients. With a cut-off value at 30% change, a high NPV was observed in both groups. When considering the subgroup of smoking patients treated with >500 μg eqivalents BDP·day−1, FeNO operating characteristics in assessing asthma control worsening are less significant.
Assessment of asthma control worsening(change in Asthma Control Questionnaire (ΔACQ) >+0.5)
In both improvement (table 5⇑) and worsening (table 6⇑) assessment of asthma control, the present authors considered a subgroup of pairs of visits with an initial ACQ score of <2 as well as a subgroup of pairs of visits without ICS dose modification. Overall, FeNO characteristics were found to be only mildly affected compared with the total group.
DISCUSSION
The present study confirms that, compared with nonsmokers, FeNO is reduced in smoking asthma patients. However, this reduction does not appear to suppress its ability to reflect asthma control in smoking patients, provided changes in FeNO values detected by repeated measurements are considered.
FeNO is a reliable marker of eosinophilic airway inflammation 26 and has the potential to be useful in the management of asthma 1–5. However, tobacco smoking, which affects around 25% of the asthma population 17, leads to a decrease in FeNO 6, 8–15 and is considered to be a confounding factor. Therefore, it is generally assumed that FeNO should not be assessed in asthmatic patients who smoke.
At first glance, the present results seem to support this common paradigm. FeNO levels were in fact substantially reduced in smoking compared with nonsmoking asthma patients and to an extent that is similar to that found in previous studies 6, 14, 15. Furthermore, whereas a single FeNO value was confirmed to be significantly related to asthma control in the nonsmoking population (i.e. FeNO level >50 ppb indicates uncontrolled asthma in most cases 18), such a relationship could not be found in the smoking population.
However, in the present authors’ previous study 18, which involved nonsmoking patients, it was shown that sequential FeNO assessments are more useful than isolated measurements in demonstrating asthma control. In the current study, this was also found to hold true for smoking asthma patients. Indeed, repeated FeNO measurements do appear helpful with regard to indicating change in asthma control over time in both populations. Thus, when asthma is uncontrolled in nonsmoking patients, an FeNO reduction of >30% would predict that asthma is controlled in two out of three cases. The degree of change in FeNO that should be considered for smoking patients is different to that for nonsmoking patients: an FeNO reduction of <20% would indicate that asthma remains uncontrolled in most cases. Conversely, when asthma is controlled, an FeNO increase of <50% would indicate that asthma remains controlled in either population.
The aim of asthma treatment is to achieve full asthma control (i.e. an ACQ <0.75). In smoking patients, however, optimal control is usually more difficult to achieve 21, 27, 28; this is most likely to be due to the reduction in anti-asthma drugs effectiveness that has recently been documented in this population 27, 28. The present study confirms this: well-controlled asthma (ACQ score <0.75) was achieved in only 15% of smoking patients compared with 33% in nonsmokers (p<0.001; data not shown). Treatment adjustments resulted in asthma that could no longer be considered poorly controlled in as many as 33% of smoking patients (data not shown). For this reason, the present authors felt that an ACQ cut-off score of 1.5 (which identifies poorly controlled asthma) was more appropriate for the present data analysis and, thus, selected it for the current study. As this level of control was achieved in only 33% of patients in the present study, the authors also considered the ability of FeNO to detect any significant improvement in asthma control 22. In this respect, repeated FeNO assessments again appear to be helpful in both populations: in most cases, an FeNO reduction of <20% indicates that no significant improvement in asthma control has occurred. Conversely, FeNO increases of <30% are helpful to rule out mild deteriorations of asthma control. The results of the present study may be summarised as follows: if FeNO does not change as indicated, the level of asthma control is not modified. This seems to remain true regardless of whether the initial ACQ score was high or low, and whether or not the ICS dose was modified.
Interestingly, when patients were treated with high-to-medium ICS doses, FeNO no longer had the ability to reflect an improvement in asthma control for smoking patients, whereas for nonsmoking patients, its ability was only slightly reduced. A similar trend was observed with respect to asthma control deteriorations. These results confirm the overall reduction of the ability of FeNO to reflect asthma control in patients treated with high-to-medium ICS doses, as was documented by the present authors in their previous study 18. In addition, it appears that confounding factors, such as high ICS doses 29, 30 and tobacco smoking 6, 8–15, which are known to reduce FeNO, would have a cumulative interfering effect that may eventually suppress the ability of FeNO to reflect asthma control. This suggests that the effect of these confounding factors might have to be taken into account when using FeNO to assess asthma control. This needs to be clarified by appropriately designed studies.
In conclusion, the present study is the first to indicate that cigarette smoking does not obviate the clinical value of measuring FeNO in asthma. Indeed, it is shown that even in smokers, sequential changes in FeNO have a relationship to asthma control. The results also suggest that factors, such as smoking and inhaled corticosteroid dose, act cumulatively to influence the ability of FeNO to be used to assess asthma control. Overall, the importance of sequential FeNO measurements in both smokers and nonsmokers is to distinguish whether or not ongoing changes or a sudden change in respiratory symptoms are/is due to changes in airway inflammation, possibly requiring a change in anti-inflammatory therapy. The present data provide evidence that enables the magnitude of changes in FeNO to be more accurately interpreted when addressing this important question.
Support statement
AstraZeneca provided a grant for the exhaled biomarker laboratory.
Statement of interest
A statement of interest for this study can be found at www.erj.ersjournals.com/misc/statements.dtl
Acknowledgments
The authors would like to thank E. Juniper (McMaster University, Hamilton, ON, Canada) for kindly allowing the authors to use the Asthma Control Questionnaire. The authors also thank M. Demosmaeker, M. Danschutter and J-P. Storms for technical assistance and D. Young (Young & Associates Ltd, London, UK) for advice.
Footnotes
This article has supplementary material accessible from www.erj.ersjournals.com
- Received October 13, 2008.
- Accepted January 3, 2009.
- © ERS Journals Ltd
References
