RESULTS

Section:

ChooseTop of pageAbstractMETHODSRESULTS <<DISCUSSIONReferencesCITING ARTICLES

Details of the subjects at study entry are shown in Table 1. There were no significant differences in markers of asthma severity between the 35 subjects for whom Borg scores were available and the 26 subjects who entered the budesonide trial, but did not have Borg scores recorded during the challenge tests. At baseline the subjects had moderately severe asthma, in that they had airway hyperresponsiveness to histamine in the moderate range, abnormal FEV₁, and significant breathing discomfort at rest, before the histamine challenge. All but three subjects had had asthma symptoms for more than 10 yr and all but six had received a diagnosis of asthma more than 10 yr ago. Half the subjects were taking inhaled corticosteroids at entry into the study, but as a group they had peripheral eosinophils and serum ECP above the normal ranges, suggesting that there was ongoing inflammatory activity. Of those subjects not taking ICS at study entry, only three had never taken ICS at any time in the past.

Effect of Budesonide Treatment on Perception of Breathlessness

Treatment with budesonide caused dramatic improvements in symptoms, lung function, airway hyperresponsiveness, blood eosinophils, and serum ECP. These treatment effects have been described in detail elsewhere (10). The improvements in all clinical and laboratory variables were significant and clinically substantial after 8 wk of treatment (Table 3). Subjects randomized to receive 3,200 μg of budesonide per day had more rapid remission of AHR than did the subjects taking 1,600 μg/d, but otherwise there were no significant differences between the groups with respect to changes in lung function, symptoms, or blood inflammatory markers.

Table 3.  MAGNITUDE OF CHANGE IN CLINICAL, LUNG FUNCTION, AND INFLAMMATORY VARIABLES BETWEEN BASELINE AND 8 wk, AND THE RELATION OF THESE CHANGES TO CHANGES IN THE PERCEPTION VARIABLES

		Change*		Slope				Intercept				PS20FEV1
				ρ†		p Value		ρ†		p Value		ρ†		p Value
Asthma score		−2.3 (−3.1, −1.5)		−0.15		0.39		0.45		0.008		0.33		0.055
FEV1, % predicted		15.1 (9.9, 20.4)		0.12		0.50		−0.56		0.0006		−0.48		0.005
DRR, doubling doses		−2.1 (−2.6, −1.6)		−0.30		0.09		0.07		0.68		−0.08		0.64
Eosinophils, %		−2.3 (−3.1, −1.6)		−0.29		0.10		0.33		0.06		0.13		0.47
ECP, μg/L		−4.45 (−7.5, −1.4)		0.02		0.90		0.07		0.7		0.12		0.50

^*Mean and 95% confidence intervals.

^†Spearman correlation coefficient (ρ).

Treatment with budesonide for 8 wk caused a significant increase in Borg/FEV₁ slope, from 0.09 (0.08–0.12) to 0.15 (0.11– 0.19) (p = 0.002), and a reduction in intercept, from 3.42 (2.77– 4.06) to 1.89 (1.33–2.44) (p < 0.0001). However, PS₂₀FEV₁ did not change significantly (5.66 [5.01–6.31] and 5.39 [4.76–6.02], p = 0.37). Figure 1 shows mean stimulus–response curves at baseline and after 8 wk of budesonide treatment constructed using individual values for slope and intercept to calculate values for each subject for the Borg score at 5, 10, 15, and 20% fall in FEV₁. There was a weak inverse correlation between the changes in Borg/FEV₁ slope and intercept (r = −0.31, p = 0.073). The magnitude of the changes in slope and intercept in the first 8 wk of treatment did not differ significantly between the two treatment groups. The mean (95% CI) change in Borg/FEV₁ slope was 0.06 (0.03–0.09) in the group taking 1,600 μg/d, and 0.07 (0.03–0.11) in the group taking 3,200 μg/d (p = 0.87).

Fig. 1. Stimulus–response curves during histamine challenge at baseline (circles) and after 8 wk of budesonide treatment (triangles) in 35 subjects with asthma. Values represent means and 95% confidence intervals of values for the Borg score at 0, 5, 10, 15, and 20% fall in FEV₁ calculated for each subject from their slope and intercept measurements.

Download Figure | Download Powerpoint

[More] [Minimize]

The significant determinants of changes in Borg/FEV₁ slope and intercept over the first 8 wk of treatment were estimated by correlation and multiple linear regression analysis. Table 3 shows correlation coefficients for the relationship between changes in the perception variables and changes in the clinical and inflammatory variables. Change in slope had a weak, nonsignificant negative correlation with change in DRR (p = 0.09) and change in circulating eosinophils (p = 0.10). However, in the multiple regression analyses neither of these factors was significant and no significant predictors of change in Borg/FEV₁ slope were found. Change in intercept had a highly significant correlation with the change in baseline FEV₁ and the change in asthma score. In the multiple regression analysis, change in intercept had significant negative associations with the change in resting FEV₁ (p < 0.0001), accounting for 26% of the variance, and with the change in DRR (p = 0.01), accounting for a further 11% of the variance (for the regression r² = 0.37, p = 0.0003). The determinants of change in PS₂₀FEV₁ were similar to those for intercept, with change in FEV₁ (p < 0.0001) accounting for 19% of the variance and change in DRR (p = 0.0002) accounting for 27% of the variance (for the regression r² = 0.46, p < 0.0001).

Multiple regression analyses were also performed to determine the predictors of perception at the 8-wk visit. The only significant predictor of slope was the DRR (for the regression r² = 0.21, p = 0.0053) and of intercept was the asthma score (for the regression r² = 0.21, p = 0.0045). There were no significant predictors of PS₂₀FEV₁.

Adaptation to Good Asthma Control

Twenty subjects completed all 72 wk of the study with perception, lung function, and inflammatory variables recorded throughout. At baseline these subjects did not differ significantly in any of the clinical, lung function, inflammatory, or perception variables from the 15 subjects who withdrew before completion of the study. Table 5 shows values for the clinical, lung function, and inflammatory variables at baseline, and 16, 48, and 72 wk after randomization. Downtitration of the budesonide dose commenced at Week 16. Repeated measures ANOVA was used to detect changes in variables during the period from 16 to 72 wk, and the results are shown in Table 5. Airway hyperresponsiveness and, to a lesser extent, the asthma score continued to improve over the course of the study, but there were no significant changes in any of the other variables during the 56 wk of budesonide downtitration, indicating that good asthma control was maintained over this period. Compared with baseline, Borg/FEV₁ slope remained significantly increased and the intercept significantly decreased at all subsequent visits, but there were no further significant changes in either index throughout the following 13 mo of treatment. Figure 2 shows stimulus–response curves generated from Borg/FEV₁ slope and intercept values at baseline and at 16, 48, and 72 wk of treatment.

Table 5.  CLINICAL, LUNG FUNCTION, INFLAMMATORY, AND PERCEPTION VARIABLES IN 20 SUBJECTS FOR THE BASELINE VISIT, AND VISITS 16, 48, AND 72 wk AFTER RANDOMIZATION

		Baseline*		16 wk*		48 wk*		72 wk*		p Value†
Asthma score		10.6 (10.1, 11.0)		5.9 (4.7, 7.1)		5.1 (4.0, 6.2)		4.45 (3.2, 5.7)		0.05
FEV1, % predicted		72.8 (64.5, 81.2)		88.5 (80.8, 96.2)		87.4 (79.3, 95.6)		86.5 (79.2, 93.8)		0.28
FEV1/FVC, %		67.9 (64.3, 71.5)		72.8 (69.5, 76.2)		73.0 (69.3, 76.7)		72.8 (69.3, 76.2)		0.94
Dose–response ratio,  % fall/μmol; +3		115.5 (65.5, 203.5)		16.4 (10.1, 26.6)		10.6 (7.5, 14.9)		10.6 (7.3, 15.2)		0.0005
Resting Borg score		3.4 (2.3, 4.5)		0.73 (0.4, 1.1)		0.75 (0.3, 1.2)		0.76 (0.3, 1.2)		0.91
Eosinophils, %		3.9 (2.9, 5.3)		1.7 (1.3, 2.3)		2.0 (1.4, 2.8)		2.5 (1.8, 3.5)		0.21
ECP, μg/L		13.8 (9.9, 19.1)		8.3 (6.0, 11.4)		9.5 (7.2, 12.5)		9.5 (6.7, 13.6)		0.80
Slope, Borg/FEV1 % fall		0.10 (0.08, 0.13)		0.18 (0.15, 0.22)		0.16 (0.11, 0.24)		0.18 (0.14, 0.22)		0.61
Intercept, Borg units		3.6 (2.6, 4.7)		1.3 (0.8, 1.7)		1.2 (0.8, 1.7)		1.2 (0.6, 1.8)		0.90
PS20FEV1, Borg units		5.8 (4.9, 6.7)		5.2 (4.3, 6.1)		5.2 (4.2, 6.1)		5.2 (4.1, 6.3)		0.99

^*Mean and 95% confidence intervals.

^†p Values are from repeated measures ANOVA on data at 16, 48, and 72 wk.

Fig. 2. Stimulus–response curves at baseline (solid circles) and after 16 (triangles), 48 (reverse triangles), and 72 (diamonds) wk of budesonide treatment, in 20 subjects who completed the full follow-up period.

Download Figure | Download Powerpoint

[More] [Minimize]

DISCUSSION

Section:

ChooseTop of pageAbstractMETHODSRESULTSDISCUSSION <<ReferencesCITING ARTICLES

This is the first study to show longitudinal changes in the perception of induced airway narrowing in subjects with asthma undergoing treatment with inhaled corticosteroids. The study has shown that within 8 wk of commencing treatment with high daily doses of budesonide there was an increase in the Borg/FEV₁ slope, indicating increased breathlessness sensitivity during airway narrowing. After treatment, the sensation of discomfort at rest was less, indicated by a decrease in intercept, but at 20% fall in FEV₁ the perception score (PS₂₀FEV₁) was the same as before treatment. We found no significant predictors of the increase in Borg/FEV₁ slope, but the decrease in intercept was explained primarily by increases in resting FEV₁ and AHR. Neither perception at baseline nor the changes in perception after treatment were related to daily ICS dose. Subjects with resting airway obstruction before treatment had lower slopes and higher intercept values than subjects without obstruction. However, there were no significant differences in the magnitude of changes in either lung function or perception variables between subjects with or without initial airway obstruction. After the initial 8-wk treatment period there were no further changes in perception during 1 yr of follow-up when the dose of ICS was downtitrated and good asthma control was maintained.

The perception measurements in these subjects were reliable. There were no significant differences between the measurements made at the beginning and end of baseline, when there were no changes in treatment or clinical status. This suggests that there was no learning effect due to the repetition of the measurement procedures. Furthermore, because the subjects commenced daily peak flow and symptom monitoring after the first visit in the run-in period, this finding suggests that it is unlikely that the monitoring itself altered perception by acting as a form of biofeedback. It is unlikely that other medications affected the perception measurements. All subjects were taking short-acting β-agonist aerosols as needed, but none were taking long-acting β-agonists or any other asthma therapy. The perception of airway narrowing is not affected by β-agonist aerosols (3). Perception measurements were obtained only from the subset of subjects recruited in the second half of the trial. These subjects did not differ in any significant way from the rest of the study group, and therefore the findings of this study are likely to be applicable to subjects meeting the inclusion criteria for the trial—that is, subjects with moderately severe, poorly controlled asthma.

In our study the mean PS₂₀FEV₁ value was consistently higher than has been reported in other studies (5, 16) and, at baseline, 25 subjects (71%) had values above 5, and could thus be defined as hyperperceivers by the criterion of Boulet and coworkers (16). These high values may have been due to the fact that the subjects recorded their Borg scores in response to the question “How is your asthma feeling right now?,” rather than to a question about the sensation of uncomfortable breathing, as used in other studies. In previous studies PS₂₀FEV₁ values have been obtained by direct interpolation of the stimulus–response curve (5, 16). Although the PS₂₀FEV₁ values in our study were obtained indirectly from the slope and intercept values this difference in methods is unlikely to explain the difference in findings, because evidence that the stimulus– response relationship is linear over this range (17, 18) suggests that they are comparable with values obtained directly. Alternatively, these subjects were volunteers for a clinical trial of inhaled corticosteroid and were recruited on the basis of symptom frequency, so there may have been some self-selection bias in favor of subjects with enhanced perception of symptoms.

Airway hyperresponsiveness was a significant predictor of Borg/FEV₁ slope both before and after treatment with budesonide, and the age of the subject made a small additional contribution at baseline. The regression model indicated that subjects with higher DRR values (more severe AHR) and greater age had lower slope values. This finding is consistent with those of a number of studies that have found that perception of increasing airway narrowing is reduced in subjects with more severe AHR (7, 8, 19) and in older subjects (20-23). Although the correlation between Borg/FEV₁ slope and circulating eosinophil numbers was of borderline significance, neither circulating eosinophil numbers nor eosinophil activation, as indicated by serum ECP levels, were significant predictors of Borg/FEV₁ slope, once DRR was taken into account. This suggests that the Borg/FEV₁ slope is more directly influenced by AHR than by inflammation. Alternatively, circulating eosinophils or eosinophil activation markers may be poor indicators of inflammatory events occurring within the airway or in the airway walls, which may be more important determinants of perception. Two previous studies have shown a negative relationship between perception and airway eosinophil numbers, measured either in bronchial biopsies (2) or in induced sputum (4). An association between Borg/FEV₁ slope and eosinophil markers cannot be excluded, however, because the sample size for the multiple regression was relatively small, and the failure of these factors to achieve significance may be due to a lack of statistical power.

The sensation of breathing discomfort at rest, indicated by the intercept value, was significantly related to asthma score both before and after treatment, with significant contributions from circulating eosinophil numbers and resting FEV₁. In the univariate analysis, change in intercept was significantly correlated with change in asthma score and change in resting FEV₁, but not with change in DRR. However, in the multiple regression analysis, after accounting for change in FEV₁, change in asthma score was no longer significant, but change in DRR had a strong negative association with change in intercept. The final regression model indicated that, as might be expected, increases in resting FEV₁ were associated with a decrease in breathing discomfort at rest. However, a decrease in the severity of AHR was associated with an increase in intercept. This observation is consistent with previous studies that have shown that perception is reduced in subjects with more severe AHR (7, 8, 19).

Resting airflow obstruction has been associated with reduced awareness of induced bronchoconstriction in several previous studies (6, 8, 9). This study confirms these findings by showing that slope is lower and intercept is higher in subjects with resting airflow obstruction than in those without. Burdon and coworkers (7) suggest that subjects with resting obstruction develop temporal adaptation to poor lung function. In the present study, treatment with ICS caused changes in both lung function and perception that were of similar magnitude in obstructed and nonobstructed subjects. This suggests that subjects who may have developed temporal adaptation to poor lung function are amenable to improvements in perception when their lung function is improved.

There have been no previous studies of the effects of a prolonged period of good asthma control on the perception of breathlessness. In nonasthmatic subjects single exposures to breathlessness induced by loaded breathing cause a reduction in the sensation of breathlessness induced several minutes later (24, 25). Prolonged exposures to breathlessness, during 4 wk of residence at high elevation, resulted in a reduction in the sensation breathlessness for at least 6 wk, but not as long as 6 mo, following a return to sea level (26). Before their enrollment in the present study, all subjects reported frequent symptoms of breathlessness and wheeze for at least 2 mo and the effects of these prior symptoms on subsequent perceptions of breathlessness might therefore have been expected to persist for longer than the 6 wk observed in the study by Wilson and coworkers (26). In fact, our study has shown that the changes in perception were maximal within 8 wk of commencing treatment, at a time when improvements in lung function and circulating inflammatory markers were also maximal. Despite further small, but significant, improvements both in AHR and in asthma score over the following 13 mo, there were no further changes in perception.

In this study we found that treatment with budesonide caused changes in Borg/FEV₁ slope and intercept, but not in PS₂₀FEV₁. Both slope (17) and PS₂₀FEV₁ (16) are commonly reported variables for measuring the perception of airway narrowing during bronchial challenge, but the findings of the present study suggest that they are not equivalent. It is possible that the slope and PS₂₀FEV₁ measure different components of perception, which may behave independently and may be affected by different physiological factors.

Corticosteroids may have effects on mood and attention (27), and on sensory nerve function (28). It is possible that ICS could affect perception by mechanisms unrelated to their anti-inflammatory activity. Although this possibility cannot be excluded, it seems an unlikely explanation for the changes in perception seen in the present study. There were no differences in perception between those subjects previously taking ICS, in whom the median dose was 800 μg/d, and those not taking ICS at the start of the study. However, this on its own is not compelling evidence against a direct effect of ICS on perception, because compliance with treatment may have been poor before the subjects entered the study, and subjects may have differed in other ways that may have had opposing effects. Stronger support comes from the fact that there were no differences in the magnitude of change in perception between the two treatment groups. The daily ICS dose was not a predictor of perception at baseline or after treatment, nor was treatment group a predictor of change in any of the perception variables. More importantly, there were no significant changes in perception over the final 12 mo of the study, when the dose of budesonide was downtitrated. The median reduction in ICS dose during the downtitration phase was 700 μg/d, a change that should have been substantial enough to alter any direct effect of ICS on sensory function.

The clinical relevance of these findings is not known and it is important to distinguish between the perception of induced airway narrowing and the perception of asthma symptoms, as judged by the relationship between spontaneous changes in lung function and symptoms scores recorded in daily diaries (29). Alterations in the ability of patients to perceive airway narrowing will almost certainly contribute to their ability to perceive symptoms in their day-to-day life, but symptom perception is also likely to be influenced by a number of factors, including personality and life style factors. The perception of symptoms is an indirect determinant of asthma treatment, particularly with symptom-based self-management plans. Subjects with enhanced perception of airway narrowing during treatment with inhaled corticosteroids may underestimate the amount of improvement obtained during treatment and be more inclined to abandon an effective treatment unnecessarily, or to use more rescue bronchodilator treatment than necessary. However, further studies are needed to determine whether the perception of induced airway narrowing is related to how patients perceive symptoms in a clinical setting (30).