METHODS

Section:

ChooseTop of pageAbstractMETHODS <<RESULTSDISCUSSIONReferencesCITING ARTICLES

Participants were children aged 5–18 years with asthma attending two academic medical center subspecialty clinics for routine evaluation of asthma. Patients were in their usual state of asthma control and were excluded if they were being seen within 1 month of an emergency department visit or hospitalization for asthma or an exacerbation of asthma requiring systemic corticosteroids. After providing consent for participation in the study, parents completed questionnaires regarding their child's asthma symptom frequency (over the last month) and medication use (over the last week) (Figure 1)

Figure 1. Questionnaire completed by parent and child prior to physician visit.

Download Figure | Download Powerpoint

[More] [Minimize]

Spirometry was performed by all participants at least 4 hours after use of short-acting β-agonists and at least 12 hours after the use of salmeterol using spirometers, that conformed to American Thoracic Society standards. All participants were required to demonstrate the ability to perform reproducible lung function tests (FVC and FEV₁ within 5% reproducibility and PEF within 10% reproducibility). Details on the methods for making these measurements are provided in an online supplement. Percent predicted values were based on the regression equations of Wang and colleagues, which provides regression models for FVC % predicted, FEV₁ % predicted, FEV₁/FVC, and FEF_25–75 % predicted for each age, sex, and race (11). Lung function was considered abnormal if values were below the fifth percentile.

RESULTS

Section:

ChooseTop of pageAbstractMETHODSRESULTS <<DISCUSSIONReferencesCITING ARTICLES

We enrolled 219 children into the study. The mean age of study participants was 10.1 ± 3.4 years. Fifty-five percent were younger than 10 years of age. Sixty-five percent of participants were male. Sixty-seven percent were white, 29% African American, 2% Hispanic, and 1% were Asian.

Patients tended to report very good levels of asthma symptom control, with 68.1% of patients being classified as intermittent or mild persistent based on symptom frequency (Table 2)

Table 3

Although there were statistically significant differences in the mean FEV₁/FVC ratio and FEF_25–75 % predicted across the classifications defined by the more severe of symptom frequency and medication use, these measures of lung function were poor discriminators of severity classification. Discriminant analysis revealed that FEV₁ % predicted correctly classified 33% of patients, FEV₁/FVC ratio correctly classified 32% of patients, and FEF_25–75 % predicted correctly classified 39% of patients. However, with four possible severity classifications, 25% of participants would be correctly classified by random assignment. Thus, the predictive value of FEV₁ % predicted, FEV₁/FVC ratio, and FEF_25–75 % predicted for discriminating between individual patients is not much greater than could be attributed to chance.

We examined the effect of age on the relationship between lung function measures among children younger than 10 years of age and those 10 years of age and older (Table 4)

We examined the relationship of lung function parameters between levels of asthma severity by sex. Neither sex demonstrated a significant difference in FVC % predicted or FEV₁ % predicted by level of asthma severity (Table 5)

Figure 2

Figure 2. FEV₁ % predicted by level of asthma severity based upon the more severe of symptom frequency and medication use.

Download Figure | Download Powerpoint

[More] [Minimize]

Figure 3. FEV₁/FVC by level of asthma severity based upon the more severe of symptom frequency and medication use. *FEV₁/FVC was considered abnormal if values were below the fifth percentile for age.

Download Figure | Download Powerpoint

[More] [Minimize]

DISCUSSION

Section:

ChooseTop of pageAbstractMETHODSRESULTSDISCUSSION <<ReferencesCITING ARTICLES

In an attempt to understand further the levels of pulmonary function that correspond to the levels of asthma severity proposed by the NAEPP/EPR2, we examined lung function measures in 219 children seeking subspecialty care for asthma. Our findings do not correspond to the levels of lung function in the NAEPP/EPR2 Guidelines. Children with asthma seen in these subspecialty asthma clinics demonstrated, on average, excellent (and normal) levels of lung function (FEV₁ % predicted) at all levels of severity, including severe persistent asthma.

It is not surprising that there is little relationship between FEV₁ % predicted and the level of asthma severity when one considers the generally poor relationship between lung function and asthma symptoms. Several studies have examined this relationship and have demonstrated very weak correlations between FEV₁ and asthma symptoms in adults (14–16) and children (17, 18), although FEV₁ % predicted has been shown to predict future asthma exacerbations (19). Our findings corroborate the findings of a lack of association between asthma symptom frequency, rescue medication use, and FEV₁ % predicted. This suggests that other determinants of airway function, such as the ratio of FEV₁/FVC or the degree of airway hyperresponsiveness, may relate better to asthma symptoms than FEV₁. A significant association between asthma symptoms (both daytime and nocturnal) and methacholine responsiveness has been noted in children with mild–moderate asthma (20, 21).

The level of asthma severity used in this report was based on the more severe of two major components: asthma symptom frequency or medication use. Medication use tended to classify patients at a higher level of severity than symptom frequency (Table 2), likely reflecting a high level of asthma symptom control among these patients. A similar pattern was noted by Colice and colleagues (16), who categorized asthma severity by symptom frequency, β-agonist use, and FEV₁ in patients participating in run-in phases for clinical trials. These authors demonstrated that using the more severe of symptom frequency, β-agonist use, and FEV₁ resulted in more patients being classified as severe persistent. However, unlike Colice and colleagues, who found that nocturnal awakening was the primary feature that led to classification at a higher level of severity, we found that medication use (both β-agonist and controller) was the major contributing parameter toward higher severity classification, likely because of the care of the children in the subspecialty clinics.

There is no gold standard for determining asthma severity. We have used the NAEPP/EPR2 Guidelines for the Diagnosis and Management of Asthma (22) as the basis of our classification scheme. This is widely accepted as a method of classification of asthma severity and incorporates two major clinical features in the determination of severity–symptom frequency and pulmonary function. However, classification of asthma severity is complex and is influenced by the variability of disease severity within a patient over time as well as being confounded by current asthma treatment (23). The NAEPP Guidelines are most appropriate for use in patients not receiving controller therapy. The Global Initiative for Asthma Guidelines recognized the importance of including medication use (both controller and reliever medications) in determining asthma severity and have developed an algorithm for incorporating symptom frequency, pulmonary function, and medication use in determining asthma severity (24). We have used a variation in this approach (exclusion of measures of lung function as this was our parameter of interest) in our analysis to determine severity.

According to the NAEPP/EPR2 Guidelines, spirometry is intended to be used in determining asthma severity classification before the initiation of long-term controller medication. Given the substantial increase in use of controller medications in the past decade (25), the applicability of the guidelines for spirometry in severity classification is limited. This is especially true among subspecialists, as fewer patients present to subspecialists without prior use of controller medications. Furthermore, these results suggest that even if the FEV₁ levels suggested by the guidelines are useful in controller-naïve individuals, they are of little value in determining the level of asthma severity among individuals receiving controller medications. However, in two recent clinical trials, children with mild–moderate asthma and persistent asthma symptoms while not receiving controller medications had well-maintained FEV₁ (5, 17).

Cockcroft and Swystun have emphasized the differentiation between asthma severity and level of asthma control (26). Our approach is consistent with this distinction in that we have incorporated the use of asthma medications into our assessment of disease activity and thus severity. However, as this was a single contact study rather than a longitudinal assessment, we were unable to assure optimization of asthma therapy. Furthermore, the assessment of asthma control in the presence of asthma therapy may be more of a test of adequacy of control than a reflection of asthma severity.

Our finding of normal levels of FEV₁ in patients with asthma is consistent with the follow-up of the Williams and McNicol cohort (7) at age 28 years (8), where the subgroup with wheezing more frequently than once per week in the previous 3 months had FEV₁ significantly lower than control subjects, but still within the 95% confidence limits. FEV₁/FVC ratios in the subgroup with persistent asthma ranged from 64 to 71%, compared with 80 to 84% in the control subjects, with clear separation of those with persistent asthma from control subjects. In addition, the FEV₁/FVC ratio in those with persistent asthma was clearly in the abnormal range. The improved sensitivity of FEV₁/FVC ratios over FEV₁ alone is explained in part by the finding that FVC is significantly higher in children with asthma than in control subjects (27). This was true for both those children with frequent symptoms and those with episodic symptoms. Most important for the study of childhood asthma was the observation that the subgroup of adults with persistent asthma and abnormal FEV₁/FVC ratios at age 28 years had abnormal FEV₁/FVC ratios at age 10 years, 79% (mean with a range 76–83%) compared with 87% (mean with a range 85–89%) for the control subjects.

The FEV₁/FVC ratio also represents exaggeration of dysanapsis that occurs in asthma. Dysanapsis was first described by Green and colleagues and Mead as variability of maximal expiratory flow–volume curves among healthy adults (28, 29). Dysanapsis in the lung is a normal phenomenon caused by disproportionate growth of airways and lung parenchyma. This effect is exaggerated in asthma with airways (represented by FEV₁) smaller than the lung parenchyma (represented by FVC) (30). The extent of dysanapsis in asthma (as represented by abnormal FEV₁/FVC) is highly correlated with the degree of airway hyperresponsiveness (20). Alternatively, the elevated FVC may be related to a component of gas trapping in obstructed airways, although we did not examine other measures of gas trapping such as residual volume-to-total lung capacity ratio by plethysmography in this study. The finding of reduced levels of FEV₁/FVC among children 10 or more years of age with increasing asthma severity may be confounded by a longer duration of asthma among the older children, especially those with more severe disease. However, the mean ages of the participants at each level of asthma severity were similar (data not shown). Unfortunately, this study did not collect data on disease onset/duration to help explore this possibility. In addition, the ratio of FEF_25–75 %/FVC has also been suggested as a marker of dysanapsis and has been related to response to eucapneic hyperventilation with cold air (31) and airway reactivity to methacholine (32). The pattern of FEF_25–75 %/FVC in this group of children was similar to that seen with FEV₁/FVC—lower levels of FEF_25–75 %/FVC with increasing asthma severity in both sexes as well as in children 10 years of age and older.

The FEF_25–75 % predicted differed significantly among levels of persistent asthma when asthma severity was classified by either medication use or the more severe of symptoms or medication use, a pattern similar to that seen with FEV₁/FVC and in contrast to that seen with FEV₁ % predicted. However, the variability of the FEF_25–75 % predicted measure is nearly twice that seen with FEV₁ % predicted and FEV₁/FVC (as reflected by their respective standard deviations), complicating the use of this measure in clinical practice. The standard deviation of the FVC % predicted and FEV₁ % predicted measurements was greatest in the severe persistent category regardless of method of severity assignment, even though the means were not different. This might indicate that although these children do not have worse lung function on average, they are a more heterogeneous group. This reinforces the notion that isolated measures of lung function are not particularly reliable in distinguishing between levels of asthma severity, especially in severe cases, but that testing must be done repeatedly and values at individual visits be compared with the “best” values obtained during clinical wellness.

There are several limitations to be considered in the interpretation of our findings. Patients in this study were all seen in tertiary-care centers. Such referral patterns may bias toward a higher level of asthma severity, different levels of compliance with asthma medications, differing socioeconomic characteristics than the general asthma population, or use of more aggressive asthma management approaches. Thus, our results are applicable to children with asthma seeking subspecialty care and may not accurately reflect a random sample of childhood asthma. Severe persistent asthma did represent more than 40% of the patients in this study when severity was based on the more severe category of symptom frequency or medication use. It is possible that the patients referred to these centers were more adherent to their prescribed regimens and thus demonstrated higher levels of lung function than patients seen in other settings. These patients may have received more aggressive (or appropriate) therapy of their asthma leading to a higher percentage of patients classified with more severe asthma by the medication criteria. The majority of patients in this study (75%) were receiving at least one asthma controller medication. Only one patient with severe persistent asthma was not receiving an asthma controller medication. Furthermore, this higher level of treatment may have resulted in higher levels of lung function. Because the overall level of asthma severity was largely driven by the severity based on controller medication use, prescription and use of severity-appropriate therapy would be expected to lead to the classification of asthma as more severe and with potentially better lung function. However, all of these patients were seen by asthma specialists and were regularly reassessed for the possibility of reducing medication dosages if adequate asthma control was present, minimizing the likelihood of overtreatment. We did not assess the presence of asthma comorbidities such as allergic rhinitis and therefore cannot directly comment on the effect of these comorbidities on asthma severity or lung function. We are unable to comment further on the effect of medication use on lung function measures, as nearly all participants were receiving some form of asthma controller therapy. In an attempt to maximize the chance of studying children at their baseline of asthma disease activity, we required the children to be at least 1 month removed from a significant asthma exacerbation (oral corticosteroids, emergency department visit, or hospitalization).

The use of controller medications by the majority of participants complicates the assignment of asthma severity. The NAEPP/EPR2 Guidelines are recommended for use in patients not receiving controller medications and do not clearly delineate methods for classifying asthma severity among patients once asthma controller therapy has been initiated. The Global Initiative for Asthma Guidelines provide an algorithm that factors in current levels of asthma symptoms and medication use to determine the level of control and need for escalation or reduction of controller medications (24), an approach that is quite similar to the method used in this analysis.

In conclusion, isolated measures of pulmonary function, especially FEV₁, are of little value in determining asthma severity, especially among patients receiving asthma controller therapy. However, there is a gradient of decline in FEV₁/FVC among children with increasingly severe asthma. Although FEV₁/FVC is more sensitive than FEV₁ in detecting differences between different severity groups, its clinical utility in guiding asthma management has not yet been examined. The longitudinal assessment of pulmonary function, with particular attention to FEV₁/FVC, should be examined to determine its utility in tracking asthma severity and guiding therapy over time.