A Single Measure of FEV^Sub 1^ Is Associated With Risk of Asthma Attacks in Long-Term Follow-Up*

Background: Clinical practice guidelines for asthma care emphasize the use of objective measures of asthma severity, and yet little data exist on the relationship between FEV^sub 1^ and asthma outcomes over long-term follow-up.

Methods: We explored the association between measures of FEV^sub 1^ percentage of predicted (FEV^sub 1^% predicted) and subsequent asthma attacks over 3-year intervals. Subjects were identified from two longitudinal cohort studies conducted in the United States and the Netherlands. Persons were included in the analysis if they reported ever having an attack of wheezing with associated shortness of breath prior to or during the follow-up period

Results: Over the course of longitudinal follow-up at 3-year intervals, 195 subjects in the Netherlands cohort contributed 510 observations, and 698 subjects in the US cohort contributed 1,268 observations (for each observation, the report of an attack since their last visit was paired with the subject’s FEV^sub 1^ recorded 3 years prior). Overall, subjects in the Netherlands cohort experienced 114 attacks (22% of the observations) and subjects in the US cohort had 517 attacks (40.6% of the observations). FEV^sub 1^% predicted was significantly associated with risk of an asthma attack over the 3 years following its measurement. After adjusting for current smoking and gender, FEV1% predicted remained an independent predictor of subsequent asthma attacks.

Conclusions: These findings support the use of spirometry as an objective measure of asthma severity and risk of adverse outcomes. (CHEST 2004; 126:1875-1882)

Key words: asthma; outcome assessment; respiratory function tests

Abbreviations: CI = confidence interval: FEV^sub 1^,% predicted = FEV^sub 1^ percentage of predicted; NAEPP = National Asthma Education and Prevention Program; OR = odds ratio; PEF = peak expiratory flow

Asthma is a chronic inflammatory disease that is usually controllable with appropriate therapy. Morbidity associated with asthma impairs quality of life and can lead to work or school absences, emergency department visits, and hospitalizations. Identification of populations at increased risk for these outcomes is thus of considerable importance to physicians, health maintenance organizations, and public health officials.

Asthma severity classification schemes have frequently incorporated multiple characteristics, including asthma symptoms, measures of lung function, and medication and resource utilization, although no uniformly accepted definition of asthma severity exists.1-3 The absence of consensus may be attributable in part to confusion over the role of a severity classification scheme. Some believe they should describe underlying disease intensity, manifested by the minimum amount of medication required to achieve adequate control.4 Others use the schemes to describe features of what might be termed asthma control, such as symptoms, need for rescue medications, and degree of airway obstruction (as measured by absolute level of or variability in peak expiratory flow [PEF] rate, or level of FEV^sub 1^).4 Additionally, some researchers have attempted to define schemes for the purpose of predicting need for hospital-based care.5,6

Spirometric measures, principally FEV^sub 1^, have long been used as markers of the degree of airways obstruction. FEV^sub 1^ is obtained in both research and clinical practice settings, and is endorsed by the National Asthma Education and Prevention Program (NAEPP) as a central component of their asthma severity classification scheme.2 FEV^sub 1^ has been validated as a measure of airways obstruction by its close-correlation with pathologic scores of airway diameter, and is an excellent predictor of mortality from chronic airway disease.7,8 FEV^sub 1^ percentage of predicted (FEV^sub 1^% predicted) in children has been found to predict adult FEV^sub 1^, and data also suggest that steeper rates of decline in FEV^sub 1^ are associated with a greater risk for the development of nonreversible airways obstruction.9,10

Surprisingly, however, little data exist on the relationship between FEV^sub 1^ and clinical asthma outcomes over long-term follow-up. We therefore sought to explore the relationship between FEV^sub 1^ and subsequent asthma attacks in two longitudinal studies of lung health.

MATERIALS AND METHODS

Populations

We analyzed data from two large population-based cohorts, details of which have been published elsewhere.11-14 The Netherlands cohort, a longitudinal study of host factors and environmental determinants of chronic obstructive lung disease, was initiated in 1965 with a random sample of residents in Vlagtwedde and Vlaardingen, the Netherlands. From this sample, those aged (at the time of the 1965 study) 40 to 44 years from Vlagtwedde and 40 to 54 years from Vlaardingen were combined with a subsequent random sampling of 15- to 39-year-old residents of Vlagrwedde, the Netherlands, and 15- to 54-year-old residents of Vlaardingen, the Netherlands, performed between 1967 and 1969. From these samples, a cohort of 4.692 was formed for longitudinal follow-up at 3-year intervals.

The US cohort included a random sample of 8,842 adults initially aged 25 to 74 years, selected between 1974 and 1977 from six US cities: Topeka, KS; Portage, WI; a portion of St. Louis, MO; Kingston-Harriman, TN; Steubenville, OH; and Watertown, MA. Subjects were reexamined at 3 years and 6 years after the initial evaluation. As described below, for the current study our analysis population was defined to include from the above cohorts all of those subjects who reported one or more asthma attacks during follow-up.

Spirometry and Questionnaires

In the Netherlands cohort, ventilatory function was examined every 3 years. Lung function was measured using a water-sealed spirometer (Lode Spirograph D53; Lode Instruments: Groningen, the Netherlands) with subjects seated and wearing nose clips. For the US cohort, participants performed a forced expiratory maneuver with a water-filled, 8-L, recording survey spiroineter (Warren E. Collins; Braintree, MA). Measurements were corrected to body temperature and standard pressure, saturated with water.

For both cohorts, respiratory questionnaires were administered at the time of each spirometry measurement.15,16 The original cohorts were drawn from general populations, without restriction to patients with asthma. Since the intent of the present study was to explore the relationship between lung function and subsequent attacks of wheezing in patients with asthma, we restricted our analysis to the subset of the original cohorts who reported on the questionnaire ever having an attack of wheezing and shortness of breath prior to or during the follow-up period. These subjects were considered likely to have asthma.17 The follow-up period for our analysis began subsequent to their first reported attack.

Only the US questionnaire asked whether or not the subject had received a doctor’s diagnosis of asthma. This allowed us to compare, in the US subjects, the relationship between FEV^sub 1^ and attacks for a potential asthma cohorts defined by the following: (1) self- report of an attack of wheezing and shortness of breath, and, alternatively, (2) self-report of doctor’s diagnosis of asthma. The Netherlands questionnaire did not ask about doctor’s diagnosis of asthma; therefore, for the purposes of comparability between the Netherlands and US cohorts, our primary analysis was conducted on the subset of the cohorts defined by self-report of an attack of wheezing and shortness of breath.

Observations where subjects were < 18 years of age were excluded. We also excluded from our analysis those subjects who had a > 10 pack-year smoking history prior to their first asthma attack, since at least some of these subjects may have had COPD.

Definition of Asthma Attack and Statistical Methods

For the US cohort, an asthma attack was defined as an affirmative response to the question, “Have you had an attack of wheezing and shortness of breath in the last 3 years?” Unlike the US questionnaire, the Netherlands questionnaire did not specify a time period in its question about the occurrence of an attack. The Netherlands questionnaire did, however, ask the respondent to report their age at the time of their last attack. For the Netherlands cohort, we therefore defined an asthma attack as an affirmative response to the question, “Have you ever had an attack of wheezing and shortness of breath?” if the subject’s age at the time of their last attack indicated the attack had occurred within the last 3 years prior to the questionnaire.

For each observation, the report of an attack in the last 3 years was paired with the FEV^sub 1^ value recorded at the field survey 3 years prior. FEV^sub 1^ was analyzed as FEV^sub 1^% predicted.18

FEV^sub 1^% predicted was grouped into discreet categories, as recommended by the NAEPP (< 60%, 60 to 80%, and > 80%), and compared with a more refined grouping of seven categories (< 50%, 50 to 59%, 60 to 69%, 70 to 79%, 80 to 89%, 90 to 100%, and > 100%).

Multiple observations of FEV^sub 1^ and report of asthma attack were utilized in available subjects. Multivariate regression analyses, utilizing a general estimating equation approach, were used to control for the correlation between re\peated measurements among individuals (Proc GENMOD, SAS version 6.12; SAS Institute; Cary, NC). All analyses were performed using the SAS statistical software program. (SAS Institute).

RESULTS

The original Netherlands and US cohorts included 4,692 subjects and 8,842 subjects, respectively. The current analysis was limited to the 195 subjects in the Netherlands cohort and the 698 subjects in the US cohort who reported an “asthma attack” (one or more attacks of wheezing and shortness of breath), at some time during the follow-up and who had a cumulative smoking history of ≤ 10 pack-years.

Table 1 shows baseline characteristics of the individual subjects included in the current analysis. In the Netherlands cohort, the 195 subjects contributed 510 observations (each observation consisted of an FEV^sub 1^ measurement paired with the follow-up questionnaire obtained 3 years later). In the US cohort, the 698 subjects contributed 1,268 observations. Participation in the original cohort studies was voluntary, and not all subjects were available for the same duration of follow-up. In the Netherlands cohort, 46% contributed one or two observations, 36% contributed three to five observations, and 18% contributed six or seven observations. None of the subjects in US cohort contributed more than two observations, since follow-up at 3-year intervals was performed only twice in this cohort.

The two populations were similar in terms of gender and pulmonary function (Table 1). More individuals in the Netherlands population were smokers at the time of their initial observation (34.4% vs 16.7% in the US cohort). The Netherlands population was younger (mean age, 37 years vs 47 years). Although the age ranges for the two populations were similar, the youngest subjects included in our analysis of the Netherlands cohort were 18 years old, compared to 25 years of age for the US cohort.

FEV^sub 1^% predicted was significantly associated with risk of an asthma attack over the 3 years following its measurement (Table 2). Overall, subjects experienced 114 attacks in the Netherlands cohort (22% of the observations) and 516 asthma attacks in the US cohort (41% of the observations). The prevalence of attacks in the observations in the US cohort restricted to subjects with a doctor’s diagnosis of asthma was only slightly higher (44%, or 184 of 412 observations).

Table 1-Characteristics of Individual Subjects at the Time of First Included Observation, by Study Population*

Table 2-Relationship Between Subsequent Asthma Attacks and Patient Characteristics

Although the overall prevalence of asthma attacks was greater among female than male subjects, the difference was not significant. Nonsmokers in the Netherlands cohort had a significantly higher risk of an attack than current smokers (Table 2), although the difference in risk was not significant after controlling for FEV^sub 1^% predicted (p = 0.12). In the US cohort, current smokers and nonsmokers had a similar risk of an asthma attack.

In both cohorts, risk of asthma attack increased significantly with decreasing category as defined by FEV^sub 1^% predicted. Among subjects with an FEV^sub 1^ < 60% predicted, 57.5% and 65.3% in the Netherlands and US cohorts, respectively, had at least one attack over the 3 years following measurement of FEV^sub 1^. For those with an FEV^sub 1^% predicted > 80, the risk was substantially lower: 13% and 34.1% in the Netherlands and US cohorts, respectively. In the US cohort, the relationship between FEV^sub 1^ and risk of asthma attack was similar when self-report of a doctor’s diagnosis of asthma, rather than self-report of a history of an asthma attack, was used to define the presence of asthma (data not shown).

We used logistic regression analysis to examine the association of FEV^sub 1^ and subsequent asthma attack, controlling for the potential confounding variables current smoking and gender. As was true in the crude analyses, the risk of asthma attack increased as FEV^sub 1^% predicted decreased, with the magnitude of the odds ratios (ORs) similar in the two populations in these multivariate models (Table 3). In both populations, there was a more than threefold increase in the odds of an attack for those with an FEV^sub 1^ < 60% predicted compared with those with an FEV^sub 1^ > 80% predicted. Compared to an FEV^sub 1^% predicted of > 80%, an FEV^sub 1^ of 60 to 80% predicted approximately doubled the odds of an attack in both cohorts (OR, 2.3; confidence interval [CI], 1.4 to 3.9%; and OR, 1.7; CI, 1.4 to 2.2) for the Netherlands and US cohorts, respectively.

We tested whether categorizing FEV^sub 1^% predicted into three broad categories (< 60%, 60 to 80%, and > 80%) was sufficient to capture important variation in risk. We found that when FEV^sub 1^ was categorized more finely, the risk of subsequent asthma attack did in fact vary between those with an FEV^sub 1^ between 60% and 70% compared with 70 to 80% predicted, although the risk appeared to be similar between those with an FEV^sub 1^ of 80 to 90% predicted compared with 90 to 100% (Fig 1). In the Netherlands population hut not in the US population, the risk of a subsequent asthma attack was reduced among those with an FEV^sub 1^ > 100% compared with those with an FEV^sub 1^ of 80 to 100%.

DISCUSSION

The NAEPP has endorsed the use of objective measures of lung function to assign a seventy rating to patients with asthma.2 The purpose of such a severity rating is to guide asthma therapy, and the major assumption underlying it is that different categories of FEV^sub 1^% predicted suggest different risks of adverse disease outcomes. Despite its intuitive appeal, little data exist in the literature to support this assumption. Our study suggests that in two separate cohorts, a single measure of FEV^sub 1^ is associated with the risk of an asthma attack over 3 years of follow-up. Moreover, this association persists after adjustment for current smoking status and sex.

Table 3-Risk of Asthma Attack Over 3-yr Period, by FEV^sub 1^ Category

We demonstrate that FEV^sub 1^ can be used to stratify patients on the basis of risk of an asthma outcome even over a relatively long time horizon, thus emphasizing the value of this simple measure of lung function. Our findings complement those of Li et al,5 who constructed a risk factor model to predict hospitalizations among adult patients with moderateto-severe asthma, and found that degree of FEV, impairment was an important factor in identifying patients at increased risk for hospitalizations over 1 year of follow-up.

A fundamental goal of a seventy classification scheme is to create categories that capture important differences in risk. We compared the FEV, cut points disseminated by the NAEPP and widely used clinicalk (categories of < 60%, 60 to 80%, and > 80%) to categories that were more narrowly defined. In our two populations, the simpler, threecategory scheme used by the NAEPP arguably captured much of the variation. Among those with more severe obstruction, however, finer categorization of lung function captured significant additional variations in risk of asthma attacks (eg, those with an FEV^sub 1^% predicted of 60 to 69% had a greater risk of an attack than those with an FEV^sub 1^% predicted of TO to 79%).

We do not think it likely that asthma severity can he fully described by FEV^sub 1^% predicted alone. Reflecting its limitations, the NAEPP seventy classification scheme includes not only objective measures of lung function (FEV^sub 1^ and PEF rate), but also symptoms, activity limitations, attack frequency, and medication use. The limitations of FEV^sub 1^% predicted are exemplified by data from the Childhood Asthma Management Program study.19 At baseline, the mean FEV^sub 1^% predicted was > 93% among this pediatric population, yet approximately one third of the children had been treated with oral steroids in the 6 months before enrollment. Although FEV^sub 1^% predicted may differentiate less well among those with a normal FEV^sub 1^ in this cohort, its ability to predict risk of asthma attacks across severity groups in pediatric patients has been shown elsewhere.20

Despite its limitations, FEV^sub 1^% predicted has a number of advantages as a marker of asthma severity, including its objectivity and reproducibility.21,22 Because many downstream consequences of asthma are the direct result of airway obstruction, it is intuitively appealing to employ a test that measures baseline obstruction level. In addition, FEV^sub 1^% predicted is frequently reported in clinical trials, and the efficacy of new therapies is often expressed in terms of their impact on FEV^sub 1^. Review of the clinical trials literature examining asthma therapies suggests that improvements in FEV^sub 1^ are paralleled by improvements in clinical asthma outcomes such as symptoms, health-related quality of life, rescue medication use, and health-care utilization.23-29 Among a population of asthmatic children involved in the Pediatric Asthma Care-PORT trial,30 baseline spirometry was effective for prediction of asthma symptom day-14 score over a 1-year period of observation.

FIGURE 1. Asthma attack incidence, by FEV^sub 1^% predicted.

In addition, a relationship between asthma and the natural history of lung function has been demonstrated.31 The development of irreversible obstruction, presumably as a consequence of airway remodeling, is seen in some patients with asthma.32-33 Thus, FEV^sub 1^ can be seen as an important measure of asthma from both a clinical and a broader, public health perspective.

For the purposes of prediction and planning, such as might be undertaken by a health system or health maintenance organization, a disadvantage of the NAEPP multifaceted approach to rating seventy is that it does not separate asthma outcomes (eg, symptoms) from objective markers of disease severity. Our analysis focused on the validation of \one potential marker of risk of subsequent asthma attacks. Of interest, a study by Tattersfield et al34 examined multiple factors for their association with exacerbations occurring in a cohort of subjects with persistent asthma. They reported that in 12 months of follow-up, female gender, increasing age, PEF variability during the run-in period, and inhaled corticosteroid dose before study onset were positively associated with the risk of exacerbation. The authors report that for a 1% increase in PEF variability, a 100-g increase in prestudy inhaled corticosteroid dose, and a 1-year increase in age, the ORs for having an exacerbation were 1.023 (95% CI, 1.011 to 1.035), 1.056 (95% CI, 1.016 to 1.096), and 1.011 (95% CI, 1.001 to 1.023), respectively. The study published by Tattersfield et al34 differed from ours, however, in that the measure of lung function examined was PEF and not FEV^sub 1^% predicted, and their outcome was not self-report of an asthma attack but rather an exacerbation defined as a worsening of asthma control requiring oral corticosteroids or as an episode in which morning PEF fell by > 30% from baseline on 2 consecutive days.

A limitation of our study is that we did not define subjects by doctor’s diagnosis of asthma for our primary analysis since we lacked this data for the Netherlands cohort. The accuracy of a single, self-reported attack of wheezing and shortness of breath as an indicator of asthma is uncertain. Although surveys seeking self- report of asthma symptoms are considered by many to be a valid means of establishing the prevalence of asthma in populations and to correlate well with doctor diagnosis, it is possible that some of the included subjects may not have had asthma.17,35,36

Among the FEV^sub 1^% categories, < 60%, 60 to 80%, and > 80%, we found that 21%, 35%, and 49% of the subjects, respectively, did not report an additional asthma attack over the follow-up period, after their initial inclusion event. Some of these subjects may not have reported subsequent attacks because they may not have had asthma. Alternatively, the questionnaires, which only asked about attacks of wheezing with shortness of breath, probably failed to capture all instances in which an asthma event of any sort had occurred. Subjects may have failed to recall attacks or been uncertain about their timing (surveys were administered once every 3 years), or reported only the more severe episodes of wheezing associated with shortness of breath (since they were asked to recall “attacks” of wheezing associated with shortness of breath). The average number of observations (follow-up visits) was 2.6 for the Netherlands cohort and 1.8 for the US cohort, an additional limitation to the sensitivity of the measure.

If misclassification did occur and were more prevalent among those with higher FEV^sub 1^% predicted values, it may have led to an overestimate of the risk associated with lower FEV^sub 1^% predicted values. Despite this concern, it should be noted that when we restricted the analysis for the US cohort to those subjects with a doctor’s diagnosis of asthma, our effect estimates were similar. Moreover, we found that in the US cohort, asthma attacks were only slightly more prevalent among the subset with a doctor’s diagnosis of asthma than among those who were defined as having asthma only on the basis of reporting a history of an attack of wheezing with shortness of breath (occurring in 44% vs 41% of observations).

Several additional limitations of the present study should be noted. First, a number of environmental, demographic, and treatment features have been shown to influence risk of asthma outcomes, including allergen exposure, race or ethnicity, socioeconomic status, living in the inner city, access to care, and use of inhaled corticosteroids.37-41 None of these variables was considered in our analysis. We think it probable, however, that the primary means by which these factors influence risk of an adverse outcome is their effect on airway obstruction, captured by FEV^sub 1^.

Second, we were only able to examine self-reported asthma attacks. Other outcomes, such as quality of life, interference with work or school, need for rescue medication use, and physician visits or hospitalisations for asthma are obviously of interest, although they are likely to be correlated with occurrence of asthma attacks.

Third, since our observational study did not constrain treatment, it is possible that measurement of FEV^sub 1^ led to differential treatment of subjects and that this affected the relationship between FEV^sub 1^ and asthma attacks. If this occurred, however, it would almost certainly have resulted in patients with lower FEV^sub 1^ values receiving more treatment. The result would therefore have been to diminish our estimate of the effect of FEV^sub 1^ on risk of attack, and it thus makes our findings more persuasive.

Fourth, asthma attack was defined by sell-report of an attack of wheezing and shortness of breath. The number of attacks during the follow-up interval and the intensity and duration of attacks was not reported, and almost certainly varied between individuals, and within individuals, across attacks. It would have been of interest to examine the relationship between lung function and specific kinds of attacks, as well as the number of attacks; our data did not permit such an analysis.

Of interest, we found that nonsmokers in the Netherlands cohort were at greater risk for asthma attacks than smokers. The association was weaker and not significant when we adjusted for FEV^sub 1^% predicted, and may suggest a healthy smoker effect.42 In essence, poorer respiratory health may have led some subjects not to smoke. The association between smoking status and risk of attacks was not seen in the US cohort, potentially because the much lower prevalence of smoking in the US cohort made it more difficult to observe such an effect.

We found a higher prevalence of asthma attacks among those in the US cohort than among the Netherlands cohort. There are several potential explanations for the differences observed, including greater misclassification in the Netherlands cohort (leading to the inclusion of more subjects without asthma in the Netherlands cohort), better asthma treatment in the Netherlands cohort, or differences in environmental or genetic factors. The English translation of the questions used in both surveys is the same. However, slight differences in interpretation of the questions after translation cannot be excluded, and may have contributed to the differences between the two groups. Independent of these differences in the absolute prevalence of reported asthma attacks, our results demonstrate a significant association between FEV^sub 1^% and asthma attacks in two separate populations, further supporting the robust nature of this relationship.

Though we think an important feature of our study was its demonstration of a relationship between FEV^sub 1^ and attacks over 3 years of follow-up, in practice planning agencies or health maintenance organizations would likely operate on a shorter time horizon. Whether or not the relationship between FEV^sub 1^ and attacks is similar for a shorter the time interval is unknown.

Finally, our use of data collected as part of longitudinal cohort studies reflected our desire to study the relationship between lung function and outcomes over the relatively long time horizon. The longitudinal data collected in these studies on both lung function and asthma outcomes gave us the ability to efficiently address this relationship.

In conclusion, a strong association exists between FEV^sub 1^% predicted and risk of asthma attacks over the subsequent 3 years among a potential asthmatic population. Further research is needed to explore the predictive accuracy of FEV^sub 1^ over a shorter time horizon and over a broad range of outcomes such as resource utilization. Nonetheless, these data support the use of objective measures of lung function when assessing risk for adverse asthma outcomes.

* From the Division of Pulmonary and Critical Care Medicine (Dr. Kitch) and Channing Laboratory (Drs. Weiss and Fuhlbrigge), Brigham and Women’s Hospital and Harvard Medical School, Boston, MA; Department or Epidemiology and Public Health (Dr. Paltiel), Yale School of Medicine, New Haven, CT; Departments of Health Policy and Management (Dr. Kuntz) and Environmental Health (Dr. Dockery), Harvard School of Public Health, Boston, MA; and Departments of Epidemiology & Statistics (Dr. Schouten), University of Groningen, Groningen, the Netherlands.

REFERENCES

1 British Thoracic Society. Guidelines for the management of asthma: a summary. BMJ 1993; 306:776-782 [published erratum appears in BMJ 1993; 307:1054]

2 National Heart, Lung, and Blood Institute. National Asthma Education and Prevention Program Expert Panel. Clinical practice guidelines: expert panel report 2; guidelines for the diagnosis and management of asthma. Bethesda, MD: National Heart, Lung, and Blood Institute, 1997; 1-153; publication No. 97-4051; available at: http:/ /www.nhlbi.nih.gov/ guidelines/asthma/asthgdln.htm; accessed June 2004

3 National Heart, Lung, and Blood Institute. Global strategy for asthma management and prevention: NHLBI/WHO workshop report. Bethesda, MD: National Heart, Lung and Blood Institute, 1995; 1-48

4 Cockcroft DW, Swystun VA. Asthma control versus asthma severity. J Allergy Clin Immunol 1996; 98:1016-1018

5 Li D, German D, Lulla S, et al. Prospective study of hospitalisation for asthma: a preliminary risk factor model. Am J Respir Crit Care Med 1995; 151:647-655

6 Osborne ML, Vollmer WM, Pedula KL, et al. Lack of correlation of symptoms with specialist-assessed long-term asthma severity. Chest 1999; 115:85-91

7 Hogg JC, Macklem PT, Thurlbeck WM. Site and nature of airway obstruction in chronic obstructive lung disease. N Engl J Med 1968; 278:1355-1360

8 \Burrows B. Course and prognosis of patients with chronic airways obstruction. Chest 1980; 77(suppl):250-251

9 Ulrik CS, Backer V. Nonreversible airflow obstruction in life- long nonsmokers with moderate to severe asthma. Eur Respir J 1999; 14:892-896

10 Roorda RJ, Gerritsen J, van Aalderen WM, et al. Follow-up of asthma from childhood to adulthood: influence of potential childhood risk factors on the outcome of pulmonary function and bronchial responsiveness in adulthood. J Allergy Clin Immunol 1994; 93:575- 584

11 Ferris BG, Speizer FE, Spengler JD. et al. Effects of sulfur oxides and respirable particles on human health: methodology and demography of populations in study. Am Rev Respir Dis 1979; 120:767- 779

12 Ware JH, Dockery DW, Louis TA, et al. Longitudinal and cross- sectional estimates of pulmonary function decline in never-smoking adults. Am J Epidemiol 1990; 132:685-700

13 Rijcken B, Schouten JP, Weiss ST, et al. The relationship of nonspecific bronchial responsiveness to respiratory symptoms in a random population sample. Am Rev Respir Dis 1987; 136:62-68

14 van der Lende R, Kok TJ, Reig RP, et al. Decreases in VC and FEV^sub 1^ with time: indicators for effects of smoking and air pollution. Bull Eur Physiopathol Respir 1981; 17:775-792

15 Medical Research Council. Committee on research into chronic bronchitis; instructions for the use of the questionnaire on respiratory symptoms. London, UK: Medical Research Council, 1966

16 Ferris BG. Epidemiology standardization project (American Thoracic Society). Am Rev Respir Dis 1978; 118:1-120

17 O’Connor GT, Weiss ST. Clinical symptom measures. Am J Respir Crit Care Med 1994; 149:S21-S28

18 Crapo RO, Morris AH, Gardner RM. Reference spirometric values using techniques and equipment that meet ATS recommendations. Am Rev Respir Dis 1981; 123:659-664

19 Long-term effects of budesonide or nedocromil in children with asthma. The Childhood Asthma Management Program Research Group. N Engl J Med 2000; 343:1054-1063

20 Fuhlbrigge AL, Kitch BT, Paltiel D, et al. FEV^sub 1^ is associated with risk of asthma attacks in pediatric population. J Allergy Clin Immunol 2001; 107:61-67

21 Enright PL. Lebowitz MD, Cockroft DW. Physiologic measures: pulmonary function tests: asthma outcome. Am J Respir Crit Care Med 1994; 149:S9-s18: discussion S19-S20

22 Enright PL, Johnson LR. Connett JE, et al. Spirometry in the Lung Health Study: 1. Methods and quality control. Am Rev Respir Dis 1991; 143:1215-1223

23 Noonan MP, Chervinsky W, Busse W, et al. Fluticasone propionate reduces oral prednisone use while it improves asthma control and quality of life. Am J Respir Crit Care Med 1995; 152(5 pt 1):1467-1473

24 Okamoto LJ, Noonan M, DeBoisblanc BP, et al. Fluticasone propionate improves quality of life in patients with asthma requiring oral corticosteroids. Ann Allergy Asthma Immunol 1996; 76:455-461

25 Agertoft L, Pedersen S. Effects of long-term treatment with an inhaled corticosteroid on growth and pulmonary function in asthmatic children. Respir Med 1994; 88:373-381

26 Otulana BA, Varma N, Bullock A, et al. High does nebulized steroid in the treatment of chronic steroid-dependent asthma. Respir Med 1992; 86:105-108

27 Ogirala RG. Sturm TM, Aldrich TK, et al. Single, high-dose intramuscular triamcinolone acetonide versus weekly oral methotrexate in life-threatening asthma: a double-blind study. Am J Respir Crit Care Med 1995; 152:1461-1466

28 Tamaoki J, Kondo M, Sakai N, et al. Leukotriene antagonist prevents exacerbation of asthma during reduction of high-dose inhaled corticosteroid. The Tokyo Idai Asthma Research Group. Am J Respir Crit Care Med 1997; 155:1235-1240

29 Corren J. Harris AG, Aaronson D. et al. Efficacy and safety of loratadine plus pseudoephedrin in patients with seasonal allergic rhinitis and mild asthma. J Allergy Clin Immunol 1997; 100:781-788

30 Fuhlbrigge A, Carey V, Lozano P, et al. Performance characteristics of periodic asthma symptom reports for outcomes research [abstract]. Am J Respir Crit Care Med 2001; 163: A316

31 Lange P, Panier J, Vestbo J, et al. A 15-year follow-up study of ventilatory function in adults with asthma. N Engl J Med 1998; 339:1194-1200

32 Backman KS, Greenberger PA, Patterson R. Airways obstruction in patients with long-term asthma consistent with “irreversible asthma.” Chest 1997; 112:124-140

33 Redington AE, Howarth PH. Airway wall remodeling; in asthma. Thorax 1997; 52:310-312

34 Tattersfield AE, Postma DS, Barnes PJ, et al. Exacerbations of asthma: a descriptive study of 425 severe exacerbations. The FACET International Study Group. Am J Respir Crit Care Med 1999; 160:594- 599

35 Jenkins MA, Clarke JR, Carline JB, et al. Validation of questionnaire and bronchial hyperresponsiveness against respiratory physician assessment in the diagnosis of asthma. Int J Epidemiol 1996; 25:609-616

36 Cerveri IC, Bruschi M, Ricciardi L, et al. Epidemiological diagnosis of asthma: methodological considerations of prevalence evaluation. Eur J Epidemiol 1987; 3:202-205

37 Donahue JG, Weiss ST, Livingston JM, et al. Inhaled steroids and the risk of hospitalization for asthma. JAMA 1997; 277:887-891

38 Gottlieb DJ, Beiser AS, O’Connor GT. Poverty, race, and medication use are correlates of asthma hospitalization rates: a small area analysis in Boston. Chest 1995; 108: 28-35

39 Rosenstreich DL, Eggleston P. Kattan M, et al. The role of cockroach allergy and exposure to cockroach allergen in causing morbidity among inner-city children with asthma. N Engl J Med 1997; 336:1356-1363

40 Marder D, Targonski P, Orris P, et al. Effect of racial and socioeconomic factors on asthma mortality in Chicago. Chest 1992; 101(suppl):426S-429S

41 Turner MO, Noertjojo K, Vedal S, et al. Risk factors for near- fatal asthma: a case-control study in hospitalized patients with asthma. Am J Respir Crit Care Med 1998; 157:1804-1809

42 Choi BC. Definition, sources, magnitude, effect modifiers, and strategies of reduction of the healthy worker effect. J Occup Med 1992; 34:979-988

Barrett T. Kitch, MD, MPH; A. David Paltiel, PhD; Karen M. Kuntz, ScD; Douglas W. Dockery, PhD; Jan P. Schouten, PhD; Scott T. Weiss, MD, MS; and Anne L. Fuhlbrigge, MD, MS

Funding was provided by AstraZeneca Pharmaceuticals, 1-K08 HL03919-01; and National Heart, Lung and Blood Institute, National Institutes of Health grant HL027427.

Manuscript received January 31, 2003; revision accepted August 17, 2004.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (e-mail: permissions@chestnet.org).

Correspondence to: Barrett Kitch, MD, MPH, Division of Pulmonary and Critical Care Medicine, Brigham and Women’s Hospital, 75 Francis St, Boston, MA 02115; e-mail: bkitch@ partners.org

Copyright American College of Chest Physicians Dec 2004