INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Airway inflammation is widely accepted as the central mechanism in the pathogenesis of asthma (1). Inflammation in the distal airways and parenchyma has also been described in asthma (2, 3). It has been hypothesized that over time, recurrent injury and repair resulting from chronic airway inflammation lead to distortion and remodeling of bronchial architecture (4, 5). Airway remodeling is characterized by increased subepithelial matrix deposition with a shift in collagen components, hypertrophy and hyperplasia of airway smooth muscle, mucus gland hyperplasia, and vascular proliferation (6, 7). Although some of these changes may occur early in the disease process (8, 9), the long-term effects of persistent airway and alveolar inflammation remain poorly described. The airways of elderly subjects who have had persistent asthma for decades are likely candidates for airway remodeling. Because the duration of asthma in the elderly can range from several months to many decades, this is an ideal population in which to study the relationship between the duration of asthma and airway or parenchymal remodeling.

Characterization of asthma in the elderly is complicated by difficulty in differentiating asthma from other forms of chronic obstructive lung disease (10). Cigarette smoking is a major diagnostic confounder in this population, and published descriptions of asthma in the elderly are derived from cohorts that include smokers (11). Because of the confounding effects of tobacco use, it is unclear whether these descriptions are of asthma or of a heterogeneous group of obstructive lung diseases that may be pathophysiologically distinct from asthma. In order for a population of elderly asthmatic individuals to serve as a model in which to study the natural history of asthma, the confounding effects of tobacco use must be eliminated.

The consequences of longstanding asthma in elderly lifetime nonsmokers have not been well described. Because of the inherent risks of invasive studies in this cohort, we assessed physiologic outcomes as surrogate measures of underlying airway structural changes. We hypothesized that longstanding asthma would result in irreversible airflow limitation and hyperinflation, and that the severity of these abnormalities would correlate with the duration of disease. In this study we describe airflow, static lung volumes, and bronchodilator responsiveness in a well-characterized cohort of elderly lifetime nonsmokers with asthma from the Bellevue Hospital Primary Care Asthma Clinic (BHPCAC) in New York City.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

All subjects who were enrolled in the BHPCAC between 1991 and 1998, and who had completed a data-base questionnaire (n = 2,112), were evaluated for entry into the study. The BHPCAC, located in a municipal hospital in New York City, serves an indigent urban population. Subjects were eligible for inclusion in the study if they were 60 yr of age or older and had a physician diagnosis of asthma based on the National Institute of Health/National Asthma Education and Prevention Program (NIH/NAEPP) guidelines (n = 235) (18). Only lifetime nonsmokers, defined as those subjects who reported zero pack-years of cigarette use, were enrolled in the study (n = 114). Subjects who did not complete baseline and postbronchodilator spirometry were excluded from analysis (n = 38). One patient was excluded because the age of onset of asthma was unknown.

A total of 75 subjects were included in the analysis. Baseline data on demographic characteristics, tobacco use, and clinical features of asthma were obtained from the BHPCAC database questionnaire, a validated instrument completed upon entry into the clinic. Information obtained from the data base was confirmed by chart review. Age of onset of asthma was determined by subjects' own report of their age at the time of initial physician diagnosis of asthma or onset of symptoms, whichever occurred earlier. Duration of asthma was calculated from the age of onset. The median duration of asthma among the study subjects was 26 yr. The median was used to divide the total cohort into subjects with asthma of long duration (LDA; n = 38) and those with asthma of short duration (SDA; n = 37). Hence, subjects who had had asthma for  26 yr were classified as having LDA and those who had had asthma for < 26 yr were classified as having SDA.

In order to optimize lung function in the study, we prescribed inhaled corticosteroids to all subjects in accordance with the NIH/ NAEPP guidelines (18). The majority of subjects required medium to high-dose inhaled corticosteroids. There was no significant difference in the dosages of inhaled corticosteroids prescribed to the patients with LDA as compared with those with SDA.

Serum IgE and Radioallergosorbent Testing

Total serum IgE and specific serum IgE to 15 different indoor and outdoor allergens was measured in a random subset of 40 subjects. The demographic and clinical characteristics of this subgroup did not differ significantly from those of the total study cohort. The allergens selected for testing were those commonly found in urban areas in the northeastern United States and included cat epithelium, dog epithelium, mouse urine protein, house dust mite (Dermatophagoides farinae and D. pterynissinus), cockroach (Blatella germanica), grasses (timothy and orchard cocksfoot), weeds (English plantain and common ragweed), trees (oak, sycamore, and maple), and molds (Alternaria tenuis and Aspergillus fumigatus). Specific IgE was measured by radioallergosorbent testing (RAST). Total serum IgE and RAST measurements were performed at a commercial laboratory (Quest Diagnostics Inc., Teterboro, NJ) with an ImmunoCAP assay (Pharmacia, Piscataway, NJ) (19, 20). The assay was calibrated against the World Health Organization standards. The normal range for specific serum IgE was 0 to 180 U/ml. For IgE to specific allergens,  0.35 kIU/L was considered positive.

Pulmonary Function Testing

After enrollment in the BHPCAC, all subjects underwent baseline and postbronchodilator spirometry in the Bellevue Hospital pulmonary function laboratory. Inhaled short-acting ₂-agonists were withheld for at least 4 h before baseline testing. Long-acting ₂-agonists were not prescribed to study subjects before pulmonary function testing. Inhaled ₂-agonists were subsequently administered via spacer device for postbronchodilator evaluation. Maximal expiratory flow-volume curves were obtained, and FVC, FEV₁, expiratory flow at mid-lung volume (₅₀), and peak expiratory flow rate (PEFR) were determined. Airway resistance (Raw) and specific airway conductance (SGaw) were measured with a panting maneuver performed in a plethysmograph. In a subgroup of 37 subjects, baseline and postbronchodilator FRC was measured by body plethysmography (P.K. Morgan, Andover, MA) using standard methods (21, 22). TLC and RV were calculated from these measurements. Accuracy checks for mouth pressure, box pressure, and body plethysmograph mouth flow were made on a daily basis. Spirometric measurements were referenced to the data of Knudsen (23) and lung volumes were referenced to the data of Goldman and Becklake (24). The % predicted values for each physiologic measurement were calculated for each study subject.

Statistical Analysis

The demographic and clinical characteristics of the subjects with LDA and those with SDA were compared through chi-square analysis. In order to minimize the potentially confounding influences of sex, height, and age on physiologic parameters, statistical analyses were performed on the % predicted values instead of the raw measured data. The % predicted values of physiologic measurements in the patients with LDA were compared with those in the patients with SDA through t tests, analysis of variance, and Wilcoxon's rank sum analyses where appropriate (a value of p < 0.05 was considered significant).

Regression analyses were performed to further delineate the relationship between pulmonary physiology and duration of asthma. The % predicted value of each physiologic parameter was regressed on the duration of asthma in years. Standardized  coefficients were used for multiple regression models to allow comparison of the relative impact of the variables. All analyses were performed with the JMP version 3.1.5 statistical software package (SAS Institute, Inc., Cary, NC).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Demographic Characteristics of Subjects

As shown in Table 1, subjects ranged in age from 60 to 80 yr, with a median age of 65 yr. There was no significant difference in age between the groups with LDA and SDA. The majority of the cohort was female (81%); more patients with LDA than with SDA were women (92% versus 70%, respectively; p < 0.02). Most subjects were of Hispanic or African-American descent (81%); the distribution of race/ethnicity was similar among both study subgroups, and reflected the demographics of our urban community.

Clinical Characteristics of Asthma in Study Subjects

The age of onset of asthma varied greatly among study subjects; some reported lifelong asthma, whereas others reported asthma onset after age 60 yr. The median age of onset was age 40 yr. Duration of asthma ranged from several months to 74 yr. The median duration of asthma in patients with LDA was 40 yr, compared with 9 yr in the patients with SDA.

All subjects had persistent asthma upon entry into the BHPCAC as defined by the NIH/NAEPP guidelines (18). The majority of the total cohort (75%), as well as the majority of patients with LDA and SDA (76% and 73%, respectively), had severe persistent asthma as determined by daily and nocturnal symptom frequency (Table 2). A history of intubation for asthma was common in the total cohort; however, more patients with LDA than with SDA had been intubated in the past (14% versus 3%, respectively). Many of the total cohort were unable to walk more than one city block (64%) or climb more than one flight of subway stairs (57%) without stopping for breath. The patients with LDA tended to report more limited ability to walk than those with SDA; however, both had equally severe impairment of their ability to climb subway stairs.

The patients with LDA and those with SDA did not differ in terms of atopy, as evidenced by total and specific serum IgE measurements. The median total serum IgE was similar in the patients with LDA (n = 20) and those with SDA (n = 20) (71 U/ml versus 56 U/ml, respectively). Likewise, there was no significant difference in the percentages of patients with LDA and with SDA who had an increased specific IgE to one of the allergens tested (57% versus 43%, respectively).

Evaluation of Airflow

Multiple measures of large airway function demonstrated that the patients with LDA had more severe airflow obstruction than those with SDA. Both the LDA and SDA patients had substantial airflow limitation as measured by FEV₁ (1.27 ± 0.07 L and 1.65 ± 0.09 L [mean ± SEM], respectively) and ₅₀ (1.14 ± 0.13 L/s and 1.56 ± 0.13 L/s, respectively). When expressed as a percentage of predicted values, the degree of airflow limitation was more severe in the LDA as compared with the SDA group, as evidenced by FEV₁ (59.5 ± 2.6 versus 73.8 ± 3.1, respectively; p < 0.007) and a trend toward lower values of ₅₀ (38.7 ± 4.3 versus 50.5 ± 4.4, respectively; p < 0.06) (Figure 1). Similarly, baseline PEFR was significantly lower among the patients with LDA than among those with SDA (3.3 ± 0.2 L/s versus 4.4 ± 0.2 L/s, respectively; p < 0.002). Furthermore, Raw was significantly greater in the LDA than in the SDA group (3.52 ± 0.24 cm H₂O/L/s versus 2.53 ± 0.24 cm H₂O/L/s; p < 0.006), and there was a trend toward a lower SGaw in the LDA than in the SDA group (0.09 ± 0.012 L/cm H₂O/L/s versus 0.12 ± 0.012 L/cm H₂O/L/s; p < 0.09).

View larger version (26K): [in this window] [in a new window]   Figure 1.   Airflow measurements. Spirometry was done on subjects with LDA (black) and SDA (grey) at baseline and after bronchodilator administration, and is presented as mean % predicted. *p < 0.006.

Regression analyses were performed to further examine the relationship between the duration of asthma and airflow in the total cohort. An inverse correlation was observed between the duration of asthma and both the % predicted FEV₁ (r = 0.26, p < 0.03) and the % predicted ₅₀ (r = 0.27, p < 0.03) (Figures 2A and 2B). A direct relationship was observed between Raw and duration of asthma (r = 0.34, p < 0.01). Inverse relationships were also observed between the duration of asthma and both PEFR (r = 0.11, p < 0.02) and SGaw (r = 0.29, p < 0.03) (data not shown).

View larger version (13K): [in this window] [in a new window]   Figure 2.   Linear regression analyses of duration of asthma on airflow. Regression analyses were done of the relationship between duration of asthma and % predicted FEV1 (A) or 50 (B). An inverse association was noted between duration of asthma and FEV1% predicted (r = 0.26, p < 0.03,) and 50% predicted (r = 0.27, p < 0.03).

Airflow Response to Bronchodilator Administration

The degree of reversibility of airflow limitation was also assessed with multiple parameters. Postbronchodilator FEV₁% predicted was significantly lower in the group with LDA than in the group with SDA (65.4 ± 2.9% versus 80.9 ± 2.9%, respectively; p < 0.0003), as was the postbronchodilator ₅₀% predicted (41.3 ± 4.4% versus 59.6 ± 4.5%, respectively; p < 0.006) (Figure 1). Similarly, the postbronchodilator PEFR was significantly lower in the group with LDA than in that with SDA (3.6 ± 0.2 L/s versus 4.8 ± 0.2 L/s, respectively; p < 0.0006). After bronchodilator administration, Raw remained higher in the group with LDA than in the group with SDA (3.18 ± 0.27 cm H₂O/L/s versus 2.04 ± 0.29 cm H₂O/L/s, respectively; p < 0.007) and SGaw was significantly lower in the group with LDA (0.09 ± 0.01% versus 0.16 ± 0.01, respectively; p < 0.0003).

Both the patients with LDA and those with SDA showed a similar mean percentage improvement in FEV₁ (11.6 ± 2.0% and 10.9 ± 2.0%, respectively). Although the patients with LDA showed a slightly lower mean absolute improvement in FEV₁ than did those with SDA (127 ± 27 ml and 167 ± 28 ml, respectively), this difference did not achieve statistical significance. A significant bronchodilator response, defined as an improvement of  12% in FEV₁ (25), was attained by similar percentages of patients with LDA and with SDA (39% and 35%, respectively). The patients with SDA showed a significantly greater improvement in ₅₀ after bronchodilator administration than did those with LDA (23.9 ± 4.6% versus 8.3 ± 4.0%, respectively; p < 0.02).

Importantly, only 18% of the patients with LDA were able to attain a postbronchodilator FEV₁ that was  80% predicted; in contrast, 50% of the patients with SDA achieved a normal FEV₁ (p < 0.003). In fact, the patients with LDA continued to have severe airflow obstruction even after bronchodilator administration, with a mean FEV₁% predicted of 65.4 ± 2.9%.

Evaluation of Lung Volumes

As seen in Figure 3, the slow vital capacity (SVC) was at the lower limit of normal and TLC was at the upper limit of normal in both the LDA and SDA groups. FRC, RV, and RV/TLC were substantially increased in both groups. However, the degree of hyperinflation was significantly greater in the LDA than in the SDA group as evidenced by the FRC (142 ± 5.6% versus 124.1 ± 4.4% predicted, respectively; p < 0.01) and RV/ TLC (60.9 ± 1.6% versus 55 ± 1.6%, respectively; p < 0.02).

View larger version (32K): [in this window] [in a new window]   Figure 3.   Lung volume measurements. Plethysmography was done on subjects with LDA (black) and SDA (grey). Lung volumes are presented as mean % predicted at baseline or after bronchodilator administration. *p < 0.04, **p < 0.02.

Regression analysis done on the entire cohort failed to show a relationship between either SVC or TLC and the duration of asthma (data not shown). However, linear regression analysis of data from the entire cohort revealed a direct correlation between FRC and duration of asthma (r = 0.38, p < 0.002) (Figure 4).

View larger version (16K): [in this window] [in a new window]   Figure 4.   Linear regression analyses of duration of asthma on FRC. Regression analysis was done of the relationship between duration of asthma and FRC% predicted. An inverse association was noted (r = 0.38, p < 0.002).

Lung Volume Response to Bronchodilator Administration

Lung volume measurements were repeated after bronchodilator administration. As shown in Figure 3, there was no significant change in TLC or SVC in either the LDA or the SDA group. There was a trend toward a reduction in FRC after bronchodilator administration in both groups. However, even after bronchodilator administration the FRC remained significantly higher in the LDA than in the SDA group (143 ± 6% versus 117 ± 6% predicted, respectively, p < 0.01). Both RV and RV/TLC remained greater in the LDA than in the SDA group after bronchodilator administration.

Flow-Volume Relationships

To determine whether airflow abnormalities were reflections of airway resistance rather than of volume, we evaluated flow- volume relationships. Baseline FEV₁/FVC was slightly lower in the LDA than in the SDA group (64.4 ± 2.1 versus 69.9 ± 1.7, respectively; p < 0.05), and remained lower in the LDA group after bronchodilator administration (64.2 ± 2.2 versus 71.7 ± 1.4, respectively; p < 0.005). Both baseline and postbronchodilator ₅₀/FVC were similarly reduced in both the LDA and SDA groups. Postbronchodilator ₅₀/TLC was significantly lower in the LDA than in the SDA group. Regression analyses revealed an inverse correlation between duration of asthma and both FEV₁/FVC (r = 0.24, p < 0.04) and ₅₀/TLC (r = 0.24, p < 0.02) (data not shown). The decrease in flow:volume ratios with duration of asthma suggested that the airflow limitation was a reflection of resistance.

Interaction Between Flow, Volume, and Duration of Asthma

To analyze the relationship between airflow limitation and hyperinflation, we performed regression analyses on data from the total study cohort. Both FRC% predicted and RV% predicted were regressed on several parameters of airflow, including FEV₁/ FVC, FEV₁% predicted, ₅₀, and ₂₅. An inverse correlation was demonstrated between RV and FEV₁/ FVC (r = 0.41, p < 0.0005), FEV₁ (r = 0.32; p < 0.008), ₅₀ (r = 0.43, p < 0.0002), and ₂₅ (r = 0.30, p < 0.01). Similar relationships were found between FRC and FEV₁/FVC (r = 0.60, p < 0.0001), FEV₁ (r = 0.30; p < 0.02), ₅₀ (r = 0.52, p < 0.0001), and ₂₅ (r = 0.41, p < 0.0005).

These findings raised the question of whether the degree of hyperinflation was determined by the degree of airflow obstruction, the duration of disease, or the interaction of these two factors. Multiple linear regression models were used to determine the relationships between either FRC or RV and several parameters of airflow, duration of asthma, and the interaction of airflow and duration. The airflow parameters tested included FEV₁% predicted, FEV₁/FVC, ₅₀% predicted, ₅₀/FVC, ₂₅% predicted, and ₂₅/FVC. In Table 3, standardized  coefficients are presented to allow comparison of the strength of the association between FEV₁/FVC or duration of disease and RV or FRC. Using this model, we noted a significant relationship between FEV₁/FVC and RV, but not between the duration of disease and RV. In contrast, both FEV₁/FVC and duration of asthma were found to be statistically significant predictors of FRC. A similar pattern was demonstrated for each airflow parameter tested.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A large body of evidence suggests that airway inflammation in asthma leads to distortion of bronchial architecture and may lead to irreversible changes in pulmonary function (4, 6, 7). However, few studies have characterized the physiologic consequences of longstanding asthma (17, 26). Longitudinal studies spanning relatively short periods have shown that asthmatic individuals have an accelerated rate of decline in lung function as compared with the general population (27, 28). Elderly asthmatic individuals represent an ideal population in which to use cross-sectional analysis to investigate the consequences of prolonged airway inflammation. Our study evaluated airflow, lung volumes, and bronchodilator response as a function of duration of disease in a cohort of elderly lifetime nonsmokers with asthma. The degree of airflow limitation and hyperinflation was directly related to the duration of asthma. Although many other factors including environmental exposures, occupational exposures, prior therapy, and genetic predisposition may have contributed to the observed physiologic abnormalities in our subjects, the fact that we were able to identify a significant association between decrements in pulmonary function and the duration of asthma is striking. Furthermore, elderly subjects with longstanding asthma failed to achieve normal airflow after bronchodilator administration, in contrast to those with asthma of short duration. Our study confirms and extends previous observations suggesting that longstanding asthma leads to irreversible airway changes, and also suggests the involvement of distal airways or parenchymal remodeling in this process.

Most previous accounts of asthma in the elderly have used spirometry as their sole method of physiologic assessment (5, 17, 26, 27, 29, 30). In our study, multiple parameters of large airway function were assessed. As compared with those with SDA, subjects with LDA consistently displayed more severe airflow limitation in all measurements. Furthermore, our study suggested that the severe decrements in pulmonary function associated with long-term asthma may become irreversible. After bronchodilator administration, most of our subjects with LDA failed to achieve normal airflow. Strikingly, more than half of these subjects continued to display severe airflow obstruction after bronchodilator administration. In contrast, 50% of the subjects with SDA attained normal airflow values after bronchodilator administration. These observations are of particular significance because all of our subjects had been stabilized on inhaled corticosteroids before pulmonary function testing. The findings support the hypothesis that longstanding airway inflammation from asthma can lead to airway remodeling and irreversible airflow limitation. Although it is of interest, the contribution of small airways disease to the observed airflow abnormalities could not be independently assessed because of the profound abnormalities in large airway function.

To further describe abnormalities associated with the duration of asthma, we examined the relationship between duration of asthma and lung volumes. Both the LDA and the SDA groups had substantially increased RV and FRC. However, subjects with longstanding asthma had an even greater degree of hyperinflation. This profound degree of hyperinflation may lead to a mechanical disadvantage, and may contribute to the dyspnea and activity limitations reported by these patients.

Dykstra and coworkers demonstrated a strong correlation between the degree of airflow obstruction and the degree of hyperinflation in a cohort of subjects with obstructive lung diseases (31). We confirmed the association between airflow obstruction (FEV₁) and hyperinflation as measured by RV and FRC in our cohort of elderly asthmatic individuals. This raised questions about the effect of the duration of asthma on these observations. When both the duration of disease and airflow limitation were considered, RV remained strongly associated only with the degree of obstruction. In contrast, the increase in FRC was associated with both the duration of disease and the degree of airflow obstruction. Although increases in FRC can result from airflow obstruction, our observation suggests an additional mechanism, such as a loss of elastic recoil, which would manifest as hyperinflation in the end-tidal position.

Multiple mechanisms may lead to loss of elastic recoil in the elderly asthmatic population. Aging has been associated with loss of elastic recoil; however, this is unlikely to account for our findings, since the groups with LDA and SDA did not differ significantly from one another in age, and % predicted values were used for analysis. Loss of elastic recoil may also result from mechanical disruption, as has been described in acute asthma (32, 33). An alternative mechanism has been suggested by recent studies using transbronchial biopsies. In these studies, subjects with nocturnal asthma showed inflammation in the distal airway and alveolar tissue that was characterized by an increase in eosinophils, macrophages, and CD4⁺ lymphocytes (2, 3). Our observations suggest the hypothesis that persistent distal airway or alveolar tissue inflammation may lead to structural changes that can result in a loss of elastic recoil reflected by increases in FRC.

It is unclear whether previous therapy was a factor in our observations. Subjects with LDA had their disease diagnosed before the widespread use of inhaled corticosteroids. Although 50% of each of our two study groups reported using inhaled corticosteroids before enrollment in the study, neither the corticosteroid dose nor the compliance with or duration of this therapy could be confirmed. A prospective study would be necessary to determine the effect of treatment with inhaled corticosteroids on the decline in lung function over time in subjects with asthma. Furthermore, our study was done with a cohort of patients derived from a specialty clinic; this select group of patients with relatively severe asthma may serve as a model, but may not represent the general population of asthmatic individuals. It is important that despite the small size of our cohort, we were able to identify a significant association between the duration of asthma and physiologic measurements. It would be of interest to look at the effect of duration of asthma on physiologic parameters in additional cohorts, including those of younger age.

In summary, the data from our study suggest that longstanding asthma results in irreversible decrements in lung function, which may manifest as tobacco-independent chronic obstructive lung disease. On the basis of a human model and multiple physiologic outcomes, the results of the study support the notion of the development of fixed airway obstruction that may result from airway remodeling due to decades of persistent airway inflammation. Moreover, the observations in the study suggest that prolonged asthma is associated with increases in FRC, suggesting an additional effect of parenchymal remodeling. Long-term prospective studies correlating physiology with histology from lung biopsy samples would be necessary to further explore these hypotheses. These data raise the question of whether early intervention in childhood asthma would prevent progressive decline in lung function. The potential role of antiinflammatory agents in preventing this process awaits further study.