Compressive air trapping in asthma: effects of age, sex, and severity

Abstract

Air trapping due to airway closure has been associated with unstable asthma. In addition to airway closure that occurs at lower lung volumes during slow expiration, there may be further closure during a forced expiration because of airway compression. The purpose of this study was to define a reference range from a nonasthmatic population and investigate the characteristics of compressive air trapping in asthma. Spirometry and plethysmography were performed in 117 nonasthmatic subjects (ages 18–87 yr) and 153 asthma subjects (ages 12–72 yr). Air trapping was assessed as residual lung volume and the ratio of forced expiratory vital capacity (FVC) to slow inspiratory vital capacity (iVC) (FVC/iVC). There were no significant age or sex effects on the FVC/iVC ratio in the nonasthmatic subjects, and a fifth percentile lower limit of normal (LLN) of 0.93 was computed. An FVC/iVC ratio less than LLN defined compressive air trapping. Asthma subjects exhibited an age-related decline in the FVC/iVC ratio of 0.0027 per year (P < 0.0001) in a mixed effects model, with additional decreases associated with severe asthma and male sex. FVC/iVC ratios< LLN were infrequent in subjects <30 yr but evident in most asthma subjects >50 yr. Lung residual volumes followed similar patterns of greater elevations in subjects with severe asthma, older age, and male sex. Compressive air trapping occurs frequently in older asthmatics, appearing to be a feature of the natural history of asthma that is greater in severe asthma and men. This component of premature airway closure affects spirometric assessment of airway function and may contribute to asthma symptoms during physical exertion.

NEW & NOTEWORTHY Premature airway closure during exhalation is a component of airway obstruction that is associated with asthma severity and instability. Compressive air trapping is airway closure that is more extensive during a forced exhalation than with a slow, passive exhalation. We report that compressive air trapping occurs in most people > 50 yr with asthma, affects men more than women, and persists after bronchodilator treatment. This component of obstruction appears to be part of the natural history of asthma.

INTRODUCTION

Air trapping due to peripheral airway closure is a component of airflow obstruction that is associated with asthma severity and instability (8, 11, 13, 22, 23, 24). In addition to premature airway closure that occurs during slow exhalation, there may be a component of airway closure that occurs during a forced exhalation if pleural pressure exceeds airway luminal pressure sufficiently to compress the airway (16). Forced expiratory vital capacity (FVC) values less than slow expiratory VC under conditions of matched end-inspiration lung volumes and total expiratory times were reported in patients with obstructive lung disease, including some with asthma (2). The authors reasoned that compressive airway closure might occur because of the changes in airflow dynamics during the forced expiration (2). A lower FVC relative to slow VC was also reported as a feature of a subgroup of subjects with severe asthma who had eosinophils in biopsies of their airway mucosa (31) and in asthmatics with moderate-to-severe airflow obstruction (3). Although this measurement may have potential value in identifying patterns of airway obstruction that have relevance to asthma phenotypes and tolerance to physical exertion, little is known about the occurrence of compressive air trapping in the broader population of asthmatics and in older people with normal airways. The goal of this study was to determine the patterns and prevalence of compressive air trapping in a well-characterized cohort of severe and nonsevere asthma subjects compared with people of a similar age range without obstructive airway disease.

MATERIALS AND METHODS

All procedures were approved by the University of Wisconsin-Madison Heath Sciences Institutional Review Board, and written consent was obtained for each Severe Asthma Research Program (SARP) subject. Data for asthma subjects were obtained from the Madison cohorts for SARP-2 and SARP-3. Data for nonasthmatic subjects were obtained from 18 SARP-2 participants and a convenience sample of 99 patients without a respiratory disease diagnosis who were referred to the University of Wisconsin Hospitals and Clinics Pulmonary Function Laboratory. All subjects were current nonsmokers, and those who were former smokers had less than 10 pack-years history. Subjects were excluded from analysis if they had a history of vocal cord dysfunction or evidence of glottal closure during spirometry or plethysmography, and subjects in the nonasthmatic group were excluded if the ratio of forced expiratory volume in one second (FEV₁) to FVC (FEV₁/FVC) was less than the LLN. In the asthma groups, 51% of the subjects had data from longitudinal visits in addition to the index visit, totaling 353 measurements of FVC to slow inspiratory vital capacity (iVC) (FVC/iVC) in 153 subjects and spanning up to 10 yr in some subjects who were enrolled in both SARP-2 and SARP-3.

Spirometry and plethysmography were conducted according to American Thoracic Society (ATS)/European Respiratory Society (ERS) guidelines (17, 30). For asthma subjects, measurements of plethysmographic lung volumes and spirometry before and after bronchodilation were obtained at a single visit in the following order: plethysmography, spirometry, maximal bronchodilation, spirometry, and plethysmography. Asthmatic subjects were studied during a stable period relative to their asthma control, no less than 4 wk after recovery from an exacerbation or the last dose of an acute course of systemic corticosteroid therapy, and at least 2 wk after the last dose of an antimicrobial agent used for a respiratory illness. Subjects were instructed to hold short-acting β-agonists for >4 h, ipratropium for >6 h, long-acting β-agonists for >12 h, and leukotriene modifiers or ultralong-acting bronchodilator agents >24 h and to avoid caffeine before the visit. Bronchodilation was initiated with 4 actuations of albuterol aerosol via valved holding chamber 1 min apart, followed by spirometry and 2 additional actuations 15 min later. If the FEV₁ after 6 actuations increased ≥5% from the post-4 actuation level, an additional 2 actuations were administered (8 actuations total), and the highest FEV₁ and FVC among all the postalbuterol measurements were recorded as the maximums. The nonasthmatic subjects had one set of measurements, and most did not receive albuterol. For plethysmography, the subjects breathed quietly until the end-expiratory volume was stable and then panted slowly against a closed mouthpiece, followed by a slow maximal exhalation and a maximal inhalation to obtain measurements of residual lung volume (RV), slow iVC, and total lung capacity (TLC). Predicted values (Prd) and Z-scores were computed for FEV₁, FVC, and the FEV₁/FVC using the 2012 Global Lung Function Initiative (GLI) equations (19). The GLI Prd for FVC were also used as the Prd for iVC. Prd for TLC and residual lung volume relative to total lung capacity (RV/TLC) were computed using the 1995 ATS Workshop equations (25) and ATS recommendations for ethnicity adjustments of TLC Prd (18). Asthma subjects were classified as severe or nonsevere in the SARP protocols (12, 26). In SARP-2, the classification at enrollment also applied to the longitudinal visits, whereas in SARP-3 the subjects were reevaluated for asthma severity at each longitudinal visit; therefore, the classification changed for a small number of subjects among the longitudinal visits. The index visit was the first visit in SARP-3 for which complete plethysmography and spirometry measurements were available or the first visit in SARP-2 with complete measurements for those subjects not participating in SARP-3.

Multidetector high resolution computed tomography (HRCT) measurements of lung density at full inspiration were obtained after bronchodilation as described (20). The relative areas (RAs) of low attenuation were defined as percentage of the inspiratory lung volume on HRCT < −950 Hounsfield units (HU) (RA < −950 HU). Measurements of RA < −950 HU were available for 53 adults in the asthma group and used as an indicator for inspiratory hyperinflation that may result from substantial emphysema.

Data analyses.

The FVC/iVC ratio was employed as an indicator for compressive air trapping. Lacking published normative data for older ages, we evaluated the FVC/iVC in the nonasthmatic group and used those data to define the 5th percentile as the LLN. Linear mixed effects models were used to estimate effects of asthma, sex, and age, incorporating all available index and longitudinal measurements, with subject as a random effect to control for repeated measures within a subject. For presentation of vital capacities, residual volumes, and bronchodilator effects, we created subgroups of age, sex, and asthma severity, computing means and standard errors with linear mixed effects models. General linear models were used for data involving only the index visit. The Akaike information criterion was used as a guide for including variables in the multivariable analyses. SYSTAT v13.1 (SYSTAT Software, San Jose, CA) was used for statistical analyses.

RESULTS

The characteristics of the nonasthmatic, nonsevere asthma, and severe asthma study groups, as measured at the index visit, are summarized in Table 1. In data from 123 asthma subjects (50% severe) at the index visit, controller medications included daily inhaled corticosteroids (75%), long-acting β-agonists (64%), leukotriene modifiers (29%), daily systemic corticosteroids (7%), long-acting muscarinic antagonists (5%), and biologics (4%).

Table 1.

Characteristics of the study groups at the index visit

Variable	No Asthma	Nonsevere Asthma	Severe Asthma
n	117	81	72
Female, %	55	49	61
Age, yr, median  (range)	55 (18–87)	24 (12–68)	47 (13–72)
BMI, kg/m2, median  (IQR)	28 (25–32)	25 (23–28)	31 (26–35)
Ethnicity, %
White	98	86	82
Black	1	11	17
Other	1	2	1
FVC, mean (SD)
%Predicted	104 (12.6)	99 (11.8)	85 (17.6)
Z-score	0.28 (0.89)	−0.08 (0.93)	−1.10 (1.34)
FEV1/FVC,  mean (SD)
Ratio	0.79 (0.05)	0.74 (0.08)	0.68 (0.11)
Z-score	−0.23 (0.70)	−1.39 (1.00)	−1.77 (1.27)
FEV1, mean (SD)
%Predicted	102 (12.9)	88 (13.6)	72 (19.6)
Z-score	0.17 (0.93)	−0.96 (1.06)	−1.99 (1.36)
iVC, %Predicted,  mean (SD)	104 (13.0)	101 (11.6)	93 (15.1)
TLC, %Predicted,  mean (SD)	108 (12.5)	110 (11.2)	111 (14.0)
RV/TLC,  mean (SD)
Ratio	0.34 (0.07)	0.31 (0.07)	0.41 (0.08)
%Predicted	96 (14.1)	111 (20.7)	124 (18.5)
FVC/iVC, ratio,  median (IQR)	1.00 (0.98–1.02)	0.99 (0.95–1.01)	0.93 (0.85–0.97)

BMI, body mass index; FEV₁/FVC, ratio of forced expiratory volume in one second (FEV₁) to forced expiratory vital capacity (FVC); FVC/iVC, ratio of FVC to slow inspiratory vital capacity (iVC); IQR, interquartile range; RV/TLC, residual lung volume relative to total lung capacity.

In the nonasthmatic subjects, the FVC/iVC ratio had a normal distribution with no significant effects of sex (P = 0.14) or age (P = 0.063) in a general linear model (mean of 1.00 and SD 0.042) (Fig. 1). The estimated fifth percentile (mean −1.64 SD) was 0.931, which we defined as the LLN for this study.

Fig. 1.

Forced expiratory vital capacity-to-slow inspiratory vital capacity (FVC/iVC) ratio measured in 117 nonsmoker subjects with normal airways, ages 18–87, 55% females. A: scatterplot by age of individual subjects. B: histogram showing distribution of the measured variable. The lower limit of normal (LLN) is defined as the estimated 5th percentile.

In contrast to the lack of significant effect of age on FVC/iVC in nonasthmatic subjects, the subjects with asthma exhibited a decline of 0.0027 in FVC/iVC per year (95% confidence limit 0.0022–0.0033; P < 0.0001; mixed effects model). There was no significant interaction between age and sex or between age and asthma severity in the model, indicating that the age effect had a similar slope for all subgroups of asthma subjects. Men had significantly lower FVC/iVC than women (mean difference 0.077, 95% confidence limit 0.054–0.101, P < 0.0001). Additionally, subjects with severe asthma had lower FVC/iVC than those with nonsevere asthma [men mean difference 0.061 (P < 0.0001); women mean difference 0.018 (P < 0.001); sex-by-severity interaction P = 0.005]. Figure 2 illustrates the age-related declines in FVC/iVC in the asthma groups. Whereas FVC/iVC < LLN occurred infrequently in asthma subjects <30 yr of either sex, most of those >50 yr exhibited FVC/iVC < LLN, indicating that compressive air trapping is a common feature in older people with asthma.

Fig. 2.

Age-associated patterns of forced expiratory vital capacity-to-slow inspiratory vital capacity (FVC/iVC) ratio in 153 nonsmoker subjects with asthma by sex and asthma severity. Effects of age, sex, and asthma severity, determined with a mixed effects model, are shown as the gray line for each subgroup. Longitudinal measures for individual subjects are connected with lines. LLN, lower limit of normal.

We conducted an exploratory assessment of other variables of interest in the asthma subjects to determine if there were factors significantly associated with the FVC/iVC ratio when added to age, sex, and severity in a general linear model. All of the following variables had no significant contribution in addition to the age, sex, and severity effects in the model: blood eosinophils, sputum percent eosinophils, serum total IgE, fraction of expired nitric oxide, age at asthma diagnosis, duration of asthma, smoking history, body mass index, waist circumference, hip circumference, RV/TLC %Prd, and TLC %Prd. Substituting years duration of asthma for chronological age in the model resulted in a loss of precision in the model as reflected by a lower model R² and a higher Akaike information criterion.

If loss of lung elastic recoil due to emphysema were causing the observed compressive air trapping, there may be evidence of enlarged airspace at maximal inspiration. We utilized the inspiratory RA-950 HU data from the subgroup of adult asthma subjects who received HRCT studies. All the subjects had <6% of lung volume (median 1.2%; interquartile range 0.7–2.2%) identified as hyperinflation, and the measured percentage less than −950 HU was not a significant predictor of FVC/iVC ratio when added to the effects of sex, age, and asthma severity in a general linear model (P > 0.8).

Figure 3 summarizes iVC and FVC before and after bronchodilation by sex, asthma severity, and age group (based on median age of 36 yr in the asthma group at the index visit). In the younger group, the iVC and FVC were not significantly different from each other and were mostly within the normal reference range. In the older asthma subjects, the iVC and FVC were reduced relative to the younger subjects (P < 0.01) and reduced to a greater extent in severe versus nonsevere asthma (P < 0.013). As shown by the FVC/iVC ratios in Fig. 2, the reductions associated with age, male sex, and severe asthma were greater for FVC than iVC. Both iVC and FVC increased after bronchodilation with albuterol in both age groups, the changes in volume being similar for iVC and FVC (Fig. 3; P > 0.3), except for men with severe asthma, in whom the response to bronchodilation was slightly greater for FVC than iVC (Fig. 3; P = 0.004).

Fig. 3.

Comparisons of slow inspiratory vital capacity (iVC) and forced vital capacity (FVC) before bronchodilation (pre-BD) and after bronchodilation (post-BD) by age group, sex, and asthma severity. Each symbol indicates the group mean and SE, determined with a mixed effects model for each asthma subgroup, and from the single set of measurements in the nonasthmatic group.

There were no significant differences in TLC %Prd among the three study groups (P = 0.25; Table 1) and negligible, although statistically significant, changes in TLC with bronchodilation (mean 0.018 liter increase, P = 0.03), suggesting that differences in vital capacities were related to air trapping. Figure 4 shows analogous patterns of air trapping measured as RV relative to TLC. Subjects with severe asthma had elevated RV/TLC %Prd compared with both nonsevere asthma and nonasthmatic groups in men and women of both age ranges before bronchodilation (P < 0.01; Fig. 4). Subjects with nonsevere asthma had elevated RV/TLC %Prd compared with the nonasthmatic group in the older (P < 0.0001) but not younger age groups. After bronchodilation in the asthma groups, there were no significant differences in RV/TLC %Prd among severe, nonsevere, and nonasthmatic groups of age <36 yr (P > 0.15), but the differences persisted among the three groups in the older subjects (P < 0.01; Fig. 4).

Fig. 4.

Comparisons of residual lung volume relative to total lung capacity (RV/TLC) for each age group among subjects with normal airways, nonsevere asthma, and severe asthma, and the postbronchodilator (Pst) measurements in the asthma groups. Each symbol indicates the group mean and SE, determined with a mixed effects model for each asthma subgroup and from the single set of measurements in the nonasthmatic group.

Figure 5 illustrates the patterns of the compressive air trapping relative to elevations in RV at the index visit. These two indicators of air trapping are largely independent of one another, as indicated by statistically insignificant Spearman’s correlation coefficients for the younger asthma subgroup (r_s = 0.001, P = 0.99), the older asthma subgroup (r_s = −0.19, P = 0.10), the pooled asthma group (r_s = −0.12, P = 0.14), and the nonasthmatic group (r_s = 0.05, P = 0.57).

Fig. 5.

Scatter plot of forced expiratory vital capacity-to-slow inspiratory vital capacity (FVC/iVC) ratio vs. residual lung volume measured after a slow expiration (RV/TLC %Predicted) at the index visit.

DISCUSSION

The data show that people with asthma develop compressive air trapping with age and that it is more prominent in men than in women and more prominent in severe than in nonsevere asthma. Chronological age is a more robust predictor than duration of asthma in the models. The similar slope of decline in the FVC/iVC ratio with age among the sex and severity subgroups suggests that this phenomenon is part of the natural history of asthma, underlying the superimposed effects of these other factors. This is in contrast to adults without airway disease, who exhibited no significant decline of FVC/iVC into advanced age.

The mechanism of compressive air trapping likely involves a combination of loss of lung elasticity and an increase in airflow resistance in the peripheral airways. According to the equal pressure point concept presented by Mead et al. (16), a forced expiratory effort will create positive intrathoracic pressure, during which maximal airflow is determined directly by the lung elastic recoil and inversely by the resistance of the peripheral airways. As air flows through the conducting airways, the intraluminal pressure decreases because of flow resistance and acceleration, eventually reaching a point where the intraluminal and intrathoracic pressures are equal. Downstream from the equal pressure point, the intraluminal pressures are less than the intrathoracic pressure, resulting in potential compression of the airway. In normal airways, the equal pressure point is located in central airways (15, 21), which have cartilage and thicker walls that are resistant to compression, and in asthma the central airways may be even stiffer because of remodeling (1, 20). However, if lung elastic recoil is reduced or peripheral airway resistance is increased, the equal pressure point will shift to a more peripheral location, where the bronchioles are less resistant to collapse, and airway closure may result.

The same patterns of asthma, older age, and male sex are also apparent with air trapping measured as RV/TLC with a slow expiration from tidal breathing to RV (Fig. 4). However, the changes in FVC/iVC and RV/TLC are not quantitatively correlated, even within the older asthma subjects, suggesting that the underlying mechanisms and locations of closure are not necessarily related (Fig. 5). The partial reversal of the air trapping with bronchodilation indicates that airway smooth muscle tone increases the probability of premature closure in some of the airways. However, the fact that FVC continues to be less than iVC after bronchodilation in the older asthmatics suggests that reducing airway smooth muscle tone is not sufficient to move the equal pressure point into a generation of less-compressible airways during a forced expiration. The older asthma subjects also had persistent elevations of RV/TLC after bronchodilation relative to the RV/TLC of the older nonasthmatic group, indicating persistent airway instability with slow expiration as well. The persistence of air trapping after bronchodilation may be related to heterogeneous distribution of inhaled albuterol, along with changes related to airway inflammation and remodeling or loss of lung elasticity.

It is well established that lung elastic recoil decreases with age in healthy individuals (7, 14, 28). Both sexes exhibit decreases in recoil with age, but men normally have higher elastic force than women within each age range (14). A loss of recoil was also reported in younger people with uncontrolled asthma or acute exacerbations of asthma, with recovery of lung elasticity after achieving clinical improvement of asthma (9, 33). However, some of the fluctuations in transpulmonary pressure associated with loss of asthma control in those reports may be due to the effects of airway closures and trapped gas on the measurements of pleural pressures. Gelb et al. (6) have accumulated data from 25 older adults with moderate-to-severe asthma and persistent airflow limitation that reveal approximately equal contributions of peripheral airway resistance and reduced lung elastic recoil to the airflow limitation. They report mild centrilobular emphysema that was apparent in the lungs of four of the subjects who died of causes unrelated to their asthma, despite having a history of only mild inspiratory hyperinflation by HRCT and normal carbon monoxide diffusion studies (6). It is possible that analogous emphysematous lesions were contributing to the air trapping observed in the current study, although the subgroup for which inspiratory HRCT was available had hyperinflation scores even smaller than those reported in the Gelb study. Diffusing capacity studies were not included in the SARP protocols. Although the prevalence of emphysema and its quantitative contribution to lung elasticity in the broader asthma population remain undefined, an accelerated loss of elastic recoil with age in people with asthma is consistent with the patterns of slow-expiratory and compressive air trapping observed in the current study.

Mucous plugs occur commonly in subsegmental airways of adults with asthma, and FVC %Prd is reduced in subjects having ≥4 segments with one or more plugs visible on multidetector computed tomography scans (5). Although collateral ventilation in the peripheral airways may prevent atelectasis and physiological shunts in the acini served by the plugged segments (27), it is likely that alternate pathways of ventilation would add airflow resistance that might contribute to compressive airway closure.

Previous studies have reported reduced FVC relative to slow VC in people with other obstructive airway diseases. In a large sample of patients referred for preoperative evaluation of pulmonary function, predictors of FVC < slow VC were reduced FEV₁ and older age, but no difference between sexes was detected (29). In COPD subjects, FVC < iVC, along with age and FEV₁, was found to be a predictor of reduced exercise tolerance (peak V̇o₂/kg) (34). Bronchodilator reversal of obstruction in COPD patients increased FVC more than slow VC, resulting in a narrowing of the slow VC-to-FVC difference (10), which is consistent with the men with severe asthma in the current study. The FVC/iVC ratio was also reported to decrease in early stages of bronchiolitis obliterans syndrome following bilateral lung transplant, reflecting the small airways obstruction independent of changes in FEV₁ (4). The current study shows older asthmatics developing compressive air trapping, which can be a source of obstruction that may not be reversible with bronchodilators and may be a component of the asthma-COPD overlap phenomenon (32).

One limitation of the current study is that 85% of the nonasthmatic group consists of patients referred to the pulmonary function laboratory for clinically-indicated evaluation. Indications included routine preoperative evaluations and monitoring for planned exposures to potentially pneumotoxic medications, but others had reported respiratory symptoms such as cough or shortness of breath. All included subjects had no current diagnosis of respiratory disease, were nonsmokers or had limited past smoking history, and had normal FEV₁/FVC and the ability to perform spirometry and plethysmography adequately. Although these subjects may have had significant health problems, it is unlikely that they had asthma or COPD, and the LLN derived from these data should be interpreted in that context.

Other limitations of the current study are the inferences from mostly cross-sectional data that reveal a linear association of the FVC/iVC ratio with age in the asthma subjects (Fig. 2). Although longitudinal data spanning up to 10 yr are available for about half the subjects, the age effect is estimated from data representing more than six decades. The apparent linearity of the age effect revealed by the mixed effects model is a novel and unexpected finding that needs to be replicated and validated in future studies, but it may be important for the understanding of patterns of airway obstruction in older people with asthma.

In conclusion, people with asthma develop a compression component of air trapping with age that appears to be a feature of the natural history of asthma. This phenomenon may contribute to persistence of obstruction after bronchodilation and could limit the ventilatory response to physical exertion. Patterns of airway obstruction may be identified as relative components of airflow limitation and airway closure, and this study illustrates that airway closure may be apportioned further into passive and compressive components.

GRANTS

This study was supported by National Heart, Lung, and Blood Institute Grants U10-HL109168 and R01-HL069116.

DISCLOSURES

S. B. Fain serves as a consultant regarding quantitative CT imaging protocols for the COPD Foundation and receives grant funding from GE Healthcare. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

R.L.S. conceived and designed research; R.L.S., M.J.O., and S.B.F. performed experiments; R.L.S., C.K., M.J.O., and S.B.F. analyzed data; R.L.S., C.K., M.J.O., S.B.F., and N.N.J. interpreted results of experiments; R.L.S. prepared figures; R.L.S. drafted manuscript; R.L.S., C.K., M.J.O., S.B.F., and N.N.J. edited and revised manuscript; R.L.S., C.K., M.J.O., S.B.F., and N.N.J. approved final version of manuscript.