Abstract
BACKGROUND: Reversibility of obstructive lung disease is traditionally defined by changes in FEV1 or FVC in response to bronchodilators. These may not fully reflect changes due to a reduction in hyperinflation or air-trapping, which have important clinical implications. To date, only a handful of studies have examined bronchodilators' effect on lung volumes. The authors sought to better characterize the response of residual volume and total lung capacity to bronchodilators.
METHODS: Responsiveness of residual volume and total lung capacity to bronchodilators was assessed with a retrospective analysis of pulmonary function tests of 965 subjects with obstructive lung disease as defined by the lower limit of normal based on National Health and Nutritional Examination Survey III prediction equations.
RESULTS: A statistically significant number of subjects demonstrated response to bronchodilators in their residual volume independent of response defined by FEV1 or FVC, the American Thoracic Society and European Respiratory Society criteria. Reduced residual volume weakly correlated with response to FEV1 and to FVC. No statistically significant correlation was found between total lung capacity and either FEV1 or FVC.
CONCLUSIONS: A significant number of subjects classified as being nonresponsive based on spirometry have reversible residual volumes. Subjects whose residual volumes improve in response to bronchodilators represent an important subgroup of those with obstructive lung disease. The identification of this subgroup better characterizes the heterogeneity of obstructive lung disease. The clinical importance of these findings is unclear but warrants further study.
Introduction
Evaluation of obstructive lung disease includes pulmonary function tests with pre- and postbronchodilator measurements. Reversibility in obstructive lung disease is defined by a change in either FEV1 or FVC.1 Many studies focus only on FEV1 as the measure of reversibility,2–8 and some have argued that FVC is an underutilized clinical outcome of reversibility.9
FEV1 and FVC are not the only parameters measured in pulmonary function testing that change in response to bronchodilators. Lung volumes have been found to be responsive to bronchodilators independent of FEV110–17. Lung volumes have the potential to be useful parameters in determining bronchodilator responsiveness, but they are not frequently used.10
Lung volumes have important clinical implications. Static hyperinflation is a risk factor for mortality.18 Dynamic hyperinflation increases the work of breathing19 and has elements of reversibility that correlate with inspiratory capacity but not with FEV1.15,20 There is a significant correlation between inspiratory capacity and maximal oxygen consumption.21
FEV1 has important predictive value regarding symptomatology, frequency of complications, and overall prognosis.4,19 FEV1 is the most important tool in assessing obstructive lung disease, but it has limitations: Although FEV1 is a good predictor of mortality, other indices such as walk distance are better predictors.22,23
FEV1 responsiveness is not a reliable indicator of increase in exercise tolerance as a response to bronchodilators20,24 or improvement in dyspnea.7 FEV1 responsiveness has limited value in predicting long-term outcomes, such as hospitalizations and mortality.25,26 We hypothesized that there are a significant number of patients whose residual volumes and total lung capacities are bronchodilator-responsive even if their FEV1 or FVC are not.
QUICK LOOK
Current knowledge
There is disagreement about what constitutes bronchodilator reversibility in pulmonary function testing, but the focus is primarily on FEV1. A small number of studies have looked at other measures of reversibility, such as residual volume (RV) and the RV/TLC ratio.
What this paper contributes to our knowledge
A sizable number of subjects had a reduction in their RV but did not meet American Thoracic Society criteria for reversibility in bronchodilator testing. The change in RV cannot be accounted for by change in forced vital capacity.
Methods
A retrospective analysis of pulmonary function tests was performed for patients who underwent pulmonary function testing between January 1, 2005 and August 31, 2013 at Saint Louis University Hospital (St Louis, Missouri). Subjects 18–79 y old with an FEV1/FVC less than the lower limit of normal based on the National Health and Nutrition Education Survey (NHANES) III predicted equation27 were included.
Credentialed pulmonary function technologists performed pulmonary function testing. Spirometry and lung volumes by body plethysmography (MCG Diagnostics, St Paul, Minnesota) were performed, according to the 2005 American Thoracic Society and European Respiratory Society (ATS/ERS) recommendations. Subjects withheld bronchodilators for 24 h before assessment of β-agonist reversibility. To evaluate reversibility, each subject inhaled 4 puffs of albuterol using a spacer with 30-s intervals between inhalations. Spirometry was performed 15 min after bronchodilator administration, followed immediately by measurement of lung volumes. For subjects in whom there were serial tests, only data from their first visits were included.
Subjects were divided into groups based on the degree of obstruction as defined by the 2005 ATS/ERS guidelines.1 FEV1 ≥ 70% was defined as mild obstruction, FEV1 ≥ 60% but < 70% as moderate obstruction, FEV1 ≥50% but < 60% as moderately severe obstruction, FEV1 ≥ 35% but < 50% as severe obstruction, and FEV1 < 35% as very severe obstruction. FVC, FEV1, total lung capacity (TLC), residual volume (RV), and FEV1/FVC were analyzed pre- and postbronchodilator using median and interquartile ranges for outcomes in response to bronchodilators by degree of obstruction. A Mann-Whitney U test was performed to determine whether there were statistically significant differences between the degrees of obstruction of each lung volume.
For FEV1 and FVC, 12% and 200 mL were used to determine a clinically important response to the bronchodilator per the ATS/ERS criteria.1 There is currently no accepted value for clinically important bronchodilator changes for TLC or RV, and various criteria were evaluated to define a clinically important change for the purposes of our study. We attempted to create a threshold at a similar level to the defined standards of FEV1. In our population, the ATS/ERS criterion of 12% change defining clinically important response was at the 56th percentile in our patient population. Applying this to other lung volumes, the 56th percentile in RV corresponded to an 8% change in response to bronchodilators; for TLC, it corresponded to < 1% change.
To further evaluate an appropriate threshold to define clinically important change, we performed receiver operating characteristic (ROC) curves for TLC and RV compared with FEV1 and FVC. For RV, the best ROC curve against FEV1 was generated at 8%, and against FVC the best ROC was between 8 and 9%. For TLC, the ROC was not helpful for generating a suggested threshold.
We reviewed coefficients of variation to help decide on various thresholds. Upon reviewing the literature,28 we found the coefficient of variation for TLC to be in the range of 2.5–4, suggesting that thresholds of reversibility of 5 and 8% were reasonable. For RV, a handful of studies reported a coefficient of variation in the range of 6–7, which suggests a cutoff of 12 or 15%. Most of the coefficients of variation for RV were 9–12, and for that reason an RV cutoff of 20% was used.
Prior studies were reviewed to help select appropriate cutoffs. Most prior studies used a 10% change in response to bronchodilators as a threshold for clinically important response,11–14,20 and therefore this value was also included in our analysis. Twelve percent was also included, because this is the ATS criterion for FEV1 and FVC.1 In our analysis, adding a volume threshold of 200 mL did not change our number of responders. A threshold of 15% change was also analyzed, in part because older guidelines suggested a 15% threshold for FEV1 and FVC.29 Twenty percent was selected for RV because of prior studies.17,30 Analyses of both RV and TLC measurements were grouped by degrees of obstruction as defined by ATS/ERS.
The ATS criteria (either FEV1 or FVC responsiveness) were compared with both total lung capacity and residual volume responsiveness using a chi-square test to assess independence of RV and TLC responsiveness, respectively. A P value of ≤ .05 was considered statistically significant and therefore represented a response that was independent.
Using Spearman's correlation, scatter plots were created showing the percentage of change in FEV1 correlated to the percentage of change in FVC, RV, and TLC. Similar plots were created to correlate the percentage of change in FVC with the percentage of change of RV and of TLC.
Median and interquartile ranges were used in interpreting the pulmonary function test data in Table 1 rather than mean and SD to better guard against the effects of outliers. This was done because there were a few extreme outliers in our data set that we did not want to exclude, but we also wanted to guard against skewing our data set.
Pulmonary Function at Each Degree of Obstruction
Given the retrospective nature of the data analysis, informed consent was not obtained, because there was no risk to patient well-being. Safeguards were implemented to protect patient data. The Saint Louis University Internal Review Board approved this research project.
Results
There were 965 subjects, and demographics are shown in Table 2. The population had a slight male predominance, and it was predominantly white.
Baseline Characteristics
Table 1 shows the median values and percentage of change in FEV1, FVC, TLC, RV, and the FEV1/FVC. There is increasing residual volume with worsening obstruction. There was not a consistent difference in TLC with increasing obstruction.
Figure 1A shows the percentage of subjects responsive to bronchodilators as defined by ATS criteria across the ATS-defined subgroups of obstruction. Most subjects who were FEV1-responsive and RV-responsive had moderate, moderately severe, or severe obstruction, and there were fewer responders with either mild or very severe disease. There was increasing FVC response with worsening obstruction, and a similar pattern is seen with ATS criteria.
Responders and degree of obstruction. Percentage of individuals who had a clinical response to bronchodilators by degree of obstruction for FEV1, FVC, and ATS criteria (either FEV1 or FVC). Degree of obstruction was defined by 2005 ATS/ERS criteria with FEV1 ≥ 70% as mild obstruction, FEV1 ≥ 60% but < 70% as moderate obstruction, FEV1 ≥ 50% but < 60% as moderately severe obstruction, FEV1 ≥ 35% but < 50% as severe obstruction and FEV1 < 35% as very severe obstruction (A). Residual volume (RV) reversibility at different degrees of obstruction percentage of individuals who were RV-responsive at thresholds of 8, 10, 12, 15, and 20% reduction of measured lung volumes are shown across the degrees of obstruction as defined by 2005 ATS/ERS criteria (B). Total lung capacity (TLC) reversibility at different degrees of obstruction. Percentage of individuals who were TLC-responsive at thresholds of 5, 8, 10, 12 and 15% reduction of measured lung volumes are shown across the degrees of obstruction as defined by 2005 ATS/ERS criteria (C).
Figure 1B shows RV responsiveness at different thresholds for percentage change: 8, 10, 12, 15, and 20%. The higher the threshold for RV responsiveness used, the fewer patients were responsive. The same pattern persisted across the ATS degrees of obstruction at all thresholds, with most reversibility present in the intermediate obstruction range and the smallest number of responders at the ends of the spectrum.
Figure 1C shows TLC responsiveness at different thresholds for percentage change: 5, 8, 10, 12, and 15%. The higher the threshold for TLC responsiveness used the fewer patients were defined as being responsive. Overall, there is a trend toward a greater TLC response with worsening obstruction.
Figure 2A shows RV responsiveness and its independence from accepted criteria for responsiveness. The subjects were divided into groups based on response to bronchodilators. The groups are those that are RV-responsive only, dual responders, null responders, and responders to ATS criteria (either FEV1 or FVC). Chi-square testing demonstrates RV response independent of ATS criteria for responsiveness.
RV Responsiveness Compared with ATS Criteria. Residual volume (RV) responsiveness and its independence from accepted criteria for responsiveness. The bar graphs show response to bronchodilators for RV alone, dual responders, nonresponders, and responders to ATS criteria (either FEV1 or FVC). Chi-square testing demonstrates RV response independent of ATS criteria (A). Total lung capacity (TLC) responsiveness, at various thresholds, and its independence from accepted criteria for responsiveness. The bar graphs show response to bronchodilators for TLC alone, dual responders, nonresponders, and responders to ATS criteria (either FEV1 or FVC). Chi square testing was performed to determine whether independence of TLC response was statistically significant (B). P-values and the exact numerical values of each response can be found in Table 3.
RV Responsiveness Compared with ATS Criteria
Figure 2B shows TLC responsiveness and its independence from accepted criteria for responsiveness. Subjects were divided into 4 groups based on response to bronchodilators. The groups consisted of TLC responsive alone, dual responders, null responders, and responders based on ATS criteria (either FEV1 or FVC). Overall TLC response was very low. The numerical data for Figure 2A and 2B can be found in Tables 3 and 4.
TLC Responsiveness Compared with ATS Criteria
There was a strong correlation between FEV1 responsiveness and FVC responsiveness with an R value of 0.74 (P < .001). There were weak correlations between RV responsiveness and that of FEV1 (R = −0.28, P < .001) or FVC (R = −0.301, P < .001). There was no correlation in responsiveness between TLC and either FEV1 (R =−0.04, P = .24) or FVC (R = −0.02, P = .48).
Discussion
Obstructive lung diseases are heterogeneous.31 Subjects in our population currently classified as having irreversible obstruction by ATS/ERS criteria have reversible residual volumes. This cohort of subjects has been largely neglected in studies of obstructive lung disease. The presence of a weak correlation between RV and FEV1 and between RV and FVC responsiveness suggests a novel group of lung volume responders who are defined as nonresponsive by ATS/ERS criteria, as does the large proportion of subjects who are RV responsive, independent of change in either FEV1 or FVC. It was surprising to find that much of the change in RV could not be accounted for by a change in FVC. Lung volume responders may be part of a clinically important cohort of patients separate from lung volume nonresponders.15,24
Most of our population that was RV responsive was in the moderate-severe group of obstruction, and those at the extremes of both mild and very severe obstruction were less RV responsive. This was consistent for RV at all thresholds used to determine a significant change. This contrasts with previous studies that used percentage of predicted lung volumes,11,12 which found a positive correlation with the degree of obstruction and the percentage of responders. Using percentage of predicted increases, the number of responders whose obstruction is worse, skews the data toward those with the most air trapping.12 This study characterizes percentage of lung volume responders by using change in measured lung volumes rather than percentage of predicted lung volumes.
At the lowest threshold for RV (8%), 21.3% of subjects had an RV reduction with nonresponsive spirometry. This represents 35.7% of subjects currently characterized as being nonresponsive to bronchodilators. At the highest threshold for RV (20%), 4.6% of our total patient population had a reduction in RV without significant changes on spirometry. Therefore, 7.7% of subjects currently characterized as being nonresponsive to bronchodilators have a > 20% reduction in their residual volume.
The total lung capacity in our cohort showed increasing response with increasing obstruction. TLC was not consistently statistically significant independent of ATS criteria at any threshold.
Lower limit of normal was used for FEV1/FVC instead of the Global Initiative for Chronic Obstructive Lung Disease (GOLD) criteria with the intent to minimize false-positives for obstructive lung disease.32,33 GOLD criteria may misclassify healthy elderly patients, because they do not account for the natural fall in FEV1/FVC with age.27,34,35 A fixed cutoff has increased sensitivity, resulting in patients without symptoms being diagnosed with disease.32,36–38 A European study of healthy, asymptomatic never-smokers found that 35% of patients > 70 y old and 50% > 80 y old would carry a diagnosis of COPD per GOLD criteria.38
Our study made no attempt to remove subjects with an asthma component. Our intent was to include all patients with obstructive disease, as defined by the lower limit of normal based on NHANES III predicted equations. We chose to include subjects with an asthmatic component to avoid eliminating patients with asthma COPD overlap syndrome.39–42
Our study is limited by the fact that it is retrospective and from a single clinical center. Reversibility in pulmonary function laboratory testing has limitations in clinical practice. Patients have been shown to benefit clinically from bronchodilators even if their pulmonary function tests do not show a short-term benefit.43,44 There is no expectation that a prospective study would alter outcomes of objective numerical data, and it is likely that our data and patient population are generalizable.
The highest percentage of our patient population was severely obstructed, which may reflect a selection bias. This is likely representative of the population at our academic institution. Referral to an academic center for pulmonary function testing is likely to select for more severely diseased patients.
There is selection bias within our patient population, as only subjects' first pulmonary function tests at our institution were included. The first pulmonary test for our patients is likely to select for less severe pulmonary function tests; although as we mention, our patients skewed toward the most severely obstructed. A future study could avoid this bias by using a random sample of patients' pulmonary function tests over a time period instead of selecting their first pulmonary function tests.
Choosing a threshold for lung volume responsiveness is difficult. There is no clear consensus on what constitutes reversibility in patients with air-flow obstruction,1,45 and reference values are typically derived from healthy subjects using prebronchodilator values.1,46,47 This presented a challenge for our study, because we were studying only subjects with airway obstruction as defined by the lower limit of normal. We believe we are able to reasonably justify our thresholds for RV and TLC with our approach of using coefficients of variation and previous studies.
By describing a novel phenotype of residual volume responders, we suggest a new approach in interpreting bronchodilator reversibility that includes RV. Our hope is that this study contributes to a better understanding of how patients respond to bronchodilators. A study of a healthy patient population may be helpful to define a threshold for RV similar to how FEV1 was defined.47
Footnotes
- Correspondence: Ravi Nayak MD, 7th Floor Desloge Towers, 3635 Vista Avenue at Grand Boulevard, St. Louis, MO 63110-0250. E-mail: nayakrp{at}slu.edu.
The authors have disclosed no conflicts of interest.
- Copyright © 2016 by Daedalus Enterprises
