0
Point and Counterpoint |

Counterpoint: Is an Increase in FEV1 and/or FVC ≥ 12% of Control and ≥ 200 mL the Best Way to Assess Positive Bronchodilator Response? NoFEV1 for Bronchodilator Response? No FREE TO VIEW

James E. Hansen, MD, FCCP; Janos Porszasz, MD, PhD
Author and Funding Information

From the David Geffen School of Medicine at UCLA and Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center.

CORRESPONDENCE TO: James E. Hansen, MD, FCCP, Department of Medicine, Harbor-UCLA Medical Center, Box 405, Torrance, CA 90502; e-mail: jhansen @labiomed.org


FINANCIAL/NONFINANCIAL DISCLOSURES: The authors have reported to CHEST that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details.


Chest. 2014;146(3):538-541. doi:10.1378/chest.14-0437
Text Size: A A A
Published online

We respect the integrity, intelligence, and experience of the experts serving on the American Thoracic Society/European Respiratory Society (ATS/ERS) pulmonary function committees1 and Tan et al2 and the validity of their data. However, we do not agree with some of their statistical analyses and interpretations of bronchodilator responsiveness. Table 9 of the ATS/ERS document states: “An increase in FEV1 and/or FVC ≥ 12% and ≥ 200 mL constitutes a positive bronchodilator response.”1 Relevant portions of this document use the terms “meaningful” and “significant” without using the term “statistically significant” (SS).

In premise 1, we show that because the variability of FEV1 in a population is not necessarily normally distributed, the ATS/ERS recommendation is invalid when applied to individual responses. Both of these and other articles used spirometric variability of a population to set positive bronchodilator responses for individuals based on 95% confidence limits (CLs). Instead, we believe that a perceptibly positive response for an individual should (nearly always) be SS.

It is statistically incorrect to use the changes of the most variable 5% of a population as a guideline to decide whether any individual’s spirometric response is meaningful, clinically significant, or SS because the variability in the population may not be normally distributed. Four relevant publications reported the mean and SD of FEV1 and FVC in their respective populations.2-5Table 1 illustrates the problem with extracts from original publications. To obtain the key volume and percent changes in each population, these authors reported the highest measured FEV1 and FVC of each subject from three preintervention (at baseline) and, often, postintervention (after rest,4 placebo,3,5 or drug2 administration) measurements. The reported population change in mean ± SD of FEV1 and FVC are shown in Table 1 (columns 1 and 2 and columns 3 and 4, respectively). Table 1 also shows the calculated 95% CLs (columns 5-8) that the study authors considered as the minimum changes of the most responsive 5% of the population. The CLs were calculated for the one-tailed or two-tailed z scores (1.65 and 1.96, respectively) times the SD of the preintervention (and sometimes postintervention values) and usually added the change in mean of the intervention.2,3,5 Each publication suggested that for another individual to have a positive bronchodilator response, the difference between the highest postintervention and preintervention ΔFEV1 or ΔFVC should exceed the 95% CL seen in the tested population (Table 1, columns 5-8). In summary, each study2-5 used its population change in mean ± SD to statistically calculate the overall abnormal, positive, or meaningful population changes of the most variable 5% of its population. Guidelines derived from the 95% CL1,2 of a population should not be used to define an individual positive responder because this would imply that only 5% of the individuals would be meaningful responders.

Table Graphic Jump Location
TABLE 1  ] Changes After Drug or Placebo Found in Volume and % FEV1 and FVC Reported by Four Author Groups With Their Calculated Changes Required To Meet Upper 95% CLs (and Become Abnormal or Positive)

Data are presented as mean ± SD unless otherwise indicated. Upper CL calculations by authors were as follows: Sourk and Nugent,3 Δmean ± 2.0 times SD; Tweeddale et al,4 1.645 times SD; Pellegrino et al,5 2.0 times SD; Tan et al,2 Δmean + 1.645 times SD. CL = confidence limit; MDI = metered-dose inhaler.

The interpretation of bronchodilator effect should consider all three satisfactory forced maneuvers before and after the intervention rather than only the single best maneuvers. Why waste valid data when nearly always three satisfactory or near-satisfactory preforced and postforced maneuvers are available to analyze? Either t tests or easy-to-do rank order tests can be used to determine whether the postmeasurement is significantly higher than the premeasurement. To be positive at a P = .05 level using rank order statistics, each of the three postintervention FEV1 values must exceed each of the preintervention FEV1 values.6Figure 1 shows the values of individual analyses.7 In this study, the number of ATS/ERS responders using the population criteria was less than one-half of those found to be individually SS responders, confirming the importance of individual analyses. In this series, ΔFEV1 ≥ 100 mL or ≥ 6.0% were SS by unpaired one-tailed t tests (P < .05) in > 95% of cases. Rank order and t test P values for ΔFEV1 agreed > 99% and for ΔFEV3 > 96.5%. Rank order tests for ΔFEV1 and ΔFEV3 agreed > 92%.

Figure Jump LinkFigure 1  Bronchodilator responses of FEV1 in 316 patients at Harbor-UCLA Medical Center Clinical Laboratory.7 Changes in consecutive patients with three acceptable forced expiratory maneuvers before (baseline) and after receiving nebulized albuterol sulfate solution. Individual positive increases in highest FEV1 in milliliters (x-axis) and percentages (y-axis) are shown, whereas the diagonal lines indicate baseline FEV1 in liters. Data for 51 patients not shown in the figure showed a negative or no response. The 68 crosses in the upper-right unshaded area (including three outside the range of the figure) depict the patients with increases in FEV1 of ≥ 200 mL and ≥ 12.0%. The 102(mostly in the shaded area) indicate other patients with statistically significant (SS) increases in FEV1 ≥ 100 mL or ≥ 6.0% (SS by t test). The(mostly in the lower-left unshaded area) indicate 95 patients who had nonsignificant, small positive responses. Of the 316 patients, only 14 with increases in FEV1 ≥ 100 mL and/or ≥ 6.0% were not SS. All 68 with increases in FEV1 ≥ 200 mL and ≥ 12.0% also had SS increases of FEV3, whereas 94 of the other 102 SS responders also had SS increases of FEV3 volumes.Grahic Jump Location

The meaningful changes based on variability used by the ATS/ERS guidelines1 (≥ 200 mL and ≥ 12.0%) are assumed to be identical for every individual regardless of age, height, sex, state of health, or precision and accuracy of the patient-equipment-technician interaction. However, individuals being assessed for bronchoreversibility may have baseline FEV1 values from < 400 mL to > 5,000 mL. Patients’ ΔFEV1 shown in Figure 1 varied considerably between low and high FEV1, depending on the baseline volume. Because the size of individual variability of FEV1 partially depends on its baseline value (Fig 2), population 95% CL should not be used to evaluate individuals. The data of Tweeddale et al4 (Table 1, columns 2-6 and 8, rows 2-4) and Tan et al2 (all columns, rows 7-9) confirm that change in volume and percent partially depend on the degree of airway obstruction.

Figure Jump LinkFigure 2  Individual variability of FEV1 is influenced by its volume. Values of the SD of three measurements per patient are plotted against baseline mean FEV1. There is a weak, but significant correlation (R2 = 0.082, P < .0001, SE of the estimate = 0.036), with a surprisingly narrow CI. However, both the tests for normality of residuals and the constancy of variance failed (Shapiro-Wilk P < .0001). Moreover, variabilities of individual measurements of FEV1 were also not homogenously distributed (Kolmogorov-Smirnov) and depended on the size of FEV1 (P < .001) (SigmaPlot 12.5; Systat Software Inc). These findings support the notion that population values cannot reasonably be used to establish cutoff values. Data are from the same population as shown in Figure 1.Grahic Jump Location

In Table 1, the percent changes for FEV1 and FVC (columns 6 vs 8) are remarkably similar when comparing healthy individuals (row 7) or patients with COPD (rows 1-6 and 8 and 9). In contrast, the absolute ΔFVC is 20% to 110% higher than that for ΔFEV1 (columns 7 vs 5). Yet, the ATS/ERS committee1 considered changes of 200 mL and 12% to be meaningful for either ΔFEV1 or ΔFVC. Tan et al2 later endorsed these guideline values. However, it is clear that the ΔFVC volume always exceeds that of ΔFEV1. A special concern is that FVC maneuvers are not controlled to be of equal duration before and after interventions. It seems reasonable, therefore, to suggest that specified time points of forced expiration (at any particular time, including ≥ 3 s7) be used instead of FVC.

No evidence has been given to support guideline recommendations that both volume and percentage criteria should be met for a positive response. When baseline volume values have such a wide range, it is unrealistic to expect that changes will necessarily meet population cutoff values for both percentage and volume.

Reproducibility between laboratories and patients differs because it depends on patient health, equipment, quality control procedures, and patient-equipment-technician interactions. Thus, 100 mL or 6.0% may not be an SS applicable universal cutoff value. There may be cases where SS is commonly achieved in a laboratory only when ΔFEV1 exceeds 8% or 150 mL. Thus, a bronchodilator response should be judged based on whether the change in that individual is SS and is likely to attain a minimal clinically important difference.8 In an excellent review article, Donohue8 suggested that the minimal clinically important difference for ΔFEV1 of 100 mL is usually perceived by patients with COPD, correlates with fewer relapses, and is in the range usually achieved with bronchodilators approved for COPD. But in a patient with a baseline FEV1 of 600 mL, an SS increase of 80 mL and 13.3% is likely to be clinically significant. Thus, with a very severe reduction in FEV1, consistent changes of < 100 mL are detectable and may be both SS and clinically significant.

The professional interpreting the bronchodilator responsiveness should have the three highest prebronchodilator and three highest postbronchodilator FEV1 plus the FEV3 or FEV6 values available. The use of rank order testing without a calculator or t testing with a calculator allows an immediate determination of SS. When patient-equipment-technician interactions are optimal and SS is achieved (P ≤ .05), the ΔFEV1 is likely to be ≥ 6% or > 100 mL and clinically significant. When these interactions are not optimal, higher changes may be required to reach SS. Statistical concurrence of ΔFEV1 with ΔFEV3 and ΔFEV6 is nearly always found. The final report by the interpreting professional should include ΔFEV1, the SS value, and the percent change.

Other contributions: The authors appreciate the insights, intelligence, and suggestions of Richard Casaburi, MD, PhD, in his reviews of this manuscript.

Pellegrino R, Viegi G, Brusasco V, et al. Interpretative strategies for lung function tests. Eur Respir J. 2005;26(5):948-968. [CrossRef] [PubMed]
 
Tan WC, Vollmer WM, Lamprecht B, et al; BOLD Collaborative Research Group. Worldwide patterns of bronchodilator responsiveness: results from the Burden of Obstructive Lung Disease study. Thorax. 2012;67(8):718-726. [CrossRef] [PubMed]
 
Sourk RL, Nugent KM. Bronchodilator testing: confidence intervals derived from placebo inhalations. Am Rev Respir Dis. 1983;128(1):153-157. [PubMed]
 
Tweeddale PM, Alexander F, McHardy GJ. Short term variability in FEV1and bronchodilator responsiveness in patients with obstructive ventilatory defects. Thorax. 1987;42(7):487-490. [CrossRef] [PubMed]
 
Pellegrino R, Rodarte JR, Brusasco V. Assessing the reversibility of airway obstruction. Chest. 1998;114(6):1607-1612. [CrossRef] [PubMed]
 
Dixon WJ, Massey FJ Jr. Introduction to Statistical Analysis.3rd ed. New York, NY: McGraw Hill; 1969:116, 354-356, 464, 545.
 
Hansen JE, Sun XG, Adame D, Wasserman K. Argument for changing criteria for bronchodilator responsiveness. Respir Med. 2008;102(12):1777-1783. [CrossRef] [PubMed]
 
Donohue JF. Minimal clinically important differences in COPD lung function. COPD. 2005;2(1):111-124. [CrossRef] [PubMed]
 
Hansen JE. Pulmonary Function Testing and Interpretation. London, England: Jaypee Brothers; 2011:101-103, 211-216, 222-226.
 

Figures

Figure Jump LinkFigure 1  Bronchodilator responses of FEV1 in 316 patients at Harbor-UCLA Medical Center Clinical Laboratory.7 Changes in consecutive patients with three acceptable forced expiratory maneuvers before (baseline) and after receiving nebulized albuterol sulfate solution. Individual positive increases in highest FEV1 in milliliters (x-axis) and percentages (y-axis) are shown, whereas the diagonal lines indicate baseline FEV1 in liters. Data for 51 patients not shown in the figure showed a negative or no response. The 68 crosses in the upper-right unshaded area (including three outside the range of the figure) depict the patients with increases in FEV1 of ≥ 200 mL and ≥ 12.0%. The 102(mostly in the shaded area) indicate other patients with statistically significant (SS) increases in FEV1 ≥ 100 mL or ≥ 6.0% (SS by t test). The(mostly in the lower-left unshaded area) indicate 95 patients who had nonsignificant, small positive responses. Of the 316 patients, only 14 with increases in FEV1 ≥ 100 mL and/or ≥ 6.0% were not SS. All 68 with increases in FEV1 ≥ 200 mL and ≥ 12.0% also had SS increases of FEV3, whereas 94 of the other 102 SS responders also had SS increases of FEV3 volumes.Grahic Jump Location
Figure Jump LinkFigure 2  Individual variability of FEV1 is influenced by its volume. Values of the SD of three measurements per patient are plotted against baseline mean FEV1. There is a weak, but significant correlation (R2 = 0.082, P < .0001, SE of the estimate = 0.036), with a surprisingly narrow CI. However, both the tests for normality of residuals and the constancy of variance failed (Shapiro-Wilk P < .0001). Moreover, variabilities of individual measurements of FEV1 were also not homogenously distributed (Kolmogorov-Smirnov) and depended on the size of FEV1 (P < .001) (SigmaPlot 12.5; Systat Software Inc). These findings support the notion that population values cannot reasonably be used to establish cutoff values. Data are from the same population as shown in Figure 1.Grahic Jump Location

Tables

Table Graphic Jump Location
TABLE 1  ] Changes After Drug or Placebo Found in Volume and % FEV1 and FVC Reported by Four Author Groups With Their Calculated Changes Required To Meet Upper 95% CLs (and Become Abnormal or Positive)

Data are presented as mean ± SD unless otherwise indicated. Upper CL calculations by authors were as follows: Sourk and Nugent,3 Δmean ± 2.0 times SD; Tweeddale et al,4 1.645 times SD; Pellegrino et al,5 2.0 times SD; Tan et al,2 Δmean + 1.645 times SD. CL = confidence limit; MDI = metered-dose inhaler.

References

Pellegrino R, Viegi G, Brusasco V, et al. Interpretative strategies for lung function tests. Eur Respir J. 2005;26(5):948-968. [CrossRef] [PubMed]
 
Tan WC, Vollmer WM, Lamprecht B, et al; BOLD Collaborative Research Group. Worldwide patterns of bronchodilator responsiveness: results from the Burden of Obstructive Lung Disease study. Thorax. 2012;67(8):718-726. [CrossRef] [PubMed]
 
Sourk RL, Nugent KM. Bronchodilator testing: confidence intervals derived from placebo inhalations. Am Rev Respir Dis. 1983;128(1):153-157. [PubMed]
 
Tweeddale PM, Alexander F, McHardy GJ. Short term variability in FEV1and bronchodilator responsiveness in patients with obstructive ventilatory defects. Thorax. 1987;42(7):487-490. [CrossRef] [PubMed]
 
Pellegrino R, Rodarte JR, Brusasco V. Assessing the reversibility of airway obstruction. Chest. 1998;114(6):1607-1612. [CrossRef] [PubMed]
 
Dixon WJ, Massey FJ Jr. Introduction to Statistical Analysis.3rd ed. New York, NY: McGraw Hill; 1969:116, 354-356, 464, 545.
 
Hansen JE, Sun XG, Adame D, Wasserman K. Argument for changing criteria for bronchodilator responsiveness. Respir Med. 2008;102(12):1777-1783. [CrossRef] [PubMed]
 
Donohue JF. Minimal clinically important differences in COPD lung function. COPD. 2005;2(1):111-124. [CrossRef] [PubMed]
 
Hansen JE. Pulmonary Function Testing and Interpretation. London, England: Jaypee Brothers; 2011:101-103, 211-216, 222-226.
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
PubMed Articles
COPD: obstructed lungs. Nurs N Z 2016;22(5):20-4.
Guidelines
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543