A comparative study of bronchodilator response: utilizing pre-bronchodilator versus predicted normal values

Abstract

Background

A positive bronchodilator response has been defined as a 12% increase in the forced expiratory volume in one second (FEV₁) or forced vital capacity (FVC) from their respective pre-bronchodilator values, combined with at least a 0.2 L absolute change. Recent recommendations suggested the use of the percent change in FEV₁ and FVC relative to their predicted normal values without having applied them in patients with airflow obstruction. The aim of the current study was to compare the two approaches over a wide range of pre-bronchodilator FEV₁ and FVC values.

Methods

A retrospective review of consecutive patients undergoing spirometry and bronchodilator testing was completed. The change in FEV₁ and FVC with a bronchodilator was expressed relative to the pre-bronchodilator and predicted normal FEV₁ and FVC.

Results

In 1,040 patients with a non-paradoxical change in FEV₁, 19.0% had a ≥ 12% change in FEV₁ using their pre-bronchodilator value compared to 5.7% using their predicted normal value. For FVC, the respective values were 12.7% vs. 5.8%. The difference was retained in patients with a ≥ 0.2 L change in FEV₁ or FVC. In unobstructed patients, the upper threshold (two standard deviations above the mean) of the bronchodilator response was 14% for FEV₁ and 10% for FVC using predicted normal values.

Conclusions

Expressing the percent change in FEV₁ and FVC relative to predicted normal values reduces the over-estimation of the bronchodilator response, especially in patients with a very low pre-bronchodilator FEV1, including in those with a ≥ 0.2 L change in FEV₁. Irrespective of pre-bronchodilator values, a ≥ 14% change in FEV₁ and ≥ 10% change in FVC relative to the predicted normal values could be considered a positive bronchodilator response.

Background

In pulmonary function testing, the evaluation of a bronchodilator response (BDR) during spirometry involves the administration of inhaled short-acting airway smooth muscle relaxing agents, such as β2-adrenergic agonists. The previous criteria for identifying a positive BDR, as outlined by the American Thoracic Society (ATS) and European Respiratory Society (ERS) guidelines, necessitate both a 0.2 L and 12% increase in forced expiratory volume in one second (FEV1) or forced vital capacity (FVC) [1,2,3]. If these dual criteria are not met, the BDR is deemed negative. However, it is important to note that this definition of a BDR lacks sensitivity, particularly for individuals with either low or high pre-bronchodilator FEV1 or FVC values.

Airflow resistance is inversely related to the airway radius (r) which, in turn, determines airway circumference (c = 2πr). Because bronchodilators decrease airflow resistance by lengthening circumferential airway smooth muscle, any increase in airway smooth muscle circumference can be converted to a related change in the radius (r). For example, a bronchodilator-induced 5 mm increase (Fig. 1) in airway circumference (c) elicits distinct effects on radial change (Δr) when referenced to the baseline radius of the constricted airway (r₁ or r₂) or to the radius of a normal airway (r_n). Consequently, a 5 mm circumference lengthening results in a 27% and 40% radial increase relative to r₁ (3 mm) and r₂ (2 mm), respectively. Conversely, when referenced to the normal airway radius (4 mm), the same circumference lengthening yields a 20% radial increase for both airways. Similarly, in the clinical arena, bronchodilator-induced changes in FEV₁ and FVC, which are surrogates of airway caliber, are exaggerated when referencing to pre-bronchodilator values versus the predicted normal values.

Fig. 1

Effects of referencing a bronchodilator-induced increase in airway radius (r) relative to the pre-bronchodilator radius (r₁ and r₂) or the radius of a normal airway (r_n) in two constricted airways

Recently, an ERS/ATS task force revisited the existing ERS and ATS guidelines and recommended a refinement in the assessment of a BDR [4]. The task force proposed normalizing BDR by expressing the changes in FEV₁ and FVC relative to their predicted normal values [∆FEV_1(%PN) and ∆FVC_(%PN)] to account for differences in pre-bronchodilator FEV₁ or FVC [4]. Based on information on the BDR in a large cohort of healthy subjects, the task force defined a positive BDR as a ∆FEV_1(%PN) and ∆FVC_(%PN) of > 12% and > 10%, respectively [5]. Notably, this approach has yet to be applied in patients with airflow obstruction. Leveraging a cohort of patients undergoing pulmonary function testing at a tertiary academic medical center, the current study sought to discern the extent to which utilizing ∆FEV_1(%PN) and ∆FVC_(%PN) would mitigate bias induced by pre-bronchodilator FEV₁ and FVC values in the assessment of a positive BDR. Additionally, this study also aimed to establish thresholds for a positive BDR, focusing on a subset of patients with pre-bronchodilator FEV₁ and FVC values in the normal range.

Methods

Study sample and spirometry testing

The study sample consisted of sequential patients who underwent spirometry with bronchodilator testing at the Pulmonary Function Laboratory of the University of Miami Hospital from February 1, 2008, to November 30, 2021 (N = 1,637). A deidentified dataset was extracted from the pulmonary function laboratory after the study received approval from the local Institutional Review Board. Patients exhibiting a paradoxical bronchodilator response (ΔFEV1 ≤ 0 L) were excluded. Spirometry was performed by trained technicians according to the pulmonary laboratory protocols per ATS standards. Before commencing the spirometry testing, patients were given comprehensive instructions and a demonstration to ensure a proper understanding of the technique. The Vyntus BODY pulmonary function system (Vyaire Medical, Mettawa, Illinois, USA) which incorporates spirometry with measurements of functional residual capacity and diffusing capacity of the lung was used. All pulmonary function testing took place in the seated position. The standardized sequence for each forced expiratory maneuver involved tidal breathing, maximal inspiration, maximum expiration, and maximal inspiration. Technicians made up to six attempts to acquire three acceptable sets, and the set with the highest FEV₁ was selected for analysis. Real-time error detection during maneuvers prompted immediate technician intervention, aligning with ATS recommendations. Spirometry assessments were conducted both before and 10 min after the administration of 2.5 mg/3 ml albuterol solution via a jet nebulizer, with a nebulization duration of 5–7 min. Informed consent was waived by the University of Miami Institutional Review Board given the retrospective review of de-identified data.

Statistical analysis

To assess and compare the BDR in FEV₁ and FVC relative to the pre-bronchodilator or predicted normal values, bivariate scatter plots were used. Subgroup analyses on BDR were undertaken in a subset of patients with a Δ0.2 L change in FEV₁ (N = 213). Comparisons of ∆FEV_1(%Pre−B) to %ΔFEV_1(%PN) and ∆FVC_(%Pre−B) to %ΔFVC_(%PN) were conducted as a function of the pre-bronchodilator FEV₁ in all patients and the subset of patients with a ≥ 0.2 L change in FEV₁. In addition, analyses were also conducted in a subgroup of patients with a pre-bronchodilator FEV₁ or FVC > 80% of predicted normal (N = 462). All analyses were conducted using Stata 17.0 (Stata Corp, College Station, TX).

Results

Sample characteristics

Table 1 presents the demographic and spirometry data on the full sample and associated patient subsets. Of the initial 1,637 patients who underwent both pre- and post-bronchodilator spirometry, 1,040 exhibited a ΔFEV₁ > 0 L, and 928 had a ΔFVC > 0 L following the bronchodilator challenge. Within the subset of patients with a ΔFEV₁ > 0 L, 213 demonstrated a ≥ 0.2 L change in FEV₁. Moreover, among the 1,637 patients, 462 displayed pre-bronchodilator FEV₁ and FVC values exceeding 80% of predicted normal values.

Table 1 Characteristic of Study Samples

BDR based on the FEV₁

Figure 2 illustrates the bivariate scatter plots for pre-bronchodilator FEV₁, ∆FEV_1(%Pre−B) and ∆FEV_1(%PN). Regardless of the method used to reference the change in FEV₁ following bronchodilator administration, a considerable proportion of patients with a low pre-bronchodilator FEV₁ had a > 12% change indicative of a positive BDR. Employing the traditional definition with pre-bronchodilator FEV₁ as the reference, 19.0% of patients (95% CI: 16.6-21.6%) met the BDR criterion. In contrast, when using the predicted normal FEV₁ as the reference, only 5.7% of patients (95% CI: 4.3-7.3%) had a positive BDR. Consequently, a quantitative disparity emerged with ∆FEV_1(%Pre−B) classifying approximately 14% more patients as having a BDR than with ∆FEV_1(%PN) (p < 0.001). The scatterplot of ∆FEV_1(%Pre−B) versus ∆FEV_1(%PN) revealed that 93.1% of patients fell above the line of identity, confirming a systematic difference between the two approaches for referencing the ∆FEV₁ with a bronchodilator (Fig. 2; bottom left).

Fig. 2

Plot of bronchodilator-induced change in FEV₁ relative to pre-bronchodilator FEV₁ (∆FEV_1(%Pre−B); top left panel) and relative to predicted normal FEV₁ (∆FEV_1(%PN); top right panel) as a function of pre-bronchodilator FEV₁ in patients with a positive change in FEV₁ (N = 1,040). The horizontal dashed line indicates a 12% change. The bottom left panel is the scatter plot of ∆FEV_1(%Pre−B) vs. ∆FEV_1(%PN) with the diagonal dashed line as the line of identity.

In the subset of patients (N = 213) with a ≥ 0.2 L change in FEV₁, the disparity in the proportion of patients with a positive BDR persisted when comparing ∆FEV_1(%Pre−B) to ∆FEV_1(%PN). Using the ∆FEV_1(%Pre−B) to define a positive BDR, 64.3% of patients (95% CI: 57.5-70.7%) exceeded the 12% threshold (Fig. 3; top left), in contrast to 25.8% (95% CI: 20.2-32.2%) when using ∆FEV_1(%PN) to define a positive BDR (Fig. 3; top right). Consequently, compared to ∆FEV_1(%PN), the number of patients with a > 12% BDR remained 40.4% higher when using ∆FEV_1(%Pre−B) while also requiring a ≥ 0.2 L change in ∆FEV₁. The difference in the proportion of BDR-positive patients based on the two references was visually evident in the scatterplot of ∆FEV_1(%Pre−B) vs. ∆FEV_1(%PN), where 90.6% of patients were above the line of identity (Fig. 3; bottom left).

Fig. 3

Plot of bronchodilator-induced change in FEV₁ relative to pre-bronchodilator FEV₁ (∆FEV_1(%Pre−B); top left panel) and relative to predicted normal FEV₁ (∆FEV_1(%PN); top right panel) as a function of pre-bronchodilator FEV₁ in patients with a ΔFEV₁ ≥ 0.2 L (N = 213). The horizontal dashed line indicates a 12% change. The bottom left panel is the scatter plot of ∆FEV_1(%Pre−B) vs. ∆FEV_1(%PN) with the diagonal dashed line as the line of identity

BDR based on the FVC

A discernible systematic difference also emerged between the two methods of assessing BDR in the proportion of patients when referencing the change in FVC to the pre-bronchodilator value [∆FVC_(%Pre−B)] versus the predicted normal FVC value [%∆FVC_(%PN)]. Within the cohort of 928 patients exhibiting a non-paradoxical ∆FVC > 0 L, 12.7% of patients (95% CI: 10.6–15.0%) surpassed the 12% BDR threshold using ∆FVC_(%Pre−B) (Fig. 4; top left) compared to 5.8% (95%CI: 4.4-7.5%) for %∆FVC_(%PN) (Fig. 4; top right). The scatterplot comparing ∆FVC_(%Pre−B) to ∆FVC_(%PN) revealed that 70.8% of patients were above the line of identity (Fig. 4; bottom left). Even among the patients with a ≥ 0.2 L change in FVC, 51% demonstrated a > 12% change in ∆FVC_(%Pre−B) compared to 25% for %∆FVC_(%PN)(Fig. 5). The ∆FVC_(%Pre−B) vs. ∆FVC_(%PN) analyses underscored this difference, with 90% of patients falling above the line of identity (Fig. 5; bottom left).

Fig. 4

Plot of bronchodilator-induced change in FVC relative to pre-bronchodilator FVC (∆FVC_(%Pre−B); top left panel) and relative to predicted normal FVC (∆FVC_(%PN); top right panel) as a function of pre-bronchodilator FVC in patients with a ∆FVC ≥ 0.0 L (N = 928). The horizontal dashed line indicates a 12% change. The bottom left panel is the scatter plot of ∆FVC_1(%Pre−B) vs. ∆FVC_(%PN) with the diagonal dashed line as the line of identity

Fig. 5

Plot of bronchodilator-induced change in FVC relative to pre-bronchodilator FVC (∆FVC_(%Pre−B); top left panel) and relative to predicted normal FVC (∆FVC_(%PN); top right panel) as a function of pre-bronchodilator FVC in patients with a ∆FVC ≥ 0.2 L (N = 220). The horizontal dashed line indicates a 12% change. The bottom left panel is the scatter plot of ∆FVC_1(%Pre−B) vs. ∆FVC_(%PN) with the diagonal dashed line as the line of identity

BDR in patients with a pre-bronchodilator FEV₁ and FVC > 80% of predicted normal

Within the cohort of 1,040 patients demonstrating a ΔFEV₁ > 0 L, a subset of 462 patients exhibited a pre-bronchodilator FEV₁ and FVC > 80% of predicted normal (Table 1). In this subgroup, the raw mean predicted normal FEV₁ was at 94.9%, and mean predicted normal FVC was 97.6%. The average ∆FEV_1(%PN) was 5.0% (SD: 4.3), while the corresponding mean ∆FVC_(%PN) was 3.2% (SD: 3.4). Employing two standard deviations above the mean as a threshold, a ∆FEV_(%PN) of ≥ 14% and ∆FVC_(%PN) of ≥ 10% could be considered the thresholds for a positive BDR irrespective of the pre-bronchodilator FEV₁ or FVC.

Discussion

The results of this study underscore that normalizing the bronchodilator-induced change in FEV₁ and FVC relative to the predicted normal FEV₁ and FVC values reduces the proportion of patients exhibiting a positive BDR, compared to referencing their prebronchodilator values. This reduction in BDR is particularly noteworthy in patients with a low pre-bronchodilator FEV₁, even among those with a ≥ 0.2 L absolute change in FEV₁. Regardless of pre-bronchodilator values, a ≥ 14% change in FEV_1(%PN) and ≥ 10% in FVC_(%PN) could be considered a positive BDR. Expressing the changes in FEV₁ and FVC relative to predicted normal values addresses the biases introduced by using the pre-bronchodilator FEV₁ and FVC in the assessment of the BDR (1–3). Additionally, this approach compensates for inter-individual differences in predicted normal FEV₁ and FVC values, addressing a limitation associated with evaluating BDR by a fixed absolute change in FEV₁ or FVC [6].

The misclassification in BDR among patients with a low or high pre-bronchodilator FEV₁ or FVC value has been previously documented in the COPDGene study [6]. In that study, the percent change in FEV₁ using the pre-bronchodilator values [∆FEV_1(%Pre−B)] aligned with the absolute change in FEV₁ (∆FEV₁) only at a pre-bronchodilator FEV₁ of approximately 1 L. At that degree of airflow obstruction, ∆FEV_1(%Pre−B) and ∆FEV₁ were 16% and 0.16 L, respectively. As anticipated, these two parameters (i.e., ∆FEV_1(%Pre−B) and ∆FEV₁) diverged significantly at lower and higher pre-bronchodilator FEV₁ values. The authors of that study suggested that the magnitude of the BDR is best assessed by ∆FEV₁ and proposed defining a positive BDR as a ∆FEV₁ > 0.16 L, irrespective of the corresponding ∆FEV_1(%Pre−B). However, this approach neglects inter-individual differences in the predicted normal FEV₁. For example, a ∆FEV₁ of 0.16 L in a patient with a predicted normal FEV₁ of 3 L cannot be equated with the same ∆FEV₁ change in a person with a predicted normal FEV₁ of 4 L. Furthermore, the data were obtained among patients with COPD, excluding other forms of obstructive lung disease including asthma, limiting the generalizability of the results.

The bias introduced by pre-bronchodilator FEV₁ and FVC in the grading of the BDR may necessitate a re-definition of what constitutes a positive BDR. As demonstrated in the current study, the conventional ATS/ERS guideline (a 0.2 L and 12% increase in FEV₁ or FVC) is susceptible to this bias, disproportionately identifying a larger number of severely obstructed patients as having a positive BDR. Hansen et al. [6] suggested that, in grading BDR, a > 0.16 L change in FEV₁ can be considered positive, irrespective of pre-bronchodilator FEV₁, based on data from patients with COPD. However, it remains unclear whether this observation can be extrapolated to other patients with airflow obstruction, including asthma. Despite the clinical diagnosis accompanying the request for pulmonary function testing in our study, this information was not utilized due to its poor accuracy. Nevertheless, it was assumed that our dataset, originating from a tertiary care pulmonary function laboratory, included patients with various forms of obstructive lung diseases, including asthma. Therefore, the normalized approach used in BDR assessment is deemed applicable to all patients with airflow obstruction. Another potential limitation of the suggested 0.16 L change in FEV₁ to define a positive BDR is the oversight of predicted normal FEV₁ and FVC, as an absolute ∆FEV₁ and ∆FVC in liters may not account for such inter-individual differences.

The criteria for grading BDR, whether based on FEV₁ or FVC, and establishing a positive BDR warrant careful consideration. It was reasoned that a positive BDR could be defined as a ∆FEV_1(%PN) or ∆FVC_(%PN) exceeding values observed in patients with a pre-bronchodilator FEV₁ or FVC within the predicted normal range. In the subset of patients meeting this criterion (i.e., having a pre-bronchodilator FEV₁ or FVC > 80% of predicted normal), 14.0% and 10.0% exceeded two standard deviations above the mean for %∆FEV_(%PN) %∆FVC_(%PN), respectively. Assuming that this BDR reflects reversal of the “normal” bronchomotor tone [7], we propose defining a positive BDR as a ≥ 14% increase in ∆FEV_1(%PN) or ≥ 10% increase in ∆FVC_(%PN), irrespective of absolute changes in liters or pre-bronchodilator FEV₁ or FVC. Tan et al. [5] reported BDR in 3,922 healthy subjects and found a mean ∆FEV_1(%PN) of 10% (upper 95%CI: 10.5) and 9.2% (upper 95%CI: 10.5) for %∆FVC_(%PN). Based on these data, recent ERS/ATS recommendations define a positive BDR as a ∆FEV_1(%PN) or ∆FVC_(%PN) of > 10% [5]. However, this recommendation has not previously been validated in patients with airflow obstruction. The current study, conducted on a large cohort of patients from a tertiary care setting, demonstrates that applying the normalized BDR method reduces bias introduced by pre-bronchodilator FEV₁ and FVC. Regarding the definition of a positive BDR, the data herein support a ∆FEV_1(%PN) ≥ 14% or ∆FVC_(%PN) ≥ 10% in agreement with the recent ERS/ATS standards for FVC but not for FEV₁ (> 10% for the ERS/ATS and ≥ 14% in the current study). The difference could relate to the respective method of data analysis. The ERS/ATS definitions are based on data obtained in healthy non-smokers irrespective of pre-bronchodilator values whereas we obtained our limits in patients with pre-bronchodilator FEV₁ and FVC values > 80% of predicted normal, possibly including patients with mild airflow obstruction The application of the recent ERS/ATS standards for assessing the BDR has been tested in patients with established medical diagnoses to understand its impact on clinical practice. Li et al. [8] demonstrated that the new ERS/ATS approach to assessing BDR better differentiated between COPD and asthma than previous ERS/ATS standards. Furthermore, Chaiwong et al. [9] found that in the clinical characterization of obstructive lung disease the previous and recent ERS/ATS standards can be used interchangeably. Finally, Beasley et al. [10] suggested that either definition of a positive bronchodilator response has a minor role in the diagnosis of obstructive lung disease. In contrast, using the recent ERS/ATS standards may have therapeutic implications in children with asthma [11]. Perhaps the definition of a positive bronchodilator response proposed in the current manuscript may be more discriminatory in the classification of patients with obstructive lung disease.

In summary, this study demonstrates the importance of using ∆FEV_1(%PN) or ∆FVC_(%PN) for assessing the BDR in patients with airflow obstruction, particularly those with lower pre-BD FEV₁ and FVC values. Unlike the recent ERS/ATS definition of a positive BDR, derived solely from data in clinically normal subjects (5), the current study used a comprehensive pulmonary function dataset encompassing a diverse range of pre-bronchodilator FEV₁ and FVC values. In contrast to the updated ERS/ATS standards, our findings advocate for raising the threshold for a positive ∆FEV_1(%PN) from 10 to 14%, while confirming the 10% for ∆FVC_(%PN).