If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
The average maximum fall in FEV1 after bronchial thermoplasty is 9.1%.
After the first 24 h the fall in FEV1 progressively improves resolving after 3–7 days.
A significantly greater fall in FEV1 of 17.1% is seen after upper lobe treatments.
The degree of deterioration in FEV1 is linearly correlated with the number of radiofrequency activations applied.
The most common adverse effect of bronchial thermoplasty (BT) is short-term aggravation of asthma immediately following the procedure. However, the magnitude and duration of this deterioration, and its predisposing factors are yet to be quantitated. This information will be particularly important as BT is extended to include more severely obstructed patients.
In this prospective, observational study of 20 consecutive patients with very severe asthma undergoing BT, post bronchodilator FEV1 was measured in the 30 min prior to surgery, and then 24 h following the 60 procedures. In half the patients, further spirometry was conducted on day 3 and day 7 post procedure.
This study enrolled 12 males and 8 females, mean age 59.7 ± 12.8 years, with mean prebronchodilator FEV1 of 52.3 ± 15.2% predicted, mean forced expiratory ratio of 51.4 ± 12.6%, and mean improvement in FEV1 post salbutamol of 19.5 ± 15.3%. All patients were taking inhaled corticosteroids, mean beclomethasone equivalent dose 1950 ± 857 mcg, and 7 patients required maintenance oral corticosteroids for control of their asthma. Twenty four hours after BT, the mean deterioration in post bronchodilator FEV1 was 166 ± 237 mls (CI 102–224, p < 0.001) or 9.1 ± 15.2% of baseline. This deterioration was significantly greater after upper lobe procedures (p < 0.01, ANOVA repeated measures), where a mean fall in FEV1 of 17.1 ± 12.6% was observed. The change in FEV1 post procedure was significantly correlated with the number of radiofrequency activations applied, r = −0.376, p < 0.005. By multivariate analysis, the only factor other than activations predictive of the change in FEV1 was age, which was protective. When the lower lobes were treated, the postbronchodilator FEV1 had returned to baseline values by day 3, but patients took 7 days to recover after upper lobe treatments. Despite the severity of asthma in these patients, and the measured deterioration post treatment, there was only one instance of readmission in the 60 procedures.
The deterioration in lung function after BT is transient and well tolerated, but is greatest after upper lobe treatment, and is significantly related to the number of radiofrequency activations applied.
Bronchial thermoplasty (BT) is a novel therapy for the treatment of severe asthma. A radiofrequency electrode is used bronchoscopically to deliver a thermal injury to hypertrophied airway smooth muscle in order to induce smooth muscle atrophy, and thereby attenuate future bronchoconstriction [
]. As might be anticipated, the clinical trials demonstrated that the major short-term complication of BT was aggravation of asthmatic symptoms immediately following the procedure, despite pre-treatment with oral prednisolone. For example, in the AIR trial [
], 62% of patients undergoing bronchial thermoplasty experienced an increase in wheeze compared to 13% of controls. However, whilst the frequency of asthmatic exacerbation after BT has been reported, the magnitude and duration of change in FEV1 post treatment has not yet been formally quantified. This information might be useful clinically to guide treatment decisions such as (i) the suitability of the patient for treatment in the first instance, (ii) the suitability of the patient for hospital discharge after the procedure, and (iii) the length of prednisolone administration post procedure. Further, there might be clues contained in this data which inform us about how best to ameliorate deterioration after BT, and which patients are most at risk. In the words attributed to American management consultant Peter Drucker ”what gets measured, gets managed”.
Therefore, in this study, we aim to measure the changes in post-bronchodilator FEV1 from immediately prior to undertaking BT to day 7 after the procedure.
2. Materials and methods
This was a single centre, prospective, observational study conducted in a university teaching hospital. Twenty consecutive patients undergoing bronchial thermoplasty for severe asthma between April 2016 and July 2017 were enrolled. The baseline clinical characteristics of the participants were collated, including age, gender, baseline spirometry, preventative and reliever medication usage, symptom scores, and exacerbation frequency. The Asthma Control Questionnaire-5 item version (ACQ-5) was chosen as a disease specific quality of life measure as it is known to stable over time and yet, sensitive to interventions [
]. The Olympus BF-Q190 bronchoscope (Olympus Medical Systems, Tokyo, Japan) was used, together with the Alair bronchial thermoplasty system (Boston Scientific, NSW, Australia). Procedures were conducted under general anesthesia using a laryngeal mask. All patients were treated with oral prednisolone, 0.5 mg/kg, for three days before and 3 days after treatment. On the morning of the procedure, patients were treated with salbutamol 5 mg nebulized immediately before theatre, and then, in addition, received intraoperative, intravenous doses of 8 mg dexamethasone and 0.4 mg glycopyrrolate. The choice of anesthetic agents was left to the discretion of the individual anesthetist. Procedures were conducted at 8am and patients were routinely observed in hospital for 24 h before discharge. The number of radiofrequency activations deployed in each treatment session was recorded.
The primary outcome parameter in this study was change in the post bronchodilator FEV1. This was measured using the Vyntus Pneumo portable spirometer (Carefusion Australia, Seven Hills, Sydney, Australia). The spirometer was calibrated immediately prior to use with a 3 L syringe, and all measurements were made by experienced respiratory scientists. Testing was performed 15 min after the administration of nebulized salbutamol 5 mg, and with the subject in the seated position. At least three acceptable manoeuvres were obtained, with the FEV1 and FVC measurement values within 0.15 L of each other during repeated testing, in accordance with European Respiratory Society/American Thoracic Society guidelines [
]. Measurements were made on the morning immediately prior to theatre, then 24 h later, prior to discharge from hospital. In a subgroup of patients, further measurements were made on days 3 and 7 post procedure.
The secondary outcome measurements related to the impact of BT, measured at the 6 week post treatment visit and compared to baseline measurements performed when the patient was stable, on their usual medication, in the 4 weeks prior to commencement of BT treatment. The variables examined were ACQ-5 and spirometry, particularly prebronchodilator FEV1%predicted. The predicted equations used for spirometry were from the Global Lung Initiative [
]. In addition, as part of baseline assessment, the transfer factor for carbon monoxide corrected for alveolar volume (KCO) was measured using the Jaeger Masterscreen Body (Carefusion, Hoechberg, Germany). Further, in the hour immediately prior to the BT procedure, and then 24 h later, prior to hospital discharge, a venous blood specimen was drawn in a convenience subset of patients, for measurement of total and differential white cell count, measured in cells per millilitre of blood, and C-reactive protein, measured in mg/litre.
Approval to collate and audit data as part of quality assurance was provided by the Human Research and Ethics Committee at Peninsula Health. All subjects provided informed consent prior to participation. Permission to use the ACQ-5 was provided by the author, Elizabeth Juniper.
2.5 Statistical analysis
SPSS version 24 (IBM corporation, New York, USA) was used for all statistical analyses. Grouped data has been reported as mean ± standard deviation, except where not normally distributed, in which case median (interquartile range) is reported. A paired t-test was undertaken to compare all paired sets of normally distributed data, and statistical significance was taken as p < 0.05 for a two-tailed test. Analysis of variance (ANOVA) for repeated measures was used to compare data such as FEV1 measured at repeated time points across 7 days. Pearson's Correlation Coefficient was calculated for bivariate continuous normally distributed data. Multivariate linear regression was performed using a stepwise backward model.
3.1 Baseline characteristics
Twenty consecutive patients participated in this study, 12 male, 8 female. The mean age was 59.7 ± 12.8 yrs. The mean baseline prebronchodilator FEV1 was 52.3 ± 15.2%predicted, with a mean change in FEV1 post-bronchodilator of +19.5 ± 15.3%. The prebronchodilator Forced Expiratory Ratio was 51.4 ± 12.6%. The mean ACQ-5 was 3.3 ± 1.1. At baseline, all patients were being treated with long acting beta2 agonists and with inhaled corticosteroids, mean beclomethasone equivalent dose of 1950 ± 887 mcg. Seven patients required maintenance oral corticosteroids to control their asthma. Participants used a median of 7 puffs (IQR 2–10) salbutamol per day and had experienced a median of 3 (IQR 2–6) oral steroid requiring exacerbations in the previous 12 months. There were no current smokers in the group, but 8 patients had smoked in the past, 5 with less than 10 pack years. The mean KCO was 96.1 ± 16.6%predicted.
A mean of 162 ± 32 activations (range 117–255) in total per patient were delivered in the three treatments. A balanced approach to general anesthesia was evident with a volatile anesthetic agent, predominantly sevoflurane, being used in all but one of the 60 anaesthetics. Fentanyl or remifentanyl were used in 58/60 anaesthetics and propofol was used in 53. Patients were discharged home from hospital the day following treatment, except on one occasion (of 60) when the patient remained in hospital for a further 24 h owing to sinus tachycardia - which in retrospect may have been due to nebulized bronchodilator. One patient was readmitted to hospital to treat radiologically proven right upper lobe pneumonia six days after the upper lobe treatment session. That patient made a complete recovery and was discharged on the fourth day. There were no other complications.
3.3 Change in FEV1 24 h post procedure
At 24 h post treatment, the mean fall in post bronchodilator FEV1 was 166 ± 237 mls (CI 102–224 mls, p < 0.001) or 9.1 ± 15.2% of the baseline post bronchodilator FEV1 in the 60 treatments. However, this fall was not distributed evenly across the three treatments, as shown in Table 1. There was a significant effect of treatment site evident by one-way repeated measures ANOVA, Wilks' lambda = 0.610, p = 0.01, and the effect size was strong, multivariate partial eta squared = 0.39. The pairwise comparisons revealed the significant difference occurred when the upper lobes were treated, p = 0.007.
Table 1Change in post bronchodilator FEV1 24 h post BT.
Delta FEV1 (mls)
−139 ± 230
−90 ± 257
−270 ± 192
Delta FEV1 (%)
−6.7 ± 13.9
−3.6 ± 16.0
−17.1 ± 12.6
40.1 ± 7.3
42.8 ± 10.7
78.8 ± 21.5
p refers to ANOVA by repeated measures, see text for between group comparison.
Following upper lobe treatments, a fall in post bronchodilator FEV1 of 15% or greater was observed in 12 of 20 (60%) cases, whereas this was only seen in 9 of 40 (23%) lower lobe treatments (p = 0.009, Fisher's exact test). Table 1 also demonstrates that a significantly greater number of radiofrequency activations were delivered when the upper lobes were treated (one way repeated measures ANOVA, Wilks' lambda 0.175), p < 0.001, multivariate partial eta squared = 0.825). The potential relationship between radiofrequency activations and the percentage change in FEV1 24 h post treatment is explored in Fig. 1. The Pearson correlation coefficient for this relationship is r = −0.376, p = 0.003.
The consistency of change in FEV1 across treatment episodes is explored in Fig. 2 where the fall in volume in postbronchodilator FEV1 after the left lower lobe treatment is compared to the fall in volume after the right lower lobe treatment. The Pearson correlation coefficient for this relationship is r = 0.68, p = 0.001. However, no significant relationship was observed when either of the right or left lower lobe treatments were compared with the upper lobe treatment.
3.4 Duration of post-procedure wheeze
The rate of recovery of postbronchodilator FEV1 to baseline value was explored in 30 treatments in 10 subjects. Measurements of FEV1 were made on Days 1, 3 and 7, and standardized for each patient as a percentage of their individual baseline value at Day 0. These results are shown in Table 2. When the right lower lobe was treated, the FEV1 had returned to baseline values by day 3, with statistically significant further improvement by day 7 (ANOVA, repeated measures, Wilks' lambda p = 0.004, partial eta squared 0.746). Similar results were obtained after left lower lobe treatments (Wilks' lambda p = 0.038). The changes in FEV1 over the course of the week after the upper lobes were treated did not achieve statistical significance, highlighting the slower recovery when compared with the lower lobe treatments. The mean post bronchodilator FEV1 at baseline in these 30 treatments was 1610 ± 438 mls, and had risen at day 7–1699 ± 421, p = 0.02 (paired t-test).
Table 2Rate of return of postbronchodilator FEV1 to baseline after treatment.
In order to assess whether there was a carry over effect of wheeze between procedures, in all 20 patients, the post bronchodilator FEV1 was compared at 5 time points, namely (i) in the 4 weeks pre procedure, (ii-iv) immediately before each of the 3 treatments and then (v) at the 6 week follow up. There was no significant difference at any time point (ANOVA repeated measures, p = 0.272).
3.5 Predictors of acute change in FEV1 post BT
A stepwise backward multivariate linear regression model was created to examine the contributions of variables that might influence the percentage change in FEV1 from baseline to 24 h post treatment. The variables examined included: age, gender, BMI, baseline FEV1% predicted, bronchodilator response at baseline, ACQ-5 and activations. Most variables had no significant effect. In the final model, only two variables remained significant, namely activations (standardized beta coefficient = −0.326, p = 0.01) and age (standardized beta coefficient = +0.312, p = 0.02). Increasing activations led to an increasing fall in FEV1 post treatment, whilst increasing age was protective. However, this model explained only 22% of the variance in change in FEV1 immediately post treatment (r = 0.467, p = 0.008).
3.6 Blood tests
The changes in leukocyte counts, measured in cells/millilitre blood, and C-reactive protein (CRP) mg/l, after bronchial thermoplasty were measured in 23 procedures in 12 patients. In this group of patients there were 8 upper lobe procedures and 15 lower lobe procedures. The results are presented in Table 3. In the second column of the Table, the blood values at the baseline preprocedure BT assessment are presented. These were then compared to the bloods drawn on the morning of the BT procedure (Day 0), the patient having received three days of oral corticosteroids. Suppression of the eosinophil count and elevation of the neutrophil count are demonstrated. The bloods drawn 24 h after the BT procedure are then compared to the Day 0 values. Significant increases in the total white cell count (WCC) and neutrophil count were observed the day following treatment, whilst no changes were seen in eosinophil counts or CRP. The change in total white cell count after upper lobe treatments was not significantly different from lower lobe treatments.
Table 3Effect of bronchial thermoplasty on white cell counts and CRP.
p1 Pre v D0
p2 D0 v D1
8.6 ± 3.3
12.1 ± 3.2
13.9 ± 3.0
5.7 ± 2.8
8.7 ± 3.3
10.5 ± 3.2
0.3 ± 0.2
0.0 ± 0.1
0.0 ± 0.1
4.2 ± 4.4
4.8 ± 5.3
p1 = unpaired t-test, p2 = paired t-test, NS = not significant, Cell counts in cells/ml.
The response to treatment at 6 weeks post BT, measured by changes in ACQ-5 and prebronchodilator FEV1%predicted are shown in Table 4. A significant mean improvement in ACQ-5 of 0.9 units was observed (CI 0.2–1.6, p = 0.01, paired t-test)- greater than the minimal clinically significant difference of 0.5 units. No significant effect on FEV1 was observed.
The aim of this study was to quantify, for the first time, the major side effect of bronchial thermoplasty treatment, namely aggravation of asthma. The study population comprised a severe group of asthmatics, with a mean baseline FEV1 of 52% predicted-more severe than reflected in the randomized control trials of BT [
]. Despite the lack of lung function reserve in these patients, the mean fall in FEV1 24 h post procedure was only 9.1% across all procedures, and it was even less than this following lower lobe treatment sessions. Furthermore, by day three following lower lobe treatments, on average, patients had returned to their pre-procedure baseline. This provides important reassurance in relation to the safety of BT in more severe patients. It also suggests that our approach of keeping all patients in hospital overnight following BT might be too conservative and that, after lower lobe treatments, some of these severe patients could be discharged home the same day.
However, the same cannot be said in relation to upper lobe treatments. After upper lobe treatments, patients were much more wheezy, with a mean fall in FEV1 of 17.1% at 24 h, and patients required a full week to recover to baseline. Since 60% of patients had a fall in FEV1 of 15% or more after the upper lobe treatment session, and given a starting mean FEV1 of 52% predicted, we would not be recommending that these type of patients be discharged home the day of the procedure. The greater impact of the upper lobe treatment underscores the concept in the original design of BT to schedule the upper lobe treatments in the third and final treatment session, with the protection of the previously treated lower lobes.
The obvious explanation for the greater fall in FEV1 after upper lobe treatment lies in the greater volume of lung treated as measured by the number of radiofrequency activations applied. In this study we have been able to demonstrate a significant association between these two variables (r = −0.376,p = 0.003). In the multivariate analysis of factors predicting the change in FEV1 post procedure, only two variables had significant influence: activations and age. It is not clear exactly why age may have been protective, possibly the explanation lies in a remodelled airway which is less sensitive to insult. However, bronchodilator responsiveness to salbutamol at baseline, which might be a marker for a remodelled airway and fixed airway obstruction, had no measurable effect in the predictive model. Importantly, the baseline FEV1 was not a predictive factor either, showing us that patients with well-preserved lung function are just as likely to experience a substantive fall in FEV1 as patients with more severe obstruction.
Since activations appear to be important in the genesis of wheeziness post BT, and since the upper lobe treatment sessions have the highest activations and the greatest fall in FEV1 post procedure, one might consider dividing the upper lobe treatments into two separate sessions. This would mean undertaking BT in 4 rather than 3 treatments but with an added margin of safety. This is speculative, but worthy of consideration, especially when providing BT to patients whose baseline FEV1 is between 30 and 50% predicted, a group we now frequently treat.
Similarly, if we were able to better prevent the fall in FEV1 post procedure, we might be able to consolidate BT into two rather than three treatments. This would be both more convenient for patients, and more cost effective for the health service industry. In this current study it seems likely that patients were adequately pretreated with oral corticosteroids, given both the eosinophil depletion and the neutrophilia observed when bloods were drawn immediately prior to the BT procedure. Even so, the slight but significant improvement in FEV1 at day 7 post procedure when compared to day 0, is noted and is likely to be due to the oral corticosteroids administered. In addition to oral corticosteroids, the administration of drugs known to block neutrophilic inflammation, such as roflumilast [
], may be an approach worthy of future testing to assess if the deterioration post BT can be ameliorated. The concordance in an individual patient between the responses to treatment of the right and left lower lobes would suggest that a randomized controlled trial could be conducted comparing putative future agents against standard of care in the same patient.
We have also considered the potential role that anesthetic agents might have in ameliorating bronchoconstriction following BT. This is an untested area in BT. One approach uses an intravenous anesthetic protocol with propofol and an opiate [
]. It is of interest that in this study, in almost every case, the consultant anesthetist present chose to use the volatile agent, sevoflurane, known to be an excellent bronchodilator.
It is acknowledged that this is a single centre observational study and that practice variations in other centres could either exacerbate or ameliorate the fall in FEV1 following BT. In this respect we would encourage other centres to collect and report this peri-procedural information in order to further the discussion about how best to deliver this therapy.
In a previous paper, we have demonstrated the importance of undertaking enough activations in order to generate adequate response to BT [
]. This current paper shows the other side of that coin, namely that proceduralists need to be mindful of the quantum of radiofrequency treatment they are administrating in order to minimize immediate post procedure aggravation of asthma. Even so, the fall in FEV1 post procedure is generally modest, and resolves in 3 days after lower lobe treatments. Upper lobes aggravate asthma to a greater extent and take longer to recover.
BT: bronchial thermoplasty.
ACQ-5: Asthma control questionnaire-5 item version.
FEV1: Forced expiratory volume in 1 s.
KCO: transfer factor for carbon monoxide, corrected for alveolar volume.
ANOVA: analysis of variance.
IQR: Interquartile range.
WCC: white cell count.
CRP: C-reactive protein.
The authors would like to thank Kim Bennetts and Annie Tran for lung function testing, Ceri Banks for patient assessment, and Drs Nicolette Holt and Jennifer Mann for case review.
Availability of data and materials
Please contact the primary author for data requests.
This study was designed by D.L., who had access to all study data and takes responsibility for data integrity and analysis. Statistical analyses were performed by D.L and W.W. All authors contributed to manuscript preparation.
Ethics approval and consent to participate
Approval to collate and audit data as part of quality assurance was provided by the Peninsula Health Human Research and Ethics Committee.
Consent for publication
The authors declare they have no competing interests.
1Department of Thoracic Medicine, Frankston Hospital, Peninsula Health, Victoria, Australia; 2Faculty of Medicine, Nursing and Health Sciences, Monash University; 3Department of Respiratory Medicine, Box Hill Hospital, Eastern Health, Victoria, Australia.
A prospective feasibility study of bronchial thermoplasty in the human airway.