Key Findings and Strength of Evidence

We identified three primary randomized controlled trials (RCTs, n=432)^18–20 of bronchial thermoplasty (BT), as well as their associated extension studies (n=245).^21–23 One retrospective comparison²⁴ and eight descriptive studies (n=55)^25–32 also reported outcomes associated with BT. Two multicenter RCTs compared BT with standard care (medical management),^19,20 and one multicenter RCT compared BT to a sham bronchoscopic procedure with standard care continued in both groups.¹⁸

Compared with standard care alone, the evidence from two RCTs suggests that BT with standard care improved asthma control (defined by the Asthma Control Questionnaire [ACQ], rates of mild exacerbations change from baseline to 12 months), utilization of rescue medication and quality of life (low strength of evidence [SOE]), but the clinical importance of the findings for each of these outcomes is unclear. Rates of hospitalizations for respiratory symptoms were not different for these comparators during the post-treatment period (6 weeks after the third BT treatment through 12-month followup) (SOE: low). The evidence base was insufficient to draw conclusions about BT’s effects on severe exacerbations or pulmonary physiologic measures compared with standard care.^19,20

Compared with the sham bronchoscopic procedure and standard care, the intention-to-treat analysis in a single RCT suggests that BT with standard care had no effect on asthma control (defined as improvement in ACQ from baseline), hospitalizations for respiratory symptoms, rescue medication usage, pulmonary physiologic measures, or other asthma symptom scores (low SOE). Reduced risk of severe exacerbations was suggested (low SOE), but the clinical importance of the degree of the reduction was unclear. Rates of emergency department visits for exacerbations during the post-treatment period were significantly lower in patients receiving the BT and standard care than in those who received the sham bronchoscopic procedure and standard care (moderate SOE).¹⁸ Serious adverse events attributed to BT occurred during the 12-week treatment period, and no deaths were reported.

Findings Compared to What Is Already Known

A 2014 Cochrane review³⁴ of BT examined the same three RCTs as we did and described the benefits of BT as modest but not clinically significant. Unlike the Cochrane review authors, we did not pool the trial results because of differences in study designs and populations. We also had greater concern about risk of bias in the trials, and therefore graded outcomes more conservatively. These findings are consistent with other systematic reviews and technology assessments.^35–37

Current clinical practice guidelines suggest a cautious approach to use of BT. For example, the American Thoracic Society (ATS)-European Respiratory Society (ERS) guidelines on definition, evaluation and treatment of severe asthma recommends BT be “performed in adults with severe asthma only in the context of an Institutional Review Board-approved independent systematic registry or a clinical study.”³⁸ The authors state that they placed greater weight on avoiding adverse effects and increased use of resources, while noting the uncertainty of benefit in terms of symptoms and quality of life, the lack of data on patients with more severe asthma (forced expiratory volume in 1 second [FEV_1] <60% of predicted value), and the need for ways to determine which patients might benefit.³⁹ Although BT has been approved for use in Japan since 2015, the Japanese Society for Allergology guidelines recommend further study of BT. Describing the AIR 2 study, they note that despite the 5-year results, longer term efficacy and safety should still be examined.³⁸

The United Kingdom National Institute for Health and Care Excellence (NICE) guidance also calls for longer term safety data, and requires clinicians performing BT to submit patient details to a centralized registry. They state that clinicians should “Ensure that patients understand the uncertainty about the procedure’s efficacy and long-term safety, and the possibility of initial worsening of their symptoms, and provide them with clear written information.”⁴⁰ Similarly, the 2011 British Thoracic Society guideline describes the procedure as “a possible treatment option in selected patients with severe persistent asthma already on maximal therapy, although its place in the treatment of asthma remains to be established.”⁴¹ The 2014 Scottish Intercollegiate Guidelines Network also describes BT as a procedure that “may be considered for patients with poorly controlled asthma despite optimal therapy.”⁴²

Finally, the Guideline from the Global Initiative for Asthma states:

[F]or highly-selected adult patients with uncontrolled asthma despite use of recommended therapeutic regimens and referral to an asthma specialty center (Step 5), bronchial thermoplasty is a potential treatment option in some countries. (Evidence level B [RCTs. Limited body of data]) Caution should be used in selecting patients for this procedure as the number of studies is small, and people with chronic sinus disease, frequent chest infections or FEV₁<60% predicted were excluded. (Evidence D [panel consensus judgment])”⁴³

Applicability

Although BT is approved for treating patients with severe persistent asthma, there was heterogeneity in the severity of asthma in the populations included in the three RCTs. The evidence from one of the RCTs is applicable to adult patients with asthma who require ≤30 mg per day of prednisone for maintenance while on high-dose inhaled corticosteroids (ICS) and long-acting beta-2 agonists (Research in Severe Asthma Trial [RISA]).²⁰ One small trial was designed to compare “real-world” patients, including those with high rates of exacerbations and with no limitation on medication use, and reported that the clinical response was lower and more variable than in the RCTs.²⁴ The number of patients in the comparison was too small to draw firm conclusions about the relative benefits of BT in patients who are sicker than those in the RCTs or about the possibility of selection bias in RCTs.

One study enrolled patients on lower doses of ICS (200 mcg/day), but the mean dose at baseline was similar to that in the other trials (Asthma Intervention Research Trial [AIR]).¹⁹ Some studies restricted enrollment to patients on <10mg per day of oral corticosteroid, fewer than three hospitalizations (AIR-2),¹⁸ fewer than four episodes requiring systemic steroids in the prior year (AIR) and use of no more than four puffs of short-acting beta-2 agonist per day (AIR and AIR-2).

The BT procedure itself appears to be similar across these studies, and was performed in several countries, in settings comparable to those in most bronchoscopy centers. Additional information about the concomitant medical therapy in the BT groups as well as in the control groups would be helpful for translating the findings to practice.

Implications for Clinical and Policy Decisionmaking

Clinicians whose patients are potential candidates for BT may want to consider the evidence presented in this review, including the highly selected and yet heterogeneous study populations, limited improvement in outcomes, and rates of adverse events (including asthma worsening and respiratory tract infections during the treatment period) when determining BT’s appropriateness for their patients. Only one small RCT included patients taking more than 10 mg per day of oral corticosteroids and patients with FEV₁ as low as 50 percent predicted.

Current guidelines stress the importance of assessing patients’ adherence to prescribed therapies and their technique in using inhalers prior to considering BT. One of the U.S. Food and Drug Administration (FDA) presenters to the Anesthesiology and Respiratory Therapy Devices Panel noted that patients in the AIR 2 trial at sites in Brazil, who made up 30 percent of the study population, were all given maintenance medications at no cost. The intention-to-treat analysis of Asthma Quality of Life Questionnaire (AQLQ) results for the patients in Brazil found greater improvement in the group receiving the sham bronchoscopic procedure than in the group undergoing BT.³³

The primary efficacy endpoint used for BT’s regulatory filings with FDA was the AQLQ score change from baseline to the average of 6-, 9-, and 12-month followup from the pivotal AIR 2 sham-controlled trial.³³ The improvement did not meet the prespecified level of statistical significance in the intention-to-treat population, but did in the per protocol analysis. Nonetheless, the degree of improvement was less than the MID, so it may not be clinically important. While the sham bronchoscopic procedure in this study was associated with some post-procedure respiratory events, there were more adverse events and more events requiring hospitalization in the BT group during the treatment period, although the latter was not a statistically significant difference. Clinicians and patients must balance this risk of adverse events in the treatment period against the evidence for later modest improvements in rates of severe exacerbations and more robust evidence for reduction in emergency department visits for exacerbation.

Limitations of the Systematic Review Process

The scope of this review may have introduced some important limitations. First, we included only trials with concurrent controls when assessing BT’s effectiveness. Similarly, conference abstracts without subsequent full publications were excluded because it is difficult to assess study risk of bias and selective outcome reporting from abstracts, and because results presented in abstracts are frequently different in final publications.

Limitations of the Evidence Base

This evidence base contains several limitations. Only one of three trials was a blinded, sham-controlled trial. As seen in our evidence analysis, this study did not show similar findings to the nonblinded, RCTs comparing BT to standard care. Outcome measures were not ascertained using the same assessment procedures. In one study, for example, exacerbations were counted only during 2-week periods when patients were asked to abstain from use of maintenance long-acting beta agonist therapy.

Evidence Gaps

Several types of evidence gaps could be addressed in future research. Our conclusions for several outcomes were limited by the small number of studies and low numbers of patients. In some instances, the effect sizes had fairly wide confidence intervals (exceeding minimally important differences) which led us to assess the evidence as imprecise. While the RCTs were multicenter, there was only one RCT using a sham control. Given the difference in findings when the sham control was used, and the subjective nature of the majority of outcomes, it would be informative to have data from additional studies enrolling similar patients and using the same design.

Two of the three RCTs were assessed as having medium risk of bias, and one low risk of bias. The risk of bias in future studies could be improved by describing both appropriate allocation of treatment and concealment, blinding patients and outcome assessors, and clarifying the role of funders.

Studies using a standard care or medical management control could be improved by reporting concomitant therapies more clearly. Greater uniformity in outcome measures, particularly for exacerbations, would be helpful. In addition, studies could address clinically important differences in outcomes, using data that have been validated by other investigators not involved in the trial (as in Appendix D).

As noted above, only one sham-controlled trial of BT has been published. Given that BT is an invasive procedure and that patients receiving the sham treatment appeared to improve on certain outcomes, further studies using a sham comparison are needed to strengthen the evidence base and help guide appropriate use of BT. Studies could also be undertaken to test BT in other populations, especially patients with poor asthma control who experience higher rates of exacerbations. The studies included in this analysis also did not examine the efficacy of BT stratified by asthma phenotype (e.g., eosinophilic, neutrophilic, paucigranulocytic). Future studies might better elucidate whether BT is more effective in certain asthma phenotypes. As of July 19, 2017, ClinicalTrials.gov listed 18 trials investigating BT that are ongoing, planned, or of unknown status. (See Appendix E.) Most are single-arm observational studies. The randomized trials examine BT administered with traditional bronchoscopy or with other image-guided interventions. No sham-controlled trials are currently registered.

Conclusions

Three RCTs and several descriptive studies meeting our inclusion criteria have evaluated BT. Based on the available literature, BT may be modestly beneficial in some patients with asthma, but is not without risks in any population. The risk of adverse events is higher during the treatment period and for several weeks afterward. Benefit is typically observed weeks to months after therapy and can last for at least 5 years, after which the duration of effect is unknown.