Crossover Patient Outcomes for Targeted Lung Denervation in Moderate to Severe Chronic Obstructive Pulmonary Disease: AIRFLOW-2

Chronic obstructive pulmonary disease (COPD) is a chronic respiratory disease characterized by fixed airway obstruction with respiratory symptoms usually caused by exposure to noxious particles or gases, commonly cigarette smoking [1]. It is a significant cause of morbidity and mortality. It is estimated to affect around 300 million people worldwide [2], with over 3 million people dying from the condition annually [3]. It comprises a variety of symptoms which can include breathlessness, cough, sputum production, and fatigue. The disease course of COPD varies between individuals. It is a chronic, progressive condition, with intermittent exacerbations.

COPD exacerbations are described as “acute worsening of respiratory symptoms, associated with a variable degree of physiological deterioration” [4]. They are usually characterized by shortness of breath, worsening cough, and there may be changes to the colour, consistency, or volume of sputum. They often have an infective precipitant (bacterial or viral), which leads to airways and systemic inflammation, and hyperinflation [5]. Patients who experience frequent exacerbations have a lower quality of life [6], higher hospitalization rate [7], and higher mortality [8] than those with a similar disease severity who do not suffer from frequent exacerbations. Exacerbations are also thought to contribute to decline in lung function [9]. Therefore, exacerbations contribute significantly to healthcare costs, decline in lung function, and mortality. COPD treatment remains an unmet need and new treatments are required. Non-pharmacological therapy in COPD, including smoking cessation and pulmonary rehabilitation, is vital in preserving lung function and reducing exacerbation frequency. Inhaled therapy is the mainstay of maintenance treatment and includes bronchodilators and inhaled corticosteroids. However, there are problems with current therapies, including the pharmacodynamics, and adherence, with under 40% of new users of tiotropium continuing treatment for 12 months [10]. Adverse effects including susceptibility to pneumonia from inhaled steroids further hamper their use. Current therapy is not disease modifying, and has only a modest role in reducing exacerbations, and despite pharmacological therapy, many patients remain symptomatic [11].

Targeted Lung Denervation (TLD) is a novel bronchoscopic therapy designed to reduce the clinical consequences of neural hyperactivity in obstructive lung diseases such as COPD through disruption of nerve signalling to and from the lung [12, 13]. In COPD, vagal activity is dysregulated, thus causing excessive bronchoconstriction, bronchial hyperreactivity, and mucus hypersecretion. These features commonly manifest in symptoms of breathlessness, sputum and cough, and exacerbations. The aim of TLD is to disrupt this neural aberrancy by delivering radiofrequency energy bronchoscopically to the airways using the denervation system (Nuvaira, Inc.).

The AIRFLOW-1 and AIRFLOW-2 study (NCT02058459) have seen patients undergo bilateral TLD [14]. Eighty-two patients were enrolled into the AIRFLOW-2 trial (patient inclusion criteria are discussed below), with moderate to severe symptomatic COPD. Data from the randomization phase of the trial showed that those treated with TLD had fewer respiratory-related adverse events and exacerbations requiring hospitalization during the randomized, sham-controlled AIRFLOW-2 clinical trial, compared to the sham arm [15]. Time-to-first severe COPD exacerbation was significantly lengthened in the TLD arm (p = 0.04, HR = 0.38) at 2 years post-TLD therapy and trended towards similar attenuation for moderate and severe COPD exacerbations (p = 0.18, HR = 0.71) [15]. TLD also appears to slow decline in lung function in patients with COPD [16]. The aim of this work was to evaluate the patients undergoing TLD as part of the crossover cohort, to establish whether the effects from the randomization phase showing reduced respiratory events post-TLD in COPD are supported.

The AIRFLOW-2 trial (NCT02058459) was a multicentre, double-blind, randomized, sham-controlled, crossover study. The study design and recruitment is previously described [15]. Study recruitment was conducted between July 2016 and May 2017. Subjects provided written informed consent to study participation and the study protocol was approved by the relevant Ethics Committees. Please see online supplementary Material for Ethics Committees and approval numbers for each participating site (for all online suppl. material, see www.karger.com/doi/10.1159/000527455). One hundred ninety-four subjects were screened and 82 subjects were randomized. Subjects aged between 40 and 75 with spirometry-confirmed COPD with a post-bronchodilator ratio of forced expiratory volume in 1 second (FEV₁)/forced vital capacity <0.70 and FEV₁ of 30–60% predicted were enrolled. Patients had to be symptomatic, assessed as either a COPD Assessment Test (CAT) score ≥10 or modified Medical Research Council dyspnea scale (mMRC) score ≥2. There was no baseline exacerbation requirement for patients. Patients with significant respiratory or cardiac co-morbidities were excluded.

Blinded patients were randomized in a ratio of 1:1 receiving either sham bronchoscopy or bronchoscopic treatment with TLD (Nuvaira, Inc., USA) [15]. Following the 12.5-month follow-up visit, eligible patients in the sham group were offered TLD treatment forming the crossover group. The impact of TLD on exacerbations of COPD was analysed in these crossover patients 1 year following TLD treatment. A standard log-rank statistic was used to compare time-to-event data between groups.

Of 39 sham subjects unblinded at 1- year follow-up, only 6 subjects continued follow-up without crossing over to TLD treatment. Three sham subjects exited the study prior to the crossover opportunity and three sham subjects exited the study early to seek another interventional treatment. Seven sham subjects failed to meet crossover eligibility criteria, leaving 20 crossover subjects who received TLD treatment and were followed for an additional year (50.0% male, age 64.1 ± 6.9 years). The mean FEV₁% predicted of this cohort was 41.4% ± 7.3%. The demographics of the population are displayed in Table 1. The occurrence of COPD exacerbations for these subjects was compared between year 1 follow-up as sham patients during AIRFLOW-2 (sham) and year 2 follow-up as patients treated with TLD (crossover).

After TLD treatment, the proportion of patients experiencing a moderate or severe AECOPD (exacerbation requiring treatment with antibiotics or systemic steroids, with or without hospitalization) decreased from 75% (sham, year 1) to 50% (crossover, year 2). The proportion of patients experiencing a severe AECOPD (exacerbation requiring hospitalization) decreased from 25% (sham, year 1) to 10% (crossover, Year 2). Time-to-first-event analyses were performed to compare the risk of AECOPD before and after TLD. Risk for moderate or severe AECOPD (Fig. 1) after TLD was not significantly reduced but was lower in the treatment group (HR: 0.65, 95% CI: 0.29–1.45, p = 0.28). Risk of severe AECOPD also did not show significant reduction after TLD but was lower in the treatment group (HR: 0.38, 95% CI: 0.08–1.97, p = 0.23) (Fig. 2).

Health-related quality of life measures including St George’s Respiratory Questionnaire for COPD (SGRQ-C) and COPD Assessment Test (CAT), along with lung function (FEV₁), were not statistically different over the two follow-up periods, shown in Table 2. Adverse events were also noted and provide additional support to the AIRFLOW-2 trial regarding the safety of TLD in patients with moderate to severe COPD. Gastrointestinal safety was supported with no serious adverse events during either follow-up period. The number and percentage of patients experiencing serious adverse events by category is demonstrated in Table 3, confirming a favourable safety profile of TLD.

COPD exacerbations are associated with poor disease prognosis, increased mortality risk, and increased healthcare costs [17]. Here, we present a crossover cohort with fewer subjects experiencing a COPD exacerbation following TLD, which is consistent with previously published data. The hazard ratios comparing AECOPD risk before versus after TLD for the crossover cohort were similar to the hazard ratios comparing sham versus TLD in the randomized cohort [15]. While underpowered to show statistical significance, the similar numerical effect might indicate reproducibility of treatment effect in terms of impact on exacerbations.

Due to COPD being a progressive condition, and lung function deteriorating with age, it is notable that these outcomes were observed despite subjects’ disease states advancing at least 16 months (and as many as 30 months) between sham and crossover treatment dates. The group of sham patients who did not elect to crossover continued to exacerbate over the second year of follow-up, and the original TLD group continued to show a lower rate of severe AECOPDs at 2 years follow-up despite many more patients at risk for their first exacerbation [18].

A reduction in exacerbation frequency may reduce decline in lung function, reduce mortality, and improve quality of life. As exacerbations incur a large financial burden, if TLD is found to be effective at reducing respiratory-related hospitalizations, this could have a significant beneficial effect on healthcare resource utilization and financial benefits.

The primary limitation of this study is the small number of subjects and the known heterogeneity in the timing of COPD exacerbation events within a given group of patients [19]. Although this analysis of crossover subjects was not powered for any outcome, the observed hazard ratios for time-to-first-event are consistent with previous studies of TLD in a similar patient population. For moderate or severe AECOPD, the current study’s hazard ratio of 0.65 is similar to the 0.66 value observed during the randomized phase of AIRFLOW-2. The severe AECOPD hazard ratios are also similar with values 0.38 and 0.35, respectively [15]. A strength of this analysis is that subjects act as their own controls, comparing AECOPD during a 1-year blinded period (no active treatment) against a later 1-year open period (active treatment).

A further limitation is challenges in recording methods of exacerbations. As patients are frequently provided with a “rescue pack” of antibiotics, they often self-medicate for what they perceive to be an exacerbation without a medical consultation. This is a common limitation to many studies evaluating exacerbations. There is also recall bias as many patients do not remember how many times they have needed their rescue pack in the preceding year. Beyond patients acting as their own controls, the study attempted to further mitigate risk of recall bias, by providing patients with a memory aid where they were advised to document the details of any exacerbations or infections. Medication utilization was also recorded by the patient and researcher. Furthermore, the patients were regularly contacted by the study team who also kept a record of patient-reported exacerbations.

The findings of this study support previously reported COPD exacerbation risk reduction following TLD. Fewer crossover patients experienced an AECOPD in the year following treatment compared to their own prior year exacerbation history. Exacerbation risk has been consistently observed during the randomization phase of the AIRFLOW-2 trial and the 2nd year of follow-up in AIRFLOW-2 [15, 18]. We believe the paired outcomes from this small study, which confirm previously presented results of a randomized, double-blind study, provide added support for the impact of TLD on risk reduction for moderate or severe COPD exacerbations. A prospective randomized, double-blind, sham-controlled study [20] is currently underway and recruiting a larger numbers of patients (AIRFLOW-3) which will further explore the efficacy of TLD in reducing COPD exacerbations.

The authors thank Alexander Peterson for his contribution to this research; Dr. Martin Mayse, the inventor of TLD therapy; NAMSA for statistical analysis support; the AIRFLOW-2 research team; and the volunteer study participants.

All participants have given their written informed consent to participate in the study. The study protocol was approved by all relevant ethics boards, see online supplementary Table.

All clinical trial expenses were reimbursed by the study sponsor (Nuvaira, Inc.). AV and DJS are the co-principal investigators for this study. All other authors declare that they have no conflicts of interest to disclose.

All clinical trial expenses were reimbursed by the study sponsor (Nuvaira, Inc.).

Francesca Conway, James Tonkin, Arschang Valipour, Christophe Pison, Christian Schumann, Peter I. Bonta, Romain Kessler, Wolfgang Gesierich, Kaid Darwiche, Bernd Lamprecht, Dirk Skowasch, Philip J. Johnson, Dirk-Jan Slebos, and Pallav L. Shah were involved in the work reported, whether that was in study design, patient recruitment, data acquisition, data analysis, or interpretation of findings. All authors took part in drafting, revising, or critically reviewing the article and approved the final manuscript.

Data are available on demand at Nuvaira, Inc., Minneapolis, MN, USA. Further enquiries can be directed to the corresponding author.

Francesca Conway and James Tonkin contributed equally to this work.