This was a phase IV randomized, double-blind, double-dummy parallel trial performed at a single study center in Incheon, Republic of Korea (Figure 1). After obtaining informed consent, unblinded staff randomly enrolled subjects (1:1) into either of the two groups using an Excel program (Microsoft, Redmond, WA, USA) with a block size of 10. The groups were: (1) one puff of active FORM/BDP (Foster [formoterol fumarate dihydrate 6 μg/beclomethasone dipropionate 100 μg], Chiesi Pharmaceuticals, Parma, Italy) and one puff of placebo FORM/BUD twice daily for 8 weeks (FORM/BDP group); and (2) one puff of active FORM/BUD (Symbicort Rapihaler [formoterol fumarate dihydrate 4.5 μg/budesonide 160 μg], AstraZeneca, London, UK) and one puff of placebo FORM/BDP administered twice daily for 8 weeks (FORM/BUD group). Four weeks after randomization, investigators called subjects by phone at a predetermined time and assessed asthma control and AE. Subjects were asked to increase the dose of each inhaler (one active and the other placebo) from one puff twice daily to two puffs twice daily if the following conditions were satisfied, which indicated uncontrolled asthma: (1) daytime symptoms ≥3/wk, (2) reliever uses ≥3/wk, (3) nocturnal symptoms, and (4) limitation of activity due to asthma. AE was indicated if persistent asthma symptoms required admission to the general ward or intensive care unit, visiting the emergency room, unscheduled visits to the outpatient clinic, or use of oral corticosteroids.

Participants and investigators assessing treatment outcomes and adverse events were blinded to the treatment; blinding was maintained until all assessments and recordings were completed.

Subjects enrolled in the study were adult asthmatics using low-dose ICS/LABA. The inclusion criteria were: (1) age ≥19 years; (2) typical asthma symptoms of dyspnea and wheezing aggravated by exercise, a cold, climate or diurnal change in temperature, and exposure to allergens, air pollution, tobacco smoke, or foul odor; (3) bronchodilator or treatment response proven by repetitive lung function tests (an increase in forced expiratory volume in 1 second [FEV₁] of 200 mL and 12% from baseline after the inhalation of salbutamol or appropriate asthma treatment) or airway hyperresponsiveness (AHR) verified by methacholine challenge test (provocative concentration of methacholine causing a 20% drop in FEV₁ [PC₂₀] ≤16 mg/mL); (4) inhalation of low-dose ICS and LABA for asthma treatment [2]. The exclusion criteria were: (1) malignancy or other serious comorbid diseases that could influence the study; (2) previous use of FORM/BDP or FORM/BUD; (3) pregnancy or lactation; (4) inadequate ICS/LABA, such as poor compliance to the inhaled medication or side effects of ICS or LABAs; (5) lack of patient consent to participate in the study.

Baseline parameters of patient’s medical history, including symptom duration, time of asthma diagnosis and treatment, comorbid allergic diseases, asthma control, AE, and current medications, were recorded. Subjects were also asked to respond to questionnaires including ACT, ACQ, and QLQAKA, and to undergo lung function tests and measurement of fractional exhaled nitric oxide (FeNO) [18-20]. These parameters were measured again at the end of the 8-week study drug treatment. Local or systemic adverse events were also recorded at the end of the 8-week treatment with the study drug.

Before the start of the study drug treatment, subject skills in handling pMDIs were evaluated as reported previously, and any errors were corrected immediately after evaluation. These skills were evaluated again after the 8-week study treatment [21]. Since active and placebo FORM/BDP used in this study did not have dose counters, the remnant doses of active and placebo FORM/BUD, which had dose counters, were recorded to determine compliance with study drugs in the FORM/BUD and FORM/BDP groups, respectively.

At the end of the 8-week study, subjects were asked to respond to inhaler preference questionnaires, as described in previous reports [5,6,8,11,22]. The subjects were asked about design, the connection between the cap and the body, size, grip, ease of pressing, general ease and convenience, flume velocity, irritation, taste, smell, feeling of the drug reaching the trachea and bronchus, overall satisfaction, and willingness to obtain a prescription for an inhaler. At the end of the study, subjects were asked to select one of the two pMDIs or other inhalers.

We compared continuous variables between groups using the Mann-Whitney U test and analyzed categorical variables using the chi-square test or Fisher’s exact test. Data are shown as medians (interquartile ranges [IQRs]) for continuous variables and frequencies (%) for categorical variables. An adjusted p<0.05 was considered statistically significant. All statistical analyses were performed with SPSS version 19 (IBM Corp., Armonk, NY, USA).

A power calculation was performed with a two-sample t test for the primary outcome variable of QLQAKA. We expected a difference of 10.0 in the QLQAKA scores between the two groups 8 weeks after treatment. We estimated that the SD would be 10.0. Aiming for a power of 0.80 and using a type I error rate of 0.05 and 40% loss to follow-up, we calculated the sample size at 40.

The study protocol was reviewed and approved by the affiliated institutional review board (IRB approval number: GFIRB2018-232). The study was registered at Clinical Research Information Service, Republic of Korea (CRIS KCT0003538). All subjects were enrolled after informed consent obtained.

Patients were screened between August 2018 and February 2019. We assessed 45 subjects for eligibility; three subjects were not eligible, and two declined to participate after being introduced to the study. The remaining 40 subjects were randomly allocated into the FORM/BDP group (n=18) or the FORM/BUD group (n=22) (Figure 2). Of the 40 participants, 24 (60.0%) completed the 8-week treatment period; 12 (30.0%) dropped out for not showing up on the day of the follow-up visit, and four (10.0%) withdrew because of adverse events. A total of 10 patients in the FORM/BDP group and 14 in the FORM/BUD group completed the study. Asthma treatment outcomes, lung function, adverse events, compliance with inhalers, subject preferences regarding inhalers, and skills in handling inhalers were analyzed.

Overall, baseline characteristics were balanced between the treatment groups (Table 1). Participants had a median age of 63.0 years (IQR, 49.5-72.0), and 47.5% of participants were male. Median height, weight, and BMI were 158.0 cm (IQR, 154.3-169.5), 57.9 kg (IQR, 53.4-71.6), and 24.0 (IQR, 22.0-26.0), respectively. A total of 22 (55.0%) and 18 (45.0%) patients were never-smokers, and current/former smokers, respectively. Among smokers, the median cumulative dose of smoking was 20.0 pack-years (IQR, 14.0-73.0). The median asthma symptom duration was 4 years (IQR, 1.38-10.0), and the time since first diagnosis and treatment was 1.8 years (IQR, 0.1-4.8). Five subjects (12.5%) showed AHR in the methacholine challenge test. One subject (1.5%) and 34 others (85.0%) showed bronchodilator response and treatment response, respectively, with an FEV₁ increase of 200 mL and 12% from baseline after the inhalation of salbutamol or appropriate asthma treatment. Asthma was well-controlled, partlycontrolled, and uncontrolled in 11 (27.5%), 23 (57.5%), and six (15.0%), respectively, according to the Global Initiative for Asthma (GINA) criteria. Median ACT, ACQ, and QLQAKA scores were 15.5 (IQR, 12.0-19.8), 1.6 (IQR, 1.1-2.6), and 54.5 (IQR, 46.0-66.0), respectively. Median frequencies of AE ever in life and in the past year were 1.0 (IQR, 0.0-1.0) and 0.0 (IQR, 0.0-0.8), respectively. In terms of comorbid allergic diseases, 18 (45.0%), three (7.5%), four (10.0%), seven (17.5%), two (5.0%), and one (2.5%) had allergic rhinitis, atopic dermatitis, urticaria, drug allergy, food allergy, and contact dermatitis, respectively. The baseline median percent-predicted forced vital capacity (FVC), FEV₁, and forced expiratory flow at 25% to 75% of forced vital capacity (FEF_25-75) values were 78.0 (IQR, 66.0-89.0), 71.0% (IQR, 62.0%-80.0%), and 47.0% (IQR, 30.0%-56.0%), respectively. Baseline FEV₁/FVC and FeNO values were 69.0% (IQR, 62.0%-73.0%) and 53.0 ppb (IQR, 37.5-92.0), respectively. Leukotriene antagonist and methylxanthine were prescribed as other asthma controllers in 24 (60.0%) and three (7.5%) subjects, respectively. Ten subjects (25.0%) showed errors in handling pMDIs; however, only two (5.0%) showed errors after instruction and immediate correction.

Among intention-to-treat and per protocol populations, no differences in baseline characteristics were found between the FORM/BDP and the FORM/BUD groups.

Asthma treatment outcomes are shown in Table 2. There were no differences in asthma control at 4 and 8 weeks or in asthma symptom scores, including ACT, ACQ, and QLQAKA scores, between the two groups 8 weeks after treatment. In four subjects (two in the FORM/BDP group and two in the FORM/BUD group) with uncontrolled asthma 4 weeks after treatment, the inhalation doses of both active and placebo drugs were escalated from 1 puff twice daily to 2 puffs twice daily. In two subjects (one in the FORM/BDP group and one in the FORM/BUD group), asthma remained uncontrolled 8 weeks after treatment. In terms of lung function, changes in FVC and FEV₁ were not different between the two groups, whereas those in FEF_25-75 and FEV₁/FVC were significantly higher in the FORM/BDP group than in the FORM/BUD group 8 weeks after treatment (17.9% of baseline [IQR, 3.5-128.2] vs. 2.9 [IQR, -16.4 to 17.9], p=0.037; 3.0% [IQR, 1.0-13.5] vs. 1.0 [IQR, -3.8 to 3.8], p=0.043, respectively) (Figure 3). Changes in FeNO, compliance with inhalers, and skills in handling inhalers did not differ between the two groups.

Adverse events during the study are presented in Supplementary Table S1. The most common adverse event was dry throat (37.5%), followed by cough (25.0%), voice difficulty (20.8%), hoarseness (20.8%), dyspnea (12.5%), a feeling of humid throat (12.5%), other abnormal feelings in the throat (8.3%), chest discomfort (8.3%), tremor (8.3%), headache (8.3%), and palpitations (4.2%). These adverse events did not differ between the two groups.

Subjects’ preference and their final choice of inhaler at the end of the study are shown in Table 3 and Supplementary Figure S1. There were no differences in preference between FORM/BDP and FORM/BUD in terms of design, the connection between the cap and body, size, grip, convenience, ease, and smell. Subjects responded that the taste of FORM/BDP was stronger than that of FORM/BUD (median, 5.0; IQR, 0.0-65.0 vs. median, 0.0; IQR, 0.0-10.0; p=0.030); however, there were no differences in preference between the two pMDIs with regard to taste. Subjects responded that the flume velocity of FORM/BDP was higher (median, 80; IQR, 50.0-100.0 vs. median, 50.0; IQR, 32.5-50.0; p=0.011), but more adequate (median, 92.5; IQR, 50.0-100.0 vs. median, 50; IQR, 32.5-77.5; p=0.013) and that inhaled FORM/BDP reached the trachea and bronchus more than inhaled FORM/BUD (median, 87.5; IQR, 50.0-100.0 vs. median, 30.0; IQR, 0.0-50.0; p<0.001) and consequently was more irritating (median, 55; IQR, 50.0-100.0 vs. median, 12.5; IQR, 0.0-50.0; p=0.001). The overall satisfaction did not differ between the two pMDIs. At the end of the study, 11 (45.8%), 12 (50.0%), and one (4.2%) chose FORM/BDP, FORM/BUD, and another ICS/LABA (fluticasone/vilanterol), respectively, as inhalers for use (Supplementary Figure 1A). However, based on the assumption that FORM/BDP had a dose counter, 15 (62.5%), eight (33.3%), and one (4.2%) chose FORM/BDP, FORM/BUD, and another ICS/LABA (fluticasone/vilanterol), respectively, as inhalers that they would be willing to use (Supplementary Figure 1B).