Clinical Bioequivalence of Wixela Inhub and Advair Diskus in Adults With Asthma

Background: Wixela® Inhub® is a dry powder inhaler approved as a generic equivalent to Advair® Diskus® (fluticasone propionate [FP]/salmeterol fixed-dose combination) for patients with asthma or chronic obstructive pulmonary disease (COPD). This study aimed at confirming the local (lung) therapeutic equivalence of both the FP and salmeterol components of Wixela Inhub (test [T]) to Advair Diskus (reference [R]) after inhalation. Methods: This randomized, double-blind, double-dummy, placebo-controlled, parallel-group study in patients ≥18 years with mild-to-moderate persistent asthma compared the local therapeutic equivalence (using forced expiratory volume in 1 second [FEV1]) of FP/salmeterol (100/50 μg) after inhaled delivery via T and R. Results: Randomized patients (N = 1127) received T (n = 512), R (n = 512), or placebo (n = 103). T and R significantly increased day 1 FEV1 area under the effect curve over 12 hours of the change from baseline (AUC[0–12]) and day 29 trough FEV1 over placebo, indicating that these endpoints were sufficiently sensitive for evaluation of bioequivalence. On day 1, T and R each increased FEV1 AUC(0–12) over placebo (3.134 L•h [T], 2.677 L•h [R]; each p < 0.0001). Following twice-daily dosing for 28 days, T and R also each increased trough FEV1 (measured on day 29) over placebo (235 mL [T], 215 mL [R]; each p < 0.0001). Least-squares mean T/R ratios (90% confidence intervals) for day 1 FEV1 AUC(0–12) and day 29 trough FEV1 were 1.120 (1.016–1.237) and 1.069 (0.938–1.220), respectively, indicating that T and R were bioequivalent for both co-primary endpoints. FP/salmeterol was well tolerated when administered via either T or R. Conclusions: These results demonstrate that the therapeutic effects of Wixela Inhub are bioequivalent to Advair Diskus in the lung. Wixela Inhub represents a therapeutically equivalent new FP/salmeterol treatment option for use in the treatment of asthma and COPD.


Introduction
I nhaled corticosteroids (ICS) and long-acting b 2adrenergic agonists (LABA) are widely used, safe, and effective anti-inflammatory and bronchodilator agents, respectively, for the treatment of asthma and chronic obstructive pulmonary disease (COPD). (1,2) Current guidelines recommend the administration of fixed-dose ICS/LABA combination drugs as maintenance therapy in asthma and COPD. (3)(4)(5) Advair Diskus (GlaxoSmithKline, Research Triangle Park, NC) is a widely prescribed ICS/LABA combination drug (fluticasone propionate [FP]/salmeterol [as xinafoate]; FPS) for asthmatic patients not controlled with ICS alone and for COPD patients at high risk of exacerbations. (1,2) With the expiration of the US patent for Advair Diskus in 2016, several generic versions are currently advancing toward regulatory approval. (6)(7)(8)(9) The most advanced of these, in terms of drug development stage in the United States, is Wixela Inhub, composed of FPS inhalation powder (Mylan, Inc., Canonsburg, PA) predispensed in a multidose inhaler (Inhub; Mylan, Inc.), (10,11) which was recently approved by the US Food and Drug Administration (FDA). (12) The FDA guidelines for the development of generic FPS inhalers require, as part of a weight of evidence approach (together with in vitro pharmaceutical equivalence and systemic pharmacokinetic bioequivalence), local (lung) therapeutic equivalence studies that, in total, demonstrate therapeutic equivalence to Advair Diskus. (13) Clinical development of Wixela Inhub followed these guidelines, and recent studies confirmed the pharmacokinetic bioequivalence of single doses of Wixela Inhub for each of the three authorized Advair Diskus dose strengths. (14) Here, we report the results of the FDA-mandated local therapeutic equivalence study (NCT02245672) in adult patients with asthma.
The objective of this study was to compare the clinical efficacy of the FP and salmeterol components of Wixela Inhub 100/50 lg and Advair Diskus 100/50 lg by using spirometry. To evaluate bioequivalence of the bronchodilator component (salmeterol), forced expiratory volume in 1 second (FEV 1 ) was measured repeatedly for 12 hours after the first study dose. The anti-inflammatory component (FP) was then evaluated by measuring trough FEV 1 after 28 days of twice-daily dosing.

Materials and Methods
In this article, ''test product'' (T) and ''reference product'' (R) are defined as follows: T is Wixela Inhub (FPS administered via the Inhub inhaler), and R is Advair Diskus.

Study design and conduct
This multicenter, randomized, double-blind, doubledummy, placebo-controlled, parallel-group study was conducted between October 22, 2014 and July 10, 2015 at 101 U.S. centers. The study consisted of a 21-28-day single-blind, placebo run-in period followed by a 4-week double-blind treatment period. The primary objective was to assess the local therapeutic equivalence of T and R using spirometry.
The study conformed to appropriate ethical guidelines and was conducted in accordance with the principles of the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use guidelines for good clinical practice (15) and the code of ethics of the World Medical Association's Declaration of Helsinki. (16) Quorum Institutional Review Board approved the study protocol, and all patients provided written informed consent.

Patients and treatments
Consistent with the FDA guidelines for a clinical endpoint study to assess local therapeutic equivalence of FPS products, (13) key inclusion criteria included age ‡18 years with diagnosis of asthma ‡12 weeks according to National Asthma Education and Prevention Program guidelines (3) ; a mean baseline FEV 1 of 50%-85% predicted after ‡6 hours without short-acting bronchodilator use; postbronchodilator reversibility (percent improvement) of ‡12% within 30 minutes of 360 lg albuterol; and current nonsmokers (with no smoking history within the past 12 months and a total smoking history of £10 pack-years). Patients were excluded if they had a respiratory condition or another severe progressive disease other than asthma and allergic rhinitis, were hospitalized for asthma within the past year or had an asthma exacerbation within the preceding 3 months, or had a respiratory tract, sinus, or ear infection within the preceding 4 weeks.
After completion of the placebo run-in period of 21-28 days (all subjects receiving placebo for Wixela Inhub, one inhalation twice daily), eligible patients were randomly assigned to one of three groups (T, R, or placebo) in a 5:5:1 ratio by using a subject identification number assigned via an automated interactive voice-/web-response system. Each treatment was administered in a double-blind, doubledummy manner (with placebo inhalers matched to T or R used for the placebo group and to maintain the blind in the active treatment groups). Patients were required to take one inhalation twice daily from each of their assigned inhalers for 4 weeks.
Advair Diskus and Wixela Inhub contained qualitatively and quantitatively equivalent formulations of both active pharmaceutical ingredients (a fixed-dose combination of micronized crystalline FP and salmeterol [as xinafoate]) and inactive excipients (lactose monohydrate). The Diskus and Inhub inhalers were medium resistance passive dry powder inhalers, contained 60 premetered doses of FP and salmeterol, and had the same operating procedures. (17) The pharmaceutical performance of multiple commercial batches of R (Advair Diskus) was tested to characterize the performance using in vitro methods, including measures of delivered dose and aerodynamic particle size distribution. (18) The single batch of Advair Diskus used in the study was representative of the median of the Advair Diskus commercial batch population.
The T drug (Wixela Inhub) was also tested to characterize performance by using in vitro methods. The two batches of Wixela Inhub used in the study were manufactured at commercial scale, representative of the product in terms of in vitro performance, and were age-matched to be within 3 months of the batch of Advair Diskus used in the study.
The placebo for Advair Diskus used commercial stock of Advair Diskus inhalers. Specifically, these were opened under good manufacturing practice (GMP) conditions, the blister strips containing FP and salmeterol were replaced with strips containing lactose, and the inhalers were subsequently closed and packaged for clinical trial use. The placebo for Wixela Inhub used Inhub inhalers containing lactose alone.

Assessments
Spirometry assessments were completed at screening (day -28); at run-in (day -3 to -7); at -0.5, 0, 0.5, 1, 2, 3, 4, 6, 8, 10, and 12 hours on day 1 of treatment, before dosing on day 15, and on day 29. Primary endpoints were the area under the effect curve over 12 hours (FEV 1 AUC (0-12) ) for the change from baseline (CFB) in FEV 1 on day 1, the first day of treatment, and CFB in trough FEV 1 on day 29 after 4 weeks of dosing. Safety assessments included adverse events (AEs) and laboratory safety tests, vital signs (blood pressure and pulse rate), and electrocardiograms.

Statistical analysis
The safety set was defined as all randomly assigned patients who had taken ‡1 dose of study drug and for whom postdose safety data were available. The full analysis set (FAS) was defined as all randomly assigned patients who had taken ‡1 dose of study drug and had provided data for either co-primary efficacy endpoint (FEV 1 AUC (0-12) or day 29 CFB in trough FEV 1 ). The per-protocol set (PPS) was defined as all patients in the FAS who had not violated or deviated from the protocol in a manner that could have affected the outcome of the FEV 1 assessments for both coprimary efficacy endpoints. The FAS was the primary analysis set used to establish assay sensitivity (the ability to discriminate both T and R treatments from placebo), whereas the PPS was the primary analysis set used to establish bioequivalence between T and R treatments.
The original sample size for this study was calculated by assuming a 92% between-subject coefficient of variation (CV; expected mean and standard deviation [SD] for CFB in trough FEV 1 on day 29 of 0.51 and 0.47 L, respectively). This led to an estimated sample size of 380 subjects for T and 380 subjects for R to give 90% power to demonstrate clinical bioequivalence (T/R ratio and 90% confidence interval [CI] wholly contained within the 0.80-1.25 limits) between T and R, assuming a true T/R ratio of 1.0. The original sample size for the placebo group (n = 76) was based on performing a two-sided significance test at the 5% level with 99.9% power, an SD of 0.47 L, and a true mean difference from each active arm of 0.3 L for CFB in trough FEV 1 on day 29 (allocation ratio of 5:1 for active to placebo). Therefore, the total number of subjects required to ) to allow for *10% dropout postrandomization. The process was repeated for the FEV 1 AUC (0-12) endpoint.
As sample size assumptions were based on historical reports of the effect of Advair Diskus in similar but not identical patient populations, (19)(20)(21) a blinded sample size re-estimation (BSSR), which was prespecified in the protocol, was conducted for this study when 286 subjects had completed the study. The assumptions made about the CV for the original sample size calculation for CFB in trough FEV 1 on day 29 were not supported by the aggregate data used for the BSSR. Therefore, the sample size was recalculated and revised accordingly. The total sample size to complete the study was revised from 836 to 990 subjects (450 [T] + 450 [R] + 90 [placebo]) requiring *1100 subjects to be randomized (the maximum allowable in the protocol). The revised sample size for the active treatment arms in this study was based on at least 81% power and assumed 112% between-subject CV (expected mean and SD for CFB in trough FEV 1 on day 29 of 0.26 and 0.29 L, respectively, as assessed from the BSSR results). The sample size for the placebo group (n = 90) was chosen to maintain the allocation ratio (5:5:1).
Determination of assay sensitivity was required for the bioequivalence results to be valid. To assess bioequivalence, LS means (one for T and one for R) from the ANCOVA models were used to generate T/R ratios for LS means for FEV 1 AUC (0-12) and trough FEV 1 efficacy endpoints. Overall, 90% CIs were calculated by using Fieller's theorem. (22) To demonstrate bioequivalence, the 90% CIs for the FEV 1 AUC (0-12) and trough FEV 1 T/R ratios were each required to be wholly contained within the interval 0.80-1.25 (i.e., 80%-125%).

Patients
Of the 1871 enrolled patients, 1127 (60%) were randomized and treated, with 512 patients each receiving T and R and 103 patients receiving placebo (Fig. 1). The most common reason for exclusion of enrolled patients was failure to meet baseline spirometry criteria. All 1127 patients receiving a study treatment were analyzed in the safety set. Baseline demographic and clinical characteristics were well matched across treatment groups (Table 1). In the total study population (safety set), mean age was 42.6 years, 40% of patients were male, mean duration (range) of asthma was 27.1 (0.3-79.8) years, mean (SD) FEV 1 percent predicted was 69.94 (8.76), and mean percent improvement in FEV 1 postbronchodilator was 23.84 (16.17). A total of 607 (54%) participants were taking ICS or ICS/LABA medication for maintenance of their asthma before entering the washout period of the study. Overall, 96% of patients in the safety set were compliant (i.e., within 75%-125% of per-protocol inhaler use) with treatment during the double-blind phase, and compliance was comparable for T (96%), R (95%), and placebo (97%).

Efficacy
Both active treatments substantially improved day 1 FEV 1 by the first time point measured (mean CFB, T, 270 mL; R, 237 mL at 30 minutes postdose; Fig. 2); there was minimal improvement with placebo (mean CFB 52 mL). The maximum increase in FEV 1 was observed at 3 hours postdose (mean CFB, T, 379 mL; R, 333 mL; placebo, 101 mL). T and R demonstrated similar FEV 1 responses, with overlapping 95% CIs over the 12 hours of serial spirometry measures made on day 1 with clear separation from placebo (Fig. 2). LS mean increases in day 1 FEV 1 AUC (0-12) were comparable for T and R (3.953 and 3.496 L h, respectively) and less for placebo 0.819 L h ( Fig. 3A and Table 2 [FAS]).
Both active treatments also substantially improved day 29 trough FEV 1 . The LS mean increases in CFB in trough FEV 1 after twice-daily dosing for 28 days were 293 mL (T), 272 mL (R), and 58 mL (placebo) (Fig. 3B and Table 2 [FAS]). As both T and R significantly increased day 1 FEV 1 AUC (0-12) and day 29 trough FEV 1 over placebo ( p < 0.0001; Table 2), the prespecified primary analysis criteria for assay sensitivity were met.

Safety
Treatment-emergent AEs occurred in 14.4% of patients in the safety set, with individual AEs displaying a similar incidence across the three treatment groups ( Table 3). The percentage of asthma-related AEs was higher in the placebo group (4.9%) and lower and comparable in both active treatment groups (T, 1.4%; R, 2.0%). The percentage of discontinuations was also higher in the placebo group (4.9%) compared with the active treatment groups (T, 2.5%; R, 2.3%). The most commonly reported AEs were infections and respiratory disorders. No serious AEs or deaths occurred during the study period. AEs associated with FPS, such as oral candidiasis and dysphonia, occurred with a similarly low incidence in the T and R groups (candidiasis: 0.8% vs. 0.4%; dysphonia: 0.2% vs. 0.6%, respectively) and did not occur at all in the placebo group. A very low incidence (<1%) of AEs categorized as cardiac disorders was observed, all of which were mild in intensity, did not require   intervention, and did not result in patients being withdrawn from the study. There were no clinically significant changes to laboratory safety tests, vital signs, or electrocardiograms.

Discussion
Wixela Inhub was recently approved by the FDA as a fixed-dose FPS powder for oral inhalation to provide a generic equivalent to Advair Diskus. This study, recommended by the FDA for the clinical development of generic inhaled drugs containing FP and salmeterol powder, (13) confirmed the local therapeutic equivalence of both the FP and salmeterol components of Wixela Inhub (T) after inhalation of 100/50 lg FPS, the lowest approved dose strength for Advair Diskus (R). Further, FPS administered as Wixela Inhub demonstrated a comparable safety profile to Advair Diskus.
A direct comparison of these results with those reported for the original Advair Diskus pivotal trials can be challenging, not only due to the expected limitations inherent in comparing between studies (23) but also because of fundamental changes in the management of asthma itself over the past 20 years. (24) As more asthma patients are treated with ICS and ICS/LABA therapy, the patients willing and able to participate in placebo-controlled studies may have a milder phenotype of asthma than those from 20 years ago and, hence, the opportunity to observe the magnitude of changes in lung function previously reported may be limited. A total of 54% of participants were taking ICS or ICS/LABA medication before the washout period in this study, of whom approximately one-third were taking an ICS without a LABA, and the remaining two-thirds were taking an ICS with a LABA. Thus, although the results (day 1 CFB in FEV 1 AUC (0-12) : 3.95 L h [T] and 3.50 L h [R]; day 29 CFB in trough FEV 1 : 293 mL [T] and 272 mL [R]) are lower in absolute magnitude of lung function improvement than those originally reported for the same dose of Advair Diskus (5.81 L h and 510 mL, respectively), (19) the findings are otherwise consistent.
The day 1 and day 29 spirometry data were also comparable with those reported for the OT329 SOLIS bioequivalence study, which used an almost identical study design. (7) For example, the day 1 CFB in FEV 1 AUC (0-12) for T (3.95 L h) and R (3.50 L h) were similar with respect to the magnitude of change with the corresponding day 1 values for OT329 SOLIS and Advair Diskus (3.72 and 3.55 L h, respectively). In addition, the day 29 placebo-corrected CFB in trough FEV 1 for T (235 mL) and R (215 mL) in this study are more similar to the corresponding day 29 values for OT329 SOLIS and Advair Diskus (168 and 163 mL, respectively) than to historical studies. The consistency of spirometry data across these more recently conducted studies suggests that the study design is robust, and the results are reproducible and representative of treatment effects in this population of asthma patients.
The design of this study was consistent with other FPS local therapeutic equivalence studies, (7,9,21) and they adhered to the FDA guidelines for evaluation of local therapeutic equivalence for FPS products. (13) The use of the lowest of three dose strengths of the FP component approved for Advair Diskus was appropriate, because it was the most likely to identify any treatment differences in FP between T and R and consistent with the FDA guidance. The use of higher dose strengths, which elicit maximal responses of FPS in many patients, (25) might have masked potential differences between T and R and  resulted in erroneous conclusions. We acknowledge, however, that international regulatory agencies may have different requirements for study designs and doses to be studied for the demonstration of local therapeutic equivalence. (26) Due to the difference in physical appearance of T and R, each treatment was administered twice daily for 28 days in a double-blind manner, using the double-dummy technique (27) with placebo inhalers matched to T or R. This can be considered a gold standard for clinical trials and contrasts with the bioequivalence study for OT329 SOLIS, in which the placebo treatment group received the placebo for the SOLIS inhaler only and hence the T and R were not blinded between each other. (7) Use of a double-dummy technique, which increases the robustness of conclusions of randomized trials of experimental interventions, (28) is a strength of this study. This robust double-dummy study design was also employed in local therapeutic equivalence trials of a novel dry powder FPS inhaler (AirFluSal Ò Forspiro Ò ; Sandoz International GmbH, Holzkirchen, Germany) (9) and a chlorofluorocarbon-free metered-dose FPS inhaler. (21) The results of the primary analyses were corroborated by secondary analyses (assay sensitivity [PPS] and bioequivalence [FAS]) that showed that (1) day 1 FEV 1 AUC (0-12) and day 29 trough FEV 1 endpoints were significantly superior to placebo for both T and R ( p < 0.0001) and (2) T was bioequivalent to R for both co-primary endpoints.
The demonstration of local therapeutic equivalence using spirometry endpoints in this article is also supported by previously presented data on pharmacokinetic bioequivalence to all three-dose strengths of Advair Diskus (100/50, 250/50, and 500/50 lg FP/S) (Haughie et al., 2019), in vitro equivalence (e.g., emitted dose) at all three dose strengths, (18) as well as meeting all of the FDA requirements for device equivalence. (17) In conclusion, Wixela Inhub, which was recently approved by the FDA, will represent a new generic-equivalent FPS treatment option for asthmatic patients whose symptoms are uncontrolled with ICS alone and COPD patients at high risk of exacerbations.