INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The Montreal protocol recommends that chlorofluorocarbon (CFC) be replaced by hydrofluoroalkanes (HFA) as propellants in metered-dose inhalers (MDIs) (1). There are no standardized methods to demonstrate in vivo bioequivalence of HFA-salbutamol to innovator CFC-salbutamol. The most practical method of showing therapeutic equivalence in vivo is by estimating their relative potencies (RP) in clinical efficacy studies. ₂-agonists have two clinically distinct pharmacodynamic effects in asthma: bronchodilation and prevention of bronchoconstriction (2). The majority of in vivo bioequivalence studies of ₂-agonists, including HFA-salbutamol, have compared the bronchodilator action of different drugs (3, 4). The nonbronchodilator effects may be studied by comparing the protection against bronchoconstriction caused by direct agents like methacholine or indirect agents like exercise, cold air, or allergen. In this study, we estimated the RP of an HFA salbutamol MDI (Proventil-HFA) and the innovator CFC salbutamol MDI (Ventolin-CFC), by comparing the dose-response curves of their bronchoprotective effect on methacholine bronchoconstriction with three doses, using the Finney bioassay. In addition, we provide recommendations for similar studies.

Subjects. Eighteen nonsmoking subjects with mild stable asthma and little or no current symptoms or airflow obstruction volunteered to participate in the study (Table 1). Their initial provocative concentration of methacholine causing a 20% reduction in FEV₁ (PC₂₀) was less than 4 mg/ml with a fourfold or more increase in PC₂₀ after 200 µg of salbutamol given from an MDI 2 h after the initial methacholine test at the enrollment visit. The medications were unchanged for the previous 3 mo. Exacerbating factors such as respiratory infection or allergen exposure were not present 6 wk before the study or during the study. Four patients were on inhaled steroids (median daily dose, 300 µg). The study was approved by the Hospital Research Ethics Committee, written informed consent was obtained from each subject, and the study was performed in accord with the Declaration of Helsinki.

Design. The study was of randomized, dose-level blind, placebo-controlled, crossover design, in two centers. The bronchoprotective effects of 100, 200, and 400 µg and placebo of Proventil-HFA were compared with the same doses and placebo of Ventolin-CFC. Each subject was given all eight doses over eight visits (each separated by 2 to 7 d), one dose at each visit. The allocation sequence was determined by a balanced, 8 × 8 latin square. At Visit 1, informed consent was obtained and subjects had a medical examination. They practiced their inhalation technique using a Vitalograph MDI modified spirometer (Vitalograph Ltd., Buckinghamshire, UK) to ensure good hand-mouth coordination, peak inspiratory flow of between 10 and 50 L/min, and a breath-hold of at least 5 s. At Visit 2, the subjects rested for 30 min and baseline FEV₁ and PC₂₀ methacholine were determined. The FEV₁ was allowed to recover spontaneously to more than 90% of baseline value and measured after 115 min. The subjects then inhaled 2 × 100 µg/dose of salbutamol MDI and methacholine challenge was initiated after 10 min so that a PC₂₀ was determined approximately 30 min postsalbutamol. The effect of methacholine was reversed with salbutamol if necessary. The protective effect of salbutamol was calculated as PC₂₀ postsalbutamol/ PC₂₀ baseline. At Visits 3 to 10 the bronchoprotective effect of both drugs was examined. Subjects attended the laboratory at the same time as Visit 2. Between visits, salbutamol MDI was allowed but not for 8 h before. FEV₁ was measured and had to be within 10% of the baseline at Visit 2. Pulse rate (PR) and blood pressure (BP) were recorded. The MDI inhalation technique was practiced. The MDI was shaken and actuated five times into a plastic bag. Then one of the four doses of either Ventolin-CFC or Proventil-HFA was inhaled within 3 min. The inhalations were administered by a third person to maintain blinding. Five minutes after the last inhalation, PR and BP were measured. After 10 min, methacholine challenge was commenced to obtain a PC₂₀. Adverse experiences to methacholine (chest tightness, cough, wheeze, shortness of breath, headache, throat irritation, nose irritation, hoarseness, flushing) were noted on a three-point Likert scale.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The FEV₁ and vital capacity were measured according to American Thoracic Society (ATS) guidelines (5). The methacholine chloride concentration of each vial was checked by high-pressure liquid chromatography before using it for the study and again after 3 mo for stability. Methacholine inhalation tests were performed as described by Juniper and coworkers (6). Doubling concentrations of methacholine between 0.06 and 256 mg/ml were inhaled by tidal breathing for 2 min at Visit 2. For subsequent visits, the initial dose of methacholine was two concentrations below the PC₂₀ at Visit 2. The fall in FEV₁ was recorded as the difference between the lowest postsaline value and the lowest postmethacholine value. PC₂₀ was expressed in noncumulative units.

Analysis

Demographic data were summarized using descriptive statistics. Log (base 2) transformed PC₂₀ values were compared using a repeated- measures analysis of variance (ANOVA) with treatment sequence assignment, patient within sequence, drug, dose, drug-dose interaction, and study day as within-subject factors in the model. Using the ANOVA model, the linear slope of the log dose-response curve was estimated for both drugs using orthogonal components for the 100-, 200-, and 400-µg data (7). RP was estimated using the estimate of the slope and the root mean square error (RMSE) from the mixed model, using the Finney bioassay methods, and a 90% confidence interval (CI) was estimated using Fieller's theorem (8). The validity of the estimate was also checked by two nonlinear mixed effects E_max models for Ventolin-CFC and Proventil-HFA (7) and the RP was expressed as the ratio of mean effective dose of HFA (ED_50HFA) to mean effective dose of CFC (ED_50CFC). The adverse events caused by methacholine were analyzed by descriptive and qualitative analysis. The effects of the drugs on PR, systolic (SBP) and diastolic (DBP) blood pressures were analyzed by repeated-measures ANOVA. The dose-response relationships were examined using SAS 6.12 PROC MIXED (SAS Institute, Cary, NC). The nonlinear E_max models were fitted using the SAS macro NLINMIX.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The mean (± SD) percentage change in baseline FEV₁ at each visit was 5.2 ± 7.1%. The methacholine concentrations were within 95% of the target concentrations (relative SD of 6%). Duplicate samples from the calibration standards analyzed 3 mo apart were within 99% of the average. Both Ventolin-CFC and Proventil-HFA significantly increased geometric mean PC₂₀ mg/ml (% SEM) in a dose-dependent manner. The values were 1.9 (1.0), 14.5 (1.1), 27.5 (1.1), 31.6 (1.1) after placebo, 100, 200, and 400 µg, respectively, of Ventolin-CFC and 1.7 (1.1), 14.8 (1.1), 26.1 (1.0), and 39.3 (1.1), respectively, after Proventil-HFA (p = 0.0001) (Figure 1). The effects on the dose-response of sequence in which the drugs were administered and different study days were not significant. The dose effect was independent of the formulation; i.e., there was no difference in the slopes of the dose-response curves between the two formulations, justifying the assumption of parallelism of the dose-response regression lines. This was confirmed by a nonsignificant difference in slopes of the two dose-response regression equations (p = 0.76, paired t test). The RP (90% CI), estimated by the Finney method using the 100-, 200-, and 400-µg data, was 1.08 (0.81-1.46) indicating that each actuation of the Proventil-HFA delivered the same quantity of salbutamol to the lung as 1.08 actuations of Ventolin-CFC. To compare the precision of this estimate with other estimates using only two doses, RP were also calculated using three different pairs of dosages: 100, 200; 100, 400; and 200, 400 (Table 2). The narrowest CI, indicating the most precise estimate was obtained with the 100-µg and 200-µg doses. The RP (90% CI) obtained by the E_max model (ED₅₀ for Proventil-HFA and Ventolin-CFC, 0.63 and 0.66, respectively) was 1.05 (0.71-1.39). The changes in heart rate, SBP, and DBP with the different doses of both drugs were not clinically or statistically significant. There were no significant adverse events to methacholine.

View larger version (16K): [in this window] [in a new window]   Figure 1.   Dose-response curves of salbutamol (µg) plotted against PC20 methacholine (log2) for Proventil-HFA (closed squares) and Ventolin-CFC (open circles).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have demonstrated that the HFA- and CFC-salbutamol are bioequivalent in their bronchoprotective action on methacholine bronchoconstriction with an RP of 1.08. This method of establishing bioequivalence in vivo is feasible and valid based on comparison of the Finney and E_max methods of bioassay. The most precise estimate of relative potency is obtained when 100 and 200 µg of the two inhalers are compared.

There are recommendations to conduct bronchodilator bioequivalence studies of ₂-agonists (9). Currently, there are no guidelines to perform bronchoprotective bioequivalence studies. In the present study, we made every effort to control factors that might influence the dose delivered to the airways, interpretation of the methacholine PC₂₀ and consequently the results of the study. We selected a range of doses routinely employed in clinical practice, to establish a dose-response curve. We ensured that the concentration of methacholine used was accurate and stable during the period of the study.

Bioassay of the bronchoprotective effect of inhaled ₂-agonists has been performed in asthmatic subjects comparing two salbutamol formulations (10), single dose of CFC- and HFA- salbutamol (11), salbutamol and metaprotenol (12) and two delivery devices of terbutaline (13). A bioassay of the bronchodilator effect of inhaled ₂-agonists in normal subjects who were bronchoconstricted with histamine has also been reported (14). These studies were not performed as rigorously as the present study; of particular importance was the lack of a range of doses, quality control of methacholine, and ensuring the inhaler techniques of the subjects.

Our results show that each Proventil actuation delivers to the airway significantly more than 81%, and significantly less than 146%, salbutamol as does one actuation of Ventolin. This range of clinical efficacy is within the recommendation of the U.S. Food and Drug Administration for establishing bioequivalence of inhaled medications (15). The RP was similar to that obtained for the bronchodilator effect by Kleerup and coworkers (3) (1.08; 90% CI, 0.95-1.23), providing validity to our estimate. The tightest CI (0.77-1.31) was obtained with the analysis using 100- and 200-µg doses, and this was close to the CI of the RP estimated by the E_max models. This suggests that the optimal doses for a precise estimation of RP of bronchoprotective effect of two different inhaled salbutamol preparations should compare 100- and 200-µg doses, which are on the linear parts of the dose-response curve, similar to the dose-response for the bronchodilator effects (16). However, if the dose-response curve has not been previously established, at least three doses need to be studied.

One of the problems with bioequivalence studies has been a lack of statistical power owing to small sample sizes (17). We conducted a post hoc power calculation using Monte Carlo simulations of the ratio of the RMSE of the dose effect from the ANOVA model (s) and the slope of the dose-response curve (b) (s/b ratio = 0.39). Our study had 80% power to detect RP with a 90% CI of 0.67 to 1.50. Fifty-one patients would have given a more precise estimate (narrower 90% CI of 0.80 to 1.25). Analysis of the cumulative bronchodilator data from Kleerup and coworkers (3) suggests that only 28 patients are required to estimate the RP for the bronchodilator effect with similar precision. This result, plus the easier design, suggests that this design is the preferred method.

We conclude that the relative potencies of inhaled short-acting beta-agonists can be determined by comparing their bronchoprotective effects on methacholine bronchoconstriction. By this method, the relative potency of HFA-salbutamol compared with CFC-salbutamol was 1.08 (90% CI, 0.81-1.46). 100- and 200-µg doses of salbutamol MDI in 18 patients are sufficient to establish bioequivalence of two different formulations with 80% power.