Discriminating measures of bronchodilator drug efficacy and potency

Introduction

Bronchodilator drugs, used alone or in conjunction with inhaled corticosteroids or other anti-inflammatories, are essential to the effective management of asthma [1, 2]. The three main categories of effective bronchodilators are inhaled β₂-adrenoceptor agonists and anticholinergics, and methylxanthines, particularly oral theophylline. The incidence of asthma continues to rise, however, enhancing the need to develop new bronchodilator drugs and products and to evaluate their therapeutic efficacy, potency and cost-effectiveness in relation to existing treatments [3]. The short-acting β₂-adrenoceptor agonist salbutamol, in a pressurized metered dose inhaler (pMDI), has long been the most widely used bronchodilator, and a key factor influencing the requirement for new products is the phasing out of chlorofluorocarbon (CFC) pMDI propellants to curtail damage to the stratospheric ozone layer [4]. The pharmaceutical industry is also responding to a decade of off-patent availability of effective asthma drugs including salbutamol and the anticholinergic ipratropium bromide, which has encouraged the development of novel generic bronchodilator/inhaler device combinations [5]. This combination of factors has led to a pressing need to develop the most economical and precise methods of evaluating efficacy and potential clinical equivalence of inhaled bronchodilators with the traditional pMDI and current alternatives in order to satisfy both clinicians and regulatory authorities.

Clinical equivalence of drugs relates to both potency and safety since the lowest effective dose must be used in order to minimize systemic side-effects. Measures of bronchodilator drug efficacy have in the past relied on direct measurement of the induced bronchodilatation by lung function testing. This often does not, however, allow potency discrimination between drug products, since a single dose can result in maximal bronchodilatation. This has led both clinicians and pharmaceutical professionals to develop more sensitive, repeatable, and therefore more discriminating methodologies employing alternative patient populations and alternative efficacy endpoints. Dose–response data are fundamental to the accurate measurement of both efficacy and systemic effects. Importantly, regulatory authorities are also addressing these issues [6–8].

This review focuses on β₂-adrenoceptor agonist studies to highlight the discriminating measures of bronchodilator drug efficacy and potency which should enable a valid comparison of novel and traditional inhaled asthma treatments. (Measures and methodologies are summarized in Table 1.)

Table 1.

Comparison of methodologies for testing of bronchodilator efficacy.

Methodology	Reference*	Population	Advantages	Disadvantages	Nominal rating
Spirometry – FEV1	13	Mild to moderate reversible airways disease	Simple technique	Plateau effect	+
			Single dose comparison
Spirometry – FEV1	21	Mild to moderate reversible airways disease	Simple technique	Variability of disease (baseline and responsiveness)	++++
Separate day dose–response comparison			No carry over effects
				Increased number of study days required
Spirometry – FEV1	24	Mild to moderate reversible airways disease	Simple technique	Plateau effect	++
Cumulative dose–response comparison			Single study day	Carry over effects
Spirometry – FEV1	30,32	More severe reversible airways disease	Simple technique	Carry over effects	+++
Cumulative dose–response comparison			Clinically relevant population
			Reduced chance of plateau effect
Plethsymography – sGaw	34	Healthy volunteers	No variability due to disease state	Specialized equipment requirement	+
			Potentially larger response	Not tested in target population
Challenge – PC20/PD20	46,47	Mild to moderate reversible airways disease	Dose–reponse more easily demonstrated	Complex technique requiring experience	++++
				Clinical relevance of model
			Smaller number of patients than FEV1	Patient safety
Pharmacokinetic – drug levels in blood	60	Healthy volunteers and mild to moderate reversible airways disease	Early levels indicate relative lung dose	Invasive	+
			Simple, requires no specialized lung function testing	Indirect measure of efficacy
			Small number of subjects
Pharmacokinetic – drug levels in urine	64,65	Healthy volunteers and mild to moderate reversible airways disease	Early levels indicate relative lung dose	Indirect measure of efficacy	+
			Simple, requires no specialized lung function testing	No information on lung distribution
			Non–invasive
			Small number of subjects
Gamma scintigraphy	67	Healthy volunteers and mild to moderate reversible airways disease	Lung distribution information (confounded by 2 dimensional images)	Complex technique requiring experience	+
				Specialized equipment requirement
			Small numbers of subject	Indirect measure of efficacy
Long term clinical studies	68	Mild to moderate reversible airways disease	Clinically relevant population and treatment period	Variability issues	+++

^*Reference to an example of a study using such a design/methodology.

Spirometry – the major endpoint

Measurements of lung function are fundamental to both the diagnosis of asthma and the demonstration of treatment effects. Acute drug effects such as bronchodilatation may be measured spirometrically at intervals following drug inhalation to establish the peak and duration of effect. The forced expiratory volume in the first second (FEV₁) which like other spirometric indices is affected by airway obstruction, is generally accepted as the most useful measure of airway patency since it is considered to be a repeatable index. Peak expiratory flow (PEF) measurement is a convenient way of monitoring longer-term drug effects on lung function during clinical trials (e.g. twice-daily self-testing by patients using a mini-Wright peak flow meter at home). PEF is more effort-dependent than the relatively prolonged FEV₁ and although average changes in FEV₁ and PEF are similar during bronchodilatation or constriction, the variability of PEF is greater. Both values are made more reliable by taking the best of at least three attempts, but accuracy is better assured during spirometry where a low forced vital capacity (FVC) reading, if this is inconsistent, indicates a poor effort which can validly be discounted. Rigorous adherence to standardized spirometry techniques is fundamental to the accurate use of these endpoints [9, 10].

When 25% of the vital capacity (VC) has been expired, the larger airways will be largely compressed and flow-limiting. From this point, the forced expiratory flow (FEF) can provide some information about resistance of the small airways, which depends on the elastic properties of the lung tissue and the magnitude of VC. FEF values between 25% and 75% of VC (e.g. FEF₂₅, FEF₅₀, FEF₇₅, FEF_25-75) can be calculated from spirometry and used to gain added insight into bronchodilator [11] and anti-inflammatory [12] drug effects. This may be particularly important in patients with a low bronchodilator reversibility of FEV₁ who are not typically selected for Phase 3 drug studies.

Despite its limitations, FEV₁ has been widely used to compare both different bronchodilator drugs and delivery devices, since the method is easy to perform and relatively noninvasive [13]. The sensitivity of FEV₁ to differentiate in these comparative studies has, however, depended upon the selection of the dose used and, importantly, upon the population of patients. The acceptability of FEV₁ as an efficacy endpoint, to both clinicians and regulatory authorities, is therefore limited to carefully designed discriminating studies. Some of these approaches are discussed below. It should be noted that this issue applies not just to spirometric indices but to other measures of bronchodilator efficacy, potency and safety.

Study design issues

Appropriate study design can greatly increase the discriminatory power of bronchodilator studies, avoiding a great deal of wasted effort and conflicting results [14]. This is recognized in current regulatory guidelines for demonstrating bioequivalence, although the standards for new drug/device comparability may not be clearly defined [7, 8, 15]. The most common study design flaw responsible for inconsistencies in past comparisons of inhaled β-adrenoceptor agonists has been identified as being the use of only a single dose of each product [14]. Most often this would be the standard therapeutic dose, which might or might not exert a maximum effect under the particular study conditions [16]. Two bronchodilators showing no statistically significant differences in such a single-dose comparison could reflect either a lack of statistical power due to insufficient sample size (type II error) or merely the fact that each drug was in excess of the amount required for maximum bronchodilatation in that specific patient population. Dose-equivalence problems are compounded if the two drugs also have different time courses, as when comparing short-acting (e.g. salbutamol) with long-acting (e.g. salmeterol) β₂-adrenoceptor agonists [17].

Dose–response curves

To obviate the bronchodilator saturation effect which will bias the study outcome if the plateau of the individual dose–response curve has been attained, it is imperative to measure the responses to at least two, and preferably three or more, doses of each drug/device under test, establishing significant dose–response relationships for bronchodilator efficacy. Dose–response curves of the systemic drug effects are equally important for establishing full therapeutic equivalence [17]. Comparison of the ascending bronchodilator dose–response curves can determine the relative doses of each agent which produce similar bronchodilatatory effects in the lung [18]. The ratio of doses giving equivalent effects (potency ratio) indicates the relative potency of the two products. Employment of a standard reference product, often the brand leader such as Ventolin pMDI (GlaxoSmithKline, UK), further allows for indirect comparison between studies and new products in terms of their potencies relative to such a reference. However, the lowest manufactured dose of an inhaled bronchodilator will often produce a good response, and developing a full dose–response curve by including smaller, experimental doses is unrealistic [6]. Adequate statistical methods such as the Finney two-by-two bioassay are therefore recommended to validate dose–response comparative studies and provide relative potency values with confidence intervals [8, 15, 19]. It has been indicated by the US regulatory authorities that a valid and sensitive comparison of the relative potency ratio of two products can be determined not only by examination of the response ratio for linear dose response relationships but also from the dose ratio for nonlinear dose response curves [20]. The principles behind the recommendation of this type of analysis are shown in Figure 1.

Figure 1.

Calculation of relative potency ratio based on the dose and response scale. The use of the dose scale improves the sensitivity of the comparison. Figure adapted from Singh 1996 [20].

Single dose/separate day regimen

Bronchodilator dose–response studies generally employ a double-blind (often double-dummy) randomised, crossover protocol with appropriate washout periods between study days. Two or more doses of each drug, and placebo if included, may be administered each on a separate day, making a total of at least 4 study days [21]. This has the disadvantage that the participating asthmatic patients must ideally show the same baseline lung function on each occasion. Whilst the majority of patients with mild/moderate stable asthma will reproduce a baseline FEV₁ ≥ 65% of predicted, they are less likely to maintain FEV₁ reversibility ≥ 15%, so that the potential for improvement can be insufficient for discriminating measurements in some patients on certain days. Increasing patient numbers can reduce the potential impact of a variable baseline on study outcome. The sensitivity of FEV₁ and other endpoints may be increased by expressing each drug-induced improvement from baseline as a percentage of the maximum improvement observed for that individual throughout the study, rather than a percentage of the predicted normal value. This approach reflects the degree to which the patient's ventilation returns to maximum achievable function after a bronchodilator treatment, giving steeper dose–response curves and emphasizing differences when the baseline obstruction is mild [22].

Cumulative dose regimen

An alternative method is the use of cumulative dosing which allows estimation of a bronchodilator dose–response within 1 day, thus reducing some of these confounding effects. Starting with a potentially submaximal dose of 100 µg of salbutamol, for example, this is followed by a second 100 µg at 30 min, and a further 200 µg at 60 min. Since each dose adds on to the residual effect of the preceding one, the dose–response is less clear-cut [6]. A comparison of cumulative and noncumulative techniques in the same subjects showed a significantly greater FEV₁ response to isoprenaline following cumulative dosing [23]. This method is affected by timing, which should take account of the duration of action of each drug: for example, with short-acting β₂-adrenoceptor agonists some of the early effects will have worn off before the last dose, making the overall dose uncertain. This makes for difficulties regarding bronchodilators with a different onset or duration of action, such as salbutamol and salmeterol [17]. However, cumulative dosing is a useful and convenient technique for comparison of drugs with a similar time course of action, or for the same drug in different formulations. If conducted in appropriate patients, this method reliably provides two key points – e.g. for salbutamol, the submaximal effect of 100 µg and the cumulative effect of 400 µg from the same baseline – and it has been used to demonstrate clinical efficacy in a number of studies [24].

Patient selection

This represents the second key factor to improving the discriminating power of studies based on the measurement of FEV₁. Subject selection for bronchodilator testing has generally taken a cautious approach, avoiding asthmatic patients with more severely impaired lung function (< 65% predicted). However, this tends to limit the scope of the dose–response relationship, which may be particularly important when comparing drugs at dose levels used in clinical practice. The presence of more severe airway obstruction would shift the bronchodilator dose–response curve to the right, requiring higher doses of drug to achieve the same degree of improvement. Use of patients with severe airway obstruction, in addition to increasing the discriminatory potential of the study, also provides important information in a clinically relevant population.

Acute asthma

In a hospital-based study, patients manifested a greater response to a higher dose of salbutamol only when they had acute asthma [25]. Thus, 30 min after initial dosage, the mean FEV₁ had improved by approximately 34% in patients treated with 2.5 mg of salbutamol via a jet nebuliser, as compared with only 6% improvement in patients receiving 0.27 mg of salbutamol via a pMDI with spacer. After 24 h, no difference was apparent in the recovery rate of these two groups, irrespective of the earlier scale of lung β₂-receptor activation and bronchodilatation. Nebulisers, though variable, have been shown to deliver a similar proportion of the nominal dose of salbutamol to the lungs as the pMDI [26, 27]. Therefore, although different inhalers were employed, this study was able to distinguish a tenfold difference in the delivered dose of β₂-adrenoceptor agonist during relatively severe airway obstruction. Hospitalized patients with acute bronchial obstruction (mean FEV₁ 37% of the predicted normal) have also been shown to respond equally well to cumulative doses of salbutamol totalling either 2 mg from the Turbuhaler DPI or 4 mg from a pMDI plus spacer [28]. This nominal dose-difference is in keeping with improved lung delivery of drug from the Turbuhaler in comparison with the pMDI [13, 29].

Patients with relatively impaired lung function (mean FEV₁ 58% of the predicted normal) have also been used outside the hospital setting, for example, to compare the FEV₁ response to salbutamol delivered via the Ventolin pMDI and a novel DPI [30]. Since these patients again started with a relatively low baseline FEV₁, a greater degree of improvement was required before they reached the plateau of their dose–response. A potentially confounding issue in utilizing patients with acute asthma is that there may be an improvement in lung function over time which is independent of bronchodilator therapy.

Nocturnal asthma

Patients with nocturnal asthma demonstrate increased severity of obstruction during the night, and ethical studies are possible if such volunteers are continually observed and drugs are compared on a single-blind (to the subject) basis [31]. The cumulative dosing method is suited to this situation, and the procedure has been used as a practical bioassay in a two-night crossover study demonstrating the therapeutic equivalence of a new salbutamol DPI with the Ventolin pMDI [32]. The mean baseline FEV₁ of these subjects upon waking or being woken during the night was 43% ± 10% of predicted.

Whole-body plethysmography – an alternative endpoint

An alternative method of investigating bronchodilator drug efficacy is to record a flow-volume or pressure-volume loop using whole-body plethysmography. The specific airway conductance (sGaw=conductance or 1/airway resistance/lung volume) gives a measure of airway patency that is broadly independent of lung size and has been found more sensitive than FEV₁ as a measure of response to bronchodilators in normal, nonasthmatic subjects [33]. Although it may be considered less repeatable than FEV₁, more recent studies have demonstrated the use of sGaw as a sensitive pharmacodynamic endpoint able to establish the therapeutic equivalence of two salbutamol inhalers: a standard pMDI and a novel DPI [34]. This comparison was carried out in a preselected group of healthy volunteers who showed sGaw reversibility (≥ 30%) in the absence of significant changes in spirometry. The potential use of healthy volunteers is advantageous in reducing the effect of the variability in baseline airway calibre and bronchodilator responsiveness occurring in patients with asthma. On the other hand, the specialized equipment is expensive and necessitates thorough training of volunteers and operators. Whole-body plethysmography may be a useful technique which is more sensitive than spirometry, but it is less readily available and has yet to be considered in regulatory authority guidelines.

Challenge methods – a further option

Airway hyperresponsiveness (AHR) is a key feature of the asthmatic condition which persists during stable asthma and increases during exacerbations due to allergen exposure [35]. The persistence of AHR in patients with stable asthma has encouraged development of nonimmunological bronchoprovocation testing as an objective diagnostic tool [36]. A variety of bronchoconstrictor agents exist, including histamine, methacholine, sodium metabisulphite, adenosine monophosphate (AMP), cold air and hypertonic saline, but histamine and methacholine are the most reliable and widely validated [37]. Compared with normal subjects, the majority of whom show little response to inhaled bronchoconstrictors, patients with asthma demonstrate a reduction in the challenge dose of methacholine required to trigger a fall in FEV₁ [35]. This suggests a potentially efficient method for comparisons of inhaled bronchodilator drug efficacy; however, the detailed conditions required for accurate discrimination remain undecided [8, 9, 37, 38].

Bronchoconstrictors

Bronchoconstrictor agents are routinely delivered in doubling doses by tidal breathing via a nebuliser or using a dosimeter device, which may help standardize dosage and improve repeatability of the results [39, 40]. The bronchoconstrictor dose–response curve is used to establish the concentration or dose of the agent required to provoke a 20% fall in FEV₁ (PC₂₀ or PD₂₀). These endpoints are more sensitive but less repeatable than FEV₁ measurement per se, and it should also be noted that challenge methods may be less acceptable to patients due to the potentially unpleasant bronchoconstrictive effects of the challenge agents. However, this diagnostic methodology provides a useful measure of therapeutic efficacy in clinical trials of anti-inflammatory treatments for asthma, when an increase in methacholine or histamine PC₂₀/PD₂₀ after some weeks of treatment indicates a reduction in AHR [41–44]. Increasingly, it is being applied in acute bronchodilator studies to thereby increase the discriminatory capacity of the bronchodilator or bronchoprotective dose–response curve [15, 22, 37–39, 45–48].

Methacholine challenge

Methacholine or histamine challenge testing is recognized and recommended by regulatory authorities [6, 8]. Methacholine has been used in a number of recent comparative studies in which discriminatory power was demonstrated in relatively small numbers of patients [15, 45, 48]. Parameswaran has demonstrated that for inhaled β₂-adrenoceptor agonists in vivo therapeutic equivalence is adequately established, given repeatability and quality control of outcome measurements, by comparing relative potencies of both bronchodilation (requiring 12 subjects for a cumulative dose–response study) and bronchoprotection (requiring 18 subjects in a crossover challenge study) [48, 49]. In common with other methods, the inclusion of more severe patient populations would serve to increase the sensitivity of challenge testing as a measure of bronchodilator efficacy. Although patient admission to challenge tests is largely restricted to exclude those with a poor baseline lung function, the methodology of methacholine testing has been shown to be safe in asthmatic subjects with a mean baseline FEV₁ only 45% of predicted [38, 50]. This would in turn ease problems of recruiting sufficient patients with the required degree of PC₂₀ dose-responsiveness to one and two actuations of the reference salbutamol pMDI, as previously requested by regulatory authorities [7, 51].

Exercise challenge

Exercise-induced asthma (EIA) or bronchoconstriction (EIB) offers a more clinically relevant way of measuring AHR which can be applied to assessments of short- and long-acting bronchodilators, generally following a single-dose/separate day regimen [36]. Exercise challenge may be ethically preferable in paediatric studies, but it is less easily reproducible than methacholine or histamine challenge [14]. In common with other indirect provocations, exercise reflects more complex mechanisms of airway obstruction including cellular and neurogenic components [12, 37].

Surrogate measures

Lung deposition in vitro

Pharmacodynamic (PD) measurements of comparative bronchodilator efficacy are based on the relative number of receptors stimulated in the lung, which in turn is linked to the amount of drug reaching the appropriate regions of the lung. This opens up the possibility of using pharmacokinetic (PK) measurements as surrogates for the potentially more expensive and time-consuming PD methods [52, 53]. Lung delivery of inhaled drugs is related to the particle size (fine particle or respirable fraction) which is aerosolized by the particular inhaler device [29, 54]. In vitro measurement of the respirable fraction is limited by the ability of the laboratory apparatus to mimic airway anatomy and in vitro predictions of the fine particle fraction tend to overestimate lung deposition, although some studies have successfully correlated in vitro performance to in vivo efficacy [55, 56]. Current recommendations include the use of multistage impingers or impactors to characterize particle size distribution adequately. Measurements should be made at a range of flow rates and profiles to reflect clinical use of the inhaler device. Development of an apparatus with a more anatomically correct ‘throat’ is also encouraged and such laboratory studies have been conducted [57, 58].

Lung deposition in vivo

Two in vivo methods of assessing lung deposition are currently recognized by the British Association for Lung Research (BALR) – a PK method, using plasma or urine concentrations of drug to represent whole-lung deposition and absorption, and a gamma-scintigraphic method [52, 57]. These methodologies are currently only recognized as ‘supplementary’ by the regulatory authorities. The PK method is valid for drugs which are absorbed but not metabolized locally in the lung and undergo little gastrointestinal absorption. The latter can be blocked by coadministration of charcoal allowing determination of lung deposition, or lung absorption can be distinguished by timing of sample collections [59–62]. Plasma levels of salbutamol over the first 20 min after inhalation represent drug absorbed from the lungs [60]. Urinary salbutamol during the first 30 min, a noninvasive and simpler method, also represents drug absorbed from the lungs and indicates relative lung bioavailability when expressed as a proportion of the 24 h urinary excretion [62, 63]. A clear linear dose–response relationship has been demonstrated over 1–5 actuations of salbutamol (100 µg) pMDI [64]. This method assumes that the amount of drug absorbed rapidly is a good measure of the amount available for bronchodilation. Although the optimal particle size distribution for lung absorption may not be the same as for bronchodilation, PK parameters have been correlated with the methacholine challenge response in patients [65]. The chief disadvantage of the PK methods is that they lack the ability to detect differences in regional lung deposition; however, a relationship has been shown between lung deposition pattern and PK profile [66]. The other in vivo BALR-recommended method, gamma-scintigraphy, can assess differences in regional deposition and lung penetration (following inhalation of ^99mTc-radiolabelled drug) by scanning the chest with a gamma camera to produce a computerized image. This method is the subject of a further review in the current series [67].

Clinical trials

Most comparisons of efficacy of inhaled bronchodilator products are concentrated on acute drug effects but formal, long-term (4- and 12-week) clinical efficacy and safety studies of short- and long-acting β₂-adrenoceptor agonists are still required by regulatory authorities [8]. In addition to subjective and clinical assessments of asthma severity, the usual objective efficacy measurements are daily (morning and evening) PEF, and periodic spirometry (FEV₁) performed prior to morning inhalation of the trial drug, comparator or placebo. A recent 12 week clinical trial has employed maximum percentage changes from baseline in FEV₁, area under the curve and duration of bronchodilation, measured for 6 h following the first daily dose at 4 weekly intervals, to demonstrate therapeutic equivalence of a new salbutamol DPI with the standard salbutamol pMDI [68]. Such periodic assessment of spirometric changes can be weakened by baseline variation since, for example, an increase in baseline FEV₁ during treatment produces an apparent diminution of bronchodilator effect [68]. Periodic tests for bronchodilator responsiveness must also take into account the duration of action of the trial drugs in order to obviate residual bronchodilatation carried over from the previous dose [69]. The potential for development of tolerance to β₂-adrenoceptor agonists is another important consideration in longer-term clinical studies, especially where regular dosing is involved [70]. Assessment of protection against bronchoconstrictors is a more sensitive measure of β₂-adrenoceptor agonist activity than bronchodilatation alone, therefore changes in acute β₂-adrenoceptor agonist responsiveness during the clinical use of trial bronchodilators can be more precisely monitored by means of methacholine or histamine PC₂₀ measurements [69, 70].

Conclusion

The burgeoning selection of new bronchodilator/inhaler devices calls for definitive standards by which to establish efficacy and safety relative to the current major treatment, the CFC-driven salbutamol pMDI. These standards are being set under regulatory authority guidance with a view to phasing out the environmentally unsafe CFC-pMDIs as soon as acceptable alternative treatments are established [7, 8, 71]. Cost and convenience are major considerations, which encourage the development of study designs reliant on short-term PD and surrogate measures rather than lengthy therapeutic studies. These alternative methodologies are often more sensitive, and therefore more discriminating for comparison of novel products and generics with the standard treatment. FEV₁ remains the key measurement, and the provision of adequate bronchodilator dose–response curves can be improved by selecting more severely asthmatic volunteers, or by employing this spirometric endpoint within a methacholine or histamine challenge study design.

Development of more sensitive, repeatable, and therefore more discriminating, methodologies will enable a clinically valid ranking of new therapies, new delivery systems and traditional inhaled asthma treatments.