Long Acting Bronchodilatators in COPD. Drug Selection by Means of the SOJA Method

Objectives: The increasing number of direct acting anticoagulant drugs [DOACs] makes it almost impossible to have sufficient knowledge of each individual medicine and device, especially for general practitioners. Reducing the number of medicines different DOACs, based on rational criteria, allows physicians and pharmacists to build experience with a more limited set of medicines and to optimise patient information. Methods: In this study DOACs are compared by means of the SOJA method. The following selection criteria were applied: approved indications, available formulations, variability of the AUC, drug interactions, clinical efficacy, side effects, dosage frequency and documentation. Results: Limited differences in scores were found between apixaban, dabigratran and rivaroxaban. Edoxaban showed a lower score, mostly because of its more limited clinical evidence and documentation. The ranking between the top 3 depends mostly on the assigned weight to the individual selection criteria. Acquisition cost was not taken into account, because this varies with time. In practice acquisition cost is of course an important selection criterion, especially because there are very limited differences between the medicines from a clinical perspective. Exclusion of this criterion also makes this comparison more internationally applicable. Conclusion: All DOACs are suitable for formulary inclusion, followed by a selection of the most suitable for a DOAC in individual patients, based on patient characteristics.

The combination formoterol/fluticasone propionate [Flutiform] was not included in the manuscript, because this combination is only approved in asthma.
After the authors had determined the set of selection criteria, Medline, Embase and the Cochrane database were searched and references from review articles obtained.

Selection criteria
The following selection criteria were applied.

Number of approved indications
This was judged as follows [data derived from the Summaries of Product Characteristics [SPC].
From a formulary perspective, it is an advantage that a medicine is approved for both COPD and asthma.
When a medicine is approved for COPD alone, it scores 70%.
When it is approved for asthma as well, it is awarded 100%.

Drug interactions
This criterion is only of importance in formulary decision making as the vast majority of patients treated with long acting bronchodilators will not experience any drug interactions. Drug inter-actions may result in a reduction of clinical efficacy of the COPD medicine in question or in a reduction of the clinical efficacy of the other drug, with which the interaction occurs. Interactions may also give rise to increased toxicity of one or both compounds. The more frequent these interactions occur and the more serious the consequences are, the lower the score for the drug in question.

Clinical efficacy
The judgement of the relative clinical efficacy should ideally be based on a large number of direct comparative studies between the combinations using clinically relevant endpoints. Unfortunately, the number of direct comparative studies is quite limited.
Double-blind comparative studies with dry powder inhalers have hardly been done, because of the necessity of the companies to participate in such a double-blind, double-dummy study.
The following rules of play were used: • Only studies in COPD patients between 2 or more medicines of from placebo-controlled studies], meta-analyses and database studies.

Tolerability
Tolerability is an important selection criterion for each individual group of medicines, especially for those medicines which are used chronically. Tolerability [the lack of occurrence of non-se- vere reactions] was determined from clinical studies [direct comparisons between 2 or more medicines of from placebo-controlled studies], meta-analyses and database studies.

Dosage frequency
The dosage frequency plays an important role in patient compliance. Compliance is not usually a problem in patients taking drugs once or twice daily, but decreases considerably in the event that 3-4 dosages are to be taken daily. The method of evaluation of this criterion corresponded with that of all of the other SOJA scores: • Once daily: 100% • One or twice daily: 90% • Twice daily: 80% • Three times daily: 40%.

Documentation [clinical studies]
The • Five percent of the maximum score was assigned for each study of a specific drug. As a result, the score for 20 studies is 100%.
• The number of patients in these studies [50%] • For every 10 patients participating in these studies 1% of the maximum score was assigned. As a result, the score for 1000 patients is 100%.

Documentation [clinical experience]
The two sub criteria are indicative of the overall clinical experience with the drug. These sub criteria may introduce a bias to the advantage of older drugs, but this is done intentionally. were excluded. Both the number of patients that has been treated on a worldwide bases and the period that a certain drug has been available are of importance, as it may take time until adverse reactions occur: • The number of years on the market [50%].
• Every year a certain drug has been on the market represents 10% of the score. If a drug has been on the market for at least 10 years, the score is 100%.
• Every one million of patient days of experience represents 1% of the score. If the number of patients days of experience exceeds 100 million, the score is 100%.
tive beta-adrenergic blockers should be preferred, although they should be administered with caution.

Hypokalaemic treatment
Concomitant hypokalaemic treatment with methylxanthine derivatives, steroids, or non-potassium sparing diuretics may potentiate the possible hypokalaemic effect of beta2-adrenergic agonists, therefore caution is required.

Valid for all anticholinergic agents
Co-administration of aclidinium bromide with other anticholinergic-containing medicinal products has not been studied and is not recommended. Co-administration with erythromycin was not associated with any serious adverse effects.

Tiotropium
No formal drug interaction studies have been performed with tiotropium, but in clinical studies the drug was combined without problems with sympathomimetics, methylxanthines and oral and inhaled corticosteroids.

Umeclidinium
Clinically significant interactions mediated by umeclidinium bromide at clinical doses are considered unlikely due to the low plasma concentrations achieved after inhaled dosing.

Metabolic and transporter based interactions
Umeclidinium bromide is a substrate of cytochrome P450 2D6  Combinations of LABA and LAMA are awarded 70%.

Interaction with P-glycoprotein inhibitors
Combinations of LABA and inhaled corticosteroids are also awarded 70%, with the exception of combinations with salmeterol or budesonide, which may show interactions with inhibitors of CYP450. These formulations score 60%.
Triple combinations are also awarded 70%.

Clinical efficacy
All medicines have been compared to placebo. These studies are not discussed in detail in this manuscript when the scope of The study details regarding the design and results of the studies is provided in a separate document .

Placebo-controlled studies
Aclidinium was compared to placebo in several double-blind studies. Aclidinium was consistently more effective concerning FEV1 and other lung function parameters as well as on exercise endurance [109][110][111][112][113][114][115]. No significant effect on exacerbation frequency was found in individual studies.
A Cochrane meta-analysis was performed on all available aclidinium studies, which included 12 multicenter RCTs randomly assigning over 9,500 participants with stable COPD. All studies were industry-sponsored. There was no difference between aclidinium and placebo in all-cause mortality and number of patients with exacerbations requiring a short course of oral steroids or antibi-

Aclidinium vs formoterol
Two studies compared aclidinium and formoterol [as well as combination of both medicines]. Although aclidinium tended to have slightly more favourable effects on lung function parameters, no significant differences in clinical efficacy were seen between both medicines [40,41]. These studies were not taken into consideration for the documentation of formoterol, because the Genuair device was used. Formoterol is not available as such as monotherapy in the Genuair device.

Formoterol vs glycopyrronium
Three studies compared formoterol to glycopyrronium, both administered as MDI. The combination of both medicines was also included in these studies. These studies are not included in the tables, because a different inhaler [MDI] was used [127][128][129].

Formoterol vs olodaterol
Four studies compared formoterol Aerolizer to olodaterol

Active controls
Indacaterol was compared to tiotropium in three studies, with salmeterol in two and with salmeterol/fluticasone in one study.

Salmeterol versus theophylline
Salmeterol was more effective than theophylline measured by pulmonary function tests in one study [156] and the combination of both agents was more effective than monotherapy in another study [157].

Placebo controlled studies
Tiotropium was more effective than placebo regarding lung

Tiotropium versus ipratropium
Comparative studies between tiotropium and ipratropium were not included in Tables 1-6, because these studies fall outside the scope of this analysis. These studies were taken into consideration for the criterion documentation.  Tiotropium was more effective than ipratropium [188][189][190] and improved symptoms when added to theophylline [191].
Tiotropium doses of 5 and 10 mcg delivered by Respimat was as effective as 18 mcg delivered by Handihaler [192].

Placebo-controlled studies
Addition of umeclidinium to fluticasone/vilanterol resulted in significantly better clinical efficacy than placebo regarding FEV1 and SGRQ [193]. The same results were obtained in another study comparing umeclidinium or placebo added to inhaled corticosteroid/long acting beta 2 agonist [194].

Comparison with salmeterol/fluticasone
Umeclidinium Ellipta [62.5 microg] was more effective than tiotropium Handihaler regarding effects on FEV1 at day 85 in a direct comparative study [89].

Placebo-controlled studies
Vilanterol was more effective than placebo in improving lung function in one study [195]. definition], but a significant reduction was seen compared to place-

bo. When the EXACT [EXAcerbations of Chronic pulmonary disease
Tool respiratory symptoms questionnaire] definition of an exacerbation was used, the rate of exacerbations was significantly lower for the 400/12 mcg combination compared to aclidinium and pla-

cebo. The time to first exacerbation [any of the two definitions] was
significantly prolonged for the 400/12 mcg combination compared to placebo, but not to aclidinium or formoterol [198].

Comparison with salmeterol/fluticasone
The

Glycopyrronium/indacaterol vs tiotropium
The combination was compared to tiotropium in three studies.
The effects on studied endpoints, such as lung function parameters, SGR, TDI and rescue medication were similar for both medicines [66][67][68]. One study showed more favourable effects on FEV1 and patient-reported dyspnoea for the combination [68].

Glycopyrronium/indacaterol vs umeclidinium/vilanterol
The combinations were compared in two randomised, controlled, cross-over studies. The purpose was to demonstrate non inferiority of glycopyrronium/indacaterol. This goal was not met, although no clinically relevant differences in any lung function parameter were found 72].

Glycopyrronium/indacaterol vs salmeterol/fluticasone
The combinations were compared in three studies. The combination was as effective when given in one container or in separate containers [207].

Comparison with individual components
The combination was more effective in improving lung function tests and SGRQ total scores than the individual components [75,[208][209][210].

Comparison with salmeterol/fluticasone
The combination olodaterol/tiotropium was more effective in improving lung function tests than salmeterol/fluticasone [211].

Comparison with individual components
The combination was significantly more effective in improving lung function tests and quality of life than the individual compo- The effects on exacerbation rates compared to salmeterol were not completely consistent. An advantage was found in most [81,86,87,92], but not all studies [83].

Placebo-controlled studies
The combination was more effective than placebo in improving health-related quality of life [221].

Comparison with individual components
The combination was more effective in improving lung function tests than the individual components. Effects on TDI and SGRQ and exercise endurance were also superior to umeclidinium and vilanterol separately. No significant differences in the effect on exacerbations were found, however [90][91][92][93][94].

Comparison with tiotropium
The combination was more effective in improving lung function tests, TDI and SGRQ than tiotropium in two direct comparative studies [94,95].

Comparison with tiotropium/olodaterol
The combination umeclidinium/vilanterol was non inferior to tiotropium/olodaterol in a per protocol population and superior in an intent-to-treat analysis regarding the primary endpoint [FEV1 change at 8 weeks] [76].
care in a randomised manner in 75 general practices, including moderate to severe exacerbations among patients who had an exacerbation within one year before the trial. The rate of moderate or severe exacerbations was significantly lower by 8.4% with the combination than with usual care [224].

Beclomethasone/formoterol/glycopyrronium
The triple combination was compared to budesonide/formoterol in the Trilogy study. Pre-dose and 2 hours post dose FEV1 were significantly improved by the triple combination. Adjusted annual exacerbation frequencies were also improved by the triple combi- The side-effects profile of Aerolizer and Nexthaler were similar in a direct comparative study [228].

Indacaterol
A Cochrane meta-analysis on all available studies, which included 13 multicenter RCTs randomly assigning over 9,600 participants with stable COPD was published early 2015 [141].
Nine trials contributed data on serious adverse events with indacaterol and placebo. No statistically significant difference in the incidence of serious adverse events or mortality was noted for [any dose level of] indacaterol and placebo. The confidence intervals were however too wide to rule out clinically important differences in serious adverse events between indacaterol and formoterol or salmeterol [141].
Four trials with active controls contributed data on serious adverse events. Dosages of 150 mcg, 300 mcg and 600 mcg indacaterol were compared to formoterol and salmeterol No statistically significant difference in serious adverse events was reported between each dose of indacaterol and both active comparators.
The confidence intervals were however too wide to rule out clini-

Umeclidinium/Vilanterol
The effects of umeclidinium and the combination with vilanterol on QTc interval were comparable to placebo [243]. An analysis of studies with budesonide did not show an increased risk of pneumonia compared to non-users of budesonide [245].

Salmeterol/fluticasone
The large scale Torch study compared the combination to its components and placebo, The general tolerability profile was simi- There are no relevant differences in safety of monotherapies and combinations of LABA and LAMA. All medicines are awarded 80%.
There is a higher incidence of pneumonia for combinations including inhaled corticosteroids. These are awarded 70%. Combinations using fluticasone propionate were scored 5% lower because of its relatively high incidence of pneumonia. These were scored 65%.

Tolerability Aclidinium
A Cochrane meta-analysis was performed on all available aclidinium studies, which included 12 multicentre RCTs randomly assigning over 9,500 participants with stable COPD. There was no significant difference between aclidinium and placebo or tiotro- No differences in the tolerability profile of formoterol and tiotropium were found in a direct comparative study [44].
The tolerability profiles of formoterol and olodaterol were similar in four comparative studies [54,55].
Studies comparing formoterol to the combination formoterol/ budesonide or formoterol/beclomethasone showed no relevant differences in tolerability [51].

Tiotropium
Tiotropium was well tolerated in placebo-controlled trials.

Umeclidinium
Umeclidinium was well tolerated. The incidence of adverse events was comparable to placebo [194].

Aclidinium/formoterol
Two studies compared the aclidinium/formoterol combinations 400/6 mcg and 400/12 mcg bid to the individual components. No statistically significant differences were observed in the incidence of adverse events between the comparators [40,41]. These results were similar at longer follow-up [43].

Formoterol/budesonide
Studies comparing formoterol to the combination formoterol/ budesonide or formoterol/beclomethasone showed no relevant differences in tolerability [51].

Glycopyrronium/indacaterol
The combination was well tolerated in clinical trials with few, if any, differences with placebo or the individual components of the combination. Large scale comparative studies with tiotropium and salmeterol/fluticasone did not show major differences in tolerabil-ity or safety [65][66][67][68][69][70][71].

Olodaterol/tiotropium
The combination was well tolerated in clinical trials. The incidence and severity of adverse events was comparable to the individual components [75,209].

Comparison with individual components
The combination as well tolerated as the individual components [97,98].

Comparison with salmeterol/fluticasone
The tolerability profile of vilanterol/fluticasone furoate was comparable to salmeterol/fluticasone twice daily in four studies, summarised in two publications [101,102].
There are no relevant differences in the tolerability profiles of the formulations. All medicines are awarded 80%.

Dosage frequency
The dosage frequency as indicated in the SPC's is as follows: Formoterol, Salmeterol, aclidinium [and combination] and the triple combination beclomethasone/formoterol/glycopyrronium are given twice daily and score 80%.
The other medicines/combinations are given once daily and are awarded 100%.

Documentation [clinical studies]
The documentation [clinical studies] is as follows.

The documentation [clinical experience] is as follows
[

Discussion and Conclusion
Score.

Applied methodology
This was done by means of the SOJA method, which is a wellestablished rational and transparent way of selecting medicines within a therapeutic class from a formulary perspective.

Outcome
Only very limited differences in scores were found between the medicines within classes [LAMA, LABA, LABA/LAMA, LABA/ ICS and triple combinations]. The highest and lowest scores were within a 5% margin.

Strength and limitations of the methodology
It should be taken into consideration that this analysis is limited to pharmacological aspects. In clinical practice, patient related factors play an important role, such as inhalation flow, hand-eye coordination and personal preferences of the patient.
The evaluation of criteria in the SOJA method is highly standardised in order to promote unbiased judgement of drugs from various pharmacotherapeutic categories based on clinically relevant criteria. There will always be room for debate whether or not the correct scoring system was used for each criterion and judgement may be arbitrary for most, if not all, criteria. This is the case with any method used to quantify properties of drugs. The SOJA method is intended as a tool for rational drug decision making, forcing clinicians and pharmacists to include all relevant aspects of a certain group of drugs, thereby preventing formulary decisions being based on only one or two criteria. Besides this, possible "hidden criteria" are excluded from the decision making process. The outcome of this study should be seen as the basis for discussions within formulary committees and not as an absolute truth.
Obviously, the score depends on the relative weight that is assigned to each individual selection criterion. Therefore an interactive program is available, which makes it easy for local and regional formulary committees to assign personal weights to each selection criterion by individual members of the committees. If a physician or pharmacist considers individual criteria as totally irrelevant, this criterion may be assigned 0 points, thereby ignoring this criterion.
There is extensive experience with the SOJA method, which made clear that almost all physicians and pharmacists assign a high relative weight to clinical efficacy, safety and tolerability and ease of use. There were very limited differences between the individual medicines within each class.
Acquisition cost was not taken into account, because this varies with time. In practice acquisition cost is of course an important selection criterion, especially because there are very limited differences between the medicines from a clinical perspective. Exclusion of this criterion also makes this comparison more internationally applicable.
The device was not taken into consideration for the calculation of the scores. There are no relevant differences between the properties of the medicines. Therefore the Working Party decided that the selection of inhaled medication for COPD should be based on the properties of the devices, rather than on the properties of the medicines.

Conclusion
The scores for the individual medicines within each class of inhaled medicines were quite similar. This makes it logical to base formulary decisions on the properties of the devices. This analysis will be published in a separate article.