Effectiveness of combined ICS/LABAs (text extract)

This is the text extract for Effectiveness of combined ICS/LABAs, browse documents here.

Effectiveness of combined ICS/LABAs delivery devices versus concurrent ICS and LABA via separate inhalers In summary: • • • There is little evidence that combination LABA/ICS products improve compliance over concurrent use of separate ICSA and LABA inhalers. There is newly emerging evidence that combination LABA/ICS products may be physiologically more effective than concurrent use of ICSs and LABAs. However, the extent of these improvements in physiological measures (which are but intermediate or surrogate outcomes) has been overstated. The true extent of peak expiratory flow rate (PEFR) reduction is 11% (using relative risk, not odds ratio). In addition, there were no differences in clinically relevant outcomes for combination LABA/ICS products. Separate analysis of withdrawals and adverse events (not performed in the industryfunded Seretide® meta-analysis) shows significantly higher rates of reported adverse events with Seretide®. Compared with salmeterol (and even fluticasone), Seretide® has much smaller changes in physiological effects, reflected in nil clinical improvements. Note that neither Seretide® nor fluticasone have been able to demonstrate clinically significant improvements in pooled analyses (figure 1).

• •

•

Figure 1

Fluticasone, salmeterol and Seretide pooled RCTs - physiological vs clinical improvements

12% 11% 10% % clinical improvement (treatment vs controls) 9% 8% 7% 6% 5% 4% 3% 2% 1% 0% 0 -1% -2%

Seretide®: PEFR difference/baseline = 1.6%

exacerbations reduced symptoms improved overall* (agents with no significant clinical improvements)

*combines mean % days w/o symptoms, mean % nights w/o symptoms, mean % days w/o rescue Rx, mean % nights w/o rescue Rx

Seretide® (Nelson etc 2003)

fluticasone plle studies (Cochrane review 2003)*

*for clinical measures, FP data are incomplete (incompatibility issues) and based on 2-11 RCTs

salmeterol (MIASMA 2000)

PEFR treatment vs controls (L/min)

FP plle studies*: PEFR difference/baseline = 3.9% salmeterol: PEFR difference/baseline = 6.5%

•

Hence, the advice from the British, GINA and New Zealand asthma guidelines still applies, that there is no difference in clinical efficacy between combination and concurrent (separate devices) LABA/ICS.

P51-0-0 #73191 PHARMAC 11.08.03

Compliance with combination LABA/ICS devices The recently updated joint British Thoracic Society/Scottish Intercollegiate Guidelines Network asthma guidelines (BTS/SIGN guidelines) (section 11.2.2, page i59) state “Combination inhalers have not been shown to improve compliance in the medium to long term.”1 The BTS/SIGN guide lines cite as evidence for the above statement a double -blind RCT comparing nedocromil/salbutamol combination with nedocromil alone 2 , an open-label RCT of combination vs. concurrent terbutaline (short-acting beta2 agonist) and budesonide (ICS)3 , and 70% long-term compliance in a fluticasone RCT4 (a somewhat curious citation, but perhaps because compliance was higher than in the other two studies). In three out of four identified RCTs for Seretide® (Bateman etc 1998, Chapman etc 1999, Van den Berg etc 2000 below), mean compliance (actual use/expected) was respectively reported at 91%, 96% and 93% for combination LABA/ICS, versus 89%, 95% and 93% for concurrent LABA/ICS use via separate inhalers. These formally combine to give a nonsignificant 1% difference in (already high) compliance (Peto one-step relative risk (RR) 1.01, 95% confidence interval (CI) 0.99-1.03) (detailed in table 3 below). In the other RCT (Aubier etc 1999 below), compliance was stated to be high in all patients (regardless of regime). These Seretide® RCT data were however complicated by Seretide® patients needing to use concurrent placebo inhaler, in order to maintain participant’s blinding (double dummy design). Hence they do not allow for differences in convenience,5 nor fully answer the question “Does the use of a single combination LABA/ICS inhaler improve compliance, beyond that gained using concurrent two separate inhalers?” That said, compliance rates in the control groups in the Seretide® clinical trials were high. This suggests that the need to use two inhalers at the same time is not necessarily a key cause of poor compliance with asthma preventive inhalers. The 2002 revised Global Strategy for Asthma Management and Prevention (GINA) guidelines state “Fixed combination inhalers are more convenient for patients, may increase compliance, ensure that the long-acting beta2-agonist is always accompanied by a glucocorticosteroid” 6 Note however that no evidence was given to support the claims of either greater convenience (although this does seem plausible) nor increased compliance.

British Thoracic Society; Scottish Intercollegiate Guidelines Network. British guideline on the management of asthma. Thorax 2003;58 Suppl 1:i1-94. http://www.sign.ac.uk/guidelines/published/support/guideline63/download.html

Braunstein GL, Trinquet G, Harper AE. Compliance with nedocromil sodium and a nedocromil sodium/salbutamol combination. Compliance Working Group. Eur Respir J. 1996 May;9(5):893-8.

Bosley CM, Parry DT, Cochrane GM. Patient compliance with inhaled medication: does combining beta-agonists with corticosteroids improve compliance? Eur Respir J. 1994 Mar;7(3):504-9.

van Grunsven PM, van Schayck CP, van Deuveren M, van Herwaarden CL, Akkermans RP, van Weel C. Compliance during long-term treatment with fluticasone propionate in subjects with early signs of asthma or chronic obstructive pulmonary disease (COPD): results of the Detection, Intervention, and Monitoring Program of COPD and Asthma (DIMCA) Study. J Asthma. 2000 May;37(3):225-34.

Eisen SA, Miller DK, Woodward RS, Spitznagel E, Przybeck TR. The effect of prescribed daily dose frequency on patient medication compliance. Arch Intern Med. 1990 Sep;150(9):1881-4.

National Heart, Lung, and Blood Institute , National Institutes Of Health. Global Strategy for Asthma Management and Prevention, revised 2002. (Scientific information and recommendations for asthma programs. NIH Publication No. 02- 3659). p108. http://www.ginasthma.com/workshop.pdf

P51-0-0 #73191

International guidelines re efficacy of combination LABA/ICS devices The recent New Zealand adult asthma guidelines7 (page 40) state “Dry powder devices that combine both LABA and ICS in one unit are now available. Such combination dry powder devices have similar, but not improved, clinical effectiveness as giving the same medication via separate devices [1+]” (citing one RCT8 ). “Although combination dry powder devices may appear more convenient, the fixed dosing of such devices makes titration of the ICS portion of the dose more difficult.” The BTS/SIGN guidelines (section 4.4.3, page i22) cite grade 1++ evidence in an evidence table 9 detailing four identified RCTs for Seretide® (Bateman etc 1998 10 , Chapman etc 199911 , Van den Berg 2000 12 , Aubier etc 199913 ). On this basis the guidelines state “There is no difference in efficacy in giving inhaled steroid and long-acting beta2 agonist in combination or in separate inhalers”. Note however that there was no attempt with the BTS/SIGN guidelines to pool the results of the four Seretide® RCTs. The 2002 revised GINA guidelines (p 108) state “Controlled studies have shown that delivering glucocorticosteroids and long-acting beta2- agonists together in a combination inhaler is as effective as giving each drug separately (Evidence B).” Note the evidence cited for the GINA statement is misreferenced as three montelukast publications.14 15 16 GINA also states “fixed combination inhalers … are usually less expensive than giving the two drugs separately.” In New Zealand, Seretide® is priced 2-3 times that of the total price of separate fluticasone and salmeterol inhalers.

Best Practice Evidence-Based Guideline: The Diagnosis and Treatment of Adult Asthma. New Zealand Guidelines Group, 2002. http://www.nzgg.org.nz/library/gl_complete/asthma/index.cfm

Bateman E, Britton M, Carrillo J, Almeida J, Wixon C. Salmeterol/fluticasone combination inhaler. A new, effective and well tolerated treatment for asthma. Clin Drug Invest 1998;16(3):193-201

Evidence table 4.22: Combined therapy of inhaled steroids and long acting B2 agonist http://www.sign.ac.uk/guidelines/published/support/guideline63/table4.22.html

10 11

Bateman et al 1998, op cit.

Van den Berg NJ, Ossip MS, Hederos CA, Anttila H, Ribeiro BL, Davies PI. Salmeterol/fluticasone propionate (50/100 microg) in combination in a Diskus inhaler (Seretide) is effective and safe in children with asthma. Pediatr Pulmonol 2000;30(2):97-105.

Chapman KR, Ringdal N, Backer V, Palmqvist M, Saarelainen S, Briggs M. Salmeterol and fluticasone propionate (50/250 microg) administered via combination Diskus inhaler: as effective as when given via separate Diskus inhalers. Can Respir J 1999;6(1):45-51.

Aubier M, Pieters WR, Schlosser NJ, Steinmetz KO. Salmeterol/fluticasone propionate (50/500 microg) in combination in a Diskus inhaler (Seretide) is effective and safe in the treatment of steroid-dependent asthma. Respir Med 1999;93(12):876-84.

Bleecker ER, Welch MJ, Weinstein SF, Kalberg C, Johnson M, Edwards L, et al. Low-dose inhaled fluticasone propionate versus oral zafirlukast in the treatment of persistent asthma. J Allergy Clin Immunol 2000;105:1123-9.

Laviolette M, Malmstrom K, Lu S, Chervinsky P, Pujet JC, Peszek I, et al. Montelukast added to inhaled beclomethasone in treatment of asthma. Montelukast/Beclomethasone Additivity Group. Am J Respir Crit Care Med 1999;160:1862-8.

Lofdahl CG, Reiss TF, Leff JA, Israel E, Noonan MJ, Finn AF, et al. Randomised, placebo controlled trial of effect of a leukotriene receptor antagonist, montelukast, on tapering inhaled corticosteroids in asthmatic patients. BMJ 1999;319:87-90. Cochrane Database Syst Rev 2000;2.

P51-0-0 #73191

Recently published meta-analysis of Seretide® (Nelson et al 2003) More recently, an industry co-written and funded meta-analysis has been published (Nelson et al 200317 ) pooling the results of the four Seretide® RCTs behind the above BTS/SIGN statement of no difference in efficacy. With pooling, this meta-analysis reported a significant advantage with Seretide® combination therapy over concurrent salmeterol and fluticasone therapy in morning peak expiratory flow rates (PEFR). “Odds of achieving a greater than 15 or greater than 30 L/min improvement with combination therapy were increased by approximately 40% compared with those after concurrent therapy, representing an additional 7% to 9% and 5% to 14% more patients, respectively, on combination therapy responding compared with those on concurrent therapy.” In the meta-analysis, >30 L/min improvements in PEFR with combination treatment occurred in 9% more patients (54% minus 45%), being a 19% improvement relative to concurrent controls (9%/45%). Likewise, >15 L/min improvements occurred in an additional 7% of combination treatment patients (73% versus 66%), a relative increase of 11% (7%/66%). These changes were reflected similarly in improvements to mean baseline morning PEFRs, where combination treatment caused on average an extra crude 5.8 L/min in PEFR over concurrent controls (38.2 vs. 32.8 L/min improvements, 5.4 L/min difference formally reported (Nelson et al 2003)). Limitations with the Seretide® meta -analysis However, the Seretide® meta-analysis had important limitations in its data and interpretation, materially affecting its key reported findings: 1. Selective reporting of key findings Statistically significant results in the Seretide® meta-analysis were confined to most but not all physiological variables, and there were no significant differences in more clinically-relevant outcomes (tables 1, 2 and 4, below): • There were statistically significant changes in self-reported mean morning and evening PEFR over weeks 1-12 (continuous variable), and achieving greater than 15 and 30 l/min improvements in morning PEFR (dichotomous variables); There was no significant change in mean change in objective clinic FEV1 at week 12 (continuous variable); There were no significant changes in the clinically relevant secondary outcomes of median % days and nights symptom- and reliever-free (median

• •

Nelson HS, Chapman KR, Pyke SD, Johnson M, Pritchard JN. Enhanced synergy between fluticasone propionate and salmeterol inhaled from a single inhaler versus separate inhalers. J Allergy Clin Immunol. 2003 Jul;112(1):29-36.

P51-0-0 #73191

% days symptom-free, median % nights symptom-free, median % days reliever-free, median % nights reliever-free). Table 1

Combination (Seretide) vs concurrent ICS/LABA - results of individual RCTs in Nelson et al 2003 (fixed effects model, Peto one-step method)

PEFR % measures of effect combination concurrent combinati concurrent difference OR(+/RR (+/95%CI) 95%CI) LABA/ICS separate on separate LABA/ICS LABA/ICS LABA/ICS RRI ARI* NNT (-ve = NNH)

physiological measures baseline PEFR >30 l/min morning PEFR increase clinical impact of >30 l/min increase >15 l/min morning PEFR increase clinical impact of >15 l/min increase

347.9

341.7 54.0% 8.7% 72.6% 4.4% 7.1 5.4 1.4% 6.5 4.7 65.4% 7.2% 45.2% 8.7% 1.42

(1.13-1.78)

1.19

(1.07-1.32)

19%

8.7%

1.41

(1.10-1.80)

1.11

(1.03-1.18)

11%

7.2%

adjusted mean change from baseline in mean morning PEFR over weeks 1-12, ITTA 39.9 32.8 WMD as published 38.1 32.8 % change in PEFR 11.0% 9.6% (adjusted mean change from baseline in mean morning PEFR over weeks 1-12, per protocol) 43.1 36.6 WMD as published 41.2 36.6 clinical measures days w/o sympts

median % days symptom free

1.22 1.16

22% 16%

2.1% 1.6%

1.18 1.13

18% 13%

1.9% 1.4%

Pharmac calculations WMD as published Pharmac calculations WMD as published Pharmac calculations WMD as published Pharmac calculations WMD as published

41.0% 39.5% 57.4% 52.6% 49.7% 47.2% 69.3% 65.3%

39.5% 39.5% 53.7% 53.7% 47.6% 47.6% 65.4% 65.4%

1.5% 0.0% 3.7% -1.2% 2.1% -0.4% 3.9% -0.1%

1.04

1.5%

1.00

1.07

0.0%

3.7%

nights w/o sympts

median % nights symptom free

0.98

1.05

-2%

-1.2%

2.1%

-87

days w/o rescue Rx

median % days reliever free

0.99

1.06

-1%

-0.4%

3.9%

-278

nights w/o rescue Rx

median % nights reliever free

1.00

-0.1%

-909

source: Pharmac analysis of: Nelson HS, Chapman KR, Pyke SD, Johnson M, Pritchard JN. Enhanced synergy between fluticasone propionate and salmeterol inhaled from a single inhaler versus separate inhalers. J Allergy Clin Immunol. 2003 Jul;112(1):29-36.

(a larger version of this table can be found at the end of this paper) • There were significantly higher rates of adverse events reported for Seretide®, otherwise no differences in withdrawal rates/withdrawals for adverse effects.

The withdrawal and adverse events were not measured in the Seretide® meta-analysis, requiring other analysis (table 2) (further details below). Table 2

Combination (Seretide) vs concurrent ICS/LABA - results of individual RCTs in Nelson et al 2003 (fixed effects model, Peto one-step method)

% measures of effect combinati concurrent difference OR(+/RR (+/95%CI) 95%CI) on separate LABA/ICS LABA/ICS RRI ARI* NNT (-ve = NNH)

compliance compliance

(Pharmac calculations)

93.0%

92.1%

0.9%

1.16

1.01

0.9%

107

(0.99-1.03)

withdrawals and adverse events total withdrawals

(Pharmac calculations)

14.6%

12.6%

2.0%

1.18

1.16

16%

2.0%

(0.84-1.57) 8.4% 7.1% 1.2% 1.19

1.17 17% 1.2% 81

withdrawals from adverse events

(Pharmac calculations)

(0.76-1.79) 59.0% 46.7% 12.4% 1.59

1.26 26% 12.4% 8

reported adverse events (+/- considered by investigators to be Rx-related)

(Pharmac calculations)

(1.09-1.44)

(a larger version of this table can be found at the end of this paper) Note that the Seretide® meta-analysis stated a prospective sole primary outcome measure, that of morning PEFR. However, PEFR was not stated to be the primary efficacy measure for one of the RCTs (Aubier etc 1999). Secondary measures in all the RCTs included FEV1 , and the clinically relevant outcomes.

P51-0-0 #73191

Withdrawals and adverse events were reported for all the individual RCTs. Withdrawals affect analysis of physiological and clinical outcomes, are themselves clinical outcomes, and should similarly have been reported. Withdrawals and adverse effects were all reported in meta-analyses of other asthma treatments, e.g. the Cochrane review of fluticasone.18 2. Reporting overstates the magnitude of physiological effects In addition, Nelson et al’s reporting of odds ratios (ORs) in the Seretide® meta-analysis overstates the magnitude of PEFR improvement; using relative risk (RR) gives much lower estimates of true relative effect: • ORs (as reported by Nelson et al) are the measure used when combining results of individual trials into a weighted summary measure able to demonstrate statistically significant effects. Odds and ORs however do not necessarily equate to risk and relative risk (RR). If the OR is interpreted as a RR it will always overstate any effect size – particularly when baseline risk is high. 19 Such potential for overstatement due to high baseline risk certainly occurs with the PEFR improvements in the Seretide® meta-analysis, with 45-73% prevalence rates for both treatments and controls. RRs can be derived from adjusted baseline event rates and pooled odds ratios, with associated confidence limits, according to the formula 20 :

RR = (1-aEc).(1-OR) 1 - [aEc.(1-OR)] where RR = relative risk; aEc =adjusted baseline event rate (i.e. control incidence rate, weighted according to inverse variance); OR = pooled odds ratio (weighted according to inverse variance) 1 -

•

Recalculating relative risks from the published ORs and calculating baseline risks gives an 11% likelihood of patients gain ing =15 L/min improvement in PEFR using Seretide® rather than concurrent ICS/LABA (RR 1.11 (95% confidence interval (CI) 1.03-1.18)). The likelihood for a =30 L/min PEFR improvement becomes 19% (RR 1.19 (1.07-1.32)). These results using RRs contrast sharply with the ORs of 1.40 and 1.42 as reported by Nelson et al.

Adams N, Bestall JM, Jones PW. Fluticasone versus beclomethasone or budesonide for chronic asthma (Cochrane Review). In: The Cochrane Library, Issue 3, 2003. Oxford: Update Software. CD002310

19 20

Davies HT, Crombie IK, Tavakoli M. When can odds ratios mislead? BMJ. 1998 Mar 28;316(7136):989-91.

algebraic transformation of formulae in Sackett D, Straus S, Richardson WS, Rosenberg W, Haynes B. Evidence-based medicine: how to practice and teach EBM, 2nd edition. Oxford: Churchill Livingstone, 2000. p136 Table 5.1 Formulae to convert odds ratios (ORs) and relative risks (RRs) to NNTs.

P51-0-0 #73191

3. No difference in clinically relevant outcomes The lack of statistically-significant clinically-relevant outcomes in the Seretide® metaanalysis reflects both (1) low variation relating to relatively small numbers of events and (2) little difference in clinically relevant outcomes: • A 15 L/min improvement in PEFR represented just a 4.4% increase in PEFR over baseline (where patients had PEFRs averaging 344 L/min at baseline). A 30 L/min represented an 8.7% increase over baseline. In other words, an extra 9% of patients had a 9% or more improvement in morning PEFR through using combination treatment (and an extra 7% had a 4% or more improvement). Likewise, the magnitude of average improved PEFR was in the region of just 1.4% ([5.4 L/min crude mean difference in PEFR between combination and concurrent] / [344 L/min mean baseline PEFR]). These seemingly low physiological changes may be reflected in the very low rates of reductions in days or nights without symptoms or reliever drugs (0.0 to 1.15% reductions, all statistically insignificant).

•

4. No analysis of withdrawals and adverse events, where these show significantly higher rates of reported adverse events with Seretide® The Seretide® meta-analysis could equally have analysed then reported (but did not) all withdrawals being prima facie 16% higher in combination than concurrent LABA/ICS users. Likewise, withdrawals due to adverse effects were prima facie higher at 17% with Seretide®. Although neither measure of withdrawal showed differences that were statistically significant (see table 3), reported adverse events in Seretide® patients were significantly higher. Seretide® users had prima facie one quarter (26%) more reported adverse events or events considered by investigators to be drug-related (289 in Seretide® patients (59%) vs. 263 in concurrent LABA/ICS users (47%), RR 1.26, 95% CI 1.09 – 1.44) (table 3). [To be consistent with the Nelson etc reporting of PEFR improvements in the Seretide® meta-analysis as odds ratios, the odds of reporting an adverse effect with Seretide® were increased by 59% prima facie compared with those after concurrent therapy (OR 1.59, 95% CI 1.16-2.17). The use of ORs is not being advocated here. Rather the prima facie 1.59 OR for reported adverse events places the use of 1.40-1.42 ORs for PEFRs in context, re-emphasising how these overstate actual increases in likelihood.]

P51-0-0 #73191

Table 3

Combination (Seretide) vs concurrent ICS/LABA - results of individual RCTs in Nelson et al 2003 (fixed effects model, Peto one-step method)

duration (weeks) no. patients combination concurrent LABA/ICS separate LABA/ICS no. % measures of effect RR (+/95%CI) RRI ARI* NNT (-ve = NNH) (variance wgts) combinati concurrent combinati concurrent difference OR on separate on separate LABA/ICS LABA/ICS LABA/ICS LABA/ICS

compliance compliance Bateman etc 1998 Chapman etc 1999 Aubier etc 1999 Van den Berg etc 2000 total or weighted average withdrawals and adverse events total withdrawals Bateman etc 1998 Chapman etc 1999 Aubier etc 1999 Van den Berg etc 2000 total or weighted average withdrawals from adverse events Bateman etc 1998 Chapman etc 1999 Aubier etc 1999 Van den Berg etc 2000 total or weighted average

12 28 12 16.7

121 180 125 426

121 191 132 444

110 173 116 399

108 181 123 412

91% 96% 93% 93.0%

89% 95% 93% 92.1%

2% 1% 0% 0.9%

1.02 1.01 1.00 1.01

(0.99-1.03)

2% 1% 0% 1%

1.7% 1.3% -0.4% 0.9%

61 74 -262 107

1.16

0.3965 0.2968 0.0000 0.3067 1.0000

12 28 28 12 22.8

121 180 167 125 593

121 191 171 132 615

18 20 31 5 74

17 16 28 5 66

14.9% 11.1% 18.6% 4.0% 14.6%

14.0% 8.4% 16.4% 3.8% 12.6%

0.8% 2.7% 2.2% 0.2% 2.0%

1.18

1.06 1.33 1.13 1.06 1.16

(0.84-1.57)

6% 33% 13% 6% 16%

0.8% 2.7% 2.2% 0.2% 2.0%

121 37 46 471 50

0.2482 0.2689 0.4033 0.0796 1.0000

121 180 167 125 593

121 191 171 132 615

11 12 16 2 41

9 9 16 2 36

9.1% 6.7% 9.6% 1.6% 8.4%

7.4% 4.7% 9.4% 1.5% 7.1%

1.7% 2.0% 0.2% 0.1% 1.2%

1.19

1.22 1.41 1.02 1.06 1.17

(0.76-1.79)

22% 41% 2% 6% 17%

1.7% 2.0% 0.2% 0.1% 1.2%

61 51 446 1,179 81

0.2585 0.2785 0.4076 0.0554 1.0000

withdrawals from asthma adverse events Bateman etc 1998 Chapman etc 1999 Aubier etc 1999 Van den Berg etc 2000 total or weighted average

121 180 167 125 593

121 191 171 132 615

4 5 0 1 10

3 5 0 1 9

3.3% 2.8% 0.0% 0.8% 2.7%

2.5% 2.6% 0.0% 0.8% 2.4%

0.8% 0.2% 0.0% 0.0% 0.4%

1.33 33% 1.06 6% #DIV/0! ###### 1.06 6% 1.16 1.15 15%

(-95% CI-+95% CI)

0.8% 0.2% 0.0% 0.0% 0.4%

121 625 2,357 276

0.3677 0.5251 0.0000 0.1072 1.0000

reported adverse events (+/- considered by investigators to be Rx-related) Bateman etc 1998 121 121 Chapman etc 1999 180 191 Aubier etc 1999 167 171 Van den Berg etc 2000 125 132 total or weighted average 593 615

88 160 28 13 289

69 164 24 6 263

72.7% 88.9% 16.8% 10.4% 59.0%

57.0% 85.9% 14.0% 4.5% 46.7%

15.7% 3.0% 2.7% 5.9% 12.4%

1.59

(1.16-2.17)

1.28 1.04 1.19 2.29 1.26

(1.09-1.44)

28% 15.7% 4% 3.0% 19% 2.7% 129% 5.9% 26% 12.4%

6 33 37 17 8

0.3499 0.2598 0.2788 0.1115 1.0000

source: Nelson HS, Chapman KR, Pyke SD, Johnson M, Pritchard JN. Enhanced synergy between fluticasone propionate and salmeterol inhaled from a single inhaler versus separate inhalers. J Allergy Clin Immunol. 2003 Jul;112(1):29-36.

(a larger version of this table can be found at the end of this paper) 5. Seretide®’s physiological and clinical improvements were comparatively low Contextually, Seretide®’s physiological and clinical improvements were low relative to those seen in other relevant ICS/LABA meta -analyses. Patterns of Seretide® clinical improvements relative to physiological improvement were consistent with the relative patterns seen for salmeterol LABA and fluticasone (when respectively compared with increasing the dose of ICS and compared with BDP or budesonide): • The above 5.4 L/min morning PEFR improvement with Seretide® was much less than the 22.4 L/min improvement occurring with salmeterol (when compared with increased doses of ICS – MIASMA meta-analysis 21 ) or even the 13.3 L/min improvement calculable for fluticasone (when compared with BDP or budesonide) (calculating from data in figures in Cochrane review22 ). In the salmeterol MIASMA meta-analysis, the above 22.4 L/min magnitude of PEFR improvement was associated with a 5 to 20% relative reduction in days or nights without symptoms or reliever drugs over 3 or 6 months. If we assume that baseline PEFRs were similar in both MIASMA and the Seretide® meta-analyses, then the 16% relative increase in PEFR in the

•

Shrewsbury S, Pyke S, Britton M. Meta-analysis of increased dose of inhaled steroid or addition of salmeterol in symptomatic asthma (MIASMA). BMJ. 2000 May 20;320(7246):1368-73.

Adams N, Bestall JM, Jones PW. Fluticasone versus beclomethasone or budesonide for chronic asthma (Cochrane Review). In: The Cochrane Library, Issue 3, 2003. Oxford: Update Software. CD002310

P51-0-0 #73191

Seretide® meta-analysis 23 would translate to a 69% relative improvement with salmeterol in MIASMA 24 . • Yet such a relatively large putative relative improvement for salmeterol PEFR (RRI 69%) translates to a much smaller clinical effect (salmeterol 11% overall reduction in days or nights without symptoms or reliever drugs at 3 months, range 5 to 17%). This discrepancy between salmeterol’s physiological and clinical effects is consistent with the contrasts seen with Seretide® between its statistically significant 5.4 L/min added PEFR improvement and its negligible clinical impacts. Incidentally, both Seretide®’s PEFR and clinical effects were appreciably lower than those of salmeterol in MIASMA. Likewise with fluticasone, the above 13.3 L/min improvement in PEFR translates to maybe 3% reduction in exacerbations and 6% improvement in symptom/reliever-free days (with neither overall clinical outcome statistically significant) (figure 2, table 4).

•

Figure 2

Fluticasone, salmeterol and Seretide pooled RCTs - physiological vs clinical improvements

22% 20% 18% % clinical improvement (treatment vs controls) 16% 14% 12% 10% 8% 6% 4% 2% 0% 0 -2% -4% PEFR treatment vs controls (L/min) 2 4 6 8 10 12 14 16 18 20 22 24

Seretide®: PEFR difference/baseline = 1.6% FP plle studies*: PEFR difference/baseline = 3.9% salmeterol: PEFR difference/baseline = 6.5%

exacerbations reduced symptoms improved* (trend, symptoms improved, intercept 0) (actual trend, symptoms improved)

*mean % days w/o symptoms, mean % nights w/o symptoms, mean % days w/o rescue Rx, mean % nights w/o rescue Rx

Adams N, Bestall JM, Jones PW Fluticasone versus beclomethasone or budesonide for chronic asthma (Cochrane Review). In: The Cochrane Library, Issue 3, 2003. Oxford: Update Software. Shrewsbury S, Pyke S, Britton M. Meta-analysis of increased dose of inhaled steroid or addition of salmeterol in symptomatic asthma (MIASMA). BMJ. 2000 May 20;320(7246):1368-73. Nelson HS, Chapman KR, Pyke SD, Johnson M, Pritchard JN. Enhanced synergy between fluticasone propionate and salmeterol inhaled from a single inhaler versus separate inhalers. J Allergy Clin Immunol. 2003 Jul;112(1):29-36.

fluticasone plle studies (Cochrane review 2003)*

*for clinical measures, FP data are incomplete (incompatibility issues) and based on 2-11 RCTs

Seretide® (Nelson etc 2003) (agents with no significant clinical improvements) salmeterol (MIASMA 2000)

(a larger version of this figure can be found at the end of this paper)

relative increase in PEFR with Seretide® = ([5.4 L/min difference between combined and concurrent ICSA/LABA use] / [32.8 L/min improvement with concurrent use]) = 16%

relative increase in PEFR with salmeterol = ([16% Seretide® PEFR RRI] x [22.4 L/min PEFR added PEFR improvement in MIASMA] / [5.4 L/min added improvement with Seretide®]) = 69%

P51-0-0 #73191

Table 4

Comparison between FP vs. BDP/bud RCTs (Cochrane review), Salmeterol vs. increased ICS RCTs (MIASMA), and Seretide® combination vs. concurrent FP/Salmeterol RCTs (Nelson 2003 meta-analysis) difference (ARR/ARI) relative effects (RRR/RRI) FP plle salmeterol Seretide® FP plle salmeterol Seretide® studies* studies* physiological measures baseline morning PEFR (L/min) 340.4 (n/avail) 344.8 mean difference PEFR tmt vs. cntrl (L/min) at 3 months 22.4 5.35 68% 16% at 6 months 27.7 not stated 13.3 41% mean difference FEV 1 tmt vs. cntrl (ml) at 3 months 100 40 at 6 months 80 not stated 120 extrapolated, using Seretide® 32.8 and 38.1 L/min improvements in cntrl and trmt groups (5.35 L/min difference, 16% RRI) clinical measures all exacerbations moderate/severe exacerbations withdrawals due to exacerbations mean % days w/o symptoms

-3.2% -1.2% at 3 months at 6 months not stated at 3 months at 6 months at 3 months at 6 months not stated, change at 3 months at 6 months

-2.7% -2.4% 12% 15% 0.4% 0.00%

-20% -22%

-9% 15% 0.00%

21% 15%

4.9% 5% 5% 17% 20% 6.9% 9% 8% -3.2% 6% 4.5% -2.7% 11% 6.7% -0.11% -1.15% -0.36%

mean % nights w/o symptoms mean % days w/o rescue Rx

19% 34% 108% 45%

-2.14% -0.76%

mean % nights w/o rescue Rx

-0.17%

all exacerbations unweighted average symptom improvement (at 3 months, or not stated) composite exacbns/symptoms

-0.41% -0.2%

-20% 61% 40.5%

-9% 15% 12.2%

-0.38% -0.2%

clinical vs. PEFR difference PEFR/baseline 3.9% 6.5% 1.6% clinical/PEFR 0.003 0.003 0.000 *for clinical measures, FP data are incomplete (incompatibility issues) and based on 2-11 RCTs

1.000

0.179

-0.012

Note that of the three ICS/LABA meta -analyses, only salmeterol has demonstrated significant clinical improvements.25 6. Quality of individual RCTs and meta -analysis The above analysis has not critically appraised the four individual contributing RCTs, nor systematically appraised the meta-analysis.26 The four RCTs were each stated to be double -dummy double -blind randomised parallel group controlled trials. Patients were children and adults with similar levels of asthma severity. The RCTs all measured both physiological and clinically relevant outcomes and included withdrawals and adverse events. However, there was some ambiguity with reporting around the quality of the individual RCTs:

Although clinically relevant improvements were not statistically significant with fluticasone, there were problems with study incomparability, hence possible type 1 error i.e. falsely ascribing no effect when a true effect exists.

formal appraisal tools from EPIQ http://www.health.auckland.ac.nz/comhealth/epiq/epiq.htm for individual RCTs are at http://www.health.auckland.ac.nz/comhealth/ElectronicGateInterV12.doc, for meta-analyses at http://www.health.auckland.ac.nz/comhealth/epiq/GateSRChklstV3.doc

P51-0-0 #73191

• • •

None of the RCTs clearly described blinding; Three of the four RCTs did not describe the process of randomisation nor whether (and how) allocation was concealed; The randomisation process was described in one RCT (Bateman etc 1998), but concealment was not explicit and must be inferred from computer block randomisation; All patients were accounted for, but whether analysis of effectiveness was by intention-to-treat or on-treatment was unclear for one RCT (Bateman 1998), with strong inference that it was on-treatment analysis. This would mean missing out patients withdrawing because of asthma exacerbations, hence unable to contribute to data re PEFR changes; All were multicentre trials, but no details were given on oversight/controls across sites.

•

The Nelson etc 2003 Seretide® meta-analysis did not describe how its component RCTs were identified (formal search strategies etc.). That said, the separate search presumably undertaken for the BTS/SIGN guidelines’ evidence table and searching PubMed revealed/reveals no other relevant RCTs. It is assumed the Nelson meta-analysis, being authored by GSK employees, would have systematically and comprehensively identified all Seretide® RCTs known to the manufacturer, but this was not made explicit in the publication. Note again the Seretide® meta-analysis chose not to pool then report on withdrawal and adverse events rates.

P51-0-0 #73191

Message of international guidelines remains unchanged. Hence in view of the above limitations with the Nelson et al Seretide® meta-analysis, the advice from the BTS/SIGN, GINA and New Zealand asthma guidelines still applies, viz. that there is no difference in clinical efficacy between combination and concurrent (separate devices) LABA/ICS. Further detail of the results of the four component Seretide® RCTs and overall pooled effects are in table 5. Table 5

Combination (Seretide) vs concurrent ICS/LABA - results of individual RCTs in Nelson et al 2003 (fixed effects model, Peto one-step method)

physiological measures duration (weeks) no. patients combination concurrent LABA/ICS separate LABA/ICS baseline PEFR no. % measures of effect RR (+/95%CI)

(variance wgts) RRI ARI* NNT

combinati concurrent combinati concurrent combinati concurrent difference OR (+/95%CI) on separate on separate on separate LABA/ICS LABA/ICS LABA/ICS LABA/ICS LABA/ICS LABA/ICS

>30 l/min morning PEFR increase Bateman etc 1998 Chapman etc 1999 Aubier etc 1999 Van den Berg etc 2000 total or weighted average clinical impact of >30 l/min increase >15 l/min morning PEFR increase Bateman etc 1998 Chapman etc 1999 Aubier etc 1999 Van den Berg etc 2000 total or weighted average clinical impact of >15 l/min increase

12 12 12 12 12

121 180 167 125 593

121 191 171 132 615

368 398 359 241 347.9

365 391 345 243 341.7 8.7%

75 100 82 63 320

66 89 75 48 278

62.0% 55.6% 49.1% 50.4% 54.0%

54.5% 46.6% 43.9% 36.4% 45.2%

7.4% 9.0% 5.2% 14.0% 8.7%

1.42

(1.13-1.78)

1.14 1.19 1.12 1.39 1.19

(1.07-1.32)

14% 7.4% 19% 9.0% 12% 5.2% 39% 14.0% 19% 8.7%

13 11 19 7 11

0.1973 0.3101 0.2815 0.2112 1.0000

121 180 167 125 593

121 368 191 398 171 359 132 241 615 347.9498

365 391 345 243 341.6693 4.4%

96 135 115 87 433

87 130 106 81 404

79.3% 75.0% 68.9% 69.6% 72.6%

71.9% 68.1% 62.0% 61.4% 65.4%

7.4% 6.9% 6.9% 8.2% 7.2%

1.41

(1.10-1.80)

1.10 1.10 1.11 1.13 1.11

(1.03-1.18)

10% 10% 11% 13% 11%

7.4% 6.9% 6.9% 8.2% 7.2%

13 14 15 12 14

0.1752 0.2966 0.3000 0.2282 1.0000

adjusted mean change from baseline in mean morning PEFR over weeks 1-12, ITTA Bateman etc 1998 121 121 368 365 Chapman etc 1999 180 191 398 391 Aubier etc 1999 167 171 359 345 Van den Berg etc 2000 125 132 241 243 total or weighted average 593 615 347.9 341.7 WMD as published 0 0 347.9 341.7 % change in PEFR

42 43 35 33 39.9 38.1 11.0%

33 36 33 28 32.8 32.8 9.6%

9.0 7.0 2.0 5.0 7.1 5.4 1.4%

1.27 1.19 1.06 1.18 1.22 1.16

27% 19% 6% 18% 22% 16%

2.5% 1.8% 0.6% 2.1% 2.1% 1.6%

0.2413 0.2892 0.2529 0.2167 1.0000

(adjusted mean change from baseline in mean morning PEFR over weeks 1-12, per protocol) Bateman etc 1998 121 121 368 365 51 42 Chapman etc 1999 180 191 398 391 43 36 Aubier etc 1999 167 171 359 345 40 36 Van den Berg etc 2000 125 132 241 243 34 33 total or weighted average 593 615 347.9498 341.6693 43.1 36.6 WMD as published 0 0 347.9 341.7 41.2 36.6 *ARI for x l/min morning PEFR increase is a measure of population impact, = (% treatment group patients achieving xx increase) minus (% control group patients achieving xx increase). ARI for mean changes in baseline morning PEFR is a measure of average individual clinical impact (improvement in lung function), = (mean improvement) / (mean baseline)

clinical measures duration (weeks) no. patients combination concurrent LABA/ICS separate LABA/ICS no. %

9.0 7.0 4.0 1.0 6.5 4.7

1.21 1.19 1.11 1.03 1.18 1.13

21% 19% 11% 3% 18% 13%

2.5% 1.8% 1.1% 0.4% 1.9% 1.4%

0.1997 0.3125 0.2753 0.2125 1.0000

measures of effect RR RRI ARI* NNT (-ve = NNH)

(variance wgts)

combinati concurrent combinati concurrent difference on separate on separate LABA/ICS LABA/ICS LABA/ICS LABA/ICS

days w/o sympts Bateman etc 1998 Chapman etc 1999 Aubier etc 1999 Van den Berg etc 2000 total or weighted average WMD as published nights w/o sympts Bateman etc 1998 Chapman etc 1999 Aubier etc 1999 Van den Berg etc 2000 total or weighted average WMD as published days w/o rescue Rx Bateman etc 1998 Chapman etc 1999 Aubier etc 1999 Van den Berg etc 2000 total or weighted average WMD as published nights w/o rescue Rx Bateman etc 1998 Chapman etc 1999 Aubier etc 1999 Van den Berg etc 2000 total or weighted average WMD as published

121 180 167 125 593 median % days symptom free

121 191 171 132 615

48 39 63 75 225

52 29 65 79 225

39.7% 21.7% 38.0% 60.0% 41.0% 39.5%

43.0% 15.2% 38.0% 60.0% 39.5% 39.5%

-3.3% 6.5% 0.0% 0.0% 1.5% 0.0%

0.92 1.43 1.00 1.00 1.04 1.00

-8% 43% 0% 0% 4% 0%

-3.3% 6.5% 0.0% 0.0% 1.5% 0.0%

-30 15 65 -

0.2299 0.2171 0.3116 0.2414 1.0000

121 180 167 125 593 median % nights symptom free

121 191 171 132 615

58 84 95 114 351

69 80 94 99 342

47.9% 46.7% 57.0% 91.0% 57.4% 52.6%

57.0% 41.9% 55.0% 75.0% 53.7% 53.7%

-9.1% 4.8% 2.0% 16.0% 3.7% -1.2%

0.84 1.11 1.04 1.21 1.07 0.98

-16% -9.1% 11% 4.8% 4% 2.0% 21% 16.0% 7% 3.7% -2% -1.2%

-11 21 50 6 27 -87

0.2224 0.3364 0.3065 0.1348 1.0000

121 180 167 125 593 median % days reliever free

121 191 171 132 615

75 73 63 91 303

68 64 65 104 301

62.0% 40.6% 38.0% 73.0% 49.7% 47.2%

56.2% 33.5% 38.0% 79.0% 47.6% 47.6%

5.8% 7.0% 0.0% -6.0% 2.1% -0.4%

1.10 1.21 1.00 0.92 1.05 0.99

10% 21% 0% -8% 5% -1%

5.8% 7.0% 0.0% -6.0% 2.1% -0.4%

17 14 -17 47 -278

0.2159 0.3182 0.2935 0.1724 1.0000

121 180 167 468 median % nights reliever free

121 191 171 483

82 126 117 325

87 118 111 316

67.8% 70.0% 70.0% 69.3% 65.3%

71.9% 61.8% 65.0% 65.4% 65.4%

-4.1% 8.2% 5.0% 3.9% -0.1%

0.94 1.13 1.08 1.06 1.00

-6% 13% 8% 6% 0%

-4.1% 8.2% 5.0% 3.9% -0.1%

-24 12 20 26 -909

0.2446 0.3999 0.3555 1.0000

(a larger version of this table can be found at the end of this paper

P51-0-0 #73191

Table 1

Combination (Seretide) vs concurrent ICS/LABA - results of individual RCTs in Nelson et al 2003 (fixed effects model, Peto one-step method)

PEFR % measures of effect combination concurrent combinati concurrent difference OR(+/RR(+/95%CI) 95%CI) LABA/ICS separate on separate LABA/ICS LABA/ICS LABA/ICS RRI ARI* NNT (-ve = NNH)

physiological measures baseline PEFR >30 l/min morning PEFR increase clinical impact of >30 l/min increase >15 l/min morning PEFR increase clinical impact of >15 l/min increase

347.9

341.7 54.0% 8.7% 72.6% 4.4% 7.1 5.4 1.4% 6.5 4.7 65.4% 7.2% 45.2% 8.7% 1.42

(1.13-1.78)

1.19

(1.07-1.32)

19%

8.7%

1.41

(1.10-1.80)

1.11

(1.03-1.18)

11%

7.2%

median % days symptom free

1.22 1.16

22% 16%

2.1% 1.6%

1.18 1.13

18% 13%

1.9% 1.4%

Pharmac calculations WMD as published Pharmac calculations WMD as published Pharmac calculations WMD as published Pharmac calculations WMD as published

41.0% 39.5% 57.4% 52.6% 49.7% 47.2% 69.3% 65.3%

39.5% 39.5% 53.7% 53.7% 47.6% 47.6% 65.4% 65.4%

1.5% 0.0% 3.7% -1.2% 2.1% -0.4% 3.9% -0.1%

1.04

1.5%

1.00

1.07

0.0%

3.7%

nights w/o sympts

median % nights symptom free

0.98

1.05

-2%

-1.2%

2.1%

-87

days w/o rescue Rx

median % days reliever free

0.99

1.06

-1%

-0.4%

3.9%

-278

nights w/o rescue Rx

median % nights reliever free

1.00

-0.1%

-909

P51-0-0 #73191 PHARMAC 11.08.03

Table 2

Combination (Seretide) vs concurrent ICS/LABA - results of individual RCTs in Nelson et al 2003 (fixed effects model, Peto one-step method)

% measures of effect combinati concurrent difference OR(+/RR(+/95%CI) 95%CI) on separate LABA/ICS LABA/ICS RRI ARI* NNT (-ve = NNH)

compliance compliance

(Pharmac calculations)

93.0%

92.1%

0.9%

1.16

1.01

0.9%

107

(0.99-1.03)

withdrawals and adverse events total withdrawals

(Pharmac calculations)

14.6%

12.6%

2.0%

1.18

1.16

16%

2.0%

(0.84-1.57) 8.4% 7.1% 1.2% 1.19

1.17 17% 1.2% 81

withdrawals from adverse events

(Pharmac calculations)

(0.76-1.79) 59.0% 46.7% 12.4% 1.59

1.26 26% 12.4% 8

reported adverse events (+/- considered by investigators to be Rx-related)

(Pharmac calculations)

(1.09-1.44)

P51-0-0 #73191

Table 3

Combination (Seretide) vs concurrent ICS/LABA - results of individual RCTs in Nelson et al 2003 (fixed effects model, Peto one-step method)

duration (weeks) no. patients combination concurrent LABA/ICS separate LABA/ICS no. % measures of effect RR (+/95%CI)

(variance wgts) NNT (-ve = NNH)

combinati concurrent combinati concurrent difference OR on separate on separate LABA/ICS LABA/ICS LABA/ICS LABA/ICS

RRI

ARI*

12 28 12 16.7

121 180 125 426

121 191 132 444

110 173 116 399

108 181 123 412

91% 96% 93% 93.0%

89% 95% 93% 92.1%

2% 1% 0% 0.9%

1.02 1.01 1.00 1.01

(0.99-1.03)

2% 1%

1.7% 1.3%

61 74 -262 107

1.16

0% -0.4% 1% 0.9%

0.3965 0.2968 0.0000 0.3067 1.0000

12 28 28 12 22.8

121 180 167 125 593

121 191 171 132 615

18 20 31 5 74

17 16 28 5 66

14.9% 11.1% 18.6% 4.0% 14.6%

14.0% 8.4% 16.4% 3.8% 12.6%

0.8% 2.7% 2.2% 0.2% 2.0%

1.18

1.06 1.33 1.13 1.06 1.16

(0.84-1.57)

6% 33% 13% 6% 16%

0.8% 2.7% 2.2% 0.2% 2.0%

121 37 46 471 50

0.2482 0.2689 0.4033 0.0796 1.0000

121 180 167 125 593

121 191 171 132 615

11 12 16 2 41

9 9 16 2 36

9.1% 6.7% 9.6% 1.6% 8.4%

7.4% 4.7% 9.4% 1.5% 7.1%

1.7% 2.0% 0.2% 0.1% 1.2%

1.19

1.22 1.41 1.02 1.06 1.17

(0.76-1.79)

22% 41% 2% 6% 17%

1.7% 2.0% 0.2% 0.1% 1.2%

61 51 446 1,179 81

0.2585 0.2785 0.4076 0.0554 1.0000

withdrawals from asthma adverse events Bateman etc 1998 Chapman etc 1999 Aubier etc 1999 Van den Berg etc 2000 total or weighted average

121 180 167 125 593

121 191 171 132 615

4 5 0 1 10

3 5 0 1 9

3.3% 2.8% 0.0% 0.8% 2.7%

2.5% 2.6% 0.0% 0.8% 2.4%

0.8% 0.2% 0.0% 0.0% 0.4%

1.16

1.33 33% 1.06 6% #DIV/0! ###### 1.06 6% 1.15 15%

0.8% 0.2% 0.0% 0.0% 0.4%

121 625 2,357 276

0.3677 0.5251 0.0000 0.1072 1.0000

(-95% CI-+95% CI)

88 160 28 13 289

69 164 24 6 263

72.7% 88.9% 16.8% 10.4% 59.0%

57.0% 85.9% 14.0% 4.5% 46.7%

15.7% 3.0% 2.7% 5.9% 12.4%

1.59

(1.16-2.17)

1.28 1.04 1.19 2.29 1.26

(1.09-1.44)

28% 15.7% 4% 3.0% 19% 2.7% 129% 5.9% 26% 12.4%

6 33 37 17 8

0.3499 0.2598 0.2788 0.1115 1.0000

P51-0-0 #73191

Figure 2

Fluticasone, salmeterol and Seretide pooled RCTs - physiological vs clinical improvements

22% 20% 18% % clinical improvement (treatment vs controls) 16% 14% 12% 10% 8% 6% 4% 2% 0% 0 -2% -4% PEFR treatment vs controls (L/min) 2 4 6 8 10 12 14 16 18 20 22 24

Seretide®: PEFR difference/baseline = 1.6% FP plle studies*: PEFR difference/baseline = 3.9% salmeterol: PEFR difference/baseline = 6.5% *for clinical measures, FP data are incomplete (incompatibility issues) and based on 2-11 RCTs

exacerbations reduced symptoms improved* (trend, symptoms improved, intercept 0) (actual trend, symptoms improved)

*mean % days w/o symptoms, mean % nights w/o symptoms, mean % days w/o rescue Rx, mean % nights w/o rescue Rx

fluticasone plle studies (Cochrane review 2003)*

Seretide® (Nelson etc 2003) (agents with no significant clinical improvements) salmeterol (MIASMA 2000)

P51-0-0 #73191

Table 5

Combination (Seretide) vs concurrent ICS/LABA - results of individual RCTs in Nelson et al 2003 (fixed effects model, Peto one-step method)

physiological measures duration (weeks) no. patients combination concurrent LABA/ICS separate LABA/ICS baseline PEFR no. % measures of effect RR(+/95%CI)

(variance wgts) RRI ARI* NNT

combinati concurrent combinati concurrent combinati concurrent difference OR(+/95%CI) on separate on separate on separate LABA/ICS LABA/ICS LABA/ICS LABA/ICS LABA/ICS LABA/ICS

12 12 12 12 12

121 180 167 125 593

121 191 171 132 615

368 398 359 241 347.9

365 391 345 243 341.7 8.7%

75 100 82 63 320

66 89 75 48 278

62.0% 55.6% 49.1% 50.4% 54.0%

54.5% 46.6% 43.9% 36.4% 45.2%

7.4% 9.0% 5.2% 14.0% 8.7%

1.42

(1.13-1.78)

1.14 1.19 1.12 1.39 1.19

(1.07-1.32)

14% 7.4% 19% 9.0% 12% 5.2% 39% 14.0% 19% 8.7%

13 11 19 7 11

0.1973 0.3101 0.2815 0.2112 1.0000

121 180 167 125 593

121 368 365 191 398 391 171 359 345 132 241 243 615 347.9498 341.6693 4.4%

96 135 115 87 433

87 130 106 81 404

79.3% 75.0% 68.9% 69.6% 72.6%

71.9% 68.1% 62.0% 61.4% 65.4%

7.4% 6.9% 6.9% 8.2% 7.2%

1.41

(1.10-1.80)

1.10 1.10 1.11 1.13 1.11

(1.03-1.18)

10% 10% 11% 13% 11%

7.4% 6.9% 6.9% 8.2% 7.2%

13 14 15 12 14

0.1752 0.2966 0.3000 0.2282 1.0000

42 43 35 33 39.9 38.1 11.0%

33 36 33 28 32.8 32.8 9.6%

9.0 7.0 2.0 5.0 7.1 5.4 1.4%

1.27 1.19 1.06 1.18 1.22 1.16

27% 19% 6% 18% 22% 16%

2.5% 1.8% 0.6% 2.1% 2.1% 1.6%

0.2413 0.2892 0.2529 0.2167 1.0000

9.0 7.0 4.0 1.0 6.5 4.7

1.21 1.19 1.11 1.03 1.18 1.13

21% 19% 11% 3% 18% 13%

2.5% 1.8% 1.1% 0.4% 1.9% 1.4%

0.1997 0.3125 0.2753 0.2125 1.0000

P51-0-0 #73191

Table 5 (cont.)

clinical measures duration (weeks) no. patients combination concurrent LABA/ICS separate LABA/ICS no. % measures of effect RR RRI ARI* NNT (-ve = NNH) (variance wgts)

combinati concurrent combinati concurrent difference on separate on separate LABA/ICS LABA/ICS LABA/ICS LABA/ICS

121 180 167 125 593 median % days symptom free

121 191 171 132 615

48 39 63 75 225

52 29 65 79 225

39.7% 21.7% 38.0% 60.0% 41.0% 39.5%

43.0% 15.2% 38.0% 60.0% 39.5% 39.5%

-3.3% 6.5% 0.0% 0.0% 1.5% 0.0%

0.92 1.43 1.00 1.00 1.04 1.00

-8% -3.3% 43% 6.5% 0% 0.0% 0% 0.0% 4% 1.5% 0% 0.0%

-30 15 65 -

0.2299 0.2171 0.3116 0.2414 1.0000

121 180 167 125 593 median % nights symptom free

121 191 171 132 615

58 84 95 114 351

69 80 94 99 342

47.9% 46.7% 57.0% 91.0% 57.4% 52.6%

57.0% 41.9% 55.0% 75.0% 53.7% 53.7%

-9.1% 4.8% 2.0% 16.0% 3.7% -1.2%

0.84 1.11 1.04 1.21 1.07 0.98

-16% -9.1% 11% 4.8% 4% 2.0% 21% 16.0% 7% 3.7% -2% -1.2%

-11 21 50 6 27 -87

0.2224 0.3364 0.3065 0.1348 1.0000

121 180 167 125 593 median % days reliever free

121 191 171 132 615

75 73 63 91 303

68 64 65 104 301

62.0% 40.6% 38.0% 73.0% 49.7% 47.2%

56.2% 33.5% 38.0% 79.0% 47.6% 47.6%

5.8% 7.0% 0.0% -6.0% 2.1% -0.4%

1.10 1.21 1.00 0.92 1.05 0.99

10% 5.8% 21% 7.0% 0% 0.0% -8% -6.0% 5% 2.1% -1% -0.4%

17 14 -17 47 -278

0.2159 0.3182 0.2935 0.1724 1.0000

121 180 167 468 median % nights reliever free

121 191 171 483

82 126 117 325

87 118 111 316

67.8% 70.0% 70.0% 69.3% 65.3%

71.9% 61.8% 65.0% 65.4% 65.4%

-4.1% 8.2% 5.0% 3.9% -0.1%

0.94 1.13 1.08 1.06 1.00

-6% -4.1% 13% 8.2% 8% 5.0% 6% 3.9% 0% -0.1%

-24 12 20 26 -909

0.2446 0.3999 0.3555 1.0000

P51-0-0 #73191

Metadata

Title

Effectiveness of combined ICS/LABAs

Abstract

Effectiveness of combined ICS/LABAs delivery devices versus concurrent ICS and LABA via separate inhalers.

Page 1

Note

This text has been extracted from the source PDF document.

Also available as plain text.

Please contact webmaster to discuss alternative format options.