Exhibit 99.3

Bleecker ER(1), Bateman ED(2), Busse W(3), Lötvall J(4), Woodcock A(5), Tomkins S(6), House KW(7), Jacques L(6), Haumann B(6)

(1)Translational Sciences, Wake Forest University Health Sciences Winston-Salem, USA; (2)Department of Medicine, University of Cape Town, Cape Town, South Africa (3)Department of Medicine, University of Wisconsin, Madison, USA; (4)Krefting Research Centre, Gothenburg, Sweden; (5)School of Translational Medicine, University Hospital of Manchester, Manchester, UK (6)Respiratory Medicine Development Centre, GlaxoSmithKline, Uxbridge, UK; (7)Respiratory Medicine Development Centre, GlaxoSmithKline, North Carolina, USA

ABSTRACT

Introduction: FF (GW685698X) is a novel ICS still active at 24h and under development as a once-daily treatment in combination with the long-acting beta2 agonist (LABA) vilanterol trifenatate (GW642444M) for asthma and chronic obstructive pulmonary disease (COPD).

Objectives: To compare the relative efficacy and safety of four doses of FF (dry powder), administered using a novel single-step activation inhaler in patients >12 years with moderate asthma uncontrolled on low doses of ICS (fluticasone propionate [FP] 200mcg/day or equivalent). FP was used as an active control.

Methods: This randomised, double-blind, double-dummy, placebo-controlled, parallel group study, randomised 622 patients to one of six treatments: FF (100, 200, 300 or 400mcg) once daily, FP 250mcg twice daily or placebo for 8 weeks. The primary endpoint was change in 24h trough (pre-dose) forced expiratory volume in 1 second (FEV1) from baseline at Week 8.

Results: Primary: at Week 8 relative to placebo, all doses of FF and FP demonstrated significantly greater increases from baseline in trough FEV1 (p<0.001) and >200mL increase. There was no evidence of a dose response between FF doses. Secondary: withdrawals due to lack of efficacy were low, and peak expiratory flow (PEF) results supported the efficacy of FF 100–400mcg and FP. Oral candidiasis was low (0–4%) and the range of 24h urinary cortisol excretion ratios (Week 8/baseline) was similar across treatments, including placebo.

Conclusion: This dose-ranging study supports the use of FF as once-daily treatment for patients with asthma uncontrolled on low ICS doses.

INTRODUCTION

·                  ICS are the most effective anti-inflammatory medications currently available for long-term control of all severities of persistent asthma.(1)–(3)

·                  Most current inhaled steroids are more effective when taken twice daily rather than once daily.(3)

·                  Patients not controlled on low-dose ICS may require combination therapy with a LABA or twice-daily therapy with a higher dose of ICS.(1)–(3)

·                  FF is a novel ICS, still active 24h after dosing, which is under development for use as the ICS component of a new once-daily ICS/LABA combination for asthma and COPD.

OBJECTIVE

·                  To evaluate the dose response, efficacy and safety of four dose regimens of FF administered once daily in the evening in patients aged >12 years with persistent asthma uncontrolled on low-dose ICS therapy, in order to identify the appropriate dose of FF for further investigation.

METHODS

·                  Phase IIb, multicentre, randomised, double-blind, double-dummy, parallel-group, placebo-controlled study, involving nine clinic visits (Figure 1).

·                  Eligible patients had an FEV1 of 40–90% (for PM visit) or 40–85% (for AM visit) of the predicted normal value at visit 1 and demonstrated a >12% and >200mL reversibility of FEV1 within ~30min following four inhalations of albuterol/salbutermol inhalation aerosol. Patients had been using an inhaled ICS for >8 weeks prior to visit 1 and were able to replace their current short-acting beta2 agonist with albuterol/salbutamol at visit 1 for use during the study.

·                  FF once daily was administered in the evening via a novel dry powder inhaler given at doses of 100, 200, 300 and 400mcg versus placebo.

·                  FP 250mcg twice daily was administered via DISKUS™/ACCUHALER™ inhaler and was included for assay sensitivity and compared against placebo for the relative magnitude of response with FF doses investigated.

Efficacy and safety measures methods

·                  Primary efficacy endpoint: mean change from baseline in trough (PM pre-dose and pre-rescue bronchodilator) FEV1 at the end of the 8-week treatment period.

Figure 1. Study design.

OD = once daily; BD = twice daily

·                  Secondary efficacy endpoints: mean change from baseline in daily trough (pre-dose and pre-rescue bronchodilator) PM PEF and daily AM PEF averaged over the treatment period; mean change from baseline in the percentage of symptom-free 24h periods and rescue-free 24h periods during the treatment period; and number of withdrawals due to lack of efficacy during the treatment period.

·                  Safety assessments included: incidence of adverse events (AEs); incidence of oral candidiasis at clinic visits 1 and 3–8; urinary cortisol excretion; and laboratory tests and vital signs, before and at the end of the 8-week treatment period.

RESULTS

Efficacy study design

·                  Patient characteristics were comparable across the groups (Table 1).

Table 1. Demographic and baseline clinical characteristics (ITT population).

ITT = intent-to-treat

·                  All dosages of FF demonstrated significant (p<0.001) mean increases in FEV1 from baseline compared with placebo (Figure 2).

·                  There was no evidence of a dose response for the four FF groups.

Figure 2. Adjusted treatment differences of change from baseline in trough FEV1 (LOCF) at Week 8 (ITT population).

Note: analysis performed using an analysis of covariance (ANCOVA) with covariates of baseline, country, sex, age and treatment

·                  The results of a test of linear trend in dose response, excluding placebo group showed no significant relationship between response and dose of FF (p>0.05; 95% CI: –0.060 to 0.616). The estimate of the slope was 0.278mL/mcg. The results of the test for linear trend in dose response, including placebo were significant (p<0.001; 95% CI: 0.391 to 0.900). The estimate of the slope was 0.646mL/mcg.

·                  FF once daily PM and FP twice daily produced significantly different (improved) changes relative to placebo in daily trough (pre-dose and pre-rescue bronchodilator) PM PEF and daily AM PEF (p=0.018 to p<0.001; Figure 3).

Figure 3. PEF over Weeks 1–8
a) PM and b) AM.

·                  The change in percentage of symptom-free 24h periods were significantly different from placebo for FF 400mcg once daily PM (p=0.010) and FP 250mcg twice daily (p=0.002), but not for FF 100, 200 and 300mcg once daily PM (p=0.311, p=0.592 and p=0.097, respectively; Figure 4a). The change in percentage of rescue-free 24h periods were significantly different from placebo for FF 100, 300 and 400mcg once daily PM, and FP 250mcg twice daily (p=0.030, p=0.031, p=0.045 and p<0.001, respectively) but not for FF 200mcg once daily (p=0.06; Figure 4b).

·                  Significantly higher withdrawals due to lack of efficacy were observed with placebo (33%) compared with the active treatment groups (p<0.001 for all FF doses and p=0.002 for FP; Figure 5).

Figure 4. Symptom-free and rescue-free 24h periods over Weeks 1–8.

Figure 5. Treatment discontinuation due to lack of efficacy (cumulative incidence).

Note: patients are represented from their date of randomisation to their date of withdrawal due to lack of efficacy

Safety

·                  Overall, FF was well tolerated, with no apparent dose-related increase in the frequency of the most commonly reported AEs across the four doses of FF (Table 2).

·                  Seven patients withdrew due to AEs, including one serious AE (asthma exacerbation); three of these events were considered treatment-related (pharyngitis and throat infection [both in the FF 100mcg group]; bilateral thigh pain [FF 400mcg group]).

·                  The urinary cortisol excretion ratios were not significantly different between placebo and any dose groups of FF or FP 250mcg twice daily (Figure 6).

Table 2. Most common on-treatment AEs (>3% incidence in any treatment group; ITT population).

*Two additional patients, 1 (<1%) in the FF 300mcg OD arm and 1 (1%) in the FP 250mcg BD arm were diagnosed as having oral candidiasis, coded as ‘oropharyngeal candidiasis’

·                  No treatment-related clinically important changes were apparent for the mean and median laboratory values noted at Week 1 and Week 8 for haematology and clinical chemistry analytes, or vital signs.

Figure 6. Adjusted 24h urinary cortisol excretion ratios.*

Note: analysis performed using ANCOVA with covariates of baseline, country, sex, age and treatment and the log of baseline values;

*The urinary cortisol population consisted of 417 (68%) patients in the ITT population who had urine samples with no factors that would confound the analysis of urinary cortisol

CONCLUSIONS

·                  These results support FF as an effective and well-tolerated ICS at doses ranging from 100 to 400mcg administered once daily in the evening, in patients with asthma uncontrolled on low doses of ICS.

·                  On the basis of the efficacy and safety data, FF 100mcg and 200mcg are likely to be the most effective doses for this population of asthma patients.

REFERENCES

(1)          Global Initiative for Asthma (GINA). Global Strategy for Asthma Management and Prevention 2007. Available at: www.ginasthma.org.

(2)          National Institutes of Health (NIH). Expert Panel Report 3: Guidelines for the Diagnosis and Management of Asthma. Full Report 2007. NIH Publication No. 07-4051. Available at: http://www.nhlbi.nih.gov.

(3)          British Thoracic Society (BTS). British Guideline on the Management of Asthma. May 2008. Available at: www.brit-thoracic.org.uk.

ACKNOWLEDGEMENTS

·                  This study was sponsored by GlaxoSmithKline (ClinicalTrials.gov: NCT00603278; protocol number FFA109685).

·                  Editorial support (in the form of writing assistance, assembling tables and figures, collating author comments, grammatical editing and referencing) was provided by Anna Ireland at Gardiner-Caldwell Communications and was funded by GlaxoSmithKline.