Exhibit 99.6

Hanania NA(1), Feldman G(2), Zachgo W(3), Shim JJ(4), Crim C(5), Sanford L(6), Lettis S(6), Barnhart F(5), Haumann B(6)

(1)Baylor College of Medicine, Houston, USA; (2)South Carolina Pharmaceutical Research, Spartanburg, USA; (3)Forschungsinstitut Hohegeest, Geesthacht, Germany (4)Guro Hospital, Seoul, Korea; (5)GlaxoSmithKline, Research Triangle Park, North Carolina, USA; (6)GlaxoSmithKline, Uxbridge, UK

ABSTRACT

Introduction: VI (GW642444M) is an inhaled long-acting beta2 agonist (LABA) with inherent 24h activity under development as once-daily monotherapy for COPD and in combination with a novel inhaled corticosteroid (ICS; fluticasone furoate) for COPD and asthma.

Objective: To evaluate the efficacy of VI administered via a novel single-step activation inhaler in patients with COPD in a dose-ranging study.

Methods: In this multicentre, randomised, placebo-controlled, double-blind, stratified, parallel-group study, 605 patients were randomised to one of six treatments: VI (3, 6.25, 12.5, 25 or 50mcg) or placebo administered once daily for 28 days. The primary efficacy endpoint was the change in trough forced expiratory volume in 1 second (FEV1) from baseline on Day 29 (23–24h post Day 28 dose). Secondary endpoints included time to increase in >100mL from baseline FEV1 on Day 1.

Results: Mean baseline FEV1 ranged from 1.18–1.33L across treatment groups and the mean (standard deviation [SD]) screening % predicted FEV1 was 50.5 (9.78). Statistically significant improvements in trough FEV1 were reported on Day 29 for all doses (3, 6.25, 12.5, 25 and 50mcg) of VI compared with placebo (p<0.001) (92, 98, 110, 137 and 165mL, respectively). Time to an increase of >100mL in FEV1 on Day 1 was significantly shorter for all VI arms compared with placebo (p<0.001) (median time of 6min in the 25 and 50mcg groups).

Conclusion: Once-daily administration of VI in patients with COPD provides clinically relevant 24h improvement in lung function with a rapid onset of effect particularly with the 25 and 50mcg doses.

INTRODUCTION

·      VI is a novel, inhaled LABA with inherent 24h activity.(1) VI is in development as a once-daily monotherapy for COPD and, in combination with the ICS, fluticasone furoate, for COPD and asthma.

·      This dose-ranging study of VI in COPD patients aimed to determine the optimal dose(s) of VI to be used in further clinical trials. Here we report the efficacy data from the study; safety data are reported separately in poster number P1185.

PATIENTS AND METHODS

Study design

·      A phase IIb, randomised, double-blind, placebo-controlled, parallel-group trial conducted at 89 centres from February to October 2008. Patients were stratified by reversibility (reversible; non-reversible).

·      The study was approved by local ethics review committees and conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. Written informed consent was obtained from all patients.

Inclusion criteria

·      Males or females aged 40–80 years with COPD (American Thoracic Society/European Respiratory Society definition)(2) and current or prior history of >10 pack-years of cigarette smoking.

·      A post-salbutamol FEV1/forced vital capacity (FVC) ratio <0.70 and FEV1 >35% and <70% of predicted normal values (The National Health and Nutrition Examination Survey III)(3) at screening (study visit 1).

·      Women of child-bearing potential had to be using acceptable methods of contraception.

Exclusion criteria

·      COPD caused by alpha 1 antitrypsin deficiency.

·      Poorly controlled COPD defined as acute worsening managed by patient with corticosteroids/antibiotics or requiring physician-prescribed treatment, both within 6 weeks of visit 1.

·      Hospitalisation for poorly controlled COPD within 12 weeks of visit 1.

·      Current diagnosis of asthma, or other lung disease (active tuberculosis, lung cancer, bronchiectasis, sarcoidosis, fibrosis, pulmonary hypertension, interstitial lung disease or other active pulmonary disease) or lung volume reduction surgery within past year. Any clinically significant abnormalities on chest scan not thought due to COPD. Lower respiratory tract infection requiring antibiotics within 6 weeks of visit 1.

·      Requirement for long-term or nocturnal oxygen therapy for >12h/day; any clinically significant 12-lead electrocardiogram abnormality; patients unable to withhold salbutamol for 6h before spirometry testing; use of medications potentially affecting the study results (Table 1).

Table 1. Medications not permitted during the study and which must not have been taken for the indicated times prior to visit 1.*

*Note: use of tricyclic antidepressants, monoamine oxidase inhibitors, beta-blockers, anticonvulsants (barbiturates, hydantoins and carbamazepine) and phenothiazines was also prohibited

Methods

·      After a 2-week, single-blind, placebo run-in period, patients were randomised to one of five VI doses (3, 6.25, 12.5, 25 and 50mcg) or placebo once daily for 28 days (Figure 1).

·      All study treatments were administered in a double-blinded manner once daily in the morning using a novel dual-strip dry powder inhalation device.

·      Spirometry was assessed at visits 1 (screening), 2 (baseline), 3, 5, 6 and 7, peak expiratory flow was assessed twice daily every day, COPD symptoms were recorded using daily diary cards.

·      The primary efficacy endpoint was the mean change from baseline (Day 1) in trough (pre-dose, pre-bronchodilator) FEV1 on Day 29 (mean 23 and 24h post Day 28 dose).

Figure 1. Study design.

Trough FEV1 at the 23rd and 24th h following dosing on visit 2 and visit 6 was measured at visit 3 and visit 7, respectively, when the patients returned to the clinic the following morning

RESULTS

·      Of 851 patients who entered the 2-week run-in period, 602 were randomised and received at least one dose of study treatment (intent-to-treat [ITT] population). Demographics, baseline characteristics and screening pulmonary function were comparable (Table 2). Most patients were white males, mean age was 62 years, approximately 50% still smoked and 76% had a COPD diagnosis <10 years.

·      Across the treatment groups, mean screening pre- and post-bronchodilator FEV1 values ranged from 1.232 to 1.401L and 1.420 to 1.555L, respectively, and the mean % predicted post-bronchodilator FEV1 from 49.9 to 51.5%. Mean % reversibility ranged from 12.8 to 18.1% and absolute reversibility from 154.1 to 187.5mL.

·      Mean treatment compliance was >99% in all groups.

Table 2. Patient baseline characteristics and pulmonary function at screening (ITT population).

*Values are mean + SD; †BMI categories: underweight (<18.5kg/m2); normal (18.5kg/m2 to <25kg/m2); overweight (25kg/m2 to <30kg/m2) and obese (>30kg/m2); BMI = body mass index; MMRC = Modified Medical Research Council

MMRC dyspnoea scale

1.     Not troubled by breathlessness except on strenuous exercise

2.     Short of breath when hurrying or walking up a slight hill

3.     Walks slower than contemporaries on the level because of breathlessness, or has to stop for breath when walking at own pace

4.     Stops for breath after about 100m or after a few minutes on the level

5.     Too breathless to leave the house, or breathless when dressing or undressing

Primary endpoint

·      Improvements in trough FEV1 versus placebo on Day 29 were statistically significant (p<0.001) with all doses of VI (92, 98, 110, 137 and 165mL, for the 3, 6.25, 12.5, 25, 50mcg doses, respectively) (Figure 2a).

·      Posterior probabilities for a true >100mL increase were >90% with both the 25mcg (92%) and 50mcg (99%) VI doses, but were lower for the 3mcg (37%), 6.25mcg (47%) and 12.5mcg (64%) doses; probabilities for a true >130mL increase were 90%, 61%, 22%, 11% and 7% with the 50, 25, 12.5, 6.25 and 3mcg doses, respectively (Figure 2b).

Figure 2. Adjusted mean change from placebo in trough FEV1 at Day 29 (a) and posterior probability distribution of the treatment difference in change from baseline in trough FEV1 on Day 29 (b).Error bars are 95% confidence intervals (CIs); ITT population.

Analyses were performed using analysis of covariance with covariates of baseline, sex, age, smoking status at screening, reversibility stratum and treatment

Bayesian analysis with non-informative prior distribution; Analyses were adjusted for baseline (pre-dose on Day 1), sex, age, smoking status at screening, reversibility stratum and treatment

Secondary endpoints

·      On Days 1 and 28, there were significant (p<0.003) dose-related increases from baseline in 0–24h weighted mean FEV1 values for all VI doses versus placebo (Table 3)

·      clinically relevant differences of >100mL were observed on both days at all dose levels except for 3mcg on Day 1

·      differences of >130mL were observed on both days with the 25 and 50mcg doses, and on Day 28 with the 12.5mcg dose.

Table 3. Change from baseline 0–24h weighted mean FEV1 on Days 1 and 28 (ITT population).

*Number of patients with analysable data on one or more days; †Number of patients with analysable data on the given day ‡SE for both LSM and LSM change; Note: analysis performed using repeated measures with covariates of baseline, sex, age, smoking status (at screening), reversibility stratum, Day (nominal), treatment and Day by treatment and Day by baseline interactions; LSM = least square mean; SE = standard error

·      Time to achieve a >100mL increase in FEV1 over the first 4h post-dose on Day 1 for each VI dose group was significantly (p<0.001) shorter than for placebo, and shortest in the 25 and 50mcg dose groups (median time 6min) (Table 4)

·      the time to achieve a >12% increase in FEV1 over the first 4h post-dose on Day 1 for each VI dose group was also shorter versus placebo, supporting the time to achieve a >100mL increase findings.

Table 4. Log-rank analyses of times for patients to achieve >100mL increase from baseline in FEV1 (0–4h post-dose) (ITT population).*

*Stratified by reversibility (reversible, non-reversible); †If >50% of patients were censored then median time to a 100mL increase in FEV1 was not defined

·      For serial FEV1 measurements, adjusted mean changes from baseline FEV1 over time on Days 1 and 28 show a dose-response effect, with a sustained duration of action over the 24h period (all doses of VI versus placebo; Figure 3).

Figure 3. Adjusted treatment differences from baseline in serial FEV1 on Day 1 and Day 28; ITT population.

Analyses were performed using repeated measures with covariates of baseline, sex, age, smoking status at screening, reversibility stratum, time (nominal), treatment, time by treatment and time by baseline interactions. Note: FEV1 is plotted at 5, 15 and 30min and 1, 2, 4, 8, 12, 23 and 24h post-dose. At each timepoint treatments are offset

CONCLUSION

·      Once-daily administration of VI in patients with COPD provides clinically relevant 24h improvement in lung function with a rapid onset of effect, particularly with the 25mcg and 50mcg doses.

REFERENCES

(1)   Kempsford R, et al. Am J Respir Crit Care Med 2010;181:A4447.

(2)   Celli BR, et al. Eur Respir J 2004;23:932–46.

(3)   Hankinson JL, et al. Am J Respir Crit Care Med 1999;159:179–87.

ACKNOWLEDGEMENTS

·      This study was sponsored by GlaxoSmithKline (ClinicalTrials.gov: NCT00606684; B2C111045).

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