Tadalafil in patients with chronic obstructive pulmonary disease

Summary Background Phosphodiesterase-5 (PDE5) inhibitors improve exercise capacity and quality of life in patients with idiopathic pulmonary arterial hypertension. However, whether such beneﬁ cial eﬀ ects take place in selected populations with chronic obstructive pulmonary disease (COPD) remains uncertain. We aimed to assess the eﬀ ects of tadalaﬁ l—a PDE5 inhibitor—on exercise capacity and quality of life in patients with COPD and mild pulmonary hypertension. Methods We did a randomised, double-blind, parallel-group, placebo-controlled trial at three centres in Scotland, UK, between Sept 1, 2010, and Sept 1, 2012. Patients with moderate to severe COPD were randomly assigned (1:1), via centralised randomisation with a computer-generated sequence and block sizes of four, to receive daily tadalaﬁ l 10 mg or placebo for 12 weeks. Patients, study investigators, outcome assessors, and those administering drugs were masked to group allocation. The primary endpoint was the mean placebo-corrected diﬀ erence between the baseline and ﬁ nal 6 min walk distance after 12 weeks. We measured change in quality of life at baseline, 8 weeks, and 12 weeks, with standardised questionnaires. Analysis was per protocol and by intention to treat. This trial is registered with ClinicalTrials.gov, number NCT01197469. Findings 120 patients were randomly assigned to receive tadalaﬁ l (n=60) or placebo (n=60), of whom 56 (93%) versus 57 (95%) completed the study. At 12 weeks the diﬀ erence in 6 min walking distance between the tadalaﬁ l and placebo groups was 0·5 m (95% CI –11·6 to 12·5; p=0·937). We recorded no statistically signiﬁ cant changes in quality of life (between-group diﬀ erence on the St George’s Respiratory Questionnaire –2·64 [95% CI –6·43 to 1·15]; Research and Development version 1 short-form-36 4·08 [–1·35 to 9·52]; Minnesota Living with Heart Failure questionnaire –2·31 [–7·06 to 2·45]). 19 (32%) of 60 patients in the treatment group had dyspepsia; the severity of dyspepsia ranged from mild to severe, with four (21%) of 19 patients needing a proton-pump inhibitor. Five (8%) of 60 participants had dyspepsia in the placebo group. Headache was noted in 17 (28%) patients in the treatment group versus 5 (8%) in the placebo group, but was mild in all patients. Two (3%) patients in the treatment group had facial ﬂ ushing, which resulted in one withdrawal. Other withdrawals within the tadalaﬁ l group happened after a transient ischaemic attack and two deaths (ruptured abdominal aortic aneurysm and pneumonia).


Introduction
Chronic obstructive pulmonary disease (COPD) is a major cause of mortality and morbidity worldwide. Pulmonary vascular disease happens early in the natural history of COPD, when pulmonary vascular changes have been reported in cigarette smokers with normal spirometry values. 1 Pulmonary arterial vasoconstriction, vessel-wall hypertrophy, and loss of pulmonary blood vessels contribute to the burden of pulmonary vascular disease, resulting in the development of pulmonary hypertension.
COPD-associated pulmonary hypertension reduces exercise tolerance 2 and quality of life. Modest increases in pulmonary arterial pressure, that are less than the cutoff considered to represent pulmonary hypertension, can strongly predict mortality 3 and risk of hospital admission during an exacerbation. 4 In patients with COPD without pulmonary hypertension (mean pulmonary artery pressure <25 mm Hg), or with borderline hypertension (21-25 mm Hg), these slight changes in pulmonary arterial pressure are associated with impaired right ventricular systolic function, hypertrophy, and dilatation, 5 changes that are increasingly present as the severity of pulmonary vascular disease worsens. Furthermore pulmonary arterial pressures frequently rise abnormally during activities of daily living, even in individuals without resting pulmonary hypertension. 6 Although severe, or socalled dis proportionate, pulmonary hypertension in patients with low-severity airway obstruction does take place in patients with COPD, it is uncommon (<5% prevalence), 7 possibly representing the coexistence of idiopathic pulmonary arterial hypertension.
Phosphodiesterase-5 (PDE5) inhibitors are pulmonary arterial vasodilator drugs that are used to treat pulmonary hypertension, supported by evidence that they improve exercise capacity and quality of life. 8 Although sildenafi l-a short-acting PDE5 inhibitor-can improve pulmonary haemodynamics at rest and during exercise in patients with COPD-associated pulmonary hypertension, 9 whether pulmonary arterial vasodilation improves exercise capacity and quality of life has not been established. Concerns about the potential for clinical worsening remain on the basis of the use of other pulmonary vasodilator drugs, such as bosentan in unselected populations with severe COPD and sildenafi l in patients with severe COPD with emphysema, but without pulmonary hypertension. 10,11 In both studies, deleterious eff ects were reported on quality of life, with no benefi cial eff ect on exercise capacity. We postulated that treatment with tadalafi l-a long-acting PDE5 inhibitor-would improve exercise capacity and quality of life in selected patients with COPD and mild pulmonary hypertension.

Study design and patients
We undertook this randomised, double blind, parallelgroup, placebo-controlled trial at three centres in Scotland, UK (Dundee, Perth, and Fife) between Sept 1, 2010, and Sept 1, 2012. We recruited patients with COPD via a network of pulmonary physiotherapists, respiratory consultants, and family doctors covering two UK National Health Service boards (Tayside and Fife) where potential candidates were identifi ed through COPD databases.
Panel 1 shows the study eligibility criteria. All patients were smokers or ex-smokers with a diagnosis of COPD confi rmed according to American Thoracic Society and European Respiratory Society criteria, a post-bronchodilator forced expiratory volume in 1 s (FEV 1 ) less than 80%, and an FEV 1 to forced vital capacity ratio of less than 70%. 12 The echocardiographic inclusion criteria signify values beyond the normal range (ie, pulmonary acceleration time <120 ms or right ventricular systolic pressure >30 mm Hg). Participants had had no exacerbation of COPD for at least 1 month at study entry. All regularly prescribed COPD drugs were continued. All patients provided written informed consent. Ethics approval was obtained from the North Research Ethics Committee (reference 10/S0801/46).

Randomisation and masking
Eligible patients were randomly assigned in a 1:1 ratio, via centralised randomisation with a computer-generated sequence and block sizes of four, to receive daily tadalafi l 10 mg or matched placebo, orally. The intervention and placebo packs were identical in presentation (capsules and bottle) and double blinding was maintained throughout. Patients, study investigators, outcome assessors, and those administering drugs were masked to group allocation. Unmasking was done only after the end of the trial once all data had been collected and analysed.

Procedures
We used a MicroLoop spirometer to establish COPD diagnoses (Micro Medical, Rochester, UK). ARG undertook echocardiographs with a Philips iE33 system (Philips, The Netherlands) for participants from Dundee, and the Siemens Acuson CV70 (Siemens AG Medical   Solutions, Germany) for those from Perth and Fife. In addition to standard acoustic windows, we used an oblique subcostal window to assess the pulmonary acceleration time and the right-ventricular ejection time with pulse-wave Doppler, as previously described. 13 We then calculated estimated mean pulmonary arterial pressure with the Dabestani equation, 13,14 and right ventricular systolic pressure with the Bernoulli equation.
We estimated right atrial pressure on the basis of the size and collapsibility of the inferior vena cava. 15 We measured the thickness of the right ventricle (subcostal window acquisition) and the function (tricuspid annular plane systolic excursion) with a standard operating procedure. 15 We qualitatively assessed the presence of dilatation of the right ventricle. Exercise capacity was limited by exertional breathlessness in all participants. If eligibility criteria were met, patients were given a 50 mg test dose of sildenafi l and observed for 3 h. Resting and postural blood pressure and pulse oximetry (SpO 2 ) were measured at baseline, and at 30, 60, 120, and 180 min. Partcipants were randomised if they were free from clinically signifi cant symptoms, hypotension (systolic blood pressure <90 mm Hg) or symptomatic postural hypotension (a decrease of ≥20 mm Hg in systolic blood pressure drop during 3 min of standing) throughout the test dose observation period. Correspondence with family doctors, safety cards, and warning stickers on paper notes minimised the risk of clinically signifi cant hypotension through inadvertent prescription of nitrate-containing drugs. The 6 min walk test was done at baseline, and at weeks 8 and 12, on a fl at, straight, 30 m course with use of standardised encouragement. 16 Two walks were undertaken 1 h apart. Patients were allowed to use their usual walking aids but oxygen was not permitted. We assessed quality of life at baseline, 8 weeks, and 12 weeks with standardised questionnaires. The St George's Respiratory Questionnaire identifi es how breathing problems aff ect patients' lives scored on a scale of 0 (no impairment) to 100 (maximum impairment). The Research and Development version 1 shortform-36 questionnaire measures functional health and well-being scores within 8 domains, scored from 0 to 100 whereby higher scores represent better function. The Minnesota Living with Heart Failure questionnaire has been validated for pulmonary hypertension 17 whereby higher scores refl ect worse quality of life. We measured SpO 2 at the beginning and end of the walk test and at any temporary stops that were needed because of severe exertional dyspnoea. B-natriuretic peptide, highsensitivity C-reactive protein, and spirometry were also measured at baseline, and at weeks 8 and 12. The singlebreath diff using capacity of the lung was done in accordance with the American Thoracic Society and European Respiratory Society task force guidelines 18  ). Echo cardio graphy was done at baseline and 12 weeks. Two additional safety visits and telephone consultations were organised around the main visits. Adverse events were logged and capsules were counted at all visits to assess compliance. Blood pressure and SpO 2 were measured at all visits.

Outcomes
The primary outcome was the mean placebo-corrected diff erence between the baseline and fi nal 6 min walk distance after 12 weeks of tadalafi l. Secondary measures were change in quality of life, cardiac stress (B-natriuretic peptide), systemic infl ammation (highsensitivity C-reactive protein), lung diff usion (singlebreath diff using capacity), lung function (spirometry), and cardiac function (echocardiography).

Statistical analysis
The primary objective was to establish whether tadalafi l would improve exercise capacity in patients with COPD. Other work in this specialty suggests that a 54 m increase in 6 min walking distance should be regarded as the minimum clinically important distance at which patients Age (years) 68 (8) 70 (7) Men 42 (70%) 40 (67%)   consider themselves to feel subjectively better. 19 From the available literature we ascertained that a mean 6 min walking distance in patients with severe COPD would be roughly 285 m (SD 96). 20 A sample size of 104 patients was needed to achieve 80% power at a 5% signifi cance level. We increased the number to 120 to account for dropouts. We present data as means with SDs for continuous variables, and as numbers with percentages for categorical variables. For the primary outcome, we analysed data with the student's t test to identify the mean diff erence between groups at 12 weeks. Furthermore, a general linear model with repeated measures for multiple time points (ANOVA) was done with a post-hoc Bonferroni correction to avoid confounding of the overall α error in the two groups. We used a paired t test to test the diff erences within groups in data at baseline and 12 weeks. For the secondary outcome, we used the student's t test to examine the diff erences between the two groups. For non-normally distributed data we used the Mann-Whitney U test. We used the χ 2 test to assess the diff erences in categorical variables between the two groups. Primary analysis was per protocol, with a separate intention-to-treat analysis done with multiple imputation to address missing data.
No interim analysis was done. Analysis was done mainly on completers, with additional sensitivity analysis, and used data for the best walk test for each patient. Data analysis was independently verifi ed by the study statistician. This trial is registered with ClinicalTrials.gov, number NCT01197469.

Role of the funding source
The Chief Scientist Offi ce for Scotland (CZB/4/666) solely funded the study, with no role in the writing and fi nal decision to submit for publication. Our sponsor was the University of Dundee who had no role in study design, data collection, data analysis, data interpretation, or writing of the report. ARG and PJH had full access to the all the data in the study and ARG had fi nal responsibility for the decision to submit for publication. Figure 1 shows the trial profi le. 120 patients were randomly assigned to receive tadalafi l (n=60) or placebo (n=60), of whom 113 (94%) completed the study. Baseline characteristics were similar between groups (table 1). Patients had severe COPD and borderline to mild pulmonary hypertension with a moderate impairment of diff usion capacity (table 1). The upper range for right ventricular systolic pressure was 81 mm Hg and the lower range for pulmonary acceleration time was 69 ms (estimated mean pulmonary arterial pressure 47 mm Hg). Right ventricular thickness was measurable at baseline in 114 (95%) of 120 patients. 108 (90%) patients had echocardiographic evidence of right ventricular hypertrophy (≥5 mm right ventricular free wall thickness). In the 89 patients in whom tricuspid annular plane systolic excursion could be measured, 86 (97%) had preserved right ventricular function. The high proportion of patients prescribed combination inhaled corticosteroid and long-acting β agonists (109 [91%]) and long-acting muscarinic antagonists (104 [87%]) was typical of a moderate to severe cohort with COPD. 13 (11%) patients were receiving long-term oxygen therapy and 16 (13%) were receiving a loop diuretic. 79 (66%) patients had exercise desaturation (≥4% decrease in SpO 2 ). No patients had a history of thromboembolic disease or were receiving anticoagulants.

Total
After one 50 mg dose of oral sildenafi l we noted a signifi cant reduction in resting SpO 2 of 2·2% (95% CI 1·6-2·7; p<0·0001; fi gure 3. Despite this decrease, only one patient described a transient increase in breathlessness versus two patients who felt less breathless. 60 min after receiving the sildenafi l, the mean systolic blood pressure reduced maximally by 15·6 mm Hg (11·1-20·0; p<0·0001). Five patients had symptomatic postural hypotension during the test dose and were therefore ineligible for randomisation. Chronic dosing with tadalafi l at 10 mg once daily did not result in any signifi cant diff erences between group in SpO 2 at 12 weeks, with a mean change of -0·91% (SD 1·6) in the tadalafi l group and -0·70% (2·0) in the placebo group (p=0·543). We recorded no signifi cant group diff erences in blood pressure (data not shown).
No patients receiving long-term oxygen therapy needed an increase in oxygen dose and no new prescriptions of oxygen therapy were reported during the trial period (data not shown).

Discussion
Use of tadalafi l for 12 weeks did not improve exercise capacity, as measured by the 6 min walking distance. One possible explanation for this absence of eff ect is that in the presence of mild pulmonary hypertension, exercise capacity is mainly limited by exhaustion of the ventilatory reserve. 21,22 Of our three chosen quality-of-life measures, the St George's Respiratory Questionnaire is the most relevant. We did not see an improvement in quality of life with 12 weeks of tadalafi l treatment. The changes between groups in the activity and impacts score subdomains could mean that tadalafi l might improve quality of life in a much larger study. However, if such an eff ect was not apparent in the 113 patients in our study, the clinical relevance of any change in quality of life is questionable.
Tadalafi l doses higher than 10 mg might be more eff ective. At the time of our study design, tadalafi l had not been licensed for pulmonary hypertension at the present daily 40 mg dose that has since become the accepted starting dose for patients with idiopathic pulmonary arterial hypertension. 8 Our dose selection was informed by the anticipated pharmacokinetics in our elderly population by contrast with those in younger patients with idiopathic pulmonary arterial hypertension. The half-life of tadalafi l is increased (from 17·5 to 21·6 h) and clearance is also signifi cantly reduced in elderly people. 23,24 Information from the Electronic Medicines Compendium states that a dose adjustment in elderly patients is not warranted; however, this recommendation is based on data in healthy patients and not in those with severe COPD and pulmonary vascular disease-a largely unstudied population. In the absence of any data about the use of daily dosing with tadalafi l in pulmonary hypertension, we drew inspiration from the successful treatment of erectile dysfunction, for which comparatively low single doses of tadalafi l are used and daily dosing with 10 mg or above was not recommended. Furthermore in the pivotal randomised trial by Galiè and colleagues, 8 10 mg, 20 mg, and 40 mg dosing of tadalafi l at 12 weeks seemed to be equally eff ective for improvement of 6 min walking distance, although only the 40 mg dose was associated with a signifi cant increase in walking distance and prolonged time to clinical worsening at 16 weeks.By contrast with our own treatment-naive study population, more than 50% of the patients in Galiè and colleagues' trial were receiving bosentan, which because of interaction with cytochrome P450 3A4 could necessitate high doses to be eff ective. Our chosen dose (10 mg) of tadalafi l caused the expected pulmonary vasodilatation with no eff ect on 6 min walking distance.
We noted signifi cant increases in pulmonary acceleration time in patients in the treatment group, only in keeping with a small reduction in pulmonary vascular resistance and pulmonary pressure, the latter of which we estimated to be 3·5 mm Hg. 13,25 This fi nding supports those from other published work showing that after 3 months of sildenafi l, the pulmonary acceleration time was the only echocardiographic parameter to change signifi cantly in line with other well-established rightheart catheter measures (mean pulmonary artery pressure and pulmonary vascular resistance). 26 Acute sildenafi l dosing decreased SpO 2 on the basis of ventilation perfusion mismatch, which increased blood fl ow to poorly ventilated areas of the lung through release of hypoxic pulmonary vasoconstriction. This event was not apparent during chronic dosing with tadalafi l despite evidence of pulmonary vasodilatation, possibly because of adaptive processes with chronic dosing, mitigating any regional ventilation perfusion mismatching.
One perceived limitation of our study is our considered decision to not undertake right-heart catheterisation for patient selection and assessment of treatment response. We ruled out right-heart catheterisation for both logistical (distance of travel for patients and need for hospital admission) and ethical reasons, in that repeated catheterisation poses a small but unacceptable risk to patients, particularly because consideration of catheterisation is not routine practice in patients with COPD with suspected borderline or mild pulmonary hypertension. Although other studies in this specialty have suggested use of pulmonary vasodilator drugs on the basis of the high prevalence of exercise-induced pulmonary hypertension in patients with severe COPD, we wished to further characterise our participants with echocardiography. 10,11 Almost all the participants in our study had a measurable pulmonary acceleration time; as such, acceleration time is the key echocardiography parameter defi ning our cohort. Echocardiographyderived pulmonary acceleration time and pulmonary acceleration time to right ventricular ejection time ratio have an excellent correlation (r -0·88 to -0·90) with right catheter-derived mean pulmonary arterial pressure. 13,27 This technique is advantageous in the assessment of patients with pulmonary hypertension who are poorly echogenic, particularly when the discordance between right ventricular systolic pressure and invasive catheter measures is considered. 28 Although a precise estimation of mean pulmonary arterial pressure cannot be derived from pulmonary acceleration time, our methods were chosen to be a simple, reliable way to stratify participants. Use of all available echocardiographic methods and not solely the right ventricular systolic pressure has been emphasised in a debate of pros and cons. 29 Another potential limitation is our use of fairly modest indicators of pulmonary vascular disease for assessment of inclusion in the study, at the risk of including patients without pulmonary hypertension. Even mild increases in pulmonary pressure, not presently considered to represent pulmonary hypertension, can result in clinically signifi cant changes to the structure and function of the right heart. 5 Our selection criteria of a pulmonary acceleration time less than 120 ms was chosen to include all patients with a mean pulmonary arterial pressure of greater than 16 mm Hg. The estimated mean pulmonary arterial pressure in our fi nal cohort and the proportion of patients with right-ventricular dilation are consistent with a degree of pulmonary vascular disease. Although the upper range for right ventricular systolic pressure and the lower range for pulmonary acceleration time suggest the inclusion of those with severe pulmonary hypertension, small numbers preclude any conclusions about the eff ectiveness of tadalafi l in participants with more severe pulmonary hypertension. Tadalafi l could still conceivably improve exercise capacity in individuals with more severe pulmonary hypertension or mild airfl ow obstruction-a comparatively rare subgroup in whom further study is warranted.
Most studies of COPD that include measures of rightheart catheterisation naturally represent highly selected groups of patients with COPD who are referred to tertiary centres for consideration of lung transplant, surgery for reduction of lung volume, or investigation of disproportionate or severe pulmonary hypertension. In our cohort, the lack of clinical improvement with tadalafi l, despite enrichment of the study population through echocardiography, suggests that PDE5 inhibitors do not improve exercise capacity and quality of life in patients with moderate to severe COPD, even if there is a

Systematic review
We searched PubMed betweeen Jan 1, 1980, and Jan 1, 2014, for "phosphodiesterase-5 inhibitors", "tadalafi l", "sildenafi l", "COPD", "pulmonary hypertension", "pulmonary arterial hypertension", "exercise tolerance", and "exercise capacity". Our review of the existing literature emphasised the scarcity of robust evidence for or against the use of phosphodiesterase-5 (PDE5) inhibitors in patients with pulmonary hypertension associated with chronic obstructive pulmonary disease (COPD). We identifi ed only two randomised trials. In the fi rst, 11 in patients with COPD and emphysema, but without pulmonary hypertension, sildenafi l had no eff ect on exercise capacity and worsened quality of life. The second trial 30 showed that sildenafi l in addition to pulmonary rehabilitation did not confer an additional improvement in exercise tolerance in patients with severe COPD and mild to moderate pulmonary hypertension.

Interpretation
Our fi ndings suggest that in patients with moderate to severe COPD, selected to have borderline to mild pulmonary hypertension, tadalafi l does not improve exercise capacity or quality of life. This outcome is in keeping with evidence from randomised trials suggesting that in patients with COPD with emphysema and no pulmonary hypertension, or in those with COPD and mild to moderate pulmonary hypertension undertaking pulmonary rehabilitation, sildenafi l does not improve exercise capacity or quality of life. Our fi ndings therefore discourage use of PDE5 inhibitors in patients with moderate to severe COPD and borderline to mild pulmonary hypertension as a means to improve exercise capacity and quality of life. Further trials should focus on comparatively rare subgroups such as individuals with severe pulmonary hypertension or mild airfl ow obstruction, whereby PDE5 inhibitors could still conceivably have benefi cial eff ects.
suggestion of pulmonary hypertension on echocardiography. This message is timely because sildenafi l is no longer patented and is available on a generic tariff . Our fi ndings are consistent with the work of Blanco and colleagues 30 in which sildenafi l did not further improve exercise capacity in participants undergoing pulmonary rehabilitation, based on a cohort with echocardiographydefi ned pulmonary hypertension and, to a lesser extent, hypertension confi rmed by right-heart catheterisation. Our study diff ers in its primary assessment of exercise capacity and has been powered accordingly, necessitating double the participants. Our complementary fi ndings strengthen the case against the indiscriminate use of PDE5 inhibitors in COPD patients with borderline or mild pulmonary hypertension (panel 2).
In what represents a large cohort of patients with COPD and mild pulmonary hypertension we have shown that tadalafi l at a dose of 10 mg per day does not improve exercise capacity despite exerting its pharmacological eff ect of pulmonary vasodilation.

Contributors
ARG is the guarantor of contents and contributed to design, data acquisition, analysis, interpretation, and drafting of the manuscript. BJL contributed to the study conception, design, and revision of the manuscript. PJH contributed to the data acquisition and manuscript revision. LW contributed to the conception, design, analysis, interpretation, revision of article, and is the study statistician. ADS contributed to the conception, design, and revision of the manuscript.

Declaration of interests
We declare that we have no competing interests.