Pharmacokinetics and Tolerability of Oral Sildenafil in Adults with Cystic Fibrosis Lung Disease

Abstract

Rationale

Airway inflammation is central to cystic fibrosis (CF) pathophysiology. Pre-clinical models have shown that phosphodiesterase inhibitors (PDEi) like sildenafil have anti-inflammatory activity. PDEi have not been studied in CF subjects.

Objectives

We evaluated the pharmacokinetics, tolerability, and safety of sildenafil in subjects with CF. Sputum biomarkers were used to explore efficacy.

Methods

An open-label pilot study of oral sildenafil administration was conducted in adults with mild to moderate CF lung disease. Subjects received oral sildenafil 20 or 40 mg p.o. t.i.d. for 6 weeks.

Measurements and Main Results

Twenty subjects completed the study. Estimated elimination rate constants were statistically different in subjects with CF compared to previously published non-CF subjects. Side effects were generally mild. There were no drug-related serious adverse events. Sputum neutrophil elastase activity decreased.

Conclusions

Subjects with CF may eliminate sildenafil at a faster rate than non-CF subjects. Sildenafil administration was safe in subjects with CF, and decreased sputum elastase activity. Sildenafil warrants further study as an anti-inflammatory in CF.

1. INTRODUCTION

Cystic fibrosis is caused by mutations in the gene encoding the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR), resulting in abnormal salt and water transport in the airways.¹ Excessive neutrophil-mediated inflammation is present from infancy, and contributes significantly to the progression of the disease, as release of neutrophil-derived DNA, proteases and oxidants cause airway damage.^2,3 Even in patients with mild asymptomatic disease, neutrophil count and active elastase are substantially higher than in normal controls. ^4,5 The development of chronic infection accelerates the inflammatory response, ultimately leading to progressive bronchiectasis and death from respiratory failure in over 80% of patients.¹

Anti-inflammatory agents such as oral corticosteroids and high-dose ibuprofen have been shown to have benefit in the treatment of CF lung disease.^6,7 Several antibiotics have also been shown to reduce inflammation, either indirectly by reducing infection, or through direct anti-inflammatory properties.⁸ Together, these findings support the conclusion that anti-inflammatory treatment is beneficial, but all currently available agents are associated with significant side-effects that limit their use. Therefore, identification of additional safe and effective anti-inflammatory agent remains a priority for the advancement of CF treatment.⁹

Phosphodiesterase inhibitors (PDEi) such as sildenafil (Viagra®, Revatio®) and vardenafil (Levitra®) have been shown in vitro and in vivo to have anti-inflammatory activity in non-CF and in CF models.^10–15 (See Table E1 in the data supplement.) PDEi are postulated to have anti-inflammatory effects resulting from their correction of chloride transport through the CFTR, or through the downstream effects of increased cyclic guanosine monophosphate on defective protein glycosylation and bacterial adherence; however, the exact mechanism of the anti-inflammatory activity of PDE is unknown.^11,14,16

In addition to inflammation reduction, CFTR-specific dysfunction may be correctable with PDEi administration. Recent in vitro and in vivo studies have shown that PDEi can potentiate CFTR-mediated chloride transport activity in respiratory and digestive tissues.^14,17–20 Additionally, investigators have shown that this class of drugs corrects mislocalization of the mutant protein by displacing it from the intracellular compartment towards the apical cell compartment studied.^20,21 Thus, both the cellular localization and activity of F508del CFTR in the airway can be restored by PDEi. The PDEi sildenafil has been studied in large trials and is FDA-approved for treatment of erectile dysfunction and for pulmonary hypertension.^22,23 The laboratory evidence of anti-inflammatory properties and CFTR modulation combined with the safety profile of sildenafil makes it an attractive agent to study in patients with CF lung disease.

This study is the first to evaluate the pharmacokinetics, safety, and tolerability of sildenafil in subjects with CF. We also evaluated exploratory efficacy endpoints including sputum biomarkers of inflammation (sputum elastase and sputum IL-8), clinical measures (exhaled breath condensate pH, sputum bacterial counts, FEV₁% predicted, change in weight) and patient-reported health using the Cystic Fibrosis Questionnaire-Revised (CFQ-R).²⁴ The goal of the study is to provide pilot data for a randomized, multicenter phase IIb trial.

2. MATERIALS AND METHODS

2.1 Study Design

This study was a single site phase IIa open-label dose escalation study of sildenafil 20 mg p.o. t.i.d or 40 mg p.o. t.i.d for 6 weeks. All subjects in the study were pancreatic insufficient. The decision for dose escalation and administration with a meal and pancreatic enzyme replacement therapy (PERT) was made after preliminary data from early subjects suggested decreased sildenafil absorption in subjects with CF compared to historical controls. Recruitment occurred from July 2010 to August 2012, and was limited to clinically stable subjects ≥ 18 years of age FEV₁ ≥ 50% predicted at screening. (Detailed inclusion and exclusion criteria are found in the data supplement). The study was stopped at completion of enrollment. Subjects remained on their pre-study medications including cycling inhaled antibiotics throughout the study. Patients requiring oral or intravenous antibiotics during the study were pre-specified to be withdrawn from the study, as use of antibiotics is known to decrease sputum elastase.²⁵ Similarly, use of other anti-inflammatory medications, including oral or intravenous corticosteroids, and NSAIDs was prohibited. Because sildenafil is metabolized by the hepatic microsomal isoenzymes, CYP34A and cytochrome p450 2C9 (CYP2C9), use of concomitant medications known to be potent inducers or inhibitors of these enzymes was not allowed. Subjects with risk factors for pulmonary hypertension (severe CF lung disease defined as an FEV₁ < 50% predicted, baseline resting O₂ saturations < 93% or nighttime hypoxia) were excluded. The protocol was approved by the National Jewish Health Institutional Review board, and subjects provided written informed consent.

2.2 End points

Description of the pharmacokinetics of sildenafil in subjects with CF was a primary goal of the study. Plasma sildenafil levels were determined by high-performance liquid chromatography as previously described and compared with published normal values.^26,27 Plasma was stored at −20 °C, and shipped frozen to the laboratory of Dr. Linda Felton at the University of New Mexico (Albuquerque, NM).

Safety and tolerability were evaluated by assessment of adverse events, vital signs including blood pressure and oxygen saturation, clinical laboratory values (complete blood count, comprehensive metabolic panel and prothrombin time), physical examination and audiologic examination. Further details of the safety and tolerability assessment are presented in the data supplement. Hospitalizations and pulmonary exacerbations were captured as adverse events. Exploratory efficacy endpoints were evaluated including change in sputum elastase from baseline through 6 weeks of therapy, and change from baseline to week 6 in sputum IL-8, exhaled breath condensate pH, sputum bacterial counts, FEV₁% predicted, change in weight, and patient-reported health using the Cystic Fibrosis Questionnaire-Revised (CFQ-R).²⁴ Spirometry was performed according to American Thoracic Society (ATS) Guidelines.²⁸ Exhaled breath condensate was collected using R-tubes (Respiratory Research Institute) by ATS recommended technique.²⁹ Spontaneous or induced sputum (as per the CF TDN research protocol) expectoration was collected and placed on ice and sent overnight to the CF TDN Inflammatory Mediator Core Laboratory (Aurora, CO) for batch analysis of active elastase and IL-8 as previously described.³⁰ Sputum bacterial counts were performed at National Jewish Health.

2.3 Statistical analysis

Change in symptoms and adverse events were tabulated with descriptive statistics. Possible difference in absorption of sildenafil in patients with CF compared to non CF subjects was evaluated by comparing plasma levels of the drug obtained in subjects treated in this work with corresponding levels obtained in subjects without CF reported in the literature.³¹

Analyses for safety and efficacy were based on paired comparison from pre-study drug values to end of study drug values using paired permutation methods.³² Because there were no significant differences in efficacy outcome measures based on dose, pooled data from both dose groups is presented for all analyses except the PK analyses. Pharmacokinetic analysis included both a 2-sample permutation method ^32,33 to compare CF subjects versus normal controls,³¹ and a paired permutation method to compare values between the 20 mg and 40 mg dosing groups.³² Confidence intervals represent bias-corrected accelerated bootstrap confidence intervals.^34,35

Exploratory analyses for efficacy (sputum elastase, sputum IL-8, EBC pH, sputum bacterial counts, FEV₁% predicted, change in weight, and change in CFQ-R score) and safety (laboratory values, vital signs, and audiologic exams) were performed on subjects who completed 6 weeks of therapy. Sputum P. aeruginosa count from pre-study drug values to end of study drug values were calculated as back-transformed difference of the logarithms (base 10) of the counts (in CFU) pre-study drug and at the end of study drug. This transformation was necessary so that the theoretical distribution of the difference is roughly normally distributed.

3. RESULTS

3.1 Subjects

Thirty-six subjects were enrolled and screened, with 9 screen failures. A total of seven subjects failed to receive 6 weeks of study drug as a result of adverse events. Four subjects voluntarily withdrew from the study: 3 for headache and 1 for insomnia. Three subjects were withdrawn by the investigator per the pre-specified study protocol: 1 for pulmonary exacerbation, 1 for sinusitis and 1 for hospitalization for distal intestinal obstruction syndrome. Twenty subjects completed 6 weeks of therapy. (Figure 1) One subject presented to the final study visit with 1 week of previously unreported symptoms consistent with pulmonary exacerbation, and therefore, efficacy data was not analyzed on that subject as pre-specified in the protocol.

Figure 1. Subject Flow Diagram.

¥This exclusion criterion was later changed based on DMC review

*Patient came to final study visit with 1 week of previously unreported symptoms consistent with pulmonary exacerbation, and therefore, as pre-specified in the protocol, data was not analyzed for this subject.

The average age of the subjects was 25.8 ± 6.0 years and 53% (10/19) were female. One hundred percent of the subjects had at least one copy of F508del and 42% (8/19) were homozygous for F508del. Baseline FEV₁ was 77.4 ± 25.8 percent predicted. One hundred percent of the patients were pancreatic insufficient. Demographics were similar when considering all subjects who received at least one dose of study drug versus those that completed 6 weeks of study drug, and between those in the 20 mg and 40 mg dosing groups. (Table 1)

Table 1.

Subject demographics

	Received study drug N=27	Completed study N=20*	Completed study 20 mg dose N=6*	Completed study 40 mg dose N=13
Age (years)	25.7 SD 5.5	25.8 SD 6.0	26.2 SD 3.4	23.4 SD 1.1
Female	59% (16/27)	53% (10/19)	67% (4/6)	46% 6/13
Homozygous for F508del	40% (11/27)	42% (8/19)	50% (3/6)	38% (5/13)
At least one copy of F508del#	96% (26/27)	100% (19/19)	100% (6/6)	100% (13/13)
FEV1 % predicted	77.5 SD 19.6	77.4. SD 21.0	81.0 SD 6.7	81.4 SD 5.5
BMI	22.2 SD 2.7	22.7 SD 2.5	22.7 SD 1.0	22.7 SD 0.7
% with P. aeruginosa infection	(74%) 20/27	63% (12/19)	67% (4/6)	62% (8/13)

Standard Deviation (SD)

^*One subject presented to the final study visit with 1 week of previously unreported symptoms consistent with pulmonary exacerbation, and therefore, efficacy data was not analyzed on that subject as pre-specified in the protocol.

^#Non-F508del mutations: N1303K, R347P, L206W, unknown, G542X, 621 +1 G>T, Q39X, R1162X, 1213delT, 3905insT, Ex17a-18del

3.2 Pharmacokinetics

Pharmacokinetic analysis was performed on all subjects who received sildenafil through the visit during which blood samples were drawn (n=22) at 30 minutes, 1 hour, 2 hours, 4 hours and 6 hours. Additionally, blood samples were drawn from subjects at all subsequent visits without regard to time of dose administration. Eight subjects received 20 mg p.o. t.i.d. of sildenafil on an empty stomach and without enzymes. Sixteen subjects received 40 mg p.o. t.i.d and were instructed to take study drug with food and pancreatic enzymes. Plasma concentration versus time curves for representative patients from the two groups are shown in data supplement figure E2.

Table 2 summarizes the pharmacokinetic data for the CF patients in this study as well as previously published results in non-CF patients. Although preliminary data from early subjects suggested that absorption might be decreased in subjects with CF, when all data were analyzed, the time to maximum concentration (Tmax) and the area under the curve (AUC) were similar in subjects with CF dosed at 20mg compared to non-CF subjects dosed at 25mg³¹ (p=0.44 and p=0.71, respectively). Maximum concentration (Cmax), however, was significantly higher in subjects with CF who received 20 mg of sildenafil on an empty stomach than in subjects without CF³¹ (p=0.03). Although the value of Cmax was higher in subjects with CF who received 40 mg of sildenafil with food and enzymes than in subjects without CF who received 50 mg, the difference was not statistically significant (p=0.19). In contrast, a statistically significantly higher Tmax and lower AUC were found in CF subjects in the 40 mg group compared to subjects without CF who received 50 mg (p=0.02).

Table 2.

Pharmacokinetic parameters

	20mg (n=7)	40mg (n=15)	p-value5	25mg (n=224)6	p-value7	50mg (n=450)6	p-value8
AUC	497.0 SD	683.8 SD	0.21	464 SD	0.71	950 SD	0.02
(ng·h/ml)1	221.3	368.2		175		346
Cmax	159.3 SD	199.4 SD	0.41	84.4 SD	0.03	157 SD	0.19
(ng/ml)	59.7	122.3		80.5		49.2
Tmax	1.29 SD	2.03 SD	0.20	1.09 SD	0.44	1.03 SD	0.02
(h)	0.7	1.37		0.89		0.76
T1/2	2.21 SD	2.05 SD	0.74	3.60 SD	0.03	3.67 SD	<0.001
(h)	0.892	0.821		0.72		0.69
Ke	0.386 SD	0.378 SD	0.74	0.200 SD	0.03	0.195 SD	<0.001
(h−1)	0.2443	0.1172		0.038		0.035
Cl/F	40.0 SD	59.7 SD	0.19	58.2 SD	0.20	57.4 SD	0.79
(L/h)	31.44	30.54		8.3		8.3

Standard Deviation (SD)

¹AUC_{(0–6 hours)}

²Based on estimated half-life from the elimination phase of the plasma concentration versus time graphs.

³Assuming first order kinetics, calculated as 0.693 divided by the estimated T_1/2.

⁴Calculated as dose divided by AUC_(0-∞)

⁵Based on comparison between 20 mg and 40 mg CF subjects

⁶Values reported by Milligan et al. (31)

⁷Based on comparison between 20 mg CF subjects and published results for 25 mg non-CF subjects

⁸Based on comparison between 40 mg CF subjects and published results for 50 mg non-CF subjects

When comparing subjects with CF who received 20 mg of sildenafil on an empty stomach versus those who received 40 mg with food and enzymes, values of Cmax were not statistically different between the two dosing groups (p=0.41). There was a trend towards a longer Tmax in the 40-mg group compared to that of the 20mg group, however the differences were not statistically significant (p=0.20). AUC was greater in the 40 mg group than in the 20 mg group, although these differences were also not statistically significant (p=0.21).

3.3 Safety Profile

All subjects who received at least one dose of sildenafil were considered for adverse events (n=27). There were no deaths during the study and there were no drug-related serious adverse events. No visual changes or episodes of priapism were reported. Three serious events occurred that were judged to be unrelated to study drug, including two pulmonary exacerbations and one episode of distal intestinal obstruction syndrome (DIOS) (Table 3). The most common drug-related adverse events were mild and consistent with those reported in the package insert and in the literature.³⁶ When considering the two dosing groups, rates of headache and insomnia were similar between the two groups. Reports of increased cough, sore throat/throat clearing and increased/thicker sputum were higher in the 20 mg group. Rhinorrhea/sinus congestion, flushing, dyspepsia and myalgias were more common in the 40 mg group. (Table 3) The most common reason for drug discontinuation was headache (three subjects). In the subjects who received 6 weeks of sildenafil, there was no statistically significant change in systolic or diastolic blood pressure [114, 95% CI 110, 118; p=0.76 and 72, 95% CI 70, 75; p=0.39, respectively], oxygen saturation (94.5, 95% CI 94, 95; p=0.41), or hearing (p=0.38–1.00 across the range of Hz tested; details in data supplement). There was no statistically or clinically significant change in safety labs with the exception of a statistically significant decrease in alkaline phosphatase (−6.4 IU, 95%CI −10.8, −2.5, p=0.008). Details of the safety analyses are presented in the data supplement.

Table 3.

Adverse Events

Adverse Event	Event Frequency All N=27	20 mg group N=8	40 mg group N=19
Rhinorrhea/increased sinus congestion	9 (33%)	1 (13%)	8 (42%)
Headache	6 (22%)	2 (25%)	4 (21%)
Sore throat/throat clearing	6 (22%)	3 (38%)	3 (16%)
Increased cough	6 (22%)	3 (38%)	3 (16%)
Flushing	5 (19%)	1 (13%)	4 (21%)
Dyspepsia	4 (15%)	0	4 (21%)
Myalgia	4 (15%)	0	4 (21%)
Increased/thicker sputum	4 (15%)	2 (25%)	2 (11%)
Insomnia	3 (11%)	1 (13%)	2 (11%)
Chest tightness/congestion	2 (7%)	0	2 (11%)
Pulmonary exacerbation	2 (7%)	1 (13%)	1 (5%)
Arthralgia	1 (4%)	1 (13%)	0
Sinusitis	1 (4%)	0	1 (5%)
Elevated wbc	1 (4%)	1 (13%)	0
Anxiety	1 (4%)	1 (13%)	0
DIOS	1 (4%)	1 (13%)	0
Malaise	1 (4%)	0	1 (5%)
Subjective fever	1 (4%)	0	1 (5%)
Palpitations	1 (4%)	0	1 (5%)
Diarrhea	1 (4%)	1 (13%)	0
Asymptomatic hypotension	1 (4%)	1* (13%)	0
Abdominal pain	1 (4%)	1 (13%)	0
Sneezing	1 (4%)	0	1 (5%)
Thrush	1 (4%)	0	1 (5%)
Unilateral increased hearing thresh hold	1 (4%)	0	1 (5%)
Rash	1 (4%)	0	1 (5%)
Viral gastroenteritis	1 (4%)	0	1 (5%)
Mild transaminitis	1 (4%)	0	1 (5%)
Uterine cramping	1 (4%)	0	1 (5%)

Event frequency represents the number of subjects that reported the event out of the 27 subjects who received 15 at least one dose of study drug.

^*Subject experienced asymptomatic hypotension at dose escalation visit, so was continued in the 20 mg group.

3.4 Exploratory Efficacy Measures

There was a statistically significant decrease in mean sputum elastase activity (−57 mcg/mL, 95% CI −119, −18; p<0.03). (Figure 2) There was also a non-significant trend towards improvement in sputum IL-8 (−4.1 × 10⁴ mcg/mL, 95% CI −9.3 × 10⁴, +7.6 × 10³; p=0.13). There was no difference in EBC pH (0.09 units, 95% CI −0.29, +0.11; p=0.44). (Table 4)

Figure 2. Oral sildenafil reduces sputum elastase in subjects with CF.

Following 6 weeks of sildenafil treatment, the mean sputum elastase decreased by −57 mcg/mL, 95% CI −119 to −18; p=0.03. Error bars represent SD. One subject had insufficient sputum for analysis at visit 1; thus, 18 subjects had sputum available for paired analysis.

Table 4.

Exploratory Outcome Measures

Outcome	N	Mean Change	95% Confidence Interval	p-value
Sputum Elastase¥	18	−57 mcg/mL	−119 to −18	0.03
Sputum IL-8¥	18	−4.1× 104 mcg/mL	−9.3 × 104, +7.6 × 103	0.13
EBC pH	19	−0.09 units	−0.29, +0.11	0.44
Sputum P. aeruginosa count*	19	0.2 CFU	−1.9, +2.5	0.80
CFQ-R Respiratory Score	19	0.3 units	−4.4, +4.7	0.88
Weight	19	0.02 kg	−0.82, +0.72	0.95
FEV1	19	−2.4% predicted	−4.5, −0.4	0.04

^¥One subject had insufficient sputum for analysis at visit 1; thus, 18 subjects had sputum available for paired analysis.

^*Represented here by the fold change in sputum P. aeruginosa count from pre-study drug values to end of study drug values; this is the back-transformed difference of the logarithms (base 10) of the loads pre-study drug and at the end of study drug. This transformation is necessary so that the theoretical distribution of the difference is roughly normally distributed.

3.5 Other clinically relevant endpoints

There was no change in sputum P. aeruginosa count (0.2 log 10 CFU, −1.9, +2.5, p=0.80), weight (0.02 kg, 95% CI −0.82, +0.72, p=0.95), or in the respiratory component of the CFQ-R (0.3 units, 95% CI −4.4 to +4.7, p=0.88). There was a small decrease in FEV₁ % predicted over the course of the study (−2.4%, 95% CI −4.5 to −0.4%, p=0.04). (Table 4)

3.6 Variability in response

Twelve of the eighteen (67%) subjects who had values for sputum elastase at both time points had a decrease in sputum elastase. The range of FEV₁ % predicted in the group of subjects that demonstrated a decrease in sputum elastase was 59%–107%. Of those 12 subjects who had a decrease in sputum elastase, 9/12 (75%) also had a decrease in IL-8. There was no statistical difference in response of the sputum biomarkers between subjects that received 20 mg versus those that received 40 mg, however, 10/12 of the subjects who had a decrease in sputum elastase were in the 40 mg group.

4. DISCUSSION

This proof-of-concept open label study evaluated the safety, and pharmacokinetics as well as exploratory efficacy endpoints of oral sildenafil administration in patients with mild to moderate CF lung disease. This FDA approved drug has a well-established safety profile in pulmonary hypertension, and we have found it to be safe in subjects with CF. Pharmacokinetics in subjects with CF were generally similar to those in historical normal controls. Six weeks of oral sildenafil therapy decreased sputum elastase activity.

A major goal of the study was to evaluate the pharmacokinetics of oral sildenafil administration in subjects with CF. Pharmacokinetics often differ in patients with cystic fibrosis, due to increased volume of distribution of hydrophilic drugs³⁷, increased renal³⁸ and hepatic³⁹ drug clearance, and impaired gastrointestinal tract absorption.³⁸ In the current study, pharmacokinetic parameters in subjects with CF were somewhat different than previously published results,³¹ as shown in Table 2. The AUC and Tmax were not statistically different in subjects with CF dosed at 20 mg of sildenafil compared to those in non-CF patients dosed at 25mg of drug,³¹ however, the Cmax was higher in subjects with CF. This observation may be related to more acidic environment in CF small bowel⁴⁰ resulting in an increased amount of drug in solution. Interestingly, Tmax was statistically higher, and AUC was statistically lower when CF patients dosed with 40 mg of sildenafil were compared to non-CF patients given 50 mg of the drug (p=0.02). Insufficient sampling prevented calculation of the elimination rate (k_e) using linear regression in the plasma concentration versus time graphs. However, we estimated the half-life (T_1/2) from these graphs and used those values to calculate an estimated k_e, assuming first order kinetics. These estimated values are presented in Table 2. While these data are estimates from a small group of subjects, comparisons with previously published data found statistically significant differences, suggesting that sildenafil elimination in CF patients is faster than in non-CF patients (higher ke and shorter T_1/2). The control population did differ from the CF subjects in age, gender and weight (older, all male, and heavier, respectively; data not shown). Therefore, it is possible that demographic factors influenced differences between the two groups. However, the AUC and Tmax were not statistically different in subjects with CF dosed at 20 mg of sildenafil on an empty stomach compared to those in non-CF patients dosed at 25mg of drug. Cmax was different between the two groups, but we believe that this observation may be related to more acidic environment in CF small bowel resulting in an increased amount of drug in solution. Estimated ke was greater in the CF subjects which could be age related, however, hepatic clearance is known to be increased in subjects with CF due to increased activity of hepatic microsomal enzymes. Thus, while it is possible that the demographics were responsible for the observed differences in Cmax and estimated clearance, we feel that the results could also be explained by what is known about CF pathophysiology/pharmacokinetics.^37–39 Further testing, however, is required to more accurately determine k_e in this patient population. We also calculated the parameter Cl/F (clearance divided by the fraction of dose reaching the systemic circulation), since bioavailability of orally administered sildenafil in CF patients is not known. There was no difference in Cl/F in subjects with CF compared to that in historical controls.³¹ We were unable to calculate another commonly reported pharmacokinetic parameter, Vd/F (volume of distribution divided by fraction absorbed), because of insufficient data points to estimate the rate of absorption (ka).

A 20 mg oral dose of sildenafil administered t.i.d. is FDA approved for use in pulmonary hypertension.⁴¹ However, the optimal dose of sildenafil, or the effect of co-administration with pancreatic enzymes in subjects with CF was unknown. We administered sildenafil 20 mg p.o. t.i.d on an empty stomach for the first 8 subjects of the study, and a dose of 40 mg p.o. t.i.d. with food and enzymes to the subsequent subjects. As expected, AUC was higher in the patients who took the 40 mg dose versus those who took the 20 mg dose; however, the increase in AUC was not proportionate to the increase in dose as previously reported,³¹ and was not statistically significant. Moreover, Cmax was unexpectedly similar between the two groups. There was a trend towards a longer Tmax in the subjects who took the higher dose with food and enzymes, although the difference in Tmax between the two groups was not statistically significant. Assuming first order kinetics, values of AUC and Cmax should be proportional to dose and Tmax should not change. Thus, it appears that despite the use of pancreatic enzymes, consumption of food with dosing likely impaired absorption in subjects with CF. These findings are consistent with previous results that reported delayed absorption with food in healthy volunteers.³⁶

Assessment of safety was another major goal of the study. Because this study was the first trial of sildenafil in CF, we employed a low-threshold for subject withdrawal. Fifteen percent (n= 4) of subjects voluntarily withdrew from the study (Figure 1). The majority of the adverse events reported were mild in nature and consistent with those reported previously.³⁶ Six patients (22%) reported headaches, and 3 withdrew from the study for this reason. A history of migraines was added as an exclusion criterion following DMC review; subsequently, there were no further withdrawals for headache. Importantly, there were no study drug-related serious adverse events. Particularly, there was no sudden loss of vision or hearing, and there was no evidence of change in the ventilation to perfusion ratio (V/Q), as observed in a study of sildenafil in severe idiopathic pulmonary fibrosis.⁴² There was a small but statistically significant decrease in FEV₁ over the course of the 6 weeks of the study. Similar small, early decreases in FEV₁ that later resolve have been observed with drug trials in CF.^43,44 There was one pulmonary exacerbation in each of the dosing groups. In our adult CF population, there is a 45% rate of 1 or more pulmonary exacerbations per year which is similar to that for the adult CF population in general. Therefore, the development of pulmonary exacerbation in 2 subjects was not felt to be excessive for the adult CF population.

Relatively high rates of adverse events have limited other anti-inflammatory agents in CF. Impaired glucose metabolism and growth failure have all but eliminated the use of systemic steroid therapy, and reports of gastrointestinal hemorrhage and nephrotoxicity have reduced use of high-dose ibuprofen to only 3% of pediatric CF patients, where it has proven efficacy.^45,46 In general, self-reported adverse events are high in CF therapeutic trials. For example, in a recent trial of azithromycin, an antibiotic with significant off-target anti-inflammatory properties, headaches were reported by 23% of subjects on study drug, compared to 31% of those receiving placebo.⁴⁷ Results of our pilot study suggest that a subpopulation of CF patients, particularly those predisposed to headaches, may not tolerate the drug. Nevertheless, this does not preclude its potential usefulness for the majority of the CF population.

The exploratory efficacy objective of this study was to determine if sildenafil can decrease lung inflammation in subjects with CF. Sputum elastase activity has been used in other early phase trials to assess efficacy as this surrogate biomarker is more sensitive than lung function or time to exacerbation for detecting short-term response to therapy.^48–50 Mayer-Hamblett et al previously demonstrated that sputum elastase has a significant longitudinal association with FEV₁: a −2.9% decline in FEV₁ for every log increase in elastase.⁴⁸ More recently, Sagel et al showed that measurements of sputum elastase predicted subsequent lung function decline.⁴⁹ Thus, change in elastase is a sensitive measure of clinically important differences. In this proof-of concept study, in a group of subjects with CF whose FEV₁ % predicted ranged from 50% to 108%, administration of 6 weeks of sildenafil led to a 57 mcg/mL (corresponding to −0.09 log10 mcg/mL change) decrease in sputum elastase. Additional exploratory analysis revealed that 75% of those who had a decrease in sputum elastase also had a decrease in sputum IL-8.

4.1 Limitations

The tolerability results might have been biased by the fact that neither the subjects nor the investigators were blinded. The interpretation of the dose escalation response was complicated by the fact that administration with food and enzymes occurred simultaneously. Additionally, the length of exposure to study drug was relatively short. Although many of the common side effects that occur with use of PDEi improve with repeated exposure (e.g. headaches and flushing), it is possible that side effects could limit long-term use in some patients. Finally, the interpretation of the exploratory efficacy endpoints in this study is limited by the fact that it was a small single-site, open-label study.

5. CONCLUSIONS

This study is the first to evaluate oral sildenafil administration in patients with mild-moderate CF lung disease. We found that sildenafil was safe in subjects with CF. Oral sildenafil administration resulted in a similar Tmax and AUC, but higher Cmax in subjects with CF compared to those without CF when lower doses were administered on an empty stomach (20 mg in CF patients and 25 mg in non-CF patients). In contrast, Tmax was statistically higher and AUC was statistically lower in CF patients given 40 mg of sildenafil with food and enzymes compared to previously published data on non-CF subjects dosed with 50 mg of the drug. Estimated T_1/2 was lower and estimated k_e was higher in both groups of CF subjects compared to those published for non-CF subjects. Findings from this study suggest that administration of sildenafil with food impairs drug absorption in CF patients. In addition, these preliminary data suggest that CF patients may eliminate sildenafil at a faster rate than non-CF patients. Using sputum elastase activity as an exploratory efficacy endpoint, we showed that oral sildenafil administration reduces airway inflammation in subjects with mild to moderate CF lung disease. The safety profile and decrease in sputum elastase following short-term sildenafil administration in this open label pilot study support further study in a randomized, double-blinded, placebo-controlled trial. Such a trial will allow us to determine if there is a role for sildenafil in the treatment of CF lung disease, and to better define which patients may benefit from this agent.

Highlights.

• We evaluated oral sildenafil administration in subjects with CF lung disease.

• Sildenafil administration was safe in subjects with mild-moderate CF lung disease.

• Subjects with CF may eliminate sildenafil at a faster rate than non-CF subjects.

• Oral sildenafil reduced sputum elastase in subjects with CF lung disease.

• Sildenafil should be studied further as an anti-inflammatory in CF.

ABBREVIATIONS

ATS

American Thoracic Society

AUC

Area under the curve

BAL

Bronchoalveolar lavage

CTEAE

Common Terminology Criteria for Adverse Events

CT

Computed tomography

CFF

Cystic Fibrosis Foundation

CFQ-R

Cystic Fibrosis Questionnaire-Revised

CF TDN

Cystic Fibrosis Therapeutics Development Network

CFTR

Cystic Fibrosis Transmembrane Conductance Regulator

CFU

Colony forming units

DIOS

Distal Intestinal Obstruction Syndrome

k_e

Elimination rate constant

EBC

Exhaled breath condensate

FDA

Federal Drug Administration

FEV₁

Forced expiratory volume in 1 second

T_1/2

Half-life

IL-8

Interleukin-8

Cmax

Maximum concentration

NCI

National Cancer Institute

PRO

Patient Reported Outcome

PDEi

Phosphodiesterase inhibitor

PASS

Power Analysis and Sample Size

SOP

Standard Operating Protocol

TDN

Therapeutics Development Network

Tmax

Time to maximum concentration

V/Q

Ventilation/perfusion