Digital ulcers: sildenafil

Drug action

Sildenafil is a potent and selective inhibitor of cyclic guanosine monophosphate (cGMP)‑specific phosphodiesterase type 5 (PDE5). Inhibition of the PDE5 enzyme increases the available intracellular cGMP, which leads to vasodilation in the systemic circulation (sildenafil summaries of product characteristics).

Regulatory status

Sildenafil is licensed in the UK for treating pulmonary arterial hypertension in adults and children. Sildenafil is also licensed for treating erectile dysfunction in men. The Department of Health has amended regulations to allow unrestricted prescribing of generic sildenafil for men with erectile dysfunction.

See summaries of product characteristics of sildenafil products for more information on these indications.

No sildenafil product is licensed for treating or preventing digital ulcers, therefore using sildenafil for this indication is off‑label.

In line with the guidance from the General Medical Council (GMC), it is the responsibility of the prescriber to determine the clinical need of the patient and the suitability of using sildenafil outside its authorised indications.

Cost

Generic sildenafil is £1.09 for 4×25 mg tablets, £1.13 for 4×50 mg and £1.22 for 4×100 mg tablets (excluding VAT; costs taken from Drug Tariff, February 2015).

The sildenafil product that is licensed for managing pulmonary hypertension (Revatio) is £446.33 for 90×20 mg tablets, £186.75 for 112 ml×10 mg/ml oral suspension and £45.28 for 20 ml×0.8 mg/ml solution for injection (excluding VAT; costs taken from MIMS, February 2015).

Clinical effectiveness

The authors of the meta‑analysis (Tingey et al. 2013) found 31 RCTs (n=1989) that met the eligibility inclusion criteria. These studies compared various different therapies with placebo or active treatment for healing and preventing digital ulcers in people with systemic sclerosis, and had primary, secondary or exploratory outcomes that investigated digital ulcers. The meta‑analysis included data from 3 small RCTs of PDE5 inhibitors. There were 2 small double‑blind RCTs of sildenafil for the treatment of Raynaud's phenomenon (Fries et al. 2005 and Herrick et al. 2011) and 1 small RCT of tadalafil (Shenoy et al. 2010). Pooled data analysis reported in the meta‑analysis is not considered in detail in this evidence summary because the results for PDE5 inhibitors included the tadalafil study.

Fries et al. (2005) was a fixed‑dose, single‑centre, German crossover study in which 20 people with secondary (n=16) or primary Raynaud's phenomenon (n=2) resistant to vasodilatory treatment were randomised to either sildenafil 50 mg twice daily or placebo for 4 weeks (2 people stopped sildenafil and withdrew because of adverse effects). After a 1‑week wash‑out period, people in the trial were then switched to the alternative treatment for a further 4 weeks. Six people had digital ulcers at baseline. The primary outcomes were frequency and duration of Raynaud attacks, capillary flow velocity and evolution of trophic digital lesions.

For people who were taking sildenafil, the frequency of Raynaud attacks was statistically significantly lower (p=0.0064) and the cumulative attack duration statistically significantly shorter (p=0.0038). In all 6 people with secondary Raynaud's phenomenon who had chronic digital ulcerations, trophic lesions began to heal during treatment with sildenafil. In 2 people, ulcerations completely disappeared. The investigators stated that ulcerations reappeared or progressed again after treatment with sildenafil was stopped (no data or analysis reported). Healing of ulceration did not occur during treatment with placebo. The secondary outcome of mean Raynaud's Condition Score, which is commonly used to assess the impact of Raynaud's symptoms on a person's day‑to‑day life (Merkel et al. 2002), was also statistically significantly lower with sildenafil compared with placebo, although it is not clear if this was a clinically significant effect.

Herrick et al. (2011) was a multicentre European study that included 57 people with Raynaud's phenomenon secondary to limited cutaneous systemic sclerosis. People in the study were randomised to receive modified‑release sildenafil (n=30, initially 100 mg once daily for 3 days then increased to 200 mg once daily for 25 days) or placebo (n=27) for 28 days. The per‑protocol population consisted of 45 people (sildenafil, n=20; placebo, n=25) who received all treatments, completed diaries, and did not violate the protocol. The authors reported that modified‑release sildenafil 200 mg once daily is similar to 50 mg of immediate‑release sildenafil 3 times daily. The primary outcome was the change in the number of Raynaud's phenomenon attacks per week. Secondary outcomes included Raynaud's Condition Score, Raynaud's phenomenon pain score and mean duration of Raynaud attacks. Eight people (3 in the placebo group and 5 in the sildenafil group) had digital ulcers at baseline (Herrick et al. 2009).

The primary outcome, the mean percentage change from baseline to week 4 in the number of Raynaud attacks per week, was statistically significantly greater for modified‑release sildenafil than for placebo in the per‑protocol population (–44.0% compared with –18.1% respectively; p=0.034). Although improvement in the mean number of attacks per week was reported in both groups, this was not statistically significantly different between groups (p=0.244). Differences between groups in the secondary outcomes of Raynaud's Condition Score pain score and mean duration of attacks were not statistically significant.

The effect of sildenafil on digital ulceration improvement and healing was an exploratory outcome in Herrick et al. (2011). Some data on digital ulcer healing were reported in an abstract (Herrick et al. 2009). The number of people with digital ulcers improved with modified‑release sildenafil from 5 (25%) at baseline to 4 (20%) at day 14 and to 3 (15%) at day 28. In the placebo group, the number of people with digital ulcers was 3 (12%) at baseline, 4 (16%) at day 14, and 5 (20%) at day 28 (no analysis reported).

The authors of the meta‑analysis (Tingey et al. 2013) reported statistical data extraction and recalculation with 95% confidence intervals (CI) for digital ulcer healing and prevention in the studies they included. They reported non‑statistically significant relative risks (RR) of 5.00 (95% CI 0.29 to 86.43; p=0.27) for digital ulcer healing and 13.00 (95% CI 0.89 to 89.39; p=0.06) for digital ulcer improvement with sildenafil compared with placebo in Fries et al. (2005), and a non‑statistically significant RR of 3.33 (95% CI 0.21 to 52.68; p=0.39) for digital ulcer healing with sildenafil compared with placebo in Herrick et al. (2011).

The meta‑analysis pooled effect for PDE5 inhibitors as a class did show a statistically significant benefit compared with placebo for digital ulcer healing (RR 3.28, 95% CI 1.32 to 8.13, p=0.01) and digital ulcer improvement (RR 4.29, 95% CI 1.73 to 10.66, p=0.002 [Tingey et al. 2013]). However, the pooled effect also included data from a crossover study evaluating tadalafil (n=24, Shenoy et al. 2010) and therefore this result may not be valid when specifically considering the effect of sildenafil for treating digital ulcers.

A small, prospective, single‑centre pilot study evaluated the effect of sildenafil on digital ulcer healing in people with systemic sclerosis (n=19; Brueckner et al. 2010). In this study, 16 people with 49 digital ulcers refractory to standard treatment were treated with maximally tolerated doses of sildenafil (up to 150 mg per day) for up to 6 months. Three people did not receive sildenafil for more than 1 month and were excluded from the analysis. The primary outcome was digital ulcer healing. Secondary end points included effects on Raynaud's phenomenon by using a visual analogue scale (VAS), the prevention of new ulcers, changes in VAS scores for pain, activity and experiencing ulcers, and the occurrence of sildenafil‑related adverse effects. This study did not include a control group and, therefore, is not discussed in detail in this evidence summary. However, it does provide some data on the efficacy and safety of sildenafil as a treatment for digital ulcers, assessed as a primary outcome and over a longer period of time than the small RCTs.

The pilot study reported that 16 people (mean 3.1 ulcers per person at baseline) were treated with a mean sildenafil dose of 114 mg per day for a mean duration of 5.2 months (Brueckner et al. 2010). There was a statistically significant improvement in the number of digital ulcers at the end of sildenafil treatment (mean 1.1 ulcers per person). In all, 49 digital ulcers were present at baseline. During sildenafil treatment, 9 people developed 12 new ulcers between them. At the end of the study 17 ulcers were present (p<0.001 compared to baseline). Secondary outcomes for pain and activity statistically significantly improved during sildenafil treatment (p=0.002 and p=0.05 respectively).

Safety and tolerability

In the meta-analysis (Tingey et al. 2013), the authors report that there were significant serious adverse events with PDE5 inhibitors, although these are not described as such in the individual trials; 7 people from the 3 trials (1 tadalafil study [n=24, Shenoy et al. 2010] and 2 sildenafil studies [n=20, Fries et al. (2005); and n=57, Herrick et al. 2011]) withdrew as a result of potential treatment‑related adverse events, including headaches, myalgia, non‑painful erections, allergic reaction, chest pain, palpitations, and facial oedema.

Across both sildenafil RCTs (Fries et al. 2005 and Herrick et al. 2011) more people in the sildenafil groups experienced adverse events (60% in Fries et al. 2005, and 93% in Herrick et al. 2011) compared with the placebo groups (0% in Fries et al. 2005, and 59% in Herrick et al. 2011). However, the investigators stated that most adverse events were transient (Fries et al. 2005) and mild or moderate and not unexpected for treatment with PDE5 inhibitors (Herrick et al. 2011). The most common adverse event for people taking sildenafil was headache (20% in Fries et al. 2005, and 50% in Herrick et al. 2011). Other less frequent adverse events included dyspepsia, arthralgia and myalgia. No statistical analysis relating to adverse events was reported in either RCT. Two people (10%) in Fries et al. (2005) and 4 people (13%) in Herrick et al. (2011) stopped taking sildenafil because of treatment‑related adverse effects (mainly headache and myalgia). No deaths or serious adverse events were reported in the individual studies and the effect of sildenafil on blood pressure and heart rate was small (Herrick et al. 2011) or non‑significant (Fries et al. 2005).

In the observational study (n=19, Brueckner et al. 2010), mild adverse effects were reported in 47% of participants, which caused 2 people to stop treatment. No statistical analysis was reported on adverse effects but the most commonly reported were palpitations, 21% (4/19); facial flushing, 21% (4/19); peripheral oedema, 16% (3/19); dyspnoea (breathlessness), 5% (1/19) and reoccurrence of atrial fibrillation, 5% (1/19). In 3 people with hypertension, the antihypertensive treatment was reduced during sildenafil treatment.

The summary of product characteristics for the sildenafil product that is licensed for managing pulmonary hypertension (Revatio) advises caution in people with certain underlying conditions which could be adversely affected by sildenafil's mild‑to‑moderate vasodilatory effects (for example, people with hypotension, fluid depletion, severe left ventricular outflow obstruction or autonomic dysfunction). Post‑marketing data for sildenafil for male erectile dysfunction has reported serious adverse cardiovascular events. However, most of these people had pre‑existing cardiovascular risk factors and many events were reported to occur during or shortly after sexual intercourse or occurred shortly after the use of sildenafil without sexual activity. Sildenafil is contraindicated in people with severe hypotension, or a recent history of stroke or myocardial infarction.

See the summaries of product characteristics for licensed sildenafil products for further information on contraindications, potential interactions and adverse effects of sildenafil.

Evidence strengths and limitations

Systemic sclerosis is a rare condition and RCTs are difficult to fund and recruit significant numbers of participants to. Data relating to the effect of sildenafil on digital ulceration are therefore very limited. Large prospective studies comparing sildenafil with placebo or other treatments that specifically consider digital ulcers as a primary outcome have yet to be published (Roustit et al. 2013). Studies with few events, such as those with small numbers of participants or those that run for short periods of time, limit the statistical power of the study to be able to detect differences in efficacy outcomes. As pointed out by the authors of the meta‑analysis (Tingey et al. 2013), the 2 RCTs looking at sildenafil (Fries et al. 2005 and Herrick et al. 2011) were underpowered to detect a statistically significant difference between the sildenafil and placebo groups for digital ulcer healing and improvement and larger studies are needed. Small participant numbers also limit the safety data that these studies can provide. Less common adverse effects could have been missed and in observational studies, such as Brueckner et al. (2010), adverse effects may not have been recorded in detail for all participants.

Other limitations highlighted in the literature include a requirement to define and standardise digital ulcer indicators and end points (Steen et al. 2009). For example, in the 2 RCTs included in this evidence summary (Fries et al. 2005 and Herrick et al. 2011), investigators did not report the classification of partial ulcer healing versus fully healed ulcer, or stipulate methodology for measuring the depth and size of ulcers. The authors of the meta‑analysis reported that these factors could affect the validity of results (Tingey et al. 2013). They also highlighted that randomisation was not stratified according to baseline ulcer presence or absence, which is a potential bias. Disease heterogeneity and baseline characteristics of study populations may have some bearing on treatment effect. For example, most published studies focus on the management of Raynaud's phenomenon as a primary outcome and the study population may include a mix of people with primary or secondary Raynaud's as in Fries et al. (2005). Additionally, smoking is associated with digital vascular disease in people with systemic sclerosis and investigators suggest that this should be taken into account in study design and data analysis (Herrick et al. 2011).

There are other limitations highlighted in the studies considered in this evidence summary. In Fries et al. (2005), blinding was compromised because all participants correctly identified whether they were receiving sildenafil or placebo. This may explain why no adverse effects were reported in the placebo treatment periods. Originally, 20 people were recruited to participate in the study but 2 people discontinued because of treatment‑related adverse effects. The data analysis was, therefore, not an intention-to-treat analysis, and the observed effect estimate on the incomplete data set may be biased.

The 2 RCTs (Fries et al. 2005 and Herrick et al. 2011) used different doses of sildenafil, and data on the effective dose for preventing and healing digital ulcers are limited. The modified‑release sildenafil 100 mg tablet used in Herrick et al. (2011) is not available in the UK.

The uncontrolled study (Brueckner et al. 2010) provides some observational data on maximally tolerated doses and showed a statistically significant benefit of sildenafil for the primary patient‑oriented outcome of ulcer healing. However, all observational studies have limitations inherent in their non‑randomised design, particularly around selection bias. The characteristics of the participants selected for treatment with sildenafil in these studies may have differed from the wider population of people with systemic sclerosis from which they were drawn, which could have affected the outcomes. Other limitations of this observational study are the small numbers of participants and the absence of a control group. In addition, the paper did not provide detailed information on digital ulcer definition and how this was assessed.

Cost effectiveness

No cost‑effectiveness studies were identified that assessed using sildenafil off‑label for managing digital ulcers.

The table below gives costs for some of the drugs used in Raynaud's phenomenon and digital ulcers associated with systemic sclerosis as recommended by the European League against Rheumatism (EULAR) Scleroderma Trials and Research group in 2009 (Kowal-Bielecka et al. 2009). Only bosentan is licensed specifically for preventing digital ulcers in people with systemic sclerosis and pre‑existing ulcers (it has not been shown to heal existing digital ulcers). Iloprost injection is unlicensed in the UK but can be supplied on request from the manufacturer. Iloprost is given by infusion, which involves admission to hospital as either a day case or as an inpatient. None of the other drugs is licensed specifically for use in people with digital ulcers and use of any of these would be off‑label.

The Department of Health has amended regulations to allow unrestricted prescribing of generic sildenafil for men with erectile dysfunction.

Current drug usage

No estimate of the current UK use of off‑label sildenafil for treating digital ulcers was available at the time this evidence summary was prepared.

Expert advisers

Dr Marina Ellen Anderson, Senior Lecturer/Honorary Consultant in Rheumatology, University of Liverpool.

Professor Ian N Bruce, Professor of Rheumatology, Director, NIHR Manchester Musculoskeletal Biomedical Research Unit, Deputy Director, NIHR/Wellcome Trust Manchester Clinical Research Facility, University of Manchester.

Professor Chris Denton, Professor of Experimental Rheumatology, Consultant Rheumatologist and Joint Director of the Centre for Rheumatology, University College London Medical School, Royal Free London NHS Foundation Trust.

Declarations of interest

Dr Anderson has received honoraria and been on advisory boards for Actelion pharmaceuticals. She is also the Merseyside and Cheshire Senate member on the NHS England Specialised Rheumatology Clinical Reference Group and co-authored the 2015 Clinical commissioning policy: sildenafil and bosentan for the treatment of digital ulceration in systemic sclerosis.

Professor Bruce has received consulting fees and research grants from: UCB, GSK, Roche, Medimmune, Genzyme/Sanofi and Pfizer and speaker fees from UCB, Roche and GSK.

Professor Denton has been an investigator in clinical trials of sildenafil and bosentan. He also chairs the UK Scleroderma Study Group and is part of the EULAR Scleroderma recommendations group. He has been a consultant to Pfizer (not currently) and Actelion (current) to advise about management of digital ulcers and Raynaud's phenomenon.

Copyright

© National Institute for Health and Care Excellence 2015. All rights reserved. NICE copyright material can be downloaded for private research and study, and may be reproduced for educational and not-for-profit purposes. No reproduction by or for commercial organisations, or for commercial purposes, is allowed without the written permission of NICE.

ISBN: 978-1-4731-1078-6