Retinal Effects of 6 Months of Daily Use of Tadalafil or Sildenafil FREE

The advent of the orally administered selective phosphodiesterase type 5 (PDE5) inhibitors sildenafil citrate (Viagra; Pfizer, New York, New York), tadalafil (Cialis; Eli Lilly and Company, Indianapolis, Indiana), and vardenafil hydrochloride (Levitra; Bayer Schering Pharma, Berlin, Germany) has profoundly changed the treatment of erectile dysfunction (ED). The phosphodiesterases degrade cyclic nucleotides, including cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP). These 2 cyclic nucleotides are intracellular second messengers that modulate a variety of physiologic functions, including phototransduction in the retina and smooth muscle–cell relaxation in vascular beds.¹ Because PDE5 is the predominant type of phosphodiesterase in the penis, it is the target of selective PDE5 inhibitors in the treatment of ED. Inhibitors of PDE5 cross the blood-retina barrier and can inhibit retina-specific phosphodiesterase type 6 (PDE6), which is structurally similar to PDE5 and is involved in regulating the phototransduction cascade. Off-target inhibition of PDE6 is believed to be the mechanism for blurred vision, blue-tinged vision, and altered perception of light with POE5 inhibitors, which are typically mild and transient.^{1 - 3}

Inhibitors of PDE5 have been prescribed on an as-needed basis.^{4 - 6} Acute administration of sildenafil at high doses has produced only very mild electroretinographic (ERG) changes (most notably a lengthening of b-wave implicit time) that were reversible. Some visual complications of long-term use have been suggested, including nonarteritic anterior ischemic optic neuropathy (NAION), central serous chorioretinopathy, and vascular events.^{7 - 13} However, these have been infrequent and of unclear cause because many men who use PDE5 inhibitors have comorbid diseases that predispose them to these visual complications.

To our knowledge, ERG and retinal changes after steady chronic PDE5-inhibitor use have not been studied. We conducted the present study to assess whether any cumulative or chronic effects not previously recognized with intermittent use would develop during 3 to 6 months of daily dosing of PDE5 inhibitors. This study used a variety of visual tests, including ERG, measurement of intraocular pressure (IOP), and evaluation of visual function and anatomy to assess the visual safety of 5 mg of tadalafil and 50 mg of sildenafil, each compared with placebo, when administered daily for 6 months.

This placebo-controlled, randomized, double-masked study was conducted from August 2005 through August 2006 at 15 US clinics using standardized ERG equipment and protocols. The study was approved by institutional review boards; each subject provided written informed consent. The trial was conducted in accordance with the Declaration of Helsinki, Good Clinical Practice Guidelines, and the Health Insurance Portability and Accountability Act (HIPAA). The study was designed primarily to assess the visual safety of daily use of 5 mg of tadalafil for 6 months. This article concerns data relevant to retinal and ERG effects.

Healthy men or those with mild ED (aged 30-65 years) who had no baseline ophthalmologic abnormalities (or risk factors) were eligible, provided that they had not used any PDE5 inhibitor within 6 weeks of the screening visit or had not discontinued such treatment because of suboptimal efficacy. Exclusion criteria included a diagnosis of diabetes mellitus, elevated IOP (>23 mm Hg) or clinical diagnosis of glaucoma, high myopia (>6 diopters), visual field abnormality (≤−2 dB mean deviation), visual acuity that could not be corrected to 20/20, color vision abnormalities, and severe cataracts in either eye by the Age-Related Eye Disease Study (AREDS) Lens Grading Protocol.¹⁴ Other exclusion criteria included aphakia, clinically significant retinal abnormalities, pupillary dilation of less than 5 mm following instillation of a mydriatic agent, drusen size of more than 63 μm, pigment mottling, and retinal pigment epithelial atrophy of more than 25 μm or detachment. Use of organic nitrates, potent cytochrome P450 inhibitors or inducers, medication associated with retinal toxicity or effects on vision, and allergies to topical medications used to dilate the pupil or anesthetize the cornea were also exclusion criteria.

Any subject with any of the 6 ERG study variables outside age-adjusted normative ranges at randomization was excluded. Normative ERG data were derived by one of the authors (S.G.C.), who tested volunteers with normal ophthalmologic examinations using the same testing equipment and protocol as in the present study (unpublished data).

For this study design (Figure 1), Visits 1 and 2 were screening visits; randomization occurred at Visit 3; Visits 4 and 5 were during-therapy visits at 3 and 6 months of daily dosing, respectively; and Visit 6 was a posttreatment follow-up 4 to 6 weeks after discontinuing the study drug.

At Visit 3, subjects were randomly assigned using computer-generated tables to receive 1 of 3 study drugs: placebo, 5 mg of tadalafil, or 50 mg of sildenafil. Subjects were instructed to take the study drugs on an empty stomach once daily. The 5-mg daily dose of tadalafil was studied because this was the highest dose being evaluated for daily administration for ED, and 50 mg of sildenafil was studied because it is the recommended starting dose of sildenafil for ED. Stratification was used to ensure a 1:1:1 ratio of assignment to each treatment for 3 variables: subject age (<50 vs ≥50 years), investigator site, and the average dark-adapted, combined, standard b-wave amplitude (≤400 vs >400 μV) for both eyes. Study investigators, subjects, and the data analysis team were unaware of the treatment assignment until the data were analyzed. Study drugs were masked by overencapsulation, a widely accepted masking technique in which the study drugs are placed in identical-appearing hard gelatin capsules and the empty spaces filled with an inactive ingredient. Compliance was measured by pill counts of the study drug blister packs at each study visit.

Comprehensive ophthalmologic examination were conducted at baseline, during treatment, and after treatment. Visual function assessments included best-corrected visual acuity using the Retro-Illuminated Modified Ferris-Bailey Early Treatment Diabetic Retinopathy Study Protocol^{15 - 16} ; color discrimination by the Farnsworth-Munsell 100-Hue Color Vision Test¹⁷ ; and static perimetry mean deviation (dB) from age-matched reference ranges by the Humphrey Field Analyzer (Carl Zeiss Meditec Inc, Dublin, California) using 24-2 threshold field test patterns and the Swedish Interactive Thresholding Algorithm standard protocol.^{18 - 22}

The comprehensive ophthalmologic examination also included slitlamp examinations to evaluate the anterior chamber, including any cellular infiltrate or flare, lens opacification and/or cataracts (ie, nuclear sclerosis, cortical spoking, posterior subcapsular) using the AREDS grading protocol, IOP using Goldmann applanation tonometry or Tono-Pen (Reichert Ophthalmic Instruments, Depew, New York),²³ and dilated funduscopy.¹⁴

The ERGs were recorded at all centers on the Espion E²ERG instrument using Dawson-Trick-Litzkow (DTL)–Plus electrodes (both by Diagnosys LLC, Lowell, Massachusetts) and following the International Society for Clinical Electrophysiology of Vision (ISCEV) protocol for full-field (Ganzfeld) ERGs.²⁴ The ERG instrument used a light-emitting diode flash that had a flash-to-flash intensity variability of approximately 1%. The series of tests (eFigure) was sequenced so that ERG testing approximately coincided with peak plasma levels for tadalafil (mean [SD] level, 120 [45] minutes after dose following dark adaptation). One of the authors (S.G.C.) reviewed study ERGs from all study sites to help maintain the quality of the tracings throughout the study and identify those containing artifacts.

Adverse events were summarized using the Medical Dictionary for Regulatory Affairs (MedDRA) preferred term by severity and relationship to study drug. Treatment-emergent adverse events were defined as conditions occurring for the first time or worsening after receiving study drug.

The primary outcome measure was the mean change from baseline to endpoint in the average for both eyes of the dark-adapted combined standard b-wave amplitude (ie, the dark-adapted mixed rod-cone response to a bright ISCEV standard flash) in the tadalafil group compared with the placebo group and in the sildenafil group compared with the placebo group by ERG testing. Secondary outcomes included the mean change from baseline to endpoint in the average for both eyes of other ERG parameters, IOP, and measures of visual function (visual acuity, static perimetry, and color discrimination), as well as the proportion of subjects with changes from baseline to endpoint in ocular anatomy (anterior chamber, lens, and fundus), and proportion of subjects with adverse events.

The minimum clinically relevant mean change in the average for both eyes' dark-adapted, combined, standard b-wave amplitude used in the sample size calculation was a change from baseline of 15% or more. This study was powered to detect an even smaller change from baseline; the sample size calculation demonstrated that 60 subjects per treatment arm would result in 90% (β = .90) power to detect (at 2-tailed α = .05) a 10% change from baseline in the primary outcome. A minimum sample size of 60 subjects per treatment arm also provided sufficient power to detect events occurring in 5% or more of subjects, although not for rare events such as NAION.

Two data analyses specified a priori were conducted to assess the change from baseline to the end of therapy. The first (the Per Protocol Analysis Set) consisted of subjects who completed all 6 months of daily dosing (through Visit 5; ≥175 days taking study drug) and who were treatment compliant. Treatment compliant was defined a priori as having taken 70% or more daily doses of the study drug. The second analysis set (the At Least 3 Months Analysis Set) comprised all subjects who completed 3 or more months of the daily study drug (at least through Visit 4; ≥84 days taking study drug) regardless of compliance.

Continuous variables (including the primary outcome) were analyzed using analysis of covariance (ANCOVA), with the mean change from baseline to the end of therapy in average parameters for both eyes as the response variables. The ANCOVA models also provided Pvalues and 95% confidence intervals for the placebo-adjusted changes from baseline. For the average ERG parameters for both eyes and other retinal function parameters, Pvalues for differences between active treatment and placebo were determined by ANCOVA of the change from baseline that included treatment group, investigator site, age category (<50 vs ≥50 years), and the baseline value of the response variables. Continuous variables were also summarized by arithmetic mean and standard deviation, and categorical data were summarized by counts and percents. Differences between treatment groups in baseline continuous data were evaluated by a type III sum of squares analysis of variance and in baseline categorical data by χ²analysis. The proportion of subjects with ocular anatomic abnormalities was compared using the Fisher exact test.

Two outlier analyses defined a priori were conducted using the Fisher exact test to detect whether thresholds established for 2 ERG parameters were exceeded in individual subjects. The first assessed the number of subjects who developed an average decrease for both eyes of more than 40% from baseline to endpoint or any postbaseline assessment more than 20% below the age-adjusted normative range in the dark-adapted bright flash b-wave amplitude. The second assessed the number of subjects who developed an average increase for both eyes of 5 milliseconds or more from baseline to endpoint, a post-baseline value of 35 milliseconds or more, or a postbaseline increase of 2 milliseconds or more above the age-adjusted baseline normative value for light-adapted 30-Hz flicker implicit time.

Because so many variables were studied, an analysis that adjusted Pvalues for comparison of multiple outcomes (multiplicity) was performed using resampling with replacement.^{25 - 26} The adjusted Pvalues were generated within classes of secondary outcomes: the class of 5 secondary ERG outcomes, of 3 visual function outcomes (color vision, visual acuity, and peripheral vision), and of 3 ocular anatomy outcomes (funduscopy, anterior chamber, cataract lens).

Of 434 subjects screened, a total of 244 subjects were randomized, 212 completed 3 or more months of the daily study drug, and 194 (79.5%) completed the protocol (all 6 months of daily study drug) (Figure 2). Of the 244 total, 82 were randomized to the placebo group, 85 to the tadalafil group, and 77 to the sildenafil group (Figure 2). The treatment groups were similar at baseline in both demographic and clinical characteristics (Table 1).

A total of 9 subjects (2 in the placebo group, 1 in the tadalafil group, and 6 in the sildenafil group) discontinued because they experienced an adverse event. Of these 9 adverse events, only 1 (a retinal artery occlusion in a placebo group subject) was an ophthalmologic event. There were no cases of NAION. Study medications were well tolerated, with the proportion of subjects experiencing treatment-emergent adverse events classified as Eye Disorders (MedDRA System Organ Class) not differing significantly in either the tadalafil or sildenafil group compared with the placebo group. No subject in any treatment group experienced a treatment-emergent adverse event classified as a Neurological Disorder of the Eye (MedDRA High-Level Group Term).

The Per Protocol Analysis Set included subjects most likely to exhibit signs of retinal toxicity should toxicity exist (ie, those who completed all 6 months of daily study drug and who were compliant). The Per Protocol Analysis Set (Figure 2) consisted of 155 subjects (49 placebo, 57 tadalafil, and 49 sildenafil) and comprised subjects who completed the protocol (completed the 6-month evaluation visit on schedule and had ≥175 days taking the study drug) and who were treatment compliant (defined as having taken ≥70% of their daily doses). It excluded subjects from 2 sites who had been entered before investigative site personnel were retrained on ERG electrode placement and subjects from 1 site with an inadequate source documentation.

For the primary endpoint of dark-adapted combined standard b-wave amplitude, the mean change from baseline to endpoint was +4.0% (+13.84 μV) for placebo, +8.5% for tadalafil (+30.95 μV; P = .13 vs placebo), and +1.1% for sildenafil (+3.95 μV; P = .87 vs placebo) (Table 2). For the remaining ERG parameters, only 1 statistically significant difference from placebo in the mean change from baseline to endpoint was observed, for the comparison between the tadalafil and placebo groups for the light-adapted single white flash b-wave amplitude.

To ensure that no drug effects went undetected, we also evaluated a larger population exposed to the study drugs. This set, the “At Least 3 Months Analysis Set,” consisted of 212 subjects (69 placebo, 78 tadalafil, and 65 sildenafil) who completed 3 or more months of the daily study drug regardless of compliance (Figure 2). Data from these subjects (eTable 1) were essentially identical to the Per Protocol Analysis Set, with the following exceptions: statistically significant differences from placebo in the mean changes from baseline to endpoint were observed in the tadalafil group for the light-adapted single white flash b-wave amplitude and in the sildenafil group for the light-adapted 30-Hz flicker implicit time and dark-adapted dim white light b-wave amplitude.

For both the Per Protocol and At Least 3 Months analysis sets, where statistically significant differences compared with placebo were noted, the absolute changes were small and within the test-retest variability range associated with ERG testing. Furthermore, when adjusted for multiplicity, none of the 3 differences in the At Least 3 Months Analysis Set was statistically significant. (Adjustments for multiplicity were not performed in the Per Protocol Analysis Set.)

In addition to the comparison of mean treatment group changes from baseline, outlier analyses were conducted to detect whether thresholds established for 2 ERG parameters were exceeded in individual subjects. No individual subject in any of the study groups in either of the analysis sets exceeded the thresholds established a priori for decreases in the dark-adapted bright flash b-wave amplitude or increases in the light-adapted 30-Hz flicker implicit time.

For the other visual testing, there were no significant changes for either 5 mg of tadalafil or 50 mg of sildenafil when compared with placebo for visual function (visual acuity, static perimetry, color discrimination), IOP, or ocular anatomy (funduscopy, anterior chamber, lens opacities) (ocular anatomy data not presented) in either the Per Protocol Analysis (Table 3) or At Least 3 Months Analysis Set (eTable 2).

Our study demonstrated no evidence of decreased ERG amplitudes or increased implicit times potentially indicative of treatment-related retinal toxicity with the daily administration of 5 mg of tadalafil or 50 mg of sildenafil. There were also no events suggestive of drug toxicity on tests of visual function (visual acuity, static perimetry, color discrimination), ocular anatomy (funduscopy, anterior chamber, lens opacities), or IOP.

There are several reasons ophthalmologists need to be acquainted with the pharmacologic profiles of PDE5 inhibitors and their potential side effects. The frequency of ED, which is a form of peripheral vascular disease that impairs men's abilities to achieve and maintain an erection, increases dramatically with age and in the presence of cardiovascular risk factors. Therefore, many men who take PDE5 inhibitors to treat their ED will also be followed up by ophthalmologists for ocular disorders such as diabetic retinopathy, macular degeneration, and ocular vascular disease. Furthermore, PDE5 inhibitors can exert direct effects on the retina, and such effects probably account for many of the visual side effects such as blue-tinged vision and light sensitivity that have been reported.²⁷

Inhibitors of PDE5 are now being evaluated for indications other than ED. After this study began (June 3, 2005), sildenafil administered 3 times per day (Revatio; Pfizer) was approved to enhance exercise ability in subjects with early-stage pulmonary arterial hypertension.²⁸ In addition, all 3 PDE5 inhibitors are being studied for the treatment of lower urinary tract symptoms associated with benign prostatic hyperplasia. The recommended doses and administration of PDE5 inhibitors for these other indications may be different than for ED. This study was designed to support the prescribing of lower-dose tadalafil on a daily basis not linked to sexual activity. (Tadalafil, 5 mg, was approved for once-daily use in the United States in January 2008 and in the European Union in June 2007, approximately 3 years after the present study was designed.)

This study compared 5 mg of tadalafil with placebo and 50 mg of sildenafil with placebo. The approved doses of sildenafil for the as-needed treatment of ED are 25 mg, 50 mg, and 100 mg. By design, this study was powered to compare each of the active drugs with placebo but was not intended as a study of tadalafil versus sildenafil. The purpose of including sildenafil compared with placebo was to assess whether any visual abnormalities, if present, were unique to the tadalafil molecule or represented a PDE5 inhibitor class effect.

The ERG dark-adapted combined standard b-wave amplitude was chosen as the primary outcome (on which the sample size calculation was based) because it is considered by many retinal specialists to be the most universal ERG endpoint to monitor functions of both rods and cones and the inner and outer retina. Neither 5 mg of tadalafil nor 50 mg of sildenafil taken daily for 6 months showed any clinically important differences compared with placebo for any ERG waveform components studied or any of the other tests of visual function and ocular anatomy.

Other studies, however, have identified ERG changes following acute administration of 100-mg and 200-mg doses of sildenafil near the time of highest plasma concentration.^{29 - 32} The most consistent finding has been a reversible prolongation of b-wave implicit time in the cones (rods may be affected also).²⁹ This is evident in both full-field and multifocal ERG, and in the a-wave of the dark-adapted combined standard response. Amplitude effects have been small and somewhat inconsistent in the literature to date.

Our results indicate that there is no cumulative damage or effect of clinical significance for either 5 mg of tadalafil or 50 mg of sildenafil taken daily for 6 months. Tadalafil binds to PDE6 with an affinity that is 50 to 100 times lower than that of either sildenafil or vardenafil.³³ One might expect the risk of changes to be much lower for tadalafil than sildenafil, given a lower affinity of tadalafil for PDE6 compared with sildenafil. A recent immunohistochemistry study performed on cadaver eyes identified expression of PDE5 in retinal ganglion and bipolar cells as well as in the retinal and choroidal vasculature.³⁴ This raises the possibility that some of the reported ERG effects of PDE5 inhibitors could reflect inner retinal activity of PDE5 rather than (or in addition to) photoreceptor PDE6 effects. However, as noted, there were no recognizable ERG changes from tadalafil or sildenafil under the conditions of the present clinical study.

Our failure to observe any significant effects during chronic administration of sildenafil may reflect 2 factors. First, the dose was lower (50 mg) compared with the aforementioned single-dose studies. Second, ERG recording was timed to the mean peak plasma concentration of tadalafil, which often develops later than that of sildenafil. Tadalafil reaches peak plasma concentrations 120 minutes after a single oral dose in healthy subjects,³⁵ while sildenafil reaches peak plasma concentrations approximately 30 to 150 (mean, 88; median, 60) minutes after a single oral dose in healthy fasting subjects.^{4 ,28}

Our study has other limitations. First, the generalizability of our results to patients with coexisting systemic or ocular disease is limited because our study population included only healthy men or those with mild ED (aged 30-65 years) who had no baseline ophthalmologic abnormalities (or risk factors). Second, although there were no cases of NAION during the study period, because NAION is a rare disease, it should be noted that this study was neither specifically designed nor powered to assess the potential risk of PDE5 inhibitor therapy and rare complications such as NAION.

Our study is notable for several reasons. It is, to our knowledge, the first to assess the visual safety of the PDE5 inhibitors tadalafil and sildenafil when administered once daily and over a prolonged period. Many studies of the visual safety of drugs in humans are conducted in research units using a single dose or multiple doses over a short time frame. This study assessed 6 months of daily dosing as well as a 4 to 6 weeks of follow-up not taking the drug in a multicenter clinical trial setting. It is also one of the first to use standardized ERG equipment, training, normative ranges, and methods across study sites and demonstrates the feasibility of establishing an ERG network for future studies.

Correspondence:Raj K. Maturi, MD, 200 W 103rd St, Ste 1060, Indianapolis, IN 46290 (rmaturi@gmail.com).

Author Contributions:Dr Cordell had complete access to the data and takes responsibility for the integrity of the data as well as the accuracy of the data analysis.

Submitted for Publication:February 27, 2008; final revision received July 31, 2008; accepted August 4, 2008.

The ERG Testing During Chronic PDE5 Inhibitor Administration (ERG-PDE5i) Consortium Investigators:The authors of this article and Neal Adams, MD, Baltimore, Maryland; Michael M. Altaweel, MD, Madison, Wisconsin; Stephen M. Auerbach, MD, Newport Beach, California; Srilaxmi Bearelly, MD, MHS, Durham, North Carolina; David Birch, PhD, Dallas, Texas; James S. Cochran, MD, Dallas, Texas; Mandi D. Conway, MD, Kenner, Louisiana; Raymond A. Costabile, MD, Charlottesville, Virginia; Donnell J. Creel, PhD, Salt Lake City, Utah; Craig F. Donatucci, MD, Durham, North Carolina; Daniel Finkelstein, MD, Baltimore, Maryland; Gary E. Fish, MD, Dallas, Texas; Frank W. Garber, MD, Grand Rapids, Michigan; April E. Harris, MD, Tucson, Arizona; John M. Maggiano, MD, Santa Ana, California; Steven A. Newman, MD, Charlottesville, Virginia; Jonathan D. Nussdorf, MD, New Orleans, Louisiana; Eileen M. Palace, PhD, Metarie, Louisiana; Belur J. Patel, MD, Temple, Texas; Robert S. Rood, MD, Grand Rapids, Michigan; Robert H. Rosa, MD, Temple, Texas; Lawrence J. Singerman, MD, Cleveland, Ohio; Michael P. Teske, MD, Salt Lake City, Utah; James N. Ver Hoeve, PhD, Madison, Wisconsin; Don C. Weiser, MD, Indianapolis, Indiana; Jay M. Young, MD, Laguna Woods, California; and Hernando Zegarra, MD, Cleveland, Ohio.

Financial Disclosure:Drs Cordell, Costigan, and Sides and Ms Klise report being employees of Eli Lilly and Company. Drs Coupland, Danis, Marmor, Maturi, and Weleber report being paid consultants to Eli Lilly and Company. Drs Antoszyk and McGettigan report no financial disclosures.

Funding/Support:This study was supported by Eli Lilly and Company (Bothell, Washington, and Indianapolis, Indiana).

Role of the Sponsor:Eli Lilly and Company (Bothell, Washington, and Indianapolis, Indiana) had a role in the design and conduct of the study, collection, management, and interpretation of data, and the preparation, review, and approval of the manuscript.

Previous Presentation:Some data were presented as abstracts to the European Society of Sexual Medicine; November 26, 2007; Lisbon, Portugal; and the Sexual Medicine Society of North America; December 7, 2007; Chicago, Illinois.

Additional Contributions:The authors wish to acknowledge Marcia Brackman, Jayne Hague, Keith Herring and Gordon M. Berry, Jr, of Eli Lilly and Company and Bruce Doran, Anne Doran, and Richard Robson of Diagnosys, LLC, for their electroretinography expertise and commitment, which helped make this study possible. Assistance in manuscript preparation was provided by Stephen W. Gutkin, Rete Biomedical Communications Corp, Wyckoff, New Jersey.