Tadalafil and vardenafil vs sildenafil: a review of patient-preference studies
This article is corrected by:
- Errata: Corrigenda Volume 107, Issue 7, 1166, Article first published online: 25 March 2011
Vincenzo Mirone, Urologic Clinic, University Federico II of Naples, Via S. Pansini, 5-80132 Naples, Italy.
The immediate objective of phosphodiesterase type 5 (PDE5) inhibitor treatment is to restore the ability of a man to achieve and/or maintain an erection adequate for sexual intercourse. As erectile dysfunction (ED) generally develops in the second half of life, the ultimate objective generally is not procreation, but quality of sexual life. Indeed, ED is known to impair quality of life considerably; two-thirds of men report that ED has impaired their self-esteem and nearly a third claim that it has damaged the relationship with their partner. It follows that the therapeutic success of PDE5 inhibition has an important subjective component, which is compounded by the subjective nature and complexity of sexual life in humans. This makes it very difficult for physicians to be certain that they have selected the optimal therapy for a couple, even after a thorough evaluation. The 2007 European Association of Urology Guidelines stress the importance of educating the patient and claim that ‘the patient will choose the final drug after his own experience’. However, PDE5 inhibitors are typically used twice a week, so a patient would have to spend ≈3 months trying the various compounds and dosages to achieve adequate exposure to all three PDE5 inhibitors; this would seem an unrealistic strategy in normal clinical practice. The acknowledgement that the patient has an important role in therapeutic decisions for ED has fuelled interest in the concept of patient preference. It has been established that patient preference depends on three factors, i.e. personal characteristics, e.g. age, duration of ED, frequency and dynamics of sexual relations, and the characteristics of their partners, e.g. age, menopausal status and level of interest in sexual activity and medication profile. Medication features of interest include efficacy in terms of quality of erection, consistency of effects, rapid onset of action, long duration of action, side-effect profile and route of administration; drug costs must also be considered if the medicinal product is not reimbursed.
International Index of Erectile Function
Psychological and Interpersonal Relationship Scales
Erectile Dysfunction Inventory of Treatment Satisfaction.
After explaining why patient preference is an important endpoint for assessing a phosphodiesterase type 5 (PDE5) inhibitor, we review the quality of published patient-preference studies and the data related to the reasons for the preference of one compound over another.
WHY PATIENT PREFERENCE?
The immediate objective of PDE5 inhibitor treatment is to restore the ability of a man to achieve and/or maintain an erection adequate for sexual intercourse. As erectile dysfunction (ED) generally develops in the second half of life, the ultimate objective generally is not procreation, but quality of sexual life. Indeed, ED is known to impair quality of life considerably: two-thirds of men report that ED has impaired their self-esteem and nearly a third claim that it has damaged the relationship with their partner . It follows that the therapeutic success of PDE5 inhibition has an important subjective component, which is compounded by the subjective nature and complexity of sexual life in humans. This makes it very difficult for physicians to be certain that they have selected the optimum therapy for a couple, even after a thorough evaluation .
The 2007 European Association of Urology Guidelines  stress the importance of educating the patient and claim that ‘the patient will choose the final drug after his own experience’. However, Hatzimouratidis and Hatzichristou  point out that PDE5 inhibitors are typically used twice a week, so a patient would have to spend ≈3 months trying the various compounds and dosages to achieve adequate exposure to all three PDE5 inhibitors; this would seem an unrealistic strategy in normal clinical practice.
The acknowledgement that the patient has an important role in therapeutic decisions for ED has fuelled interest in the concept of patient preference. It has been established that patient preference depends on three factors, i.e. personal characteristics, such as age, duration of ED, frequency and dynamics of sexual relations, and the characteristics of their partners, such as age, menopausal status and level of interest in sexual activity and medication profile. Medication features of interest include efficacy in terms of quality of erection, consistency of effects, rapid onset of action, long duration of action, side-effect profile and route of administration; drug costs have to be considered, too, if the medicinal product is not reimbursed .
PREFERENCE STUDIES: METHODOLOGICAL ASPECTS
Three PDE5 inhibitors are currently commercially available; each one is characterized by particular pharmacokinetic features. In the attempt to evaluate which is the most favoured drug, several preference studies have been conducted in the last few years. Unfortunately, several of them have proved to be severely flawed because of methodological pitfalls. Mulhall  made the point that patient preference studies are not the same as the classical efficacy and safety studies, and set forth detailed recommendations for their conduct. The ideal study for preference evaluation should: (i) be randomized, double-blind, crossover; (ii) not include bias by pre-selection of patients (e.g. only nonresponders to previous treatment); (iii) make fair comparisons, adopting equivalent doses of each drug and giving each drug for equal periods; (iv) be based on an intent-to-treat analysis. He also indicated several pitfalls: drug instructions might differ considerably, e.g. because of differences in duration of action, and invalidate the blinding, or might even condition patient preference; period effects might produce improvements over time that bias results in studies with medications that require an adaptation period or training; carry-over effects might produce an overlap in the effects of medications, jeopardizing their comparison, when the wash-out interval between periods is too short. He also raised the issue of how patient preference should be assessed. A single question, such as ‘Which treatment did you prefer?’ is limited, as it does provides no context to the answer. A multi-question preference questionnaire, including questions that are related to putative contributing factors, such as time to onset of action, duration of action, quality of erection and side-effects, is preferable. Several published studies are available that evaluated patient preference for tadalafil or vardenafil vs sildenafil and that meet some, but not all, the requirements listed above. The aim of this review is to establish whether current knowledge enables any conclusive judgement.
We electronically searched Medline (from 1966 to 2007 week 18) and EMBASE (from 1988 to 2007 week 18) using the search string ‘(tadalafil or vardenafil) AND Sildenafil AND clinical trial AND erectile dysfunction’ combined with ‘drug preference OR drug choice OR patient preference OR patient attitude OR patient satisfaction’ and limited to articles in English and referring to humans. The studies identified were stringently assessed according to the Mulhall criteria.
PREFERENCE STUDIES: SELECTION
We identified seven patient-preference studies (Table 1) [5–12], comprising six interventional trials [5–10,12] and one observational trial . All of them compared tadalafil with sildenafil, except two, i.e. a large double-blind study by Rubio-Aurioles et al. that compared vardenafil with sildenafil, and a small open-label study by Tolra et al. that compared sildenafil with vardenafil and tadalafil. Three of the studies were declared as sponsored by Lilly ICOS LLC [7,8,11] and one jointly by Bayer and Glaxo-SmithKline .
Table 1. Overview of PDE5 inhibitor patient-preference studies in patients with ED
1–2 week washout SIL vs TAD
≥3 months ED
SIL naive/insufficient therapy
|SIL 50 mg|
30 min up to 4 h before sex
TAD 20 mg
30 min up to 24 h before sex
|||Open-label, fixed dose,|
1-way cross over,
1-week screening3-week SIL assess
1 week treatment-free washout
6-week TAD initiation
3-week TAD assess
≥3-month history of ED
on SIL treatment at stable dose for 6–24 weeks
Stable heterosexual relationship
|SIL continue 25–100 mg as before (100 mg max 35%)|
TAD 20 mg
|% choosing TAD for extension|
|||Randomized, Double-blind, crossover|
4-armed study drug preference
TAD – SIL or SIL – TAD
DIP TAD – SIL (drug was always TAD)
|Drug comparison, 219|
Instructions comparison, 46
≥3-month history of ED
|SIL 50 mg|
(up to 35% 100 mg) TAD 20 mg × 12 weeks
+ 12-week extension
SIL take 30 min up to 4 h before
sex; TAD take 30 min up to 24 h
Time from dosing to sex
|||Randomized||291 completers||SIL 25–100 mg||TP question|
SIL vs TAD
7–10 day washout
|30 min to 4 h before sex|
TAD 10–20 mg
30 min to 12 h before sex × 12 weeks
Correlate baseline, IIEF,
SEP and PAIRS with PP
SIL vs TAD vs VAR
≥6 months ED
Mild to moderate according to IIEF
|SIL 100 mg|
TAD 20 mg
VAR 20 mg
≥6 times over
|||Pooled data from two|
crossover with 1 week
SIL vs VAR
|931 patients (ITT population)|
4-week treatment periods
Stable heterosexual relationship
≥6 months ED
History/diagnosis of diabetes mellitus,
hypertension and/or hyperlipidaemia
|SIL, 100 mg|
VAR 20 mg
previous 8 weeks
Use of SIL or TAD within the
Need to change to other medication
|data collected on|
4, to patient
5, to partner
N (%) patients
Accompanied by partner
Who informed partner
In the light of the considerations above, the six interventional trials are good, but not ideal, studies, as all of them meet most, but not all of Mulhall’s criteria. A critical aspect was blinding. Three of the five interventional studies [6,8,9] were open-label studies because of the difficulty in conducting a blind study with these compounds, as there are major differences in their instructions on use. The value of the blinded study by Govier et al. is reduced by several other shortcomings; it included patients who were not naïve to PDE5 inhibitor use, and the comparison of the two compounds was unfair, as tadalafil was given at its regular dose, whereas only the starting dose of sildenafil (50 mg) was given to patients, without the option of titrating to the maximum dose of 100 mg; moreover, the duration of treatment was too short (4 weeks).
Thus, the two pivotal patient-preference studies are the double-blind studies by Keitz et al., which compared tadalafil and sildenafil, and by Rubio-Aurioles et al., which compared vardenafil and sildenafil. In these two studies elaborate measures were taken that put patients in conditions that differed considerably from usual clinical practice. The study by von Keitz et al. had an elaborate crossover design; patients were randomized to one of 4 arms (i) tadalafil + tadalafil instructions vs sildenafil + sildenafil instructions; (ii) sildenafil + sildenafil instructions vs tadalafil + tadalafil instructions; (iii) tadalafil + tadalafil instructions vs tadalafil + sildenafil instructions; (iv) tadalafil + sildenafil instructions vs tadalafil + tadalafil instructions.
The comparison between arm 1 and 2 was between drugs, whereas the comparison between arm 3 and 4 was between instructions. This design enabled the study to also address the key issue of the difference in instruction sets by comparing these while always giving the same drug, tadalafil. The only pitfall of this study is that 66% of the patients were not naïve to PDE5 inhibitor.
Double-blind conditions were achieved in the study by Rubio-Aurioles et al. by encapsulating sildenafil tablets, creating a formulation that does not exist on the market. This obliged them to carry out another trial to show the bioequivalence of the encapsulated formulation with the marketed one. This appears to be the only flaw of the study, which met all the main Mulhall criteria for patient-preference studies. However, the patient population only included a particular subgroup of PDE5 inhibitor users, those with cardiovascular risk factors, i.e. diabetes mellitus, hypertension and/or hyperlipidaemia ,
Particular populations were recruited also in the other studies by [6,8,9,11], which are also important, notwithstanding the lack of blinding, for this very reason. Their patient populations included: all patients naive to PDE5 inhibitor [8,9], the general population including the elderly [6,8,9], patients in a non-interventional clinical practice setting , different cultural settings in terms of country (USA/Canada and various European countries, including UK, Sweden, Italy, Spain, Germany) and different healthcare settings (general hospitals, University teaching hospitals, private practice).
A non-interventional setting that closely resembled routine clinical practice is the main strength of the large observational study by Lee et al.. This study enrolled 2425 patients in Canada, and assessed some of the partners, using more detailed questionnaires, which provide more extensive information related to the reasons for preference: the Erectile Dysfunction Inventory of Treatment Satisfaction (EDITS), the Psychological and Interpersonal Relationship Scale (PAIRS), the Drug Attribute Questionnaire, and preference questions put not only to the patient but also to the partner. Such questionnaires were also used in the studies by Eardley et al. and Tolra et al..
Tadalafil vs sildenafil
All the studies comparing tadalafil and sildenafil showed a significant preference for tadalafil than sildenafil (Table 2). The outcome of the study by von Keitz et al. was clear; 73% of patients preferred tadalafil to sildenafil (P < 0.001). The outcome of all the other trials confirmed the results of this study, in that most patients preferred tadalafil to sildenafil. However, the proportion of patients who favoured tadalafil varied considerably, from 52.2% in the study by Tolra et al. to 90.5% in the study by Stroberg et al.. However, the latter trial was conducted in patients who had taken sildenafil for <6 months. As long-term users were excluded, such recruitment could have led to the recruitment of subjects who participated because they were not fully satisfied with sildenafil, skewing the results in favour of tadalafil; in addition, the assessment period was shorter than in the other trials (3, vs 12 weeks in most of the others). The outcome of the four remaining trials [5,8,9,11] was much more similar to the study of von Keitz et al., as the proportion of patients who preferred tadalafil was 66–71%.
Table 2. Preference rates (% of patients) for the seven studies
Vardenafil vs sildenafil
In the study by Tolra et al. vardenafil was the least popular PDE5 inhibitor, as only 20% of patients preferred this drug. However, the sample was too small to provide adequate power, so this preference rate should be considered as a preliminary finding. Its results were contradicted by the large, adequately powered, non-inferiority, double-blind study by Rubio-Aurioles et al., which documented a 38.9% preference rate for vardenafil vs 34.5% for sildenafil (26.6% of patients had no preference).
SECONDARY PREFERENCE OBJECTIVES
Some studies also addressed secondary preference objectives. In the study by Eardley et al., in which patients were asked to specify whether they preferred one drug to the other ‘strongly’ or ‘moderately’, 73% of the patients who preferred tadalafil had a strong preference, vs only 45% amongst those who preferred sildenafil. In the observational trials by Lee et al. physicians rated that 72% of patients allocated to the sildenafil-tadalafil sequence and 61% allocated to the tadalafil-sildenafil sequence preferred tadalafil; 76% of partners of patients allocated to the sildenafil-tadalafil sequence and 65% allocated to the tadalafil-sildenafil sequence preferred tadalafil. This study is also unique in showing how 6–8% of patients refused both sildenafil and tadalafil, and wanted to try another medication, and 2% preferred to stop any ED medication in a setting resembling routine clinical practice.
WHY DID MOST PATIENTS PREFER TADALAFIL?
Factors for preference
The outcome in the study by von Keitz et al. was not influenced by age, country, presence of comorbidities, aetiology and severity of ED, previous treatment with sildenafil, titration of sildenafil dose or treatment sequence. The same was observed in other studies [5,6,8], in which both the pre-study sildenafil dose [5,6,8] and the duration of ED  did not influence the preference rate.
Eardley et al. also conducted a post hoc analysis of the data in an attempt to identify patient characteristics associated with their preference. The only baseline characteristic that was significantly associated with preference was hyperlipidaemia, which increased the preference for sildenafil from 27% to 49% (P = 0.01); no explanation for this finding is given by the authors.
One possibility is that patients preferred tadalafil because of its more flexible instructions for use (drug to be taken 0.5–24 h before sexual activity, vs 0.5–4 h before, for sildenafil) and efficacy not influenced by food intake. This issue was addressed by von Keitz et al., who found that significantly more patients preferred the instructions for tadalafil (67% vs 33%, P = 0.04). Thus, the difference in instructions for use contributes to the preference for tadalafil.
Another possibility is that tadalafil was preferred because it is better tolerated. In their post hoc analysis, Eardley et al. showed that the perception of side-effect severity was associated with preference (P = 0.001). However, the safety profile of the two drugs is comparable according to the Summary of Product Characteristics (Available at: http://www.emea.europa.eu/humandocs/Humans/EPAR/vigra/vigra.htm; http://www.emea.europa.eu/humandocs/Human/EPAR/cialis/cialis.htm). The most common events (≥2% of patients) are similar with both drugs: headache (the most frequent, occurring in 10.8% of patients with sildenafil and 14.5% with tadalafil), dyspepsia, flushing and nasal congestion, dyspepsia being more common with tadalafil (12.3% vs 3.0%) and flushing being more common with sildenafil (10.9% vs 4.1%); in addition, sildenafil is associated with visual alterations (3.6%), whereas tadalafil is associated with back pain (6.5% and myalgia (5.7%).
A third possibility is that tadalafil was preferred because of its efficacy. Two patient-preference studies also collected efficacy information via the International Index of Erectile Function (IIEF) questionnaire [8,9]. In both studies there was a trend towards greater improvement in IIEF scores with tadalafil; in the study by Tolra et al. the median IIEF score increased from 17 at baseline to 29 with sildenafil and to 30 with tadalafil (P = 0.09), and in the study by Eardley et al. the mean difference was 9.6 with sildenafil and 10.1 with tadalafil (P = 0.08). An analysis of the difference in IIEF domain scores showed that the trend in favour of tadalafil was due to significantly greater improvement in orgasmic function (difference, +0.28 95% CI 0.02–0.53, P = 0.03) and sexual desire (+0.19 95% CI 0.02–0.35, P = 0.02). The analysis of the association between the change in IIEF domain scores vs baseline and patient preference showed that greater changes in erectile function, sexual desire, intercourse satisfaction and overall satisfaction were significantly associated with patient preference (all P < 0.001, except sexual desire P = 0.002) . However, such numeric differences might not be regarded as clinically relevant.
The hypothesis that some patients preferred tadalafil to sildenafil because of greater efficacy is better discussed considering the three patient-preference trials in which the patients were asked the reasons for their preference. In the observational trial by Lee et al. physicians provided reasons for patient preference. According to them, the main reason for preferring tadalafil was either its longer duration of action (45%) or greater efficacy (better erections) (34–37%), but on the other side, greater efficacy (better erections in 64%) was also the main reason for preferring sildenafil. In the study by Tolra et al. patients provided the reasons for their preferences directly: the main reason for preferring tadalafil was the awareness of the opportunity of being able to have a second intercourse the following day, with the same quality of erection (48.9%), or achieving a more intense and longer-lasting erection (23.4%). In the minority of patients who did not prefer tadalafil, the main reason for preferring sildenafil and vardenafil was the achievement of a more intense and longer-lasting erection (84% and 88.9%, respectively). Tolra et al. also showed that the median EDITS score was significantly higher after tadalafil than after sildenafil (41 vs 38, P = 0.01).
The PAIRS questionnaire was administered in the studies by Eardley et al. and Dean et al.. Tadalafil improved several PAIRS domains to a significantly greater extent vs baseline than did sildenafil: sexual self-confidence (+0.17 95% CI 0.09–0.25, P < 0.05), spontaneity (+0.15, 0.09–0.22, P < 0.05) and, especially, time concerns (+0.39, 0.32–0.46, P < 0.05) ; greater changes in these PAIRS domain scores were significantly associated with patient preference (P < 0.001) . Dean et al. also evaluated the Drug Attribute Questionnaire, which asked what was the best and second-best reason for the patient’s preference. The main reason for preferring tadalafil (67% of patients) was the ability to obtain an erection, with sexual stimulation, long after the drug was taken (e.g. 16 or 24 h later), followed by the firmness of the erection (35.4%) and the erections that occurred the following morning (30.1%), whereas the main reasons for preferring sildenafil were firmness of erection (60%), early onset of action (41%) and reliable efficacy after every stimulation (40%). The 150 patients who had justified their preference for tadalafil, and the 12 who had justified their preference for sildenafil by mentioning the possibility of having an erection the following day were then asked to choose a statement that explained why this was so important. Nearly half (46%) of the patients chose the following statement: ‘There was less need to plan when my partner and I would have sex’ whereas nearly a third (33%) chose ‘I could take the pill and not worry about when my partner and I might have sex’, and a fifth (21%) ‘It made me feel that my sexual function was like it used to be before I got ED’.
None of the currently available patient-preference studies on PDE5 inhibitors are completely devoid of methodological flaws, so further trials are required. Of the 7 trials conducted so far according to Mulhall criteria, five comparing tadalafil with sildenafil consistently reported that patients preferred tadalafil to sildenafil, mainly because of the longer duration of action that increases the patient’s freedom in sexual life. Interestingly, although efficacy measures failed to detect any clinically relevant difference, in patient perception the efficacy of the two drugs might differ, and this might contribute to explaining patient-preference results for either tadalafil or, in a minority of cases, sildenafil. These observations further outline the need to seriously consider patient’s and partner’s wishes and preferences, to achieve the best clinical results. Finally, one study indicated that patients have no particular preference between vardenafil and sildenafil, and another suggested that tadalafil might also be preferred to vardenafil.
The authors thank Dr Jennifer Hartwig for her contribution to the writing and linguistic revisions of this paper and Filippo Aglietti (Eli Lilly Italy) for his invaluable help in articles selection and retrieval.
CONFLICT OF INTEREST
Riccardo Sicuteri and Andrea Rossi are full-time employees of Eli Lilly Italy. Francesco Montorsi is a speaker for Eli Lilly Italy and a paid consultant for Pfizer, Pierre Fabre Medicament, AMS, Mipharm and Bayer Schering.