Skip navigation

PDE-5 inhibitors - non-RCT evidence

Clinical bottom line

There are very considerable amounts of information available in the form of observational studies. For sildenafil, the first in class, there were more than 33,000 men in observational studies, compared with under 4,000 in randomised trials.


Evidence from systematic reviews of good quality randomised trials may be the highest form of evidence, but it is not the only evidence available to us. There may be circumstances in which randomisation is not ethical, for instance when efficacy has been established in general terms, but where evidence in special circumstances needs to be assessed. Another example is for ascertainment of rare but serious adverse events, where even the most comprehensive collection of randomised trials could lack sufficient numbers of patients or events.


To examine the amount of evidence, PubMed was searched for original studies of two main types - randomised studies not included in the systematic review for some reason, and for any cohort or cross sectional studies - involving sildenafil, tadalafil, and vardenafil published between 2000 and 2005. The results are shown in the tables below.

Randomised studies not included in systematic reviews

Ten sildenafil studies with 561 patients were not included in a recent systematic review (Table 1). All were randomised, and three were double blind. One was a withdrawal design demonstrating that beneficial effects of sildenafil were lost on stopping the drug, one in Parkinson's disease where hypotension stopped the study, and one crossover study in congestive heart failure. The other studies were either declared to be open, or were not convincingly double blind.

Table 1: Randomised studies with sildenafil not included in a recent systematic review

Study Type Reference Condition Number of men Outcome Country
R, DB, withdrawal study Christiansen E et al.
Int J Impotence Res 2000 12:177-182
ED treated with sildenafil 99 Improvements maintained with continued treatment, lost with withdrawal UK, Norway, France
R, DB, crossover Hussain IF et al.
J Neurol Neurosurg Psychiatry 2001 71:371-374
Parkinsonism (Parkinson's disease or multiple system atrophy) 18 IIEF scores improved, but hypotension stopped study in multiple system atrophy UK
R, DB, crossover Bocchi EA et al.
Circulation 2002 106:1097-1103
Congestive heart failure 23 Single 50 mg dose decreased HR and BP on standing, increased exerecise capacity Brazil
R, crossover, not effectively DB Webster LJ et al.
Arch Intern Med 2004 164:514-520
Congestive heart failure 35 IIEF, QoL and depression scores improved. No symptomatic hypotension Canada
R, open Perimenis P et al.
Int J Impotence Res 2004 16:256-260
ED 15 IIEF scores improved more with sildenafil than continuous positive airways pressure, but less than sildenafil in the general population Greece
R Mancini M et al
Int J Impot Res 2004 16:8-12
vasulogenic and nonvasculogenic
22 IIEF scores increased for both sildenafil and alprostadil, not placebo. Cavernosous arterial insufficiency improved for vasculogenic patients only Italy
R, open, crossover Perimenis P et al.
Int J Impot Res 2004 16:2-7
ED arteriogenic 43 Sildenafil better than apomorphine for successful attempts and satisfaction Greece
R, open, crossover Eardley I et al.
BJU International 2004 93:1271-1275
ED 139 IIEF and satisfaction better for sildenafil than apomorphine UK
R, open, parallel Ugarte F et al.
Int J Impot Res 2002 14:S48-S53
ED 123 IIEF scores and success better for sildenafil than phentolamine Mexico
R, crossover Lammers PI et al.
Int J Impot Res 2002 14:54-60
ED 44 No sig diff in efficacy for sildenafil and combinations of apomorphine, phentolamine and paparavine Mexico
R= randomised; DB = double blind; ED = erectile dysfunction; IIEF = international index of erectile function


These studies from Europe and central and south America, were mostly small, with seven having fewer than 50 patients. Two of the larger and two of the smaller studies compared sildenafil directly with apomorphine or phentolamine, and found sildenafil to be more effective.

One study was found for tadalafil (Table 2), with 4,262 patients in a European study examining dosing preferences. No such studies were found for vardenafil.

Table 2: Randomised study with tadalafil excluded from the systematic review, and cohort studies with tadalafil and vardenafil

Study Type Reference Condition Number of men Outcome Country
R, open, crossover Mirone V et al
Eur Urol 2005 47:846-854.
ED 4262 More preferred on-demand than 3times a week dosing Europe
Prospective cohort Del Popolo et al.
Spinal Cord 2004 42:643-648.
Spinal cord injury 28 Tadalafil effecive Italy
Prospective cohort Montorsi F et al.
Eur Urol 2004 45:339-344.
ED 1173 Open-label extension of RCT, with 42% completing 24 months. AE discontinuations 6% Europe
Prospective cohort Potempa AJ.
Eur Urol 2004 46:73-79.
ED 398 Titration study. 80-90% effective, with low asverse event rates Germany
Prospective cohort Stief C et al.
In J Clin Pract 2004 58:230-239.
ED 566 Extension of RCTs. Good sustained efficacy Germany
R= randomised; DB = double blind; ED = erectile dysfunction; IIEF = international index of erectile function


Cohort or cross-sectional studies

Twenty-five sildenafil studies with 33,708 patients were found (Table 3). There were 17 prospective cohort studies with 3,668 men, six retrospective cohorts with 29,352 men, and two cross-sectional studies with 127 men.

The prospective cohort studies were often large, and nine of them had more than 100 men. Almost all of them had measures of improvement in erectile function, most often using the standard IIEF outcomes. Improvement rates were typically above 70%, consistent with findings from the randomised trials.

With one exception, the retrospective studies had information on over 500 men each, with one having over 22,000 men. This UK study was designed to look for adverse events following the introduction of sildenafil in the UK, and is described in detail later.

Table 3: Prospective and retrospective cohort studies of sildenafil

Reference Condition Number of men Outcome Country
Prospective cohort
Sunwoo S et al.
Int J Impot Res 2005 17:71-75
ED 572 80% reported improved erections with sildenafil, while six disontinued. 71 adverse events among 56 patients Korea
Arandda P et al.
Am J Hypertens 2004 17:139-145
Hypertension 291 IIEF scores improved Spain
Fujisawa M et al.
Arch Androl 2004 255-260
ED 44 Rate of improvement higher in youinger than older men with siildenafil Japan
Sahin Y et al.
Transplant Proc 2004 36:56-58
End-stage renal failure on haemodialysis 55 Overall response rate 75% with sildenafil 50 mg Turkey
Benchekroun A et al.
Int J Impot Res 2003 15 Suppl1:S18-24
ED 71 Dose escalation 25-100 mg, showed good efficacy and improve quality of life Morocco
Fujisawa M et al.
Arch Androl 2002 48:15-21
ED 40 QoL scores improved. Correlation with ED response Japan
Sairam K et al.
BMC Urology 2002 2:4
ED 147 90% successful, 80% willing to continue. High incidence of some adverse events UK
Souverein PC et al.
Int J Impot Res 2002 14:259-265
ED 317 Over 18 months 45% discont sildenafil. Factors associated with discont identified. Pts with history of prev drug treatment more likely to switch Netherlands
Yenicerioglu Y et al.
BJU Int 2002 90:442-443
End-stage renal disease 41 IIEF and QoL scores improved Turkey
Raffaele R et al.
Eur Urol 2002 41:382-386
Parkinson's disease 33 Improved erections in 84%. Improvement in depressive symptoms in 75% Italy
Sters W et al.
In J Impot Res 2001 13:261-267
ED, extension studies of RCTs 1008 90% improved erections over 36-52 weeks. Disocntinuations due to adverse events low, and cardiovascular adverse events below 1% USA
Sanchez Ramos A et al.
Spinal Cord 2001 39:637-643.
Spinal cord injuries 170 88% had improved erections Spain
Gans WH et al.
J Spinal Cord Med 2001 24:35-40
Spinal cord injuries 17 Significant improvement in erections. 1 discontinued USA
Palumbo S et al.
Eur Urol 2001 40:176-180
ED 380 77% response overall Italy
Fagelman E et al.
Urology 2001 57:1141-1144
ED 164 Improvement in 68%-81% USA
McMahon CG et al.
J Urol 2000 164:1192-1196
27% psychogenic
73% organic
278 Erectile response and QoL scores improved Australia
Green BG, Martin S.
NeuroRehabilitation 2000 15:101-105
ED due to SCI or MS 40 Initial response: 36/40 improved erections, 15/40 complete intercourse. At 2 years 13/40 discontinued USA
Retrospective cohort
Griffiths L et al
J Clin Pharm Ther 2005 30:297-304
ED 513 78% prescibed for in accordance with Gov guidelines. Varaition between surgeries UK
Boshier A et al
BJU Int 2004 93:796-801
sildenafil prescribed 22473 Effective in 71%. Safety profile showed no unexpected events. No increase in deaths UK
Russo D et al.
J Nephrol 2004 17:291-295
Kidney transplant patients 20 IIEF scores improved Italy
Sheu JY et al.
J Chin Med Assoc 2003 66:480-486.
sildenafil prescribed 3168 Telephone interview of about half showed good efficacy, with high disontinuation because of lack of efficacy Taiwan
Jiann B-P et al.
Int J Impot Res 2003 15:412-417
ED and prescirbed sildenafil 1658 Information on usage, success rates, doses, effect of comorbidities China
Lim PHC et al.
Int J Urol 2002 9:308-315
ED 1520 83% had satisfactory erections. Differences in efficacy between races and comorbidities Singapore
Monga M et al.
Urology 2002 59:753-757
older men 73 ED prevalence inc with age. 660/976 men sexually active, 73 had used sildenafil at some time USA
Muller MJ, Benkert O
J Affect Disrord 2001 66:255-261
ED 54 Scores for depression better with sildenafil than no treatment Germany


For tadalafil there were two prospective cohorts with 1201 men, and for vardenafil two more cohorts with 964 men (Table 2). In each case one of these was an open-label extension of randomised trials.


Thus for cohorts studies alone, there were almost ten times as many patients on sildenafil (33,708) as in the included randomised trials (under 4,000 in a recent systematic review). For tadalafil and vardenafil there was information from fewer men in cohort studies than in randomised trials (Figure 1), which is likely to increase as more randomised trials are published with these newer PDE-5 inhibitors.

Figure 1: Numbers of men in randomised trials in the systematic review, and in cohort studies

What this shows, interestingly, is that for a new treatment in an area where there was previously few effective treatments available, we might expect to see a considerable amount of observational information published, which can be useful and probably should not be overlooked.