Free shipping on orders over $149
SafeRxPills - Online Pharmacy
These products are intended for personal use only. We are not licensed by the FDA.
Back to BlogMen's Health

Erectile Dysfunction Treatment When Tablets Don't Work: What Comes Next, by the Evidence

S

SafeRxPills Pharmacy Team

Our pharmacy team consists of certified pharmacists and medical writers with 10+ years of experience in pharmaceutical sciences.

May 18, 202620 min read
Last updated: October 15, 2026
Share:
In this article (21 sections)

Sildenafil, tadalafil and vardenafil help most men with erectile dysfunction, but not all of them. The American Urological Association (AUA) guideline notes that these tablets are ineffective in about 40% of men. If you are one of them, or the tablets have stopped doing what they used to, this guide covers what comes next, in the order most urologists would think about it: first checking whether the tablet was really tried properly, then the treatments that do not depend on a tablet at all.

Every number here comes from a regulator's label, a guideline or a published trial, listed at the end. We sell ED tablets such as Cenforce 100 and Vidalista 20, so we have an interest in tablets working. Where the evidence points away from tablets, we say so. For a general overview of every option, see our ED treatment options guide; this page is for men for whom the first option has not worked.

The short answer

Before giving up on tablets, check how they were used. In one study of 100 men referred as sildenafil "non-responders", 56 had used it wrongly (most never tried 100 mg, or took it right after a meal), and 31 responded once dose and timing were corrected. Other studies found that about half of earlier failures succeeded after instruction. Switching to a different tablet has trial support for vardenafil after sildenafil. If tablets still fail, the next options are alprostadil (an injection into the penis or a pellet in the urethra), a vacuum device, or a penile implant. Testosterone helps only men with proven low testosterone, and even then the effect on erections is modest. Shockwave therapy is still called investigational by the AUA. The AUA says tablets, injections, urethral alprostadil, vacuum devices and implants can all be offered from the start, chosen together with your doctor.

What counts as "tablets not working"

Doctors usually call someone a non-responder only after several correct attempts at the highest suitable dose. A man who took one 50 mg tablet after a large dinner and gave up has not had a fair trial. The AUA guideline says patients should be told that sexual stimulation is needed and that "more than one trial with the medication may be required to establish efficacy". The Viagra and Cialis labels do not set a number of attempts, so treat any fixed number you read online as a rule of thumb from studies, not a label instruction.

A fair trial usually means:

  • the highest dose your doctor considers safe for you (for sildenafil that is 100 mg in the label; see our guide to doses above 100 mg for why higher is not the answer)
  • at least four to eight separate attempts, on different days
  • sexual stimulation after taking the tablet, since these drugs do not cause an erection on their own
  • for sildenafil and vardenafil, avoiding a large or high-fat meal just before the dose, which can delay the effect, and taking it about an hour before sex

Step 1: check how the tablet was taken

This step sounds too simple to matter, but the studies say otherwise. Hatzichristou and colleagues assessed 100 consecutive men referred to a clinic as sildenafil non-responders. Their mean age was 59. They found:

  • 56 of the 100 had used sildenafil inappropriately
  • 45 had never taken the 100 mg dose
  • 32 had taken it on a full stomach, right after a meal
  • 22 had taken it just before sex, without waiting
  • 12 did not know that sexual stimulation was needed

After being shown how to take it and asked to use at least four tablets, 31 of the 100 men responded: 10 on 50 mg and 21 on 100 mg.

A Taiwanese study by Jiann and colleagues found much the same in 60 men who said sildenafil did not work: 36 had tried fewer than four times, 27 had never taken 100 mg and 18 did not know stimulation was needed. Of 41 men who were followed up after instruction and four 100 mg doses, 24 (58.5%) responded. In a review of sildenafil data sets, McCullough reported that 55% of 137 men who had previously failed became successful after re-education, and that the chance of success kept rising with each attempt before levelling off at about eight attempts.

Our guide to taking sildenafil safely sets out the timing, food and dose rules from the label.

Bar chart of response among men who had not responded to sildenafil: 55% of 137 after re-education (McCullough 2002); 58.5% (24 of 41) after rechallenge with instructions and four 100 mg doses (Jiann 2004); 31 of 100 after correct dose and timing (Hatzichristou 2005); 61.8% reported improved erections on vardenafil vs 14.7% on placebo in a 463-man randomised trial (Carson 2004). Studies used different measures of success.
Figure 1. Response after a proper retry or a switch in men labelled sildenafil non-responders. Each bar is a different study with its own definition of success. Sources: McCullough 2002, Jiann 2004, Hatzichristou 2005, Carson 2004.

Step 2: rule out the things that block tablets

Some causes of poor response can be fixed. The AUA guideline asks doctors to measure morning testosterone in every man with ED, because low testosterone is common and treatable. Other things that can weaken the response to tablets include poorly controlled diabetes, heavy drinking, some blood pressure and antidepressant medicines, and anxiety about performance. Our guide to ED causes covers which medicines are linked to ED. Do not stop a prescribed medicine on your own; ask the prescriber whether an alternative exists.

The AUA also says doctors should consider referral to a mental health professional for men with ED, because anxiety, depression and relationship strain can both cause ED and make it harder to treat.

Step 3: try a different tablet

Switching between sildenafil, tadalafil and vardenafil is common practice, but the evidence is uneven. The European Association of Urology (EAU) puts it plainly: switching "may be beneficial. However, no evidence for this has been reported in the available RCTs." There is one good randomised trial for one direction of switch.

SwitchStudyResult
Sildenafil to vardenafilCarson 2004, randomised, 463 men who had not responded to sildenafil, 12 weeksImproved erections in 61.8% on vardenafil vs 14.7% on placebo; normal erectile function scores in 30% vs 6%
Sildenafil or vardenafil to tadalafilOzgur 2009, uncontrolled, 40 menNo significant change in erection scores (11.90 to 12.67, P=0.30)
On-demand to daily tadalafilTrial of 623 men with a poor response to on-demand tablets, as summarised by the AUAAbout 40% reached a normal erectile function score vs 12% on placebo

So a switch to vardenafil (for example Vilitra 10, a generic vardenafil) has the best direct evidence after sildenafil failure, and daily low-dose tadalafil (such as Vidalista 5) is another option your doctor may suggest. For how the three tablets differ in onset and duration, see sildenafil vs tadalafil vs vardenafil.

Step 4: testosterone, but only if it is really low

Testosterone is not an ED treatment for men with normal levels. The AUA defines testosterone deficiency as a total testosterone below 300 ng/dL on at least two morning blood tests, together with symptoms. In men who meet that definition, the evidence is mixed:

  • TRAVERSE (1,161 men aged 45 to 80 with low testosterone and low libido): testosterone gel raised sexual activity and desire compared with placebo, "but not erectile function".
  • Spitzer 2012 (140 men with low testosterone and ED): once the sildenafil dose was optimised, adding testosterone gel was not better than adding placebo.
  • Pooled analyses point the other way: the AUA's own meta-analysis of four trials found that combining testosterone with a PDE5 tablet raised erectile function scores by 2.69 points more than the tablet alone, and an eight-study analysis of 913 men also favoured the combination.

The AUA's position is that men with ED and low testosterone "should be informed that PDE5i may be more effective if combined with testosterone therapy". In practice: get tested, and if testosterone is truly low, treating it may help libido and may help the tablets work a little better. It will not fix ED caused by blood vessel or nerve damage.

Step 5: alprostadil injections

Alprostadil (also called prostaglandin E1) relaxes the smooth muscle and widens the arteries inside the penis. Injected into the side of the penis with a fine needle, it acts directly on the erectile tissue rather than through the same pathway as the tablets, which is why it can work when tablets have failed. US brands are Caverject and Edex.

The first dose is given in a clinic so the doctor can find the lowest dose that works and check for a prolonged erection. In the Edex label's placebo-controlled studies, among men who had responded during this in-clinic dose finding:

  • study 1: erection sufficient for intercourse in 74% (60 of 81) vs 7% on placebo
  • study 2: 73% (103 of 141) vs 13% on placebo

In the Caverject label, over 80% of patients in two studies had an erection sufficient for intercourse, and in a six-month open study 87% of 13,762 home injections led to satisfactory sexual activity. The AUA guideline reports success rates across injection studies ranging from 53.7% to 100%.

Grouped bar chart from US labels. Edex injection study 1: erection sufficient for intercourse 74% (60 of 81) vs 7% placebo. Edex study 2: 73% (103 of 141) vs 13% placebo. MUSE urethral alprostadil home phase: at least one successful intercourse 65% vs 19% placebo. Only men who responded during in-clinic titration entered these phases.
Figure 2. Alprostadil injection and urethral pellet vs placebo in men who responded to a test dose. The two products were tested in separate trials. Sources: Edex and MUSE US labels.

Injection rules and side effects

The labels set firm limits. Caverject and Edex should be used no more than three times a week, with at least 24 hours between doses. Caverject doses above 60 micrograms are not recommended. The target is an erection lasting no more than an hour.

In the Caverject label (1,861 men for up to 18 months), penile pain was the most common side effect at 37%, compared with 2% on placebo. Prolonged erection of 4 to 6 hours occurred in 4%, priapism (over 6 hours) in 0.4%, and penile fibrosis (scarring) in 3% overall and 7.8% in one 18-month self-injection study. The Caverject label also reports needle breakage after marketing, sometimes needing surgery to remove.

The labels list these contraindications (people who should not use injections):

  • conditions that raise the risk of priapism, such as sickle cell anemia or trait, multiple myeloma or leukemia
  • penile deformity, scarring or Peyronie's disease
  • men with a penile implant

An erection lasting more than 4 hours needs emergency care. Untreated priapism can permanently damage the penis.

Bar chart of local adverse reactions with Caverject injection in 1,861 men for up to 18 months: penile pain 37%, prolonged erection of 4 to 6 hours 4%, penile fibrosis 3%, injection site hematoma 3%, injection site ecchymosis 2%, priapism over 6 hours 0.4%. Penile pain with placebo was 2%. Fibrosis reached 7.8% in one 18-month self-injection study.
Figure 3. Local side effects with alprostadil injections in the Caverject label. Fibrosis was 7.8% in one 18-month self-injection study. Source: Caverject prescribing information.

Step 6: alprostadil without a needle

Urethral pellet (MUSE). A small pellet of alprostadil is placed into the opening of the penis with an applicator. In two placebo-controlled trials, 1,511 men started in-clinic testing and 996 (66%) had an erection sufficient for intercourse. At home, 65% of those men had at least one successful intercourse vs 19% on placebo (the NEJM report gives 64.9% vs 18.6%). Penile pain affected 32% vs 3% on placebo, and urethral burning 12% vs 4%. During in-clinic testing, 3% had symptomatic low blood pressure, which is why the first dose is given under supervision. The label allows no more than two doses in 24 hours.

Cream (Vitaros). Alprostadil cream applied to the tip of the penis is licensed in the UK and some other countries but not in the US. The UK summary of product characteristics allows it no more than two to three times a week and once in 24 hours, and lists penile burning, pain and redness as common side effects. It reports that two phase 3 studies showed improvement over placebo but does not give the percentages.

Step 7: vacuum erection devices

A vacuum device is a plastic cylinder placed over the penis. A pump draws blood in, and a ring placed at the base holds the erection. It needs no medicine and can be used by men who cannot take tablets.

A 2026 meta-analysis of 18 studies (1,065 men with hard-to-treat ED, often unresponsive to tablets) found a pooled success rate of 80%. Bruising of the penis was the most common problem, at 24.3%. The AUA notes that 11 of 12 studies reporting patient satisfaction found 60% or more satisfied, and recommends devices with a vacuum limiter only.

Step 8: penile implants

A penile prosthesis is surgically placed inside the penis. The most common type is a three-piece inflatable implant with a small pump in the scrotum. It is the option most often used when nothing else has worked, and the AUA calls it "best conceptualized as irreversible", because the natural erectile tissue is changed by the surgery.

Satisfaction is high. A 2025 meta-analysis of 83 studies and 12,132 men, followed for an average of nearly four years, found that 83% of men were satisfied, with higher rates for three-piece devices. Long-term complications ranged from 3% (erosion) to 4.6% (mechanical failure). The AUA guideline gives mean satisfaction of 86.2% for inflatable implants and 75.1% for malleable (bendable) ones, and reports infection rates of 1% to 2% in most series with infection-resistant coated devices.

OptionBest evidenceMain drawbacks
Retry tablet correctly31% to 58.5% of "non-responders" respond in clinic studiesSame tablet side effects
Switch to vardenafil61.8% improved vs 14.7% placebo (randomised trial)One trial; same drug class
Alprostadil injection73% to 74% vs 7% to 13% placebo in test-dose respondersPenile pain 37%; fibrosis; priapism risk
Urethral alprostadil65% vs 19% placebo in test-dose respondersPenile pain 32%; urethral burning
Vacuum device80% pooled success; authors call for larger randomised trialsBruising 24.3%; cool, hinged erection
Penile implant83% satisfied in a 12,132-man meta-analysisSurgery; irreversible; infection and mechanical failure

The numbers in this table come from different studies with different definitions of success, so they show the size of the evidence for each option, not a ranking.

Shockwave therapy: not yet a standard treatment

Low-intensity shockwave therapy sends sound-wave pulses into the penis, with the aim of growing new blood vessels. It is widely advertised by private clinics. The trials are mixed:

  • A 2017 meta-analysis of 7 sham-controlled trials (602 men) found erectile function scores rose by 6.40 points with shockwaves vs 1.65 with a sham device, with very high variation between trials.
  • A 2019 meta-analysis of 7 trials found a 4.23-point advantage over sham at one month.
  • A 2026 meta-analysis of 12 trials concluded that current evidence "does not support its routine inclusion in ED treatment algorithms".

The AUA guideline says shockwave therapy "should be considered investigational". The EAU is a little more open, giving it a weak recommendation for men with mild ED due to blood flow problems, and for men with blood-flow-related ED who respond poorly to tablets. If you consider it, ask what evidence the clinic uses, how many sessions are planned and what it costs in total.

Weight, exercise and smoking

Lifestyle change sounds like generic advice, but there is a good trial behind it. Esposito and colleagues randomised 110 obese men aged 35 to 55 with ED (and no diabetes, high blood pressure or high cholesterol) to an intensive weight loss and exercise programme or general advice. After two years, BMI fell from 36.9 to 31.2 in the programme group, weekly exercise rose from 48 to 195 minutes, and erectile function scores rose from 13.9 to 17. In the control group scores barely changed (13.5 to 13.6). 17 of 55 men in the programme group (31%) no longer had ED, compared with 3 of 55 (5%) in the control group.

The AUA says lifestyle changes improve overall health and may improve erectile function. The EAU recommends starting them before or alongside other treatment. They also help tablets work better, because the same blood vessel damage underlies both.

Grouped bar chart from a 2-year randomised trial of 110 obese men with ED: 17 of 55 (31%) in the weight loss and exercise group reached an erectile function score of 22 or more, vs 3 of 55 (5%) in the control group. BMI fell from 36.9 to 31.2 in the intervention group.
Figure 4. Lifestyle change and ED in obese men. The intervention group lost weight (BMI 36.9 to 31.2) and raised exercise from 48 to 195 minutes a week. Source: Esposito 2004.

Treatments the guidelines advise against

The AUA guideline lists some treatments that are not recommended or are still experimental:

  • surgery to tie off penile veins (venous ligation): not recommended
  • stem cell therapy: investigational
  • platelet-rich plasma (PRP) injections: experimental

Herbal "natural Viagra" products are another warning area. The US FDA has issued many warnings about sexual enhancement supplements found to contain hidden prescription drug ingredients, often sildenafil or tadalafil. That is dangerous for anyone taking nitrates, which must never be combined with these drugs. Our guide to spotting fake Viagra covers the warning signs.

What the guidelines say about the order of treatment

The AUA guideline, published in 2018, moved away from a strict ladder. Its central statement is that "using the shared decision-making process as a cornerstone for care, all patients should be informed of all treatment modalities that are not contraindicated, regardless of invasiveness or irreversibility, as potential first-line treatments". In other words, a man who prefers an injection or an implant does not have to fail tablets first.

The EAU guideline keeps a clearer order. It recommends PDE5 tablets as first-line treatment (strong recommendation), injections as an alternative first-line or second-line option (strong), and urethral or topical alprostadil as an alternative first-line option in selected men (weak).

Special situations

After prostate cancer surgery. Nerve damage means tablets often work less well, and many men need injections, a vacuum device or eventually an implant. In the 2026 vacuum device meta-analysis, success after prostatectomy was 84.5%.

Diabetes. Men with diabetes respond less well to tablets on average, and in the same vacuum device analysis success was 73.0% in men with diabetes. See our guide to diabetes and ED.

Spinal cord injury. The Caverject label starts men with ED caused purely by nerve damage, such as spinal cord injury, on a lower test dose (1.25 micrograms instead of 2.5).

Younger men. ED in men under 40 can be linked to anxiety, but it can also be an early sign of heart disease. See ED in young men.

Safety: what never changes

  • Never take sildenafil, tadalafil or vardenafil with nitrates (such as nitroglycerin) or "poppers" (amyl nitrite), or with riociguat. The blood pressure drop can be dangerous.
  • Do not combine two ED tablets, or a tablet with an injection, unless your doctor has planned it.
  • Any erection lasting more than 4 hours is an emergency.
  • Sudden vision loss or sudden hearing loss after an ED tablet needs urgent medical attention.
  • If sex brings on chest pain, stop and seek help.

When to see a urologist

See a urologist or sexual medicine specialist if two tablets have failed at the right dose and timing, if you have a curved or painful erection, if you have had pelvic surgery or radiotherapy, or if you want to discuss injections, a vacuum device or an implant. Injections, urethral alprostadil and implants all need a doctor's assessment and, for the first dose or the operation, a clinic.

Where SafeRxPills fits

We sell generic PDE5 tablets made in India, such as Fildena 100, Cenforce 100, Vidalista 20 and Vilitra 10. These are not FDA-approved products. We do not sell injections, urethral pellets, vacuum devices or implants, which need a clinic. If a properly tried tablet has failed for you, a urologist is the right next step, not a higher-strength tablet bought online. For more on generic versions, see our guide on whether generic Viagra works as well as the brand.

Sources

  • Burnett AL et al. Erectile Dysfunction: AUA Guideline. Journal of Urology 2018;200(3):633-641. PMID 29746858. AUA
  • European Association of Urology. Sexual and Reproductive Health guidelines, Management of erectile dysfunction. EAU
  • Hatzichristou D et al. Sildenafil failures may be due to inadequate patient instructions and follow-up. European Urology 2005;47(4):518-22. PMID 15774252.
  • Jiann BP et al. Is failure of sildenafil due to suboptimal use? International Journal of Impotence Research 2004;16(1):64-8. PMID 14963473.
  • McCullough AR et al. Urology 2002;60(2 Suppl 2):28-38. PMID 12414331.
  • Carson CC et al. The efficacy and safety of vardenafil in men with erectile dysfunction caused by sildenafil non-response. BJU International 2004;94(9):1301-9. PMID 15610110.
  • Ozgur BC et al. Urology Journal 2009;6(4):267-71. PMID 20027555.
  • Pencina KM et al. Effect of testosterone replacement therapy on sexual function and hypogonadal symptoms in men with hypogonadism (TRAVERSE). Journal of Clinical Endocrinology and Metabolism 2024;109(2):569-580. PMID 37589949.
  • Spitzer M et al. Annals of Internal Medicine 2012;157(10):681-91. PMID 23165659.
  • Zhu J et al. Translational Andrology and Urology 2020;9(2):591-600. PMID 32420164.
  • CAVERJECT (alprostadil) US prescribing information, Pfizer, revised March 2023. DailyMed
  • EDEX (alprostadil) US prescribing information, Endo, revised March 2024. DailyMed
  • MUSE (alprostadil) urethral system US prescribing information, Meda, revised 5/2024. FDA label PDF
  • Padma-Nathan H et al. Treatment of men with erectile dysfunction with transurethral alprostadil. New England Journal of Medicine 1997;336(1):1-7. PMID 8970933.
  • Vitaros 3 mg/g cream, UK summary of product characteristics. emc
  • Zhang F et al. Vacuum erection devices in refractory erectile dysfunction: meta-analysis. International Journal of Impotence Research 2026;38(2):76-85. PMID 40542251.
  • Corona G et al. Andrology 2025;13(3):610-623. PMID 39010798.
  • Clavijo RI et al. Journal of Sexual Medicine 2017;14(1):27-35. PMID 27986492.
  • Campbell JD et al. Therapeutic Advances in Urology 2019;11:1756287219838364. PMID 30956690.
  • Cayan S et al. World Journal of Men's Health 2026. PMID 41714896.
  • US Food and Drug Administration. Sexual enhancement and energy product notifications. FDA
  • VIAGRA (sildenafil citrate) US prescribing information, Pfizer, Contraindications and Warnings.
  • Esposito K et al. Effect of lifestyle changes on erectile dysfunction in obese men: a randomized controlled trial. JAMA 2004;291(24):2978-84. PMID 15213209.

This guide is for information only and is not medical advice. ED can be a sign of heart disease or diabetes. See a doctor before starting or changing any ED treatment.

?Frequently Asked Questions

What is the next ED treatment if tablets don't work?

First make sure the tablet was tried properly: the highest suitable dose, several attempts on different days, sexual stimulation, and for sildenafil no large or high-fat meal just before the dose. If it still fails, options include switching tablets, alprostadil injections or urethral pellets, a vacuum device, or a penile implant. The AUA says all of these can be considered, chosen with your doctor.

How many times should I try sildenafil before saying it does not work?

The Viagra label does not set a number. The AUA guideline says more than one trial may be needed. Studies of men who thought sildenafil had failed used at least four doses at 100 mg, and one review found success kept rising until about eight attempts.

Does switching from sildenafil to tadalafil help?

There is no randomised trial of that switch in non-responders, and one small uncontrolled study of 40 men found no significant improvement. A randomised trial of 463 sildenafil non-responders did find that vardenafil improved erections in 61.8% vs 14.7% on placebo.

Will testosterone fix erectile dysfunction?

Only men with proven low testosterone (below 300 ng/dL on at least two morning tests, with symptoms) should consider it. In the TRAVERSE trial, testosterone improved sexual desire and activity but not erectile function. Some pooled analyses found it helps tablets work slightly better in men with low levels.

How well do alprostadil injections work?

In the Edex label's placebo-controlled studies, among men who responded to an in-clinic test dose, 73% to 74% had an erection sufficient for intercourse vs 7% to 13% on placebo. The AUA reports success rates of 53.7% to 100% across injection studies.

Do penile injections hurt?

Penile pain was the most common side effect in the Caverject label, reported by 37% of men compared with 2% on placebo. Only 3% stopped Caverject because of pain.

How often can I use alprostadil injections?

The Caverject and Edex labels allow no more than three injections a week, with at least 24 hours between doses. The first dose should be given in a clinic to find the right dose. An erection lasting more than 4 hours needs emergency care.

Is MUSE better than injections?

MUSE avoids a needle, but in its trials 65% of men who passed the clinic test had at least one successful intercourse at home vs 19% on placebo, and penile pain affected 32%. Injections had higher success rates in their own trials. They were not compared head to head.

Do vacuum erection devices work?

A 2026 meta-analysis of 18 studies in men with hard-to-treat ED, often unresponsive to tablets, found a pooled success rate of 80%. Bruising was the most common problem at 24.3%. The AUA recommends devices with a vacuum limiter.

Is shockwave therapy for ED worth it?

The AUA calls it investigational. Meta-analyses show a modest gain over sham treatment in erection scores, but a 2026 analysis of 12 trials concluded the evidence does not support routine use. The EAU gives it a weak recommendation for mild ED caused by poor blood flow.

How satisfied are men with a penile implant?

A 2025 meta-analysis of 12,132 men found 83% were satisfied, with long-term complications of 3% to 4.6%. The AUA reports 86.2% satisfaction with inflatable implants and infection rates of 1% to 2% with infection-resistant coated devices in most series. The surgery is considered irreversible.

S

SafeRxPills Pharmacy Team

Pharmacist, Content, B.Pharm, M.Pharm, Rajiv Gandhi University

Pharmacist writing on antiparasitic, ophthalmic and dermatological generics. Focuses on what the regulatory labelling and published evidence actually support.

Comments (0)

Leave a Comment

No comments yet. Be the first to share your thoughts!