Is it worth going through tests at all, or can I just try a tablet and see?
The sequence that produces an erection, and the four points it breaks
Six steps in order, and the tablets only touch one of them.
Anchor figures used below, with the sources named in the text
Any degree of difficulty, ages 40 to 7052 percent
Complete difficulty at 40, and at 705 percent rising to 15
Attributable to medicationUp to 25 percent of cases
Lead time before coronary symptoms2 to 3 years
Penile artery diameter1 to 2 mm against 3 to 4 mm coronary
Exercise benchmark before treatment2 flights of stairs in 10 seconds
Testosterone threshold with symptomsUnder 300 ng/dL on two morning samples
Response to on-demand oral treatmentAround 60 to 70 percent
Sexual stimulation releases nitric oxide from parasympathetic nerve endings and from the endothelium of the cavernosal tissue. Nitric oxide activates guanylyl cyclase, cyclic GMP rises, smooth muscle in the arteries and sinusoids relaxes, inflow increases sharply, and the expanding sinusoids compress the emissary veins against the tunica albuginea so the blood stays put. That last step, veno-occlusion, is passive and mechanical, and it is the one most people have never heard of.
Four things have to be present: a signal, healthy endothelium, arteries wide enough to carry the flow, and a tunica that can hold it. Desire and testosterone sit upstream, deciding whether the sequence starts.
Every PDE5 inhibitor acts at a single point in that chain, slowing the breakdown of cyclic GMP to amplify a signal already being sent. Where the nerve supply is destroyed or the arteries cannot deliver, the amplifier has less to work with, which is the whole explanation for why these drugs work brilliantly in some men and barely at all in others.
Which point has failed is what the assessment is for.
Exercise tolerance is established before treatment in every risk category
The threshold is a specific physical task, not an impression of general fitness.
The Princeton consensus process, now in its fourth iteration, sets the benchmark used in clinics. About 30 minutes of sexual activity with a usual partner corresponds to roughly 2 to 3 metabolic equivalents, the same as walking a mile on the flat in 20 minutes or briskly climbing two flights of stairs in 10 seconds without cardiac symptoms.
More vigorous activity in younger couples reaches 5 to 6 METs, about four minutes of a standard Bruce protocol treadmill test. Completing that without ischaemia means sexual activity is generally safe.
The panel is explicit that exercise tolerance should be established in every man before treatment, whatever his cardiovascular risk category. A man who cannot manage the stairs, or gets chest tightness doing it, needs a stress test before a prescription.
It takes one question, which is the argument for asking every time. The other half of that conversation is nitrates: a man at low risk carrying a spray he has not used in two years may not need it any more, and that is a decision for his cardiologist.
Fifty-two percent is the quoted number and the breakdown is the useful part
One study from the late 1980s still supplies the number that appears in every consultation.
The figure comes from the Massachusetts Male Aging Study, a community sample of 1,290 men aged 40 to 70 surveyed around Boston between 1987 and 1989. Combined prevalence of minimal, moderate and complete difficulty was 52 percent: 17 percent minimal, 25 percent moderate, 10 percent complete.
The age gradient is the part worth carrying. Probability of any degree ran 39 percent at age 40 and 67 percent at 70, while complete difficulty tripled from 5 percent to 15 across the same span. The popular reading, that half of men over 40 cannot have sex, is wrong in both directions.
After adjustment for age, difficulty was more likely with heart disease, hypertension, diabetes, the medicines used to treat them, and measured indexes of anger and depression. Smoking raised the probability of complete failure specifically in men who already had heart disease or hypertension. That list, published in 1994, is still essentially the modern causal model.
The artery that fails first is one to two millimetres wide
Erectile tissue is the smallest arterial bed in the body that anyone notices failing.
Penile arteries measure 1 to 2 mm across. Coronary arteries measure 3 to 4 mm, carotids 5 to 7, femorals 6 to 8. The same burden of endothelial disease narrows the smallest vessel to the point of limiting flow first, which is why erectile difficulty is so often the earliest symptom of a systemic vascular problem.
The interval has been measured. Erectile symptoms precede coronary symptoms by about 2 to 3 years and a cardiovascular event by 3 to 5. Among 300 men with clinically significant coronary disease, around two-thirds had erectile difficulty first, with roughly a three-year gap.
Severity tracks severity, too: worse erectile function correlates with worse coronary disease. A man in his fifties presenting with gradual, non-situational loss of rigidity has handed his clinician a cardiovascular screening opportunity with two or three years of warning attached, and a prescription written without measuring blood pressure, glucose and lipids wastes it.
The tablet, meanwhile, will often work. Vascular disease bad enough to cause symptoms is rarely bad enough to stop a PDE5 inhibitor working, which is exactly why the warning gets missed.
Where the fault is neurological, the arithmetic changes
An amplifier is only as good as the signal reaching it.
Diabetes damages the small vessels and the autonomic nerves together, which is why it produces the most treatment-resistant version of this problem. The trial data prices the penalty: in a once-daily tadalafil study in 298 men with type 1 or type 2 diabetes, endpoint success was 46 percent on 2.5 mg and 41 percent on 5 mg against 28 percent on placebo, where the general population studies show much wider gaps. Real benefit, clearly smaller, better said before treatment than after a disappointment.
Radical pelvic surgery is the other large category. Cavernous nerve injury during prostatectomy removes the signal these drugs exist to amplify, and response depends on whether the nerve bundles were spared and how much time has passed. Spinal injury, multiple sclerosis and pelvic radiotherapy sit alongside it.
The history gives these away for free. Onset dated to an operation is a different problem from a gradual decline over three years, and neither needs a test to be recognised. In the same 3,151-man series, neurogenic causes made up 4.1 percent and anatomical causes 2.8 percent, concentrated in men who already know why.
One man in four here is describing a side effect
The highest-yield document in the consultation is the repeat prescription list.
Classes worth checking against the date the symptom started
- 5-alpha reductase inhibitors: finasteride and dutasteride, by a wide margin the most reported.
- Psychotropics: tricyclics and benzodiazepines carry the strongest population signal; SSRI effects fall mainly on desire and orgasm.
- Antihypertensives: thiazides and older non-selective beta-blockers, with the thiazide evidence weaker than its reputation.
- Spironolactone through antiandrogen effects; antipsychotics through dopamine blockade and raised prolactin; opioids through suppression of the hypothalamic-pituitary-gonadal axis.
Medication accounts for up to 25 percent of cases and that share is frequently reversible, which makes the medicine list the highest-yield item in the room. It is also the one most often skipped, because a repeat prescription reads like background rather than a suspect.
An analysis of ten years of FDA adverse event reports sharpens the pattern. Of the reports naming erectile difficulty, 5-alpha reductase inhibitors accounted for 2,823, about 46 percent, and neuropsychiatric drugs for 2,442, about 40 percent. Finasteride had much the highest proportional reporting ratio at 110, dutasteride 9.4.
Reporting databases exaggerate what patients already suspect, so population data is a useful corrective. In the Boston Area Community Health survey, tricyclic antidepressants carried an adjusted odds ratio of 3.35 and benzodiazepines 2.34, while the association with SSRIs and atypical antipsychotics disappeared once other factors were accounted for. Thiazides have been blamed since early trials suggesting up to a 2.4-fold risk, and larger studies have not confirmed it.
Timing settles most of these cases without statistics. If rigidity fell away within weeks of a new tablet, that tablet is the first hypothesis and the fix is a conversation with whoever prescribed it. None of this is an instruction to stop a cardiac or psychiatric medicine unilaterally.
Preserved morning erections change the whole assessment
Three questions sort most men into a category, and none needs a laboratory.
Nocturnal erections happen in REM sleep and the normal pattern is well characterised: four to five episodes a night, mean duration around 30 minutes, rigidity above 70 percent on a strain gauge. Intact nocturnal tumescence excludes severe organic disease, not a mild organic contribution.
Almost nobody needs the formal test. The clinical version is the question about morning erections, and with two others it does most of the sorting. Sudden, datable onset, difficulty present with one partner and absent with another or absent during masturbation, and preserved morning rigidity together describe a psychogenic pattern. Gradual onset over years, consistent everywhere, with morning erections fading in parallel, describes an organic one.
Psychogenic causes are not a small category, and treating them as a diagnosis of exclusion is a mistake. In that 3,151-man series they were the largest group at 59.2 percent, with a median symptom score of 15 against 12 for organic causes.
Performance anxiety runs on a mechanical loop. A man takes a tablet, monitors himself for an effect, and the attempt fails because attention has moved away from arousal; that failure then becomes the evidence for the next attempt. Naming it in the consultation does more than any dose adjustment, and it is a genuine reason to prefer a daily regimen over a tablet taken with an eye on the clock.
Four blood tests earn their place and the rest are ordered from habit
Most first-visit test panels in this field are longer than the evidence supports.
The guideline position is unusually direct: with the possible exception of serum testosterone, glucose or HbA1c, and in some cases lipids, no routine serum study is likely to alter management. Everything beyond that is either hunting a comorbidity or filling a page.
Physical examination still matters, for specific things rather than as a ritual. Blood pressure and waist circumference, because of what the artery-size argument implies. The penis itself, for a palpable plaque or curvature, with the caveat that a plaque alone is not proof of clinically relevant deformity. Testicular size and secondary sexual characteristics where a hormonal cause is on the table. A digital rectal examination is not required to evaluate erectile difficulty, though prostate symptoms often coexist and deserve their own assessment.
Specialised investigations belong to a small minority. Penile duplex ultrasound, nocturnal tumescence monitoring and cavernosography answer questions raised by young men with a traumatic history, medico-legal work, and Peyronie's disease before surgery. Ordering them at first presentation delays a prescription that would probably have worked.
One examination finding gets over-read routinely. The bulbocavernosus reflex is absent in up to 30 percent of entirely normal men, so its absence proves nothing.
The five-item score grades severity and cannot name the cause
A score of 21 or below is the threshold that prompts a proper look.
The International Index of Erectile Function short form, also called the SHIM, asks five questions about confidence, rigidity sufficient for penetration, maintaining an erection after penetration, difficulty maintaining it to completion, and satisfaction with attempts. Each scores 1 to 5, giving a total between 5 and 25.
The bands are fixed: 22 to 25 no dysfunction, 17 to 21 mild, 12 to 16 mild to moderate, 8 to 11 moderate, 5 to 7 severe. Guidelines recommend a validated questionnaire for grading severity and for measuring whether treatment works.
What it cannot do is identify a cause, and that is where it gets misused. It does not separate vascular from neurogenic from hormonal from drug-related, it relies entirely on recall, and it assumes recent sexual activity, which makes it hard to complete honestly for a man who has been avoiding sex for a year. Those men often have the most to gain.
Its best use is as a baseline. A score written down before treatment and again at review turns a vague answer about whether things are better into a comparable number.
What happens between the assessment and the first tablet
Assessment finds things. What gets done about them separates a consultation from a prescription.
Same day
Erection past four hours. Chest pain during or after sex. Sudden vision or hearing loss on a PDE5 inhibitor.
First appointment
Medicine list with start dates, blood pressure, exercise tolerance, morning erections, desire separated from rigidity.
Within the month
Glucose or HbA1c, lipids where indicated, morning total testosterone, repeated if the first is under 300 ng/dL.
Before a prescription
Nitrates and riociguat excluded, any culprit medicine reviewed with its prescriber, deficiency treated on its own terms.
At review
The same five-item score again, so improvement is a number rather than an impression.
The gap between finishing the assessment and writing the prescription is where treatable causes get treated. A culprit medicine gets reviewed with whoever prescribed it. Testosterone deficiency confirmed on two morning samples gets managed on its own terms. Blood pressure, glucose and lipids found along the way get managed because they were found, whatever happens to the erections. Some of those men then need no tablet, and some of the rest respond better for having had the underlying problem addressed first.
When an oral drug is appropriate, the class works in roughly 60 to 70 percent of men on demand and 30 to 35 percent fail the first agent tried. Which molecule fits depends on timing, meals, the medicine list and one heart-rhythm question, all set out in the class comparison.
A handful of findings outrank the whole sequence. An erection lasting more than four hours is an emergency, because ischaemic priapism damages tissue within hours. Chest pain during or after sex is chest pain. Sudden loss of vision or hearing on a drug in this class means stopping the tablet and being seen the same day. New penile curvature with pain is Peyronie's disease and has its own pathway.
Last Updated
Mailbag
What readers asked
Answered by Dr. Marc Vidal, MD · Internal medicine and clinical pharmacology
Questions about causes arrive in two forms: men who want to skip the assessment, and men who have been given a test result without an explanation. These replies are general education and address the question rather than the person.
You can try a tablet and it will often work, which is precisely the argument for the tests. Vascular disease bad enough to cause erectile difficulty is usually not bad enough to stop a PDE5 inhibitor working, so a good result tells you nothing about your arteries. Erectile symptoms precede coronary symptoms by about two to three years and a cardiovascular event by three to five, and penile arteries are 1 to 2 mm wide against 3 to 4 mm for coronaries, which is why the small vessel complains first. The list that actually changes management is short: blood pressure, glucose or HbA1c, lipids in some men, and a morning testosterone. That is one appointment and one blood sample. Skipping it does not make the underlying problem go away, it just removes the two or three years of warning you were handed.
My husband still gets erections in the morning but nothing during sex. Does that mean it is in his head?
It points that way and it is not a dismissal. Preserved morning rigidity means the nerves, the arteries and the hormonal signal are all capable of producing the response, because that is exactly what has happened overnight. Combined with onset he can date, and with difficulty that varies by situation, that is the recognised psychogenic pattern. Psychogenic causes are not a leftover category: in one series of over three thousand men worked up systematically they were the largest single group at 59 percent. The mechanism is usually a loop rather than a mood problem. Attention moves from arousal to monitoring performance, the attempt fails, and that failure becomes the evidence for the next one. It responds to being named and discussed, sometimes to a course of therapy, and sometimes to a daily tablet that removes the moment of watching the clock.
I started finasteride for hair loss eight months ago and things have gone downhill. Coincidence?
Possibly not, and the timing you describe is the strongest evidence you have. Medication accounts for up to a quarter of cases overall, and in an analysis of ten years of FDA adverse event reports, 5-alpha reductase inhibitors made up about 46 percent of all reports naming erectile difficulty, with finasteride carrying by far the highest proportional reporting ratio in the dataset. Spontaneous reporting exaggerates what patients already suspect, so that figure is not a probability for you personally. What it does justify is a proper conversation with whoever prescribed it rather than quietly stopping. The two things worth establishing are whether the decline really began after the drug and whether desire fell alongside rigidity, since that pattern also raises the question of testosterone. Do not stop and restart repeatedly to test it; that produces confusion rather than an answer.
The GP wants him to climb two flights of stairs before prescribing anything. Is that a real test?
It is a real benchmark and it comes from the Princeton consensus process. Sexual activity of around thirty minutes with a usual partner corresponds to roughly two to three metabolic equivalents, which is about the same as walking a mile on the flat in twenty minutes or briskly climbing two flights of stairs in ten seconds without cardiac symptoms. If he can do that comfortably, the cardiac risk of sex is low. If he cannot, or he gets chest tightness doing it, the guidance is a stress test before treatment rather than a prescription. The panel asks for exercise tolerance to be established in every man before starting treatment, whatever his risk category, so this is not a sign that anybody suspects something specific about him. More vigorous activity reaches five to six METs, which corresponds to about four minutes of a standard treadmill protocol.
My testosterone came back at 280. Does that explain everything and will replacement fix it?
It is worth acting on and it is unlikely to be the whole story. Deficiency is defined as total testosterone under 300 ng/dL together with symptoms and signs, and diagnosis needs at least two morning samples, so the first thing to do with a 280 is repeat it in the morning rather than build a plan on it. Timing matters because levels fall through the day, and an acute illness can produce a misleadingly low reading. Hormonal causes account for something like 7 percent of cases in systematic series, against about 21 percent vascular, so a borderline result rarely explains a man's whole picture on its own. The symptom pattern that fits is low desire and weak morning erections alongside the erectile trouble. Where deficiency is confirmed, one useful point: a PDE5 inhibitor may work better combined with testosterone therapy than either does alone, so this is not a choice between the two.
He has type 2 diabetes. Are the tablets less likely to help him?
Less likely, and still clearly worth trying. Diabetes damages the small vessels and the autonomic nerves at the same time, so both the delivery and the signal are affected, and that shows up in the trial numbers. In a study of once-daily tadalafil in 298 men with type 1 or type 2 diabetes, endpoint success was 46 percent on 2.5 mg and 41 percent on 5 mg against 28 percent on placebo. Those are real gains and they are visibly narrower than the gaps reported in general population studies. Saying so in advance is better than letting him discover it as a disappointment. Two practical points follow. Glycaemic control is part of the treatment rather than a separate project, and a fair trial matters more here than anywhere: highest tolerated dose, correct interval, four to six attempts before anyone calls it a failure.
I am 34 and this started suddenly about two months ago. Do I need the same tests as an older man?
The same short list, interpreted differently. Sudden onset in a healthy man in his thirties, particularly if morning erections continue and the difficulty is situational, points strongly towards a psychogenic pattern, which is the largest single category across all ages. Two things still deserve checking. The medicine list, since antidepressants, benzodiazepines and finasteride for hair loss are common in your age group and all have a track record here. And a morning testosterone, because guidelines ask for it in every man presenting with this and low levels in a young man raise their own questions. What you do not need at first visit is duplex ultrasound or overnight monitoring. Those answer specific questions, usually where there is a history of pelvic or penile trauma or where surgery is being considered.
How reliable is that five-question score they made him fill in?
Reliable for what it measures, useless for what people expect from it. The five-item index grades severity on a scale from 5 to 25, with 22 to 25 counting as no dysfunction, 17 to 21 mild, 12 to 16 mild to moderate, 8 to 11 moderate and 5 to 7 severe. It is validated, it is quick, and it turns a vague answer at review into a number that can be compared with the one from before treatment. What it cannot do is identify a cause. It will not separate vascular from neurological from hormonal from drug-related, and it depends entirely on recall, so a difficult conversation shifts the answers. It also assumes recent sexual activity, which makes it awkward for men who have been avoiding sex altogether. Treat the number as a baseline and the history as the diagnosis.
He had an erection that would not go down after about five hours and it was painful. It settled eventually. Should we still do something?
Yes, and please tell whoever prescribes for him even though it resolved. An erection lasting more than four hours needs emergency assessment at the time, because the ischaemic form damages the erectile tissue within hours and the window for preserving function is measured in the same units. That it settled on its own is good news and does not remove the follow-up. What matters now is why it happened: which drug and what dose, whether an injectable treatment was involved, and whether there is an underlying condition such as sickle cell disease or a haematological disorder that predisposes to it. Men who have had one episode are at higher risk of another, so the plan for future treatment needs adjusting rather than repeating. If it happens again, that is an emergency department visit at the four-hour mark, not a wait-and-see.
Read each reply as a general teaching point, not a plan built for the person who asked. Your own situation - notes, bloods, every medicine you take - belongs in front of a prescriber who can weigh all of it together.