Prazosin for PTSD Nightmares: Independent Evidence & Safety

Table of contents

• Evidence summary
• What prazosin is
• How it works
• What it is prescribed for
• What works and what does not
• Benefits by claim
• Risks and all side effects
• All interactions
• Who should avoid prazosin
• Dosage and how to take it (doses studied and label doses)
• Follow the money: who makes it and who funded the evidence
• Related research
• Frequently asked questions
• Sources and funding notes

Evidence summary

Independent evidence and credibility scorecard

Independence tiers: 1 = regulator or government body; 2 = academic or Cochrane team with no declared commercial funding for the work; 3 = independent authors analysing mostly manufacturer-run trials, or minor declared industry ties; 4 = manufacturer-funded or manufacturer-authored. Credibility runs from A (highest) to D.

Note on clustering: nearly all the trial evidence comes from the US and from military or veteran populations. Much of it comes from one VA research network in Seattle. The meta-analyses come from the US, China, Norway and Brazil, but they re-analyse the same small set of US trials.

What prazosin is

Prazosin is a quinazoline alpha-1 blocker. Pfizer sold it in the US as Minipress capsules, approved on 23 June 1976 (Drugs@FDA). In the UK, Pfizer's brand is Hypovase (SmPC). Today most prescriptions are filled with cheap generic capsules from many companies (DailyMed).

• US: prescription only. The only labelled use is high blood pressure (label).
• UK: prescription only. It is licensed for high blood pressure, heart failure, Raynaud's phenomenon and urinary symptoms of an enlarged prostate (SmPC). NHS Highland says patients should be told nightmares are "not a licensed indication" (NHS Highland).
• France: a 2026 French military review describes the "imminent withdrawal" of prazosin there and discusses doxazosin as a replacement (Sipahimalani 2026). We could not open the French regulator's notice.

How it works

Prazosin blocks alpha-1 adrenergic receptors. These receptors respond to noradrenaline (norepinephrine), the "fight or flight" chemical messenger. In blood vessels this relaxes the vessel walls and lowers blood pressure, which is why the drug was developed (label).

The nightmare theory is that people with PTSD have an overactive brain response to noradrenaline, especially at night. Unlike some related drugs, prazosin gets into the brain and could damp that response down (Raskind 2007).

What has been shown in people:

• In a 13-person crossover trial, prazosin was reported to increase total and REM (dreaming) sleep; we could not check the exact figures.
• In a 2007 trial, dreams on prazosin rated more like ordinary dreams and less like trauma replays (Raskind 2007).
• The "overactive noradrenaline" explanation is plausible, but no test can yet pick out which patients have it. Researchers from the original team have argued for a personalised approach because of this (Raskind 2020).
• Prazosin wears off quickly. Its half-life is about 2 to 3 hours, compared with roughly 15 to 19 hours for extended-release doxazosin (Rodgman 2016). This short action is why it is usually taken at bedtime.

What it is prescribed for

Licensed uses: high blood pressure (US and UK), plus heart failure, Raynaud's and an enlarged prostate in the UK (US label; UK SmPC).

Off-label uses: mainly nightmares and sleep problems linked to PTSD. Some clinicians also use it for nightmare disorder without PTSD. The AASM's 2018 paper lists prazosin under "may be used" for both (AASM 2018). Researchers have also tested prazosin for chronic post-concussion headache (NCT02266329), but that use is outside this review.

Where it sits in national guidance:

• VA/DoD (US, 2023): we could not confirm the nightmare-specific recommendation or its strength from the guideline text, and secondary sources disagree. The guideline is reported to advise against prazosin for PTSD overall (VA/DoD 2023).
• AASM (US): 2010 grade not verified in our check (Aurora 2010). In 2018 it was downgraded to "may be used", with imagery rehearsal therapy as the only "recommended" treatment. A later review describes the downgrade as a response to contradictory efficacy evidence (AASM 2018).
• NICE (England, 2018): the PTSD guideline does not mention prazosin. Its drug options for adults who prefer medication are venlafaxine or an SSRI such as sertraline (NICE NG116).

What works and what does not

Benefits by claim

Does prazosin stop PTSD nightmares?

The case for it rests on four small placebo-controlled trials:

• 2003: 10 Vietnam veterans in a 20-week crossover. A mean bedtime dose of 9.5 mg beat placebo on nightmares, sleep and global change (Raskind 2003).
• 2007: 40 veterans over 8 weeks, with 34 evaluable, at a mean of 13.3 mg. The authors reported "large effect sizes" (Raskind 2007).
• 2008: 13 civilians, crossover design (Taylor 2008).
• 2013: 67 active-duty soldiers back from Iraq and Afghanistan, treated for 15 weeks (Raskind 2013).

A University of Pittsburgh trial in 50 veterans found that both prazosin and a behavioural sleep programme beat placebo. Sleep outcomes were reported as better than placebo in the active groups; we could not verify the percentages (Germain 2012).

Then came VA Cooperative Study 563. It randomised 304 veterans with chronic PTSD and frequent nightmares at 13 VA centres to prazosin or placebo for 26 weeks, with doses up to 20 mg a day in men and 12 mg in women. At 10 weeks there was no difference on any of the three primary outcomes:

• nightmares (CAPS item B2, scored 0–8): difference 0.2 points (95% CI −0.3 to 0.8)
• sleep quality: difference 0.1 points on the PSQI, scored 0–21 (−0.9 to 1.1)
• global clinical change: difference 0 (−0.3 to 0.3)

Nothing changed at 26 weeks either (Raskind 2018).

Why the clash? Explanations have been proposed, but we could not verify them against the full trial report. No explanation has been tested directly. The plain reading is that the average benefit is smaller and less certain than the early trials suggested.

Does it help you sleep longer or better?

The evidence is mixed:

• A network meta-analysis of 99 drug trials in PTSD ranked prazosin first for insomnia, nightmares and poor sleep quality. The authors rated confidence in these estimates as very low to moderate (Lappas 2024).
• A 2020 meta-analysis found no significant benefit on sleep quality (Zhang 2020).
• A 2025 meta-regression found a significant effect on insomnia (SMD −0.654) (Mendes 2025).

Is it as good as imagery rehearsal therapy?

Imagery rehearsal therapy (IRT) teaches people to rewrite a recurring nightmare while awake and rehearse the new version. A Dutch meta-analysis of 8 IRT and 7 prazosin trials found no significant difference between the two on any outcome (Yücel 2020). A 29-trial network meta-analysis named prazosin and IRT as the two effective options for trauma-related nightmares (Zhang 2022). The VA/DoD 2023 guideline, however, found the IRT evidence insufficient in PTSD-specific trials (VA/DoD 2023).

Does it treat PTSD itself?

No, not on current evidence. Pooled effects on total PTSD scores were small (SMD −0.31) (Reist 2021) or not significant (Zhang 2020; Mendes 2025). The VA/DoD advises against using prazosin as a PTSD treatment in its own right (VA/DoD 2023). An Australian clinical update suggests a time-limited trial only when nightmares and insomnia dominate. It says this should come after first-line therapies and after checking for treatable problems such as obstructive sleep apnoea (Maguire 2026).

What about doxazosin, terazosin or other alpha-blockers?

Doxazosin lasts longer and is the most studied alternative:

• An 8-person crossover trial of extended-release doxazosin, up to 16 mg, improved one self-rated PTSD checklist but not the total clinician-rated score (Rodgman 2016).
• A 141-veteran VA trial in people with PTSD and alcohol use disorder found doxazosin no better than placebo for PTSD (Back 2023).
• The first nightmare-focused RCT (n=65, up to 10 mg) missed its primary outcomes. It showed modest diary-based gains in staying asleep and in worst-dream severity, mainly in men (Richards 2025).
• A German trial testing clonidine and doxazosin is registered but listed as suspended (NCT05360953).
• For terazosin and tamsulosin, a 2026 review found only case reports, pilot and open-label studies (Ponomareva 2026).

Risks and all side effects

Common side effects in label trials of more than 900 people, all taking it for blood pressure (US label):

• dizziness 10.3%
• headache 7.8%
• drowsiness 7.6%
• lack of energy 6.9%
• weakness 6.5%
• palpitations 5.3%
• nausea 4.9%

Most of these settle with continued use.

Less common side effects (1% to 4%) (US label):

• Gut: vomiting, diarrhoea, constipation.
• Heart and circulation: swelling, orthostatic hypotension, breathlessness, fainting.
• Mental and nervous system: vertigo, depression, nervousness.
• Other: rash, passing urine more often, blurred vision, red eyes, nosebleeds, dry mouth, stuffy nose.

Rare or uncertain (under 1%, or reported after marketing) (US label):

• Heart and circulation: fast or slow heart rate, angina, flushing, vasculitis.
• Gut and liver: abnormal liver tests, pancreatitis.
• Mental and nervous system: tingling, hallucinations, insomnia.
• Urinary and sexual: incontinence, impotence.
• Skin and hair: hair loss, lichen planus, hives.
• Other: ringing in the ears, sweating, fever, positive ANA blood test, joint pain, enlarged breasts in men, allergic reactions, eye pain.
• The label also cites literature reports of worsening narcolepsy, with causation uncertain.

In the PTSD trials: the early trials found no difference in blood pressure between groups (Raskind 2007; Raskind 2013). In CSP 563, supine systolic pressure fell by 6.7 mmHg more on prazosin. New or worsening suicidal thoughts were reported in 8% on prazosin against 15% on placebo (Raskind 2018). Dropout for any reason did not differ from placebo (OR 1.00, 95% CI 0.62 to 1.62) (Zhang 2020).

Product quality: FDA records show Class II recalls of generic prazosin capsules for the nitrosamine impurity N-nitroso-prazosin impurity C above acceptable intake limits. They include Teva lots (recall started October 2025) and Appco Pharma lots (recall started March 2026), plus repackaged cartons from Amerisource Health Services in October 2025 (FDA enforcement data). Teva's recall covered about 580,000 bottles of 1, 2 and 5 mg capsules in total (FDA enforcement data; Fierce Pharma). FDA describes the lots as exceeding the acceptable intake limit set under its Carcinogenic Potency Categorization Approach for nitrosamines (FDA enforcement data). These limits concern long-term cancer risk, not immediate poisoning. Do not stop the medicine on your own; a pharmacist can check whether a batch is affected.

All interactions

• Other blood pressure medicines and diuretics: the blood-pressure-lowering effects add up. The label advises cutting prazosin to 1–2 mg three times a day, then re-titrating, when another drug is added (US label).
• Beta-blockers such as propranolol, and calcium-channel blockers: these can cause a marked fall in blood pressure (US label; SmPC).
• PDE-5 inhibitors (sildenafil, tadalafil and similar erectile dysfunction drugs): these can cause symptomatic low blood pressure. The label says to start them at the lowest dose (US label). The VA/DoD repeats this caution for nightmare use (VA/DoD 2023).
• Other medicines that cause dizziness on standing: the VA/DoD advises caution (VA/DoD 2023).
• Alcohol: it makes dizziness and fainting more likely. So do heat, exercise and standing for a long time (US label).
• Antidepressants: most people in the big VA trial were taking them. Whether they change prazosin's effect is unknown (AASM 2018).
• Benzodiazepines: one meta-regression linked higher benzodiazepine use in trials with a larger prazosin effect. This is an exploratory finding, not advice to combine them (Mendes 2025).
• Lab tests: prazosin can give false-positive results on screening tests for phaeochromocytoma, an adrenal tumour, by raising urinary VMA and noradrenaline metabolites (US label).
• Medicines with no adverse interaction in limited experience: the label lists digoxin, insulin and older diabetes tablets, diazepam, allopurinol, aspirin and indomethacin (US label).

Who should avoid prazosin

• Allergy: anyone allergic to prazosin, other quinazolines (such as doxazosin or terazosin) or the capsule ingredients (US label).
• Low blood pressure or falls: people with already low blood pressure, frequent fainting or a high fall risk need special care, because the main harm is a drop in blood pressure (SmPC).
• Pregnancy: safety is not established. Limited experience in 44 pregnant women treated for severe hypertension found no fetal abnormalities (US label).
• Breastfeeding: small amounts pass into breast milk, so caution is advised (SmPC).
• Children: safety and effectiveness have not been established (US label).
• Kidney or liver problems: the UK label starts at 500 micrograms a day in moderate-to-severe kidney impairment or liver dysfunction (SmPC).
• Planned cataract surgery: tell the eye surgeon in advance (US label).
• Narcolepsy: the label cites reports of worsening (US label).
• Possible sleep apnoea: treatable causes of disturbed sleep, such as obstructive sleep apnoea, should be ruled out first (Maguire 2026).
• Jobs where fainting would be dangerous: avoid driving and hazardous tasks for 24 hours after the first dose and after each increase (US label).

Dosage and how to take it (doses studied and label doses)

The figures below are label doses and doses used in trials or guidelines. They are not personal dosing advice. Prazosin for nightmares should only be started and adjusted by a prescriber.

• US label, for high blood pressure: starts at 1 mg two or three times a day. The usual range is 6–15 mg a day in divided doses, generally up to 20 mg. Treatment should always begin with the 1 mg capsule (US label).
• UK label: for high blood pressure, starts at 500 micrograms two or three times a day for 3–7 days (first dose in the evening), then 1 mg two or three times a day for a further 3–7 days, then rises gradually; the maximum is 20 mg a day (SmPC).
• VA/DoD 2023, for PTSD nightmares: dosing wording not verified from the guideline text; see the trial and label doses above (VA/DoD 2023).
• NHS Highland: 1 mg at night to prevent first-dose fainting, then 2 mg after 2 to 3 days, then 1 mg steps every 2 to 7 days depending on benefit. It notes that US veterans' guidance gives a target of 6–10 mg a day, that studies suggest veterans may need more than civilians, and that responses are described from 1 mg to more than 16 mg (NHS Highland).

Doses reached in trials:

• Mean 9.5 mg at bedtime (2003).
• Mean 13.3 mg in the evening (2007).
• Mean 3.1 mg in civilians (2008).
• Soldiers in 2013: men averaged 15.6 mg at bedtime plus 4.0 mg mid-morning; women averaged 7.0 mg plus 1.7 mg (2013).
• CSP 563 allowed up to 20 mg a day in men and 12 mg in women, in divided doses (2018).

A 2026 Australian clinical update stresses an adequate dose and duration, with blood pressure monitoring (Maguire 2026). No trial has tested how best to stop the drug. Any change should be planned with the prescriber.

Follow the money: who makes it and who funded the evidence

Who makes and sells it

• Brand owner: Pfizer (US) holds the original US application for Minipress (NDA 017442, 1976) (Drugs@FDA). Pfizer Ltd holds the UK licence for Hypovase (SmPC).
• US generics: FDA records list generic applications from Teva, Mylan, Aurobindo, Alembic, Mankind, Granules, MSN, Appco and Novitium, among others (openFDA Drugs@FDA). Several are India-based. The label we used is for capsules made in Tamil Nadu, India, and distributed by Novitium Pharma in New Jersey (DailyMed).
• India: in September 2026, Pfizer Ltd stopped selling Minipress XL, an extended-release prazosin tablet, after its US parent ended manufacture (Medical Dialogues). Novartis India agreed to buy the Minipress trademarks in India for about ₹1,250 crore. The report cites IQVIA sales of ₹228.6 crore for Minipress XL (Outlook Business). That market is for blood pressure and prostate use, not nightmares.
• Price: the VA/DoD calls prazosin "inexpensive". It adds that patients may find it more acceptable because it is classed as a blood pressure drug, with less stigma (VA/DoD 2023).

Who funded the trials

The nightmare research is unusual: there is essentially no drug-company money in it. No company stands to profit from a new use for a cheap, off-patent drug.

• Early trials: the Seattle trials were supported by NIMH grant R01MH069867 and the VA (2007; 2013; Taylor 2008). Registry entries for related studies list the DoD and VA Puget Sound as collaborators (NCT00990106).
• CSP 563: sponsored by the VA Office of Research and Development (NCT00532493).
• Pittsburgh comparison trial: funding statement not verified in our check (Germain 2012).
• Doxazosin trials: funded by the VA and DoD (NCT03339258; NCT02500602).

The conflicts that matter here are institutional and intellectual, not commercial:

• One VA research network proposed the theory, ran most of the positive trials and then ran the trial that failed.
• The VA also writes the guideline and treats the patients.

Publishing a large negative trial of its own idea is, if anything, an admission against interest.

Documented integrity events

• Nitrosamine recalls: Class II recalls of generic prazosin for impurities, by Teva (2025) and Appco (2026) (FDA enforcement data).
• Brand withdrawals: Pfizer's US parent ended manufacture of Minipress XL (Medical Dialogues), and French clinicians report a withdrawal in France (Sipahimalani 2026).
• No enforcement action found: we found no FDA warning letters, lawsuits or enforcement actions about the nightmare use itself.

Related research

For wider, evidence-graded context, see these Pure City Research guides:

• Nightmare disorder: assessment, treatments and safety
• Post-traumatic stress disorder: treatments and safety
• Complex PTSD
• Trauma-focused therapy
• EMDR therapy
• CBT-I: cognitive behavioural therapy for insomnia
• Insomnia
• Obstructive sleep apnoea
• Orthostatic hypotension
• Blood pressure medicines
• Sertraline
• Trazodone

Frequently asked questions

Sources and funding notes

What we could not verify:

• We could not verify from the guideline text: the 2023 VA/DoD nightmare wording, the 2010 AASM Level A grade, the Richards 2025 doxazosin details, the Germain 2012 percentages, and the exact Taylor 2008 sleep figures. The NICE NG116 point comes from secondary summaries, because the NICE page was not accessible. The Fierce Pharma article was not accessible, so the recall figures rest on FDA data. We could not open the full NEJM paper for CSP 563, so we do not report its authors' individual disclosures.
• We could not access the funding statements for the Reist 2021, Yücel 2020 and Zhang 2022 meta-analyses, or for the 2003 Raskind trial. PubMed lists no grant for that trial.
• We could not access full text of the 2025 doxazosin RCT. Its funding is taken from the trial registry and the PubMed conflict statement.
• We could not open the French regulator's notice on prazosin's withdrawal in France, or the UK BNF entry.
• We found no reliable figure for global prazosin sales or for how many people take it for nightmares.
• No trial has tested how to stop prazosin, or how long to continue it.

1. VA/DoD Clinical Practice Guideline for PTSD and Acute Stress Disorder, version 4.0 (2023) (guideline page) — national guideline; funded by VA Evidence Based Practice, Office of Quality and Patient Safety.
2. Prazosin hydrochloride capsules, US prescribing information (DailyMed) — FDA-approved labelling; distributor Novitium Pharma.
3. Drugs@FDA: Minipress, NDA 017442; openFDA application list — US regulator records.
4. openFDA drug enforcement (recall) records for prazosin — US regulator data.
5. Hypovase Summary of Product Characteristics (emc) — UK approved product information; licence holder Pfizer Ltd.
6. NHS Highland: Prazosin for PTSD-related nightmares (version 2, reviewed June 2024) — NHS prescribing guideline; public funding.
7. NICE NG116: Post-traumatic stress disorder (2018) — national guideline; funded by the UK government via NICE.
8. Morgenthaler et al. 2018, AASM position paper on nightmare disorder — professional society; funded by the AASM.
9. Aurora et al. 2010, AASM best practice guide for nightmare disorder (AASM notice) — professional society.
10. Raskind et al. 2018, NEJM, VA CSP 563 (NCT00532493) — RCT; VA Cooperative Studies Program.
11. Raskind et al. 2003, Am J Psychiatry — crossover RCT; no grant listed on PubMed.
12. Raskind et al. 2007, Biol Psychiatry — RCT; NIMH R01MH069867.
13. Taylor et al. 2008, Biol Psychiatry — crossover RCT; VA and NIMH MH069867; no conflicts.
14. Raskind et al. 2013, Am J Psychiatry — RCT in active-duty soldiers; NIMH R01MH069867.
15. Germain et al. 2012, J Psychosom Res — RCT; DoD CDMRP and NIH.
16. McCall et al. 2018, J Clin Psychopharmacol (NCT02199652) — pilot RCT; Augusta University.
17. Reist et al. 2021, CNS Spectrums — meta-analysis; funding not accessed.
18. Zhang et al. 2020, Sleep Medicine — meta-analysis; NSFC (China) and NIH MH64827; no conflicts.
19. Yücel et al. 2020, Sleep Med Rev — meta-analysis, IRT vs prazosin; University of Amsterdam; funding not accessed.
20. Zhang et al. 2022, Neurosci Biobehav Rev — network meta-analysis; funding not accessed.
21. Skeie-Larsen et al. 2022, IJERPH — meta-analysis; no external funding; no conflicts.
22. Mendes et al. 2025, Prog Neuropsychopharmacol Biol Psychiatry — meta-regression; no specific grant; no conflicts.
23. Lappas et al. 2024, Sleep Medicine — network meta-analysis; no competing interests declared.
24. Raskind 2020, Sleep Med Rev — commentary by the original investigator.
25. Maguire et al. 2026, Australasian Psychiatry — narrative clinical update; academic.
26. Ponomareva et al. 2026, J Clin Psychopharmacol — narrative review of alpha-1 blockers; academic and VA.
27. Richards et al. 2025, J Clin Sleep Med (NCT03339258) — doxazosin RCT; San Francisco VA and DoD; one author consulted for Otsuka.
28. Rodgman et al. 2016, J Clin Psychiatry (NCT02308202) — doxazosin pilot; Baylor College of Medicine.
29. Back et al. 2023, J Clin Psychiatry (NCT02500602) — doxazosin RCT; VA Office of Research and Development.
30. Sipahimalani et al. 2026, L'Encéphale — French military review; no competing interests declared.
31. NCT05360953, NCT00990106, NCT02266329 — trial registrations (Charité with German Federal Ministry of Education and Research; Seattle with DoD and VA; VA).
32. Fierce Pharma: Teva prazosin recall (2025) — trade press.
33. Medical Dialogues: Pfizer to discontinue Minipress XL in India (2026); Outlook Business: Novartis India to acquire Minipress trademark (2026) — trade and business press.

Educational research reviewed 6 October 2026. This article is not medical advice; talk to a qualified clinician or pharmacist before starting, stopping or combining treatments.