Clonazepam: Independent Evidence on Panic, Epilepsy, Side Effects & Withdrawal

by
Key takeaways
  • Clonazepam (Klonopin in the US, Rivotril elsewhere) is a long-acting benzodiazepine. In the US it is licensed for some seizure disorders and for panic disorder (FDA Klonopin label). The UK product licence we checked covers epilepsy only (UK SmPC 4.1).
  • The FDA label notes an unusual dose finding for panic disorder. In the fixed-dose licensing trial, 2, 3 and 4 mg a day "were less effective than the 1 mg/day dose and were associated with more adverse effects". A clear benefit over placebo was seen "consistently only for the 1 mg/day group" (FDA label).
  • In the short term it works for panic. An independent Cochrane network meta-analysis of 70 trials ranked clonazepam among the most effective drugs for response and for reducing how often panic attacks happen. The same review rated the benzodiazepine trials as low quality (Guaiana et al., Cochrane 2023).
  • Guidelines advise against it for anxiety beyond a crisis. NICE says benzodiazepines "should not be prescribed" for panic disorder (NICE CG113). The WHO makes a strong recommendation against benzodiazepines for GAD and panic, allowing emergency use for 3–7 days at most (WHO mhGAP 2023). For epilepsy, NICE lists it as a second-line option for myoclonic seizures (NICE NG217).
  • Its FDA boxed warning covers deaths with opioids, abuse and addiction, and withdrawal reactions that can be life-threatening. Clonazepam made up 24% of the estimated 92 million US benzodiazepine prescriptions dispensed in 2019, second only to alprazolam (FDA 2020).
  • Roche developed and launched the brand in 1973. In January 2021 Roche sold the commercial rights to Rivotril to CHEPLAPHARM, a family-run company in Greifswald, Germany (CHEPLAPHARM 2021).
  • Evidence grade: Moderate for short-term panic relief and for certain seizure types. Insufficient for benefit in panic beyond 9 weeks. Risk for dependence, withdrawal, falls and deaths with opioids.

Independent evidence review · Prescription medicine

Clonazepam is two medicines in one: a long-term anti-seizure drug for some people with epilepsy, and a fast-acting panic treatment that guidelines now want kept short. For panic it does beat placebo, and its manufacturer's own dose-finding trial found that 1 mg a day worked better than higher doses. But the trials lasted 6 to 9 weeks, none measured dependence over years, and independent reviewers rate them low quality. Clonazepam also stays in the body for a long time, with a half-life "typically 30 to 40 hours" (FDA label). That smooths out its effect, but it also means drowsiness and impairment can carry over into the next day.

Best evidence for short-term control of panic attacks; certain seizure types (myoclonic, absence, Lennox-Gastaut)
Main risks dependence and withdrawal (including seizures), sedation and falls, fatal breathing suppression with opioids or alcohol
Key rule for panic, lowest dose for the shortest time; never stop suddenly; no alcohol or opioids
Safety first
Clonazepam is a prescription-only medicine and a controlled substance in the US. Do not start it, stop it or change the dose without the prescriber who manages your treatment. Stopping suddenly after regular use can cause severe withdrawal, including seizures, which can be life-threatening. If you take it for epilepsy, stopping suddenly can also bring on status epilepticus (FDA label; UK SmPC 4.4). Taking it with opioids (including codeine, tramadol and methadone), alcohol or other sedatives can cause profound sedation, slowed breathing, coma and death. Do not drive or use machinery until you know how it affects you. Like all anti-seizure medicines, clonazepam carries a warning about suicidal thoughts and behaviour. If you have thoughts of harming yourself, seek urgent help now. In the UK, call NHS 111 or 999. If someone taking clonazepam is very drowsy, breathing slowly or cannot be woken, call 999 (or your local emergency number).

Table of contents

Evidence summary

Claim Evidence Source Funding / conflict Strength
Benzodiazepines beat placebo for panic disorder in the short term 24 studies, 4,233 participants. Response RR 1.65 (95% CI 1.39–1.96), NNTB 4. Remission RR 1.61. Overall methodological quality was poor, with "probable publication bias". No study looked at long-term efficacy, dependence or withdrawal. Breilmann et al., Cochrane 2019 Cochrane team (Germany/Italy). Nine of ten authors declared no conflicts. One declared lecture fees from Eli Lilly, Janssen, MSD, Pfizer and others. Moderate (low-quality trials)
Clonazepam is among the most effective drugs for panic 70 trials. Diazepam, alprazolam and clonazepam ranked as the most effective for response. "Only clonazepam and alprazolam showed a strong reduction in the frequency of panic attacks compared to placebo." The benzodiazepine comparisons were low quality. Guaiana et al., Cochrane 2023 Cochrane (UK, Italy, Canada, Japan, Germany). Several authors declared fees from Otsuka, Pfizer, Eli Lilly and others. We found no declared ties to clonazepam's makers. Moderate
1 mg a day is the best-supported panic dose Fixed-dose trial of 0.5–4 mg/day. FDA: a significant difference from placebo "was observed consistently only for the 1 mg/day group". 74% on 1 mg/day were free of full panic attacks at endpoint vs 56% on placebo. The published paper reported 1 mg and above as "equally efficacious" for panic attacks, with 1–2 mg giving "the best balance". FDA label; Rosenbaum et al., J Clin Psychopharmacol 1997 Licensing trial submitted by the NDA holder (then Roche). PubMed lists "Research Support, Non-U.S. Gov't" but does not name the sponsor. Moderate (manufacturer data)
Clonazepam as good as antidepressants for panic Across all benzodiazepines, response RR 0.99 (95% CI 0.67–1.47), from only 2 studies with 215 participants. Low-quality evidence. No clinically significant differences between individual benzodiazepines. Bighelli et al., Cochrane 2016 Cochrane (Italy). One author was an expert witness for Accord Healthcare; another declared pharma lecture fees. Insufficient
Long-term benefit in panic disorder "The effectiveness of Klonopin in long-term use, that is, for more than 9 weeks, has not been systematically studied in controlled clinical trials." FDA label Regulator-approved label Insufficient
Useful for certain seizure types US licence: Lennox-Gastaut syndrome, akinetic and myoclonic seizures, and absence seizures not responding to succinimides. "Up to 30%" of initial responders lost anticonvulsant effect, often within 3 months. NICE: second-line option for myoclonic seizures. FDA label; NICE NG217 Regulator and UK government guideline Moderate
REM sleep behaviour disorder (off-label) The AASM "suggests that clinicians use clonazepam (vs no treatment)" for isolated RBD in adults. This is a conditional recommendation. Howell et al., AASM guideline 2023 Funded by the AASM, which discloses industry support from sleep-drug and device makers (AASM disclosure). None of clonazepam's makers appeared in that list. Weak
Dependence, withdrawal and harm with opioids Boxed warning on all three. Protracted withdrawal can last "weeks to more than 12 months". US benzodiazepine-involved overdose deaths rose from 1,298 in 2010 to 11,537 in 2017. FDA label; FDA 2020 Regulator-required safety text Risk

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 professional bodies with some industry income; 4 = manufacturer-funded or manufacturer-authored. Credibility runs from A (highest) to D.

Source Who funds it Country Independence tier (1–4) Credibility (A–D) Why it would tell the truth / residual bias
NICE CG113 and NG217 UK Department of Health and Social Care UK 1 A Has no product to sell and must justify NHS spending. The panic recommendation dates from 2004. We did not check committee conflict-of-interest registers.
WHO mhGAP guideline 2023 World Health Organization International 1 A Uses GRADE and states openly that its evidence certainty is low. Written with low-resource settings in mind, which may make it more cautious about dependence-forming drugs.
FDA Klonopin label and FDA 2020 communication US government, with industry user fees for drug review US 1 (safety text) / 4 (efficacy data) A for warnings The warnings are legally required and cut against sales. The efficacy sections summarise trials run by the manufacturer.
UK SmPC (clonazepam, Neuraxpharm) Written by the licence holder, approved by the MHRA UK 1–4 A for restrictions Contraindications and warnings are regulator-approved. Written by a generic company with no incentive to overstate a brand.
NHS medicines pages UK government UK 1 A− Plain-language, no commercial interest. The pages were last reviewed on 27 January 2023 and were due for review in January 2026.
Breilmann et al., Cochrane 2019 Cochrane (commercial sponsorship not allowed) Germany / Italy 2 A Nine of ten authors declared no conflicts, and the review openly downgrades its own evidence. It can only pool the trials that exist, most of which were run by industry.
Guaiana et al., Cochrane 2023 Cochrane; one author NIHR-supported International 2–3 A− A 70-trial network meta-analysis. Some authors declared unrelated industry fees. It looked at short-term efficacy and dropout only.
Rosenbaum et al. 1997 Non-government support (sponsor not named in PubMed); the trial underpins the manufacturer's US panic licence US 4 (probable) B A large, placebo-controlled, fixed-dose design. Its published summary reads more favourably for higher doses than the FDA's account of the same design.
AGS Beers Criteria 2023 American Geriatrics Society; "There was no sponsor for this paper." US 2 A− A professional society whose remit is preventing harm in older people. Some panellists have disclosed consulting ties.
AASM RBD guideline 2023 American Academy of Sleep Medicine US 3 B Uses GRADE, and most authors reported no conflicts. The AASM receives industry support (for example from Idorsia, Eisai, Jazz and Lilly), but clonazepam's makers were not named in its disclosure (AASM disclosure).
CHEPLAPHARM company statements CHEPLAPHARM Germany 4 B for facts about its own deals; D for product claims Used here only for facts about ownership and the acquisition.

What clonazepam is

Clonazepam is a benzodiazepine, the same drug family as diazepam, lorazepam and alprazolam (NHS). It is absorbed quickly and almost completely: bioavailability is about 90%, and peak blood levels come 1 to 4 hours after a dose. It is cleared slowly, with an elimination half-life "typically 30 to 40 hours" (FDA label). This makes it one of the longer-acting benzodiazepines.

  • Brand names: Klonopin (US) and Rivotril (most other countries). The FDA's benzodiazepine list names Klonopin (FDA 2020), and CHEPLAPHARM describes Rivotril as "launched in 1973 and ... currently registered in approximately 85 countries" (CHEPLAPHARM).
  • US approval: The original Klonopin application (NDA 017533) was approved on 4 June 1975. The sponsor now listed is CHEPLAPHARM (Drugs@FDA). A 2010 US label was distributed by "Genentech USA, Inc. A Member of the Roche Group" (2010 Klonopin label). The current label says the product is distributed by H2-Pharma, LLC (Montgomery, Alabama) and "Licensed by: CHEPLAPHARM Arzneimittel GmbH", Greifswald, Germany (FDA label).
  • Generics: Available. The openFDA database lists 10 generic (ANDA) applications for clonazepam (openFDA query). In the UK, clonazepam tablets and oral solutions are sold by several generic companies, including Neuraxpharm, Aristo, Rosemont and Thame (emc listing).
  • Prescription status: Prescription-only in both countries. In the US it is a Schedule IV controlled substance ("CIV" on the label). In England and Wales it is one of the medicines covered by drug-driving law, with a blood limit of 50 µg/L (GOV.UK).

How it works

The FDA label says the exact mechanism behind clonazepam's anti-seizure and anti-panic effects "is unknown", but is believed to relate to its ability to enhance the activity of GABA, "the major inhibitory neurotransmitter in the central nervous system" (FDA label). In plain terms, the NHS says it works "by increasing levels of a calming chemical, gamma-aminobutyric acid (GABA), in your brain", which can "relieve anxiety, stop seizures or fits or relax tense muscles" (NHS).

Clonazepam is broken down in the liver, mainly by reduction of its 7-nitro group, and CYP3A enzymes "may play an important role". Less than 2% leaves the body unchanged in urine (FDA label). This explains two practical points. First, liver disease and drugs that block CYP3A can raise levels. Second, because of the 30–40 hour half-life, the drug builds up over several days of regular use and is still in the body a day or more after the last dose.

The same GABA effect that calms panic also causes the main problems. The body adapts to it, so tolerance and physical dependence can develop. The label warns that "little tolerance develops to the amnestic reactions and other cognitive impairments caused by benzodiazepines", even when tolerance to the therapeutic effect does (FDA label).

What it is prescribed for

Licensed uses

UseUS (FDA)UKGuideline position
Panic disorder, with or without agoraphobia Licensed. Efficacy was based on two trials lasting 6 to 9 weeks. Not in the UK SmPC we checked, which lists epilepsy only. The NHS page describes use for panic disorder at 1–2 mg a day (NHS). Not recommended by NICE: benzodiazepines "are associated with a less good outcome in the long term and should not be prescribed" for panic disorder (NICE CG113 1.3.20). WHO: not recommended, except emergency use for 3–7 days at most.
Seizure disorders Alone or as an add-on for Lennox-Gastaut syndrome (petit mal variant), akinetic and myoclonic seizures, and absence seizures that have not responded to succinimides. "All clinical forms of epileptic disease and seizures in infants, children and adults", especially absence seizures, tonic-clonic, focal and myoclonic seizures (UK SmPC 4.1). NICE NG217 5.4.3: if first-line treatment fails for myoclonic seizures, consider one of several options, including clonazepam, as monotherapy or add-on (NICE NG217).

Other uses

  • Muscle spasms and restless legs syndrome: The NHS page lists involuntary muscle spasms and "sometimes restless legs syndrome" among clonazepam's uses (NHS). We did not review trial evidence for these uses.
  • REM sleep behaviour disorder: The American Academy of Sleep Medicine conditionally suggests clonazepam for isolated RBD and for RBD due to a medical condition. It gives the same conditional suggestion for immediate-release melatonin (Howell et al. 2023). The 2023 Beers Criteria say benzodiazepines "may be appropriate" for RBD and seizure disorders in older adults, while advising against them in general (AGS Beers 2023).
  • Generalised anxiety disorder and PTSD: Clonazepam is not licensed for GAD in the US label or the UK SmPC we checked. NICE says not to offer a benzodiazepine for GAD "except as a short-term measure during crises" (NICE CG113 1.2.26). The US VA/DoD guideline makes a strong recommendation against benzodiazepines for PTSD (VA/DoD 2023).

What works and what does not

Claimed benefit Verdict Evidence Key caveat
Fewer panic attacks in the short term WORKS Beat placebo in two 6–9 week licensing trials (FDA label). One of only two drugs with a "strong reduction in the frequency of panic attacks" in the Cochrane network (Guaiana 2023). The difference from placebo was "approximately 1 panic attack per week". The trials were low quality and short.
Higher doses give more benefit for panic NOT SUPPORTED FDA: 2, 3 and 4 mg/day "were less effective than the 1 mg/day dose and were associated with more adverse effects" (FDA label). The published paper reported doses of 1 mg and above as equally effective at reducing panic attacks (Rosenbaum 1997). Neither found higher doses better.
Long-term panic control (months to years) INSUFFICIENT Not studied in controlled trials beyond 9 weeks (FDA label). Cochrane reviewers say the trials "did not examine the long-term efficacy nor the risks of dependency and withdrawal symptoms" (Breilmann 2019). NICE links benzodiazepines to worse long-term outcomes in panic.
Better than antidepressants for panic INSUFFICIENT No difference in response (RR 0.99), from 2 small trials (Bighelli 2016). No difference between drug classes in the network meta-analysis (Guaiana 2023). Benzodiazepines had fewer dropouts in short trials. Antidepressants do not cause benzodiazepine-type dependence.
Seizure control (myoclonic, absence, Lennox-Gastaut) WORKS for some Licensed in the US and UK. A second-line NICE option for myoclonic seizures. "Up to 30%" of initial responders lose the anticonvulsant effect, often within 3 months (FDA label). It can trigger tonic-clonic seizures in people with mixed seizure types.
REM sleep behaviour disorder MIXED Conditional AASM suggestion (Howell 2023). Most people with RBD are older and so at greater risk of benzodiazepine-related falls and confusion (Beers 2023).
Treating depression NOT A USE "Depression" was reported as an adverse event in 7% on clonazepam vs 1% on placebo in the panic trials (FDA label). The label notes that HAM-D depression scores still fell more on clonazepam than on placebo.

Benefits by claim

Panic disorder: what the licensing trials actually showed

The US panic licence rests on two double-blind, placebo-controlled trials in adult outpatients (FDA label, Clinical Trials):

  • Study 1 (9-week, fixed dose): Patients took 0.5, 1, 2, 3 or 4 mg a day, or placebo. "A significant difference from placebo was observed consistently only for the 1 mg/day group." At endpoint, 74% of patients on 1 mg a day were free of full panic attacks, compared with 56% on placebo. The difference in full panic attacks was "approximately 1 panic attack per week".
  • Study 2 (6-week, flexible dose): Patients took 0.5 to 4 mg a day, and the mean dose was 2.3 mg a day. At endpoint, 62% on clonazepam were free of full panic attacks, compared with 37% on placebo. Again, the difference was about 1 attack per week.

Two points matter for readers. First, the placebo response was large: more than half of the placebo group in Study 1 were free of full panic attacks by the end. This is common in panic trials and is one reason why effect sizes need context. Second, the dose finding is unusual. The FDA wording in the dosing section is that doses of 2, 3 and 4 mg a day "were less effective than the 1 mg/day dose and were associated with more adverse effects". The published version of the same fixed-dose design (413 patients) described things differently. It reported that "daily dosages of 1.0 mg and higher were equally efficacious in reducing the number of panic attacks", with 1.0 to 2.0 mg giving "the best balance of therapeutic benefit and tolerability" (Rosenbaum et al. 1997). Both accounts agree that higher doses did not add benefit. They differ on whether doses above 1 mg did worse. We report both as written. The FDA account is the regulator's reading of the full submitted data.

Panic disorder: the independent reviews

Cochrane's 2019 review pooled 24 trials of any benzodiazepine against placebo. The response rate was higher on benzodiazepines (RR 1.65, 95% CI 1.39–1.96), which means about four people needed to be treated for one extra person to respond (NNTB 4). Fewer people dropped out on benzodiazepines (RR 0.50). However, dropouts due to adverse effects were higher (RR 1.58), and the reviewers rated the evidence low quality. They warned that the trials' "validity ... is questionable due to possible unmasking of allocated treatments, high dropout rates, and probable publication bias" (Breilmann et al., Cochrane 2019). Their conclusion is worth quoting: the real clinical choice "is not between benzodiazepines and placebo, but between benzodiazepines and other agents, notably SSRIs".

The 2023 Cochrane network meta-analysis compared individual drugs directly and indirectly across 70 trials. Clonazepam ranked among the three most effective drugs for response, was among the drugs with clinically meaningful remission, and had one of the strongest reductions in panic symptom scales. "Only clonazepam and alprazolam showed a strong reduction in the frequency of panic attacks compared to placebo." The authors still concluded that the reliability of these findings "may be limited due to the overall low quality of the studies" (Guaiana et al., Cochrane 2023).

Why guidelines still say no for panic

The independent reviews and the guidelines do not really disagree. The reviews show short-term benefit. The guidelines weigh that against dependence, withdrawal and the lack of long-term data. The WHO made its recommendation against benzodiazepines strong, despite low-certainty evidence, and drew on the same Slee and Breilmann reviews (WHO mhGAP 2023). NICE advises antidepressants as the only long-term drug treatment for panic disorder, alongside CBT (NICE CG113). The NHS clonazepam page itself points people with panic disorder towards SSRIs such as sertraline, paroxetine or escitalopram, and towards CBT (NHS).

Generalised anxiety disorder

Clonazepam is not licensed for GAD. At class level, the largest network meta-analysis (89 trials, 25,441 patients) found that "paroxetine and benzodiazepines were effective but also poorly tolerated when compared with placebo" (Slee et al., Lancet 2019). This review received no funding. We found no clonazepam-specific GAD data in the sources we reviewed.

Epilepsy

For epilepsy, the evidence we could verify is regulatory and guideline-based rather than recent meta-analysis. The US label and the UK SmPC both license clonazepam for several seizure types. NICE places it among second-line options for myoclonic seizures, and points prescribers to its separate guideline on medicines associated with dependence for "safe prescribing and managing withdrawal of clobazam and clonazepam in adults" (NICE NG217). The important caveat is loss of effect. The FDA label says "up to 30% of patients who initially responded have shown a loss of anticonvulsant activity, often within 3 months". The UK SmPC notes that in some childhood epilepsies control may be re-established by raising the dose or pausing clonazepam for 2 or 3 weeks under careful observation (UK SmPC 4.2).

Risks and all side effects

The FDA boxed warning, summarised

The Klonopin label carries a boxed warning covering three risks (FDA label):

  • Opioids: "Concomitant use of benzodiazepines and opioids may result in profound sedation, respiratory depression, coma, and death." Prescribers should reserve combined prescribing for patients with no adequate alternative.
  • Abuse, misuse and addiction: Use of benzodiazepines, including Klonopin, "exposes users to risks of abuse, misuse, and addiction, which can lead to overdose or death". The risk should be assessed before and throughout treatment.
  • Dependence and withdrawal: "The risks of dependence and withdrawal increase with longer treatment duration and higher daily dose." Abrupt stopping or rapid dose reduction "may precipitate acute withdrawal reactions, which can be life-threatening". A gradual taper is required.

The FDA added the dependence and addiction parts to all benzodiazepine labels in September 2020. It reported that, in 2019, an estimated 92 million benzodiazepine prescriptions were dispensed from US outpatient pharmacies: alprazolam 38%, clonazepam 24% and lorazepam 20%. In 2018, "an estimated 50% of patients who were dispensed oral benzodiazepines received them for a duration of two months or longer". Benzodiazepine-involved overdose deaths rose from 1,298 in 2010 to 11,537 in 2017 (FDA 2020). In the UK, the MHRA told prescribers to "only prescribe benzodiazepines (or benzodiazepine-like drugs) and opioids together if there is no alternative" (MHRA, March 2020).

Common side effects in the panic trials

These figures come from the pooled 6- to 9-week panic trials (574 people on clonazepam, 294 on placebo) (FDA label, Tables 3 and 4). Overall, 17% stopped clonazepam because of an adverse event, compared with 9% on placebo.

Side effectClonazepamPlaceboNote
Somnolence (sleepiness)37%10%The most common reason for stopping (7% vs 1%)
Depression (reported as adverse event)7%1%Led to stopping in 4% vs 1%
Fatigue7%4%
Abnormal coordination6%0%Relevant to falls and driving
Ataxia (unsteadiness)5%0%
Dizziness8%4%
Memory disturbance4%2%
Reduced intellectual ability2%0%
Upper respiratory tract infection8%4%
Decreased libido1%0%

In epilepsy treatment, the label reports drowsiness in approximately 50% of patients, ataxia in approximately 30% and behaviour problems in approximately 25%. Some of these "may diminish with time" (FDA label). The NHS lists daytime sleepiness, disturbed sleep, dizziness or unsteadiness, and muscle weakness as common side effects (NHS).

Serious risks

RiskWhat the sources saySource
Breathing suppressionMay cause respiratory depression. Use with caution in COPD or sleep apnoea. In the UK, severe respiratory insufficiency and sleep apnoea syndrome are contraindications. Call 999 for difficulty breathing.FDA; UK SmPC 4.3; NHS
Withdrawal reactionsAcute signs include anxiety, insomnia, tremor, panic attacks and tachycardia. Severe signs include seizures, delirium, psychosis, catatonia and suicidality. Protracted withdrawal can last "weeks to more than 12 months".FDA §Dependence
Suicidal thoughts and behaviourIn a pooled analysis of 199 trials of 11 anti-seizure drugs, the risk of suicidal thinking or behaviour was about twice that on placebo (adjusted RR 1.8), roughly "one case of suicidal thinking or behavior for every 530 patients treated".FDA label; UK SmPC 4.4
Worsening of seizuresIn people with several coexisting seizure types, it may increase or bring on generalised tonic-clonic seizures. Combined with valproate, it may produce absence status.FDA
Falls, fractures, confusion in older adults"All benzodiazepines increase the risk of cognitive impairment, delirium, falls, fractures, and motor vehicle crashes in older adults." Avoid (strong recommendation).AGS Beers 2023
Memory loss (anterograde amnesia)Can occur at therapeutic doses, with risk increasing at higher doses. It may be associated with inappropriate behaviour.UK SmPC 4.4
Paradoxical reactionsAgitation, aggression, irritability, nightmares, hallucinations and psychosis. More likely in children and older people.FDA
Liver and blood problemsPeriodic blood counts and liver function tests are "advisable during long-term therapy". Contact a doctor about yellowing skin or eyes, easy bruising or nosebleeds.FDA; NHS
Increased saliva and bronchial secretionsMainly in infants and small children, who may need airway supervision.UK SmPC 4.8
Rare allergic reactionsAnaphylaxis and angioedema have been reported rarely.UK SmPC 4.8

Dependence and withdrawal in practice

The NHS says clonazepam "is not likely to be addictive if you take it for a short time (2 to 4 weeks)", and that people who have taken it for longer will have their dose reduced gradually (NHS). It lists possible effects of stopping suddenly, including confusion, seizures or fits, depression, nervousness or irritability, sweating and diarrhoea. The FDA label adds that in the panic trials patients were withdrawn over a 7-week downward titration. That discontinuance period "was associated with good tolerability and a very modest clinical deterioration". The label also cautions that there are not enough long-term data "to accurately estimate the risks of withdrawal symptoms and dependence" with longer use (FDA label).

Overdose

Benzodiazepine overdose ranges from drowsiness to coma. Overdose "in combination with other CNS depressants (including alcohol and opioids) may be fatal" (FDA label). The NHS advises contacting 111 if you take more than your prescribed dose, and calling 999 for breathing difficulty (NHS).

All interactions

Interacting drug or substanceEffectSeriousnessSource
Opioids (codeine, tramadol, morphine, oxycodone, methadone, heroin)Profound sedation, respiratory depression, coma and deathBoxed warningFDA; MHRA
AlcoholDeeper sedation and risk of not breathing properly or not waking. The UK SmPC also notes that alcohol can provoke seizures in people with epilepsy.Avoid completelyNHS; UK SmPC 4.5
Other sedatives: hypnotics, other benzodiazepines, sedating antihistamines, antipsychotics, tricyclic antidepressants, mirtazapine, baclofen, tizanidine, barbituratesAdded sedation and breathing suppressionHighUK SmPC 4.5
Other anti-seizure drugs: phenytoin, carbamazepine, phenobarbital, lamotrigineThese induce clonazepam metabolism, causing "an approximately 38% decrease" in levels. Clonazepam may change phenytoin levels, so monitoring is recommended.ModerateFDA
Sodium valproateRarely, absence status epilepticusModerateUK SmPC 4.5
CYP3A inhibitors and other enzyme blockers: azole antifungals (e.g. fluconazole), cimetidine, disulfiram, fluvoxamine, ritonavirHigher clonazepam levels and stronger effectsModerateFDA; UK SmPC
Rifampicin (enzyme inducer)May increase clearance, lowering levelsModerateUK SmPC; NHS
Blood pressure medicines (ACE inhibitors, ARBs, calcium channel blockers, beta-blockers, diuretics, alpha-blockers, moxonidine)Enhanced blood-pressure-lowering effect; with alpha-blockers or moxonidine, added sedationModerateUK SmPC 4.5
LevodopaClonazepam may possibly antagonise its effectsLow–moderateUK SmPC
TheophyllineMay possibly reduce clonazepam levelsLowUK SmPC
Sertraline, fluoxetineNo effect on clonazepam pharmacokineticsNone foundFDA
Herbal sedatives (valerian, passionflower)Added drowsiness. The NHS says not to combine them.ModerateNHS
CaffeineMay reduce the calming effectLowNHS
Cannabis, heroin, methadone; stimulantsVery deep sleep and breathing risk with cannabis, heroin or methadone. The NHS says stimulants can also add to drowsiness.HighNHS
Flumazenil (reversal agent)Can precipitate withdrawal and seizures in long-term usersHighFDA

Clonazepam does not affect any type of contraception, according to the NHS (NHS).

Who should avoid clonazepam

Contraindications

  • US label: sensitivity to benzodiazepines; significant liver disease; acute narrow-angle glaucoma (FDA).
  • UK SmPC: hypersensitivity to benzodiazepines; acute pulmonary insufficiency; severe respiratory insufficiency; sleep apnoea syndrome; myasthenia gravis; severe hepatic insufficiency. It "must not be used in patients in a coma, or in patients known to be abusing pharmaceuticals, drugs or alcohol" (UK SmPC 4.3).

Use with extra caution

The NHS asks people to tell their doctor before starting if they have lung, liver or kidney problems, spinal or cerebellar ataxia, a history of alcohol or drug problems, a recent loss or bereavement, depression or thoughts of self-harm, or a personality disorder diagnosis (NHS). The UK SmPC notes that "in cases of loss or bereavement, psychological adjustment may be inhibited by benzodiazepines". It may also cause problems in porphyria (UK SmPC 4.4).

Older adults

The 2023 Beers Criteria advise avoiding benzodiazepines in older adults (moderate-quality evidence, strong recommendation). The criteria note that older adults have "increased sensitivity to benzodiazepines and decreased metabolism of long-acting agents". They do allow that use "may be appropriate for seizure disorders, rapid eye movement sleep behavior disorder" and a few other situations. Beers also says to avoid combining opioids with benzodiazepines (AGS Beers 2023). The UK SmPC recommends that the starting dose in older people "should not exceed 0.5 mg/day" for epilepsy (UK SmPC 4.2).

Pregnancy and breastfeeding

  • Pregnancy: Use late in pregnancy can cause sedation and withdrawal in the newborn, including breathing problems, floppiness and feeding difficulties. The FDA label says observational data "do not report a clear association with benzodiazepines and major birth defects" (FDA). The UK SmPC says it should be used in pregnancy only if there is a compelling indication (UK SmPC 4.6). The NHS says it "can be taken during pregnancy, particularly if you need it to control your epilepsy", after weighing risks with a doctor (NHS). Stopping suddenly in pregnancy can worsen epilepsy.
  • Breastfeeding: The sources differ. The UK SmPC says mothers on clonazepam "should not breastfeed". The NHS says you "can usually take clonazepam" while breastfeeding if the baby is healthy, while watching for sleepiness or poor feeding, and advises not sharing a bed with your baby. The FDA label asks mothers to monitor for "sedation, poor feeding and poor weight gain". This is a decision for the prescriber and the family.

Liver and kidney disease

Significant liver disease is a US contraindication, and severe hepatic impairment is a UK contraindication. The label advises caution in kidney impairment because clonazepam's metabolites are excreted by the kidneys (FDA).

Dosage and how to take it

Your prescriber sets the dose. The ranges below are the official licensed adult ranges, provided for context only. They are not a guide to self-dosing.

UseOfficial adult dosingSource
Panic disorder (US licence)Start 0.25 mg twice daily. After 3 days, an increase to the target dose "for most patients of 1 mg/day" may be made. Maximum 4 mg/day. One dose at bedtime may reduce daytime sleepiness.FDA label
Panic disorder (NHS)1 mg to 2 mg each dayNHS
Seizure disorders (US)Start no more than 1.5 mg/day in three divided doses. Increase by 0.5–1 mg every 3 days as needed. Maximum 20 mg/day.FDA label
Epilepsy (UK)Start no more than 1 mg/day (0.5 mg/day in older people). Usual maintenance 4–8 mg/day in 3 or 4 divided doses, reached over 2–4 weeks. Up to 20 mg/day at the physician's discretion.UK SmPC 4.2

How long before it works

The NHS says that for panic disorder and restless legs syndrome, clonazepam "should take around 1 hour to start working". For seizures and muscle spasms it "might take a few days to a week" to work fully, because the dose is built up gradually (NHS).

How long to take it

For epilepsy, clonazepam is "usually prescribed long term". For other conditions, the NHS says your doctor "will want to regularly review if you still need it" (NHS). The FDA label says "there is no body of evidence available to answer the question of how long the patient treated with clonazepam should remain on it" for panic disorder.

How to stop

Never stop suddenly. For panic disorder, the US label describes discontinuation "with a decrease of 0.125 mg twice daily every 3 days, until the drug is completely withdrawn". If withdrawal reactions develop, it suggests "pausing the taper or increasing the dosage to the previous tapered dosage level" and then going more slowly (FDA label). For epilepsy, the UK SmPC says treatment "even if of short duration, must be withdrawn by gradually reducing the dose in view of the risk of precipitating status epilepticus" (UK SmPC 4.4). Any taper should be planned and supervised by the prescriber, and may take much longer than the minimum schedules above.

Driving

In England and Wales it is illegal to drive with clonazepam above 50 µg/L in blood, unless it was prescribed, taken as advised and is not impairing your driving (GOV.UK drug-driving; GOV.UK drug-driving law). Driving while impaired is an offence regardless of prescription. People with epilepsy are also subject to separate licensing rules, and the UK SmPC notes that "as a general rule, epileptic patients are not allowed to drive" (UK SmPC 4.7).

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

Who owns it

  • Originator: Roche (F. Hoffmann-La Roche, Basel, Switzerland). Rivotril was launched in 1973 (CHEPLAPHARM). Roche's US business, via Genentech, distributed Klonopin as recently as a 2010 label (2010 label).
  • Current rights holder: CHEPLAPHARM Arzneimittel GmbH, Greifswald, Germany. It announced on 11 January 2021 that it had acquired "the commercial rights for Rivotril® from Roche", describing the deal as complementing "our portfolio of globally marketed benzodiazepines" (CHEPLAPHARM 2021). Drugs@FDA now lists CHEPLAPHARM as the NDA sponsor (Drugs@FDA).
  • US distributor: H2-Pharma, LLC, Montgomery, Alabama (FDA label).
  • Business model: CHEPLAPHARM describes itself as a specialty pharmaceutical company "headquartered in Germany" whose mission is to be "a sustainable platform for well-established and trusted pharmaceutical brands". Its company timeline lists repeated purchases of older brands from Roche, Eli Lilly, AstraZeneca and others. It reports annual turnover rising from EUR 3m in its early years to EUR 1,498m in the most recent figure shown. Its CEO, Sebastian Braun, received an EY "Entrepreneur of the Year" award in the "Family Business" category (CHEPLAPHARM company page). We found no clonazepam-specific revenue figure.
  • Generics: Clonazepam has been off-patent for decades. There are 10 US generic applications (openFDA), and several UK generic licence holders (emc). Most clonazepam dispensed today is likely to be generic, so no single company earns most of the money from it. We did not verify market share.

Who funded the evidence

  • Panic licensing trials: These were submitted to the FDA by the brand holder of the time, Roche. The published fixed-dose trial lists non-government research support, but PubMed does not name the sponsor (Rosenbaum 1997). We treat it as manufacturer-generated evidence.
  • Independent reviews: The Cochrane reviews (Breilmann 2019; Guaiana 2023; Bighelli 2016) are covered by Cochrane's rule that reviews "cannot be funded or conducted by commercial sponsors or commercial sources with a real or potential vested interest" (Cochrane policy). Some authors declared unrelated pharmaceutical fees, listed above. Slee 2019 received no funding.
  • Guidelines: NICE (UK government), WHO, and the US VA/DoD are publicly funded. The Beers Criteria had no sponsor. The AASM RBD guideline was funded by the AASM, which publicly discloses support from several sleep-drug makers (AASM disclosure); Roche, Genentech and CHEPLAPHARM were not named.
  • Safety evidence: The boxed warnings come from the FDA's own review of post-marketing data (FDA 2020), and the UK co-prescribing advice from the MHRA.

What this means: The claim that clonazepam works for panic in the short term rests on manufacturer trials, but independent Cochrane reviewers who re-examined them reached the same broad conclusion while downgrading the quality. The strongest cautions (dependence, withdrawal and opioid deaths) come from regulators, not from the companies. We found no documented integrity events (fines, retracted trials) specific to clonazepam in the sources we reviewed.

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

Sibling medicine reviews:

Frequently asked questions

How long does clonazepam take to work?

For panic disorder, the NHS says it "should take around 1 hour to start working". Peak blood levels come 1 to 4 hours after a dose (FDA label). For epilepsy, where the dose is increased gradually, full effect can take a few days to a week or more (NHS).

Is clonazepam stronger than Xanax or diazepam?

"Stronger" is not a useful comparison, because doses differ between drugs. What does differ is duration. Clonazepam's half-life is typically 30 to 40 hours (FDA label), much longer than alprazolam's. In the Cochrane panic network, diazepam, alprazolam and clonazepam all ranked highly for response (Guaiana 2023). An earlier Cochrane review "failed to find clinically significant differences between individual benzodiazepines" (Bighelli 2016). The Beers Criteria say shorter-acting benzodiazepines are not safer than long-acting ones for falls in older people.

Can you drink alcohol on clonazepam?

No. The NHS says, "Do not drink alcohol while taking clonazepam." Alcohol can cause very deep sleep, and "there's a risk you may not be able to breathe properly" (NHS). In epilepsy, alcohol can also provoke seizures (UK SmPC).

Is clonazepam addictive?

It can be. The FDA boxed warning covers abuse, misuse, addiction and physical dependence. The risk rises "with longer treatment duration and higher daily dose" (FDA label). The NHS says addiction is not likely over 2 to 4 weeks, and is more likely if you have ever had problems with alcohol or drugs (NHS). Physical dependence can develop even when you take it exactly as prescribed.

How do I stop clonazepam safely?

Only with your prescriber's plan. After more than 2 to 4 weeks of use the dose is reduced gradually, because stopping suddenly can cause withdrawal reactions, including seizures (NHS). The FDA label describes the minimum pace used in the panic trials and advises pausing or slowing the taper if withdrawal symptoms appear. If you take clonazepam for epilepsy, stopping too fast can bring on status epilepticus.

Does clonazepam cause weight gain or memory problems?

Weight gain and weight loss are both listed as uncommon reported events (FDA label). Memory effects are more established. In the panic trials, memory disturbance was reported by 4% on clonazepam vs 2% on placebo. Anterograde amnesia can occur at normal doses, with the risk rising at higher doses (UK SmPC). Tolerance develops little to these cognitive effects.

Why does my prescriber keep my panic dose at 1 mg?

Because that is what the evidence supports. In the fixed-dose licensing trial, the FDA reports that higher doses of 2, 3 and 4 mg a day "were less effective than the 1 mg/day dose and were associated with more adverse effects" (FDA label). The label allows up to 4 mg a day for individuals who may benefit, but the average patient did not do better on more.

Can I drive while taking clonazepam?

Only if it is not affecting you. In England and Wales there is a legal blood limit of 50 µg/L, but a medical defence applies if it was prescribed, you took it as advised and your driving is not impaired (GOV.UK). Do not drive if you feel sleepy, dizzy, clumsy or unable to concentrate (NHS). Separate licensing rules apply to people with epilepsy.

Sources and funding notes

Last reviewed: 26 September 2026. Educational information from Pure City Research, not medical advice. Medicine decisions belong with a qualified prescriber.

Have a question — or want us to cover something?

Ask about anything on this page, or request the next deep dive: an ingredient, a supplement, or a health concern. We use published research, evidence syntheses, and regulatory guidance, with clear source links.

We store your topic, message, optional email, and this page so we can manage and reply to the request. Do not include diagnoses, medications, or other sensitive medical information. See our Privacy Policy.