Zopiclone: Independent Evidence on Insomnia, Side Effects, Driving & Withdrawal

Table of contents

• Evidence summary
• What zopiclone 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 zopiclone
• Dosage and how to take it
• 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

Tier 1 means fully public or academic with no relevant ties. Tier 4 means paid for by the manufacturer. Credibility A to D is our overall judgement of how far each source can be relied on for the claims we use it for.

What zopiclone is

Zopiclone is a prescription sleeping tablet (a hypnotic). The NHS describes it as "a type of sleeping pill that's sometimes known as a Z-drug" (NHS).

• Drug class: the UK licence classes it as a "benzodiazepine related drug" (ATC code N05CF01). Chemically it is a cyclopyrrolone, a different structure from the benzodiazepines (Zimovane SmPC).
• UK brand: Zimovane 7.5 mg film-coated tablets. The marketing authorisation holder is Aventis Pharma Limited, trading as Sanofi, in Reading. The licence number is PL 04425/0624, first authorised on 6 May 1993 (SmPC). Generic zopiclone tablets are widely available from several manufacturers. For example, an MHRA defect notice from 2020 names ratiopharm and Mylan products (MHRA).
• Origin: early pharmacology and pharmacokinetic research on zopiclone, under the code RP-27267, was published by scientists at Rhône-Poulenc Santé research centres in France (Piot et al., 1990, Gaillot et al., 1987). Independent academic groups also studied it under the same RP code (Tokunaga et al., 1987). By 1999 a 15-year clinical review described its use across more than 30 countries (Hajak, 1999).
• UK legal status: prescription-only (POM). Since 10 June 2014 it has also been a Class C controlled drug, in Schedule 4 Part 1 of the Misuse of Drugs Regulations (Home Office Circular 008/2014).
• US status: an openFDA search of Drugs@FDA returned no approved zopiclone product (openFDA).
  • The US instead has eszopiclone (Lunesta), a cyclopyrrolone with a single S-configured chiral centre. It was approved under NDA 021476 in December 2004, and the current label holder is Waylis Therapeutics (Lunesta label, openFDA).
  • The US Drug Enforcement Administration (DEA) lists "Zopiclone" in Schedule IV, and gives Lunesta as the example brand (DEA controlled substances list).

How it works

Zopiclone boosts the effect of GABA, the brain's main calming (inhibitory) chemical messenger. It does this by acting on the GABA-A receptor complex, which opens a chloride channel and makes nerve cells less active.

• The UK licence says zopiclone and other cyclopyrrolones "act on a different site to those of benzodiazepines" and cause different changes in the shape of the receptor. Even so, its effects overlap with benzodiazepines: it is hypnotic, sedative, anti-anxiety, anticonvulsant and muscle-relaxing (SmPC section 5.1).
• NICE's summary is that, although the Z-drugs differ in structure from benzodiazepines, "they are also agonists of the GABA receptor complex". They were developed to overcome disadvantages of benzodiazepines such as next-day sedation, dependence and withdrawal. (NICE TA77 section 3).

How the body handles it (pharmacokinetics):

• Zopiclone is absorbed quickly, with peak levels at 1.5 to 2 hours. The NHS says it takes around 30 to 60 minutes to start working (NHS).
• The elimination half-life is about 5 hours, rising to about 7 hours in older people (SmPC section 5.2). This is longer than zolpidem's 2.5 hours (NICE TA77), which helps explain why next-morning effects are a particular concern.
• It is broken down mainly in the liver by the enzyme CYP3A4, with some help from CYP2C8. That is why some antibiotics, antifungals, HIV medicines and enzyme-inducing drugs change its levels.
• Clearance is "clearly reduced" in people with cirrhosis.

A common side effect is a bitter taste. The UK licence lists dysgeusia (bitter taste) as common, affecting between 1 in 100 and 1 in 10 people (SmPC section 4.8). The NHS describes it as "a bitter or metallic taste in your mouth" (NHS).

What it is prescribed for

• UK licensed use: "the short-term treatment of insomnia in adults" (SmPC section 4.1). The NHS adds that a doctor may prescribe it for trouble falling asleep, waking in the night or waking too early, when this is "affecting your everyday life" (NHS).
• US: not licensed (see above).

Where guidelines place it

• NICE (UK), TA77, 2004: NICE says hypnotics should only be used after "due consideration of the use of non-pharmacological measures", and only for severe insomnia that is interfering with daily life. Then:
  • they should be prescribed "for short periods of time only";
  • because nothing clearly separates zolpidem, zopiclone and the shorter-acting benzodiazepines, the cheapest should be used;
  • switching is appropriate only for side effects linked to a specific drug;
  • people who have not responded to one of these drugs "should not be prescribed any of the others" (NICE TA77).
  In NICE's more recent daridorexant appraisal, CBT for insomnia (CBTi) is described as "the standard first treatment for people with long-term insomnia" (NICE TA922).
• European Insomnia Guideline 2023:
  • CBT-I is first-line for chronic insomnia "in adults of any age" (grade A).
  • If CBT-I does not work well enough, a drug can be offered. Benzodiazepine receptor agonists, which include zopiclone, "can be used for the short-term treatment of insomnia (≤ 4 weeks)" (grade A).
  • Longer use "may be initiated in some cases" (grade B) (Riemann et al., 2023).
• American College of Physicians (ACP) 2016: CBT-I is the initial treatment for all adults (strong recommendation). Adding a drug after CBT-I has failed is a weak recommendation based on low-quality evidence (Qaseem et al., 2016).
• Older adults (AGS Beers Criteria 2023): the Beers Criteria say to "Avoid" Z-drugs in people aged 65 and over (strong recommendation, moderate-quality evidence). The reasons given are adverse events similar to benzodiazepines, including delirium, falls, fractures, emergency visits and car crashes, together with "minimal improvement in sleep latency and duration". Because Beers is written for the US, its list names eszopiclone, zaleplon and zolpidem rather than zopiclone. The reasoning applies to the class as a whole (AGS Beers Criteria 2023).

In short: zopiclone is a second-line, short-term option once CBT-I and sleep-hygiene measures have been tried, or while waiting for them. It is not a long-term treatment.

What works and what does not

Benefits by claim

1. Short-term sleep quality (acute treatment)

The best independent estimate comes from the NIHR-funded network meta-analysis by De Crescenzo and colleagues. It included 170 trials in the systematic review, and 154 double-blind randomised trials covering 30 treatments and 44,089 participants in the network analysis (Lancet 2022).

• Zopiclone was one of the drugs more effective than placebo for acute treatment, along with benzodiazepines, doxylamine, eszopiclone, lemborexant, seltorexant and zolpidem. The group's SMDs ranged from 0.36 to 0.83, with high to moderate certainty.
• Zopiclone, zolpidem, eszopiclone and benzodiazepines were also more effective than melatonin, ramelteon and zaleplon (SMD 0.27–0.71).
• The abstract does not give a separate SMD for zopiclone alone, so we have not reported one.

Tolerability is where zopiclone looks worse. The same analysis found:

• more dropouts because of side effects than placebo: OR 2.00 (95% CI 1.28–3.13; very low certainty);
• more dropouts than eszopiclone (OR 1.82 [1.01–3.33]), daridorexant (3.45 [1.41–8.33]) and suvorexant (3.13 [1.47–6.67]), all at low certainty.

The authors concluded that several licensed drugs "can be effective in the acute treatment of insomnia but are associated with poor tolerability, or information about long-term effects is not available".

2. How big is the benefit in practical terms?

• An independent meta-analysis of FDA approval data for Z-drugs found that sleep onset in the lab was about 22 minutes faster than on placebo (95% CI 11–33 minutes) (Huedo-Medina et al., BMJ 2012).
  • That dataset covered only eszopiclone, zaleplon and zolpidem, because zopiclone was never submitted to the FDA. It is the closest audited benchmark for the class, not a measurement of zopiclone itself.
  • The authors called the drug effect and placebo response "rather small and of questionable clinical importance". They also noted that the two together produced "a reasonably large clinical response".
• Placebo matters a great deal in insomnia trials. A meta-analysis of 32 drug trials found that "63.56% of the drug responses are achieved even in the placebo groups" (Winkler & Rief, Sleep 2015). In practice, a large part of the improvement people feel on a sleeping tablet would also happen with a dummy pill and the reassurance of treatment.

3. Zopiclone compared with CBT (Sivertsen et al., JAMA 2006)

This is the landmark independent head-to-head trial. Researchers at the University of Bergen randomised adults with chronic primary insomnia (mean age 60.8) to three groups:

• CBT covering sleep hygiene, sleep restriction, stimulus control, cognitive therapy and relaxation (n=18);
• zopiclone 7.5 mg nightly (n=16);
• placebo (n=12).

Treatment lasted 6 weeks, and the two active groups were followed up at 6 months. Sleep was measured with polysomnography (sleep-lab style recordings) and sleep diaries (Sivertsen et al., 2006).

• CBT gave better short- and long-term results than zopiclone on 3 of 4 outcome measures.
• "For most outcomes, zopiclone did not differ from placebo."
• Sleep efficiency with CBT rose from 81.4% to 90.1% at 6 months. With zopiclone it went from 82.3% to 81.9%.
• The CBT group spent more time in deep (slow-wave) sleep and less time awake at night. Total sleep time was similar in all three groups.

Limitations: the trial was small, had one site and was limited to adults aged 55 and over. The allocation concealment and blinding were rigorous, though: the zopiclone and placebo pills had identical appearance, smell and flavour (full text).

4. Compared with benzodiazepines and zolpidem

NICE's assessment group reviewed 24 randomised trials, including 13 comparing zopiclone with a benzodiazepine (NICE TA77 section 4). It found:

• "no consistent pattern of superiority of 1 drug over another";
• "evidence of multiple testing of outcomes, with selective reporting of significant findings";
• no consistent differences in next-day residual effects;
• no data from the randomised trials on how often withdrawal or dependence happened.

In the one 2-week trial comparing the two Z-drugs, zolpidem 10 mg was associated with shorter time to fall asleep than zopiclone 7.5 mg (OR 1.72; 95% CI 1.04–2.84).

Risks and all side effects

Frequencies are from the UK licence: common = 1 in 100 to 1 in 10; uncommon = 1 in 1,000 to 1 in 100; rare = 1 in 10,000 to 1 in 1,000; very rare = fewer than 1 in 10,000; "not known" = cannot be estimated (Zimovane SmPC section 4.8).

Next-day impairment and driving

This is zopiclone's best-documented harm.

• What the licence says: the risk of impaired driving is higher if zopiclone "is taken within 12 hours of performing activities that require mental alertness", if a higher dose is taken, or if it is combined with alcohol, other depressants or drugs that raise zopiclone levels (SmPC section 4.7).
• On-the-road tests: zopiclone 7.5 mg is so reliably impairing that researchers use it as the "positive control" when testing new sleeping pills. In a review of eight standardised on-the-road studies (191 healthy volunteers), it consistently impaired next-morning driving (Verster et al., 2011).
  • Lane weaving (the standard deviation of lateral position) increased by an average of 3.0 cm, compared with 2.4 cm at a blood alcohol concentration of 0.05%.
  • Older and younger volunteers were affected similarly.
• Real-world accidents: across 3.1 million Norwegians, the first week after filling a prescription for zopiclone or zolpidem carried a standardised incidence ratio for road accidents of 2.3 (2.0–2.7) (Gustavsen et al., 2008).

UK law:

• It is illegal in England, Scotland and Wales to drive if any legal drug impairs your driving.
• Zopiclone is not one of the medicines with a specified blood limit under the 2015 drug-driving rules. That list includes drugs such as diazepam, temazepam and morphine.
• The "unfit to drive" offence still applies to zopiclone (GOV.UK: Drugs and driving, GOV.UK advice for healthcare professionals).

Dependence, withdrawal and the 2026 MHRA warning

In January 2026 the MHRA reviewed dependency-forming medicines, including Z-drugs. It concluded that existing product information "did not sufficiently communicate the extent of the known risks of addiction, dependence, withdrawal and tolerance" (MHRA Drug Safety Update, 8 January 2026). It asked for:

• stronger warnings in the product information, patient leaflets and outer packaging;
• a conversation about how treatment will end before it starts;
• gradual tapering that "can sometimes take weeks or months".

The current Zimovane licence already contains this wording (SmPC, revised 11 February 2026). It lists these withdrawal symptoms:

• rebound insomnia, anxiety, tremor, sweating, agitation, confusion, headache, palpitations, nightmares, hallucinations, panic attacks, muscle cramps and gut upset;
• in severe cases, derealisation, depersonalisation, sensitivity to noise, light and touch, and numbness;
• very rarely, seizures.

There is also regulatory history. In 2014 the UK made zopiclone a Class C controlled drug. The Home Office cited reported harms from misuse, including "a risk of coma, respiratory depression and death associated with use of excess doses of these drugs in combination with alcohol or other similar depressants" (Home Office Circular 008/2014).

A 2003 review of the world literature found 22 published case reports of zopiclone dependence. Most people involved had a history of drug or alcohol misuse or a psychiatric condition, and in extreme cases doses reached 30 to 120 times the recommended amount (Hajak et al., Addiction 2003).

Overdose

The licence says overdose "should not be life threatening unless combined with other CNS depressants, including alcohol". It can cause drowsiness through to coma, and also methaemoglobinaemia and haemolytic anaemia (SmPC section 4.9). If someone has taken more than their prescribed dose, the NHS advice is to call NHS 111. If they are told to go to A&E, they should not drive themselves (NHS).

All interactions

Sources: Zimovane SmPC sections 4.4–4.5, NHS zopiclone, AGS Beers Criteria 2023.

Who should avoid zopiclone

Do not use (contraindications)

The UK licence says zopiclone must not be used by people with (SmPC section 4.3):

• myasthenia gravis;
• respiratory failure;
• severe sleep apnoea syndrome;
• severe liver impairment;
• an allergy to zopiclone or to any ingredient, such as wheat starch in people with wheat allergy;
• a previous complex sleep behaviour after taking zopiclone.

It is also not for children and adolescents under 18.

Use with caution or at a lower starting dose

• History of addiction. This includes alcohol, drugs or prescription medicines, "especially opioids" (NHS). The dependence risk is higher.
• Depression or suicidal thoughts. Only small quantities should be supplied, and ongoing insomnia should prompt a review for depression (SmPC).
• Breathing problems. This covers chronic respiratory insufficiency and milder sleep apnoea. The starting dose is 3.75 mg.
• Liver or kidney impairment. The starting dose is 3.75 mg.
• Rare sugar-intolerance conditions. The tablets contain lactose, so they are not suitable for hereditary galactose intolerance, Lapp lactase deficiency or glucose-galactose malabsorption.

Older adults

• The UK licence starts older people on 3.75 mg (SmPC).
• The Beers Criteria advise avoiding Z-drugs in people aged 65 and over (strong recommendation) (Beers 2023). They also advise avoiding them in people with delirium, dementia or cognitive impairment, or a history of falls or fractures.
• An Austrian systematic review for deprescribing recommended "discontinuing Z-drugs, principally because of the high risk for falls and fractures" in older adults. It added that the quality and quantity of the evidence are low (Scharner et al., BMC Geriatr 2022).

Pregnancy

• The UK licence says use in pregnancy "is not recommended". Zopiclone crosses the placenta.
• Cohort data on more than 1,000 pregnancies exposed to benzodiazepine-like drugs in the first trimester have not shown malformations. However, some case-control studies reported more cleft lip and palate with benzodiazepines.
• Use late in pregnancy can cause "floppy infant syndrome", breathing depression and withdrawal in the newborn (SmPC section 4.6).
• The NHS says zopiclone "is not usually used during pregnancy". It may be prescribed briefly if insomnia is severe and other treatments have not worked (NHS).

Breastfeeding

The two official UK sources differ, so this is a decision for the prescriber.

• The manufacturer's licence says that although concentrations in breast milk are low, "use in nursing mothers must be avoided" (SmPC).
• The NHS says it "can be used for a short time while breastfeeding" after checking with a doctor. It also warns not to share a bed with a baby after taking zopiclone, because of the risk of sudden infant death syndrome (NHS).

Dosage and how to take it

Your prescriber sets your dose and how long you take it. The ranges below are the official UK licensed doses, for information only (Zimovane SmPC section 4.2).

• How to take it: one dose a night, swallowed whole with water, not sucked or chewed, just before bed. Never take a second dose in the same night (NHS, SmPC).
• How long it takes to work: about 30 to 60 minutes (NHS).
• Missed dose: skip it and take the next dose at bedtime the following night. Do not double up (NHS).
• How long a course lasts: "as short as possible and should not exceed four weeks including the period of tapering off". Going beyond that requires the prescriber to re-evaluate (SmPC).

How to stop

Do not stop suddenly after regular use. The NHS says "Your doctor will gradually reduce your dose so that you can stop taking it safely" (NHS). The 2026 licence wording says (SmPC section 4.4):

• the reduction schedule should be tailored to the individual;
• each step should get smaller as the dose falls;
• the taper can be paused, or the previous dose restored, if withdrawal symptoms become intolerable;
• tapering from a high dose "may take in excess of weeks or months".

Pure City Research does not give personal tapering schedules. Agree a plan with your prescriber.

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

Originator and current owner

• Origin: the earliest pharmacology, pharmacokinetic and dependence studies of zopiclone (code RP-27267) came from Rhône-Poulenc Santé research centres at Vitry-sur-Seine and Antony, France (Piot et al., 1990, Gaillot et al., 1987).
• Corporate trail: Sanofi's SEC filing describes the chain of ownership (sanofi-aventis Form 20-F, 2006). The same filing gives the company's principal executive offices as Paris, France.
  • Rhône-Poulenc was formed in 1928 from the merger of two French companies.
  • In December 1999, Rhône-Poulenc and Germany's Hoechst combined to form Aventis.
  • Sanofi-Synthélabo took control of Aventis in August 2004. Aventis merged into sanofi-aventis on 31 December 2004.
• Today: the UK Zimovane licence is still held by Aventis Pharma Limited, "trading as Sanofi" (SmPC).

Generics and revenue

• Zopiclone has long been off-patent, and cheap generics dominate UK supply.
• When NICE appraised it in 2004, an adult dose cost £0.16, the same as zolpidem (NICE TA77).
• We did not find a product-level revenue figure for zopiclone or Zimovane in the Sanofi annual report we reviewed, so no revenue is stated here.

Who funded the evidence

• Company research: the early company-authored research was largely reassuring. A 1990 mouse study by Rhône-Poulenc Santé scientists suggested that zopiclone "may not produce physical dependence" (Piot et al., 1990). A 1999 review, with no funding stated in the abstract, called the risk of withdrawal reactions at therapeutic doses "very low", and said tolerability apart from bitter taste was "similar to that of placebo" (Hajak, 1999).
• Later regulatory view: three decades later, the UK regulator concluded that dependence can occur even with short-term use at recommended doses, and that previous warnings were insufficient (MHRA 2026). We record this difference as a fact. It does not by itself show that the earlier work was done in bad faith. Animal models and early trials are often poor at detecting dependence in people.
• Independent evidence: the independent evidence used in this article came from:
  • Norwegian public and university grants (Sivertsen 2006, Gustavsen 2008);
  • the UK NIHR (De Crescenzo 2022, which has disclosed industry ties among its authors);
  • Cardiff academics with no competing interests (Donnelly 2017);
  • the unsponsored Beers panel.
  These sources are more cautious about zopiclone than the early company-linked literature.
• NICE appraisal: the appraisal behind NICE TA77 was prepared by an academic group, the Liverpool Reviews and Implementation Group. Aventis, Sanofi Synthelabo, Wyeth and generic makers were consultees. NICE noted that the dependence studies the manufacturers submitted "were not considered to be methodologically robust" (NICE TA77 section 9, section 4).
• Integrity events: we found no documented regulatory sanctions or research-misconduct findings specific to zopiclone in the sources we reviewed.

Related research

• Sleep prevention guide: the evidence-based plan for better sleep
• Sleep supplements: independent evidence review
• Melatonin: evidence, safety and dosing
• Magnesium: evidence review
• L-theanine: evidence review
• Ashwagandha: evidence and safety
• St John's wort: evidence and interactions (it lowers zopiclone levels)
• Stress, anxiety and depression prevention guide
• Sibling medicine reviews: zolpidem, temazepam, trazodone, mirtazapine, daridorexant, suvorexant

Frequently asked questions

Sources and funding notes

• NHS: Zopiclone, a medicine used to treat insomnia (last reviewed 22 June 2026). UK public health service; no commercial funding.
• Zimovane 7.5 mg SmPC, electronic medicines compendium (revised 11 February 2026). Manufacturer's regulator-approved product information; the marketing authorisation holder is Aventis Pharma Ltd, trading as Sanofi (UK, parent company in France).
• NICE TA77: zaleplon, zolpidem and zopiclone for the short-term management of insomnia (28 April 2004), including section 3, section 4 and section 9. UK public body; the assessment was prepared by an academic group; manufacturers were consultees.
• NICE TA922: daridorexant. UK public body; evidence submitted by Idorsia (Switzerland). Used here only for NICE's statement on CBTi.
• NICE Annual Report and Accounts 2024 to 2025. The source of NICE's funding figures.
• MHRA Drug Safety Update: improving information supplied with gabapentinoids, benzodiazepines and z-drugs (8 January 2026). UK government regulator, advised by the Commission on Human Medicines.
• Home Office Circular 008/2014. UK government; zopiclone controlled as Class C from 10 June 2014.
• GOV.UK: Drugs and driving, the law and Drug driving and medicine: advice for healthcare professionals. UK government.
• De Crescenzo et al., Lancet 2022, network meta-analysis of insomnia drugs. Funded by the UK NIHR Oxford Health BRC. Author declarations include employment at Boehringer Ingelheim, fees from Angelini Pharma, and leading Janssen-sponsored seltorexant trials.
• Sivertsen et al., JAMA 2006, CBT vs zopiclone (full text). Norway; funded by the University of Bergen, the Meltzer Fund and the Norwegian Foundation for Health and Rehabilitation; no financial disclosures reported.
• Riemann et al., European Insomnia Guideline 2023, J Sleep Res. European Sleep Research Society; Europe PMC lists NIHR grant support; full conflict-of-interest statements not accessed (publisher page blocked).
• Qaseem et al., ACP guideline 2016, Ann Intern Med. American College of Physicians (USA).
• American Geriatrics Society 2023 Beers Criteria. USA; "There was no sponsor for this paper"; minor panel consultancies disclosed.
• Huedo-Medina et al., BMJ 2012, Z-drugs meta-analysis of FDA data. USA/UK; no external support; no relevant financial relationships. Did not include zopiclone.
• Winkler & Rief, Sleep 2015, placebo response in insomnia trials. Germany, University of Marburg; funding not shown in the abstract.
• Verster et al., Curr Drug Saf 2011, zopiclone and driving. Netherlands/international academic authors; funding and conflicts not shown in the PubMed record.
• Gustavsen et al., Sleep Med 2008, hypnotics and road accidents. Norwegian Institute of Public Health (public).
• Donnelly et al., PLoS One 2017, benzodiazepines, Z-drugs and hip fracture. Cardiff University / NHS Wales academics; no competing interests.
• Scharner et al., BMC Geriatr 2022, Z-drugs in older adults. Austria (Medical University of Vienna, Paracelsus Medical University); no competing interests declared.
• Hajak et al., Addiction 2003, abuse and dependence with zolpidem and zopiclone. Germany; funding not shown in the abstract.
• Hajak, Drug Saf 1999, 15 years of zopiclone. Germany; single-author narrative review; funding not shown in the abstract.
• Piot et al., Neurosci Lett 1990, Gaillot et al., Sleep 1987 and Tokunaga et al., 1987. The first two are by Rhône-Poulenc Santé authors (France; manufacturer); the third is a Japanese academic pharmacology study using the RP-27267 code.
• sanofi-aventis Annual Report on Form 20-F (fiscal year 2006), US SEC. Company filing (France); used for corporate history.
• openFDA Drugs@FDA query for zopiclone and Lunesta (eszopiclone) label. US FDA public data.
• DEA Controlled Substances, alphabetical order. US government.
• MHRA Class 4 medicines defect notice, zopiclone tablets (2020). UK regulator; cited only to show that generic suppliers exist.

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