Why this poison is interesting

The Z-drugs were invented to be benzodiazepines without the parts prescribers disliked: shorter-acting, subtype-selective, marketed as less habit-forming. In overdose they inherit the one benzodiazepine property that matters most — the ceiling. Because they are positive allosteric modulators, enhancing the action of GABA rather than opening the chloride channel by themselves, there is a limit to how much respiratory depression a lone overdose can produce. The zopiclone label states it directly: overdose should not be life threatening unless combined with other CNS depressants, including alcohol.¹

So the library's benzodiazepine page and its ceiling argument ought to cover them, and at the receptor it does. But the plan for this library flagged the Z-drugs as an entry the benzodiazepine page's argument does not straightforwardly reach — and the reason is in the mortuary data. In an English study of self-poisoning between 2005 and 2012, zopiclone/zolpidem overdose was nine times more likely to end in death than diazepam (fatal-toxicity index) and twelve times more likely by case fatality, differences that alcohol involvement… was unlikely to account for.³ A drug with a receptor ceiling should not be killing people at nine times the rate of another drug with the same ceiling.

The resolution is the whole point of the page, and it is a lesson in choosing the comparator. Diazepam is a long-acting, low-toxicity benzodiazepine. Measured instead against benzodiazepines as a class — which includes the genuinely more toxic alprazolam and temazepam — zopiclone's fatal toxicity was not significantly different… when adjusted for usage.⁴ The Z-drugs are more dangerous than the safest benzodiazepine and about as dangerous as the average one, and in both datasets they are contributory factors rather than primary substances⁴ — the deaths are co-ingestion deaths. The ceiling is real; it just was never a promise of safety in company.

A poison is a drug whose kinetics have escaped its pharmacology.

The toxic principle

The toxic principle is the same as the therapeutic one: enhancement of GABA-mediated inhibition at the GABA-A receptor, through the benzodiazepine recognition site rather than through a channel the drug opens alone. Zolpidem is an imidazopyridine which selectively binds the omega-1 receptor subtype² — the α1-containing GABA-A receptors that mediate sedation — which is why it is a hypnotic with comparatively little of the anxiolytic, anticonvulsant or muscle-relaxant action of a classical benzodiazepine. Zopiclone, a cyclopyrrolone, has hypnotic, sedative, anxiolytic, anticonvulsant and muscle-relaxant actions and acts on a different site to those of benzodiazepines including different conformational changes in the receptor complex¹ — a distinct binding site on the same macromolecular complex.

Because both are positive modulators, the overdose is a graded deepening of sedation — varying degrees of central nervous system depression ranging from drowsiness to coma according to the quantity ingested¹ — that reaches a plateau rather than a cliff. This is the ceiling, and it is the same mechanism that makes lone benzodiazepine overdose usually benign. The subtype selectivity of zolpidem means its overdose is even more purely hypnotic: the label describes impairment of consciousness ranging from somnolence to coma² with little of the additional syndrome a broader GABA-A modulator might add. Established

The reason the Z-drugs are more toxic than diazepam but not clearly more toxic than the benzodiazepine class is not a special property of the molecule; it is that diazepam is an unusually safe benzodiazepine. The fatal-toxicity ranking in the English data placed temazepam (10×) and zopiclone/zolpidem (9×) close together and both far above diazepam³, while the New Zealand data — comparing against all benzodiazepines and finding alprazolam and chlormethiazole to be the toxic outliers — put zopiclone in the middle of the pack.⁴ Same drug, two rankings, decided by the reference drug. Established

Toxicokinetics

The two drugs differ in their kinetics more than in their pharmacodynamics, and the differences matter mainly for duration. Both are rapidly absorbed short-acting hypnotics cleared by hepatic metabolism; zolpidem is the shorter-acting and the more highly protein-bound, zopiclone the slightly longer-acting with an active metabolite. Neither is a candidate for extracorporeal removal.

Zopiclone and zolpidem — two short-acting hypnotics, neither removable by dialysis
Parameter	Therapeutic	In overdose	Why it changes
Absorption	Zopiclone: Cmax 1.5–2 h, unaffected by food¹. Zolpidem: Cmax 0.5–3 h, bioavailability 70%, ~35% first-pass²	Rapid for both — the onset of sedation is fast	Fast absorption means the overdose declares itself early rather than after a latent interval; there is no metabolic delay to a toxic species and no slow-release preparation in routine UK use.
Distribution	Zopiclone: protein binding ~45%, weak and non-saturable, Vd 91.8–104.6 L (≈1.3 L/kg)¹. Zolpidem: protein binding 92.5%, Vd 0.54 L/kg²	Zolpidem is tightly bound but small-volume; zopiclone is loosely bound	Zolpidem's high protein binding and small volume together make it a poor dialysis target — the label confirms it is not dialyzable². Zopiclone's loose binding is offset by a metabolic clearance route that dialysis cannot improve on.
Metabolism	Zopiclone: CYP3A4 (and CYP2C8) to an N-oxide (active in animals) and N-desmethyl (inactive)¹. Zolpidem: CYP3A4 (+CYP1A2); all metabolites… pharmacologically inactive²	Hepatic; CYP3A4 the shared main route	Both are CYP3A4 substrates, so CYP3A4 inhibitors raise the concentration and inducers lower it. Zolpidem produces no active metabolite; zopiclone's N-oxide is active in animals but contributes little to the human overdose picture.
Elimination	Zopiclone: mainly metabolic; renal clearance of unchanged drug only 8.4 mL/min vs plasma clearance 232 mL/min¹. Zolpidem: metabolites in urine (56%) and faeces (37%)²	Metabolic, not renal — renal failure alone does not cause accumulation	Half-lives are short — ~5 h (zopiclone) and mean 2.4 h (zolpidem)¹² — so a single-agent overdose is self-limiting over hours. The long tail belongs to the co-ingestant, not to the Z-drug.
Dialysability	—	No. Zolpidem is not dialyzable, and dialysis in renal-failure patients on therapeutic doses showed no reduction in levels². Zopiclone has no EXTRIP recommendation and a metabolic clearance route	There is nothing for a dialyser to do: zolpidem is too highly bound and too small in volume, and both are cleared metabolically faster than a circuit would clear them. Extracorporeal treatment has no place here.

Metabolism and the metabolites

Neither drug is bioactivated to a poison, and the metabolic story is short. Zolpidem is metabolised by CYP3A4 with a CYP1A2 contribution to pharmacologically inactive products.² Zopiclone is metabolised, chiefly by CYP3A4, to an N-oxide that is active in animals and an N-desmethyl metabolite that is inactive¹ — but the human overdose is the parent drug's, and neither metabolite drives it. The only toxicological consequence of the metabolism is the CYP3A4 dependence, which makes both drugs sensitive to interacting inhibitors and inducers.

The Z-drugs — GABA-A modulation, hepatic clearance, no toxic metabolite

Zolpidemω1-selective GABA-A modulator; half-life ~2.4 h²
ZopicloneCyclopyrrolone; distinct GABA-A site; half-life ~5 h¹
GABA-A receptor (benzodiazepine site) — positive allosteric modulationEnhances GABA rather than opening the channel alone — the source of the overdose ceiling
CYP3A4 (± CYP1A2 / CYP2C8) — the shared clearance routeZolpidem → inactive metabolites; zopiclone → an animal-active N-oxide and an inactive N-desmethyl¹². No poison is produced
Urinary and faecal excretion of metabolitesMetabolic clearance dominates; only ~8.4 mL/min of zopiclone leaves unchanged in urine¹

Elimination and accumulation

Elimination is metabolic and quick. Zopiclone's plasma clearance is 232 mL/min with a half-life near 5 hours and no accumulation on repeated dosing¹; zolpidem's half-life is a mean of 2.4 hours². A single-agent overdose therefore resolves over hours, and the patient who is going to wake up does so within the same emergency-department stay. The accumulation risks are not renal but pharmacokinetic-interaction and co-ingestion risks.

There is no enterohepatic recirculation of consequence and no active human metabolite to prolong the effect. The short half-lives are why prolonged coma from a Z-drug alone should prompt a search for a co-ingestant or another cause rather than being attributed to the hypnotic. Inferred

Target organs — and why those

Brain — the α1 GABA-A receptor

TargetPositive allosteric modulation at the benzodiazepine site, ω1/α1-selective for zolpidem

Why hereThe intended and the toxic target. Enhancement of GABA at the α1-containing receptors produces sedation, and in overdose a graded depression of consciousness to coma.¹² The ceiling — the reason a lone overdose plateaus rather than progressing to apnoea — is a direct property of positive allosteric modulation: with no GABA present, the drug does little on its own. This is the organ the drug was designed to act on, doing more of the same. Established

At the bedsideDrowsiness, confusion, ataxia, hypotonia progressing to coma; recovery over hours in a single-agent ingestion.¹ Zolpidem's picture is more purely hypnotic because of its ω1 selectivity.²

The respiratory centre — only in company

TargetGABA-A modulation summating with other CNS depressants below each drug's ceiling

Why hereIncluded because it is where the deaths occur, and because the mechanism is summation rather than an intrinsic property. A Z-drug alone rarely depresses ventilation to a fatal degree — the ceiling holds — but combined with alcohol, an opioid or another sedative, the additive effect crosses a threshold neither drug would reach alone.¹³ The organ is targeted not by the drug but by the combination, which is why the population death rate outruns the receptor pharmacology. Established

At the bedsideRespiratory depression, chiefly in mixed overdose or in a debilitated patient.¹ Prolonged or deep respiratory depression from a Z-drug alone should prompt a search for a co-ingestant.

The red cell — zopiclone only, in severe overdose

TargetNot established; the label reports methaemoglobinaemia and haemolysis without a mechanism

Why hereIncluded because it is genuinely anomalous and because a reader who sees methaemoglobinaemia in a sedative overdose needs to know the label predicts it. The zopiclone SmPC lists methaemoglobinaemia, haemolysis/haemolytic anaemia among severe-overdose findings¹; the occurrence is label-documented but the mechanism is not given, and zolpidem does not share it. The card exists to record a documented, unexplained toxicity, not to assert a pathway. Established

At the bedsideCyanosis unresponsive to oxygen, a saturation gap, or haemolysis in a severe zopiclone overdose¹; see methaemoglobin inducers for how that syndrome is recognised and treated.

Timeline of effects

A short, self-limiting course when the Z-drug is alone — and an open-ended one when it is not

Time

What you seeWhat is happening

0–2 hOnset
What you seeRapid drowsiness, ataxia, slurred speech; deepening toward coma with a large ingestion.¹²
What is happeningFast absorption — Cmax at 1.5–2 h (zopiclone) and 0.5–3 h (zolpidem).¹² Sedation tracks the concentration; there is no latent phase.
2–8 hPeak and plateau
What you seeMaximum CNS depression; in single-agent overdose this reaches a ceiling and holds rather than progressing to apnoea.¹
What is happeningPositive allosteric modulation has a limit without additional GABA drive. In mixed overdose the plateau is breached — the summation with alcohol or opioids is what turns a benign course dangerous.¹³
HoursResolution — if alone
What you seeWaking over hours; a patient who does not wake as expected has probably taken something else.
What is happeningShort half-lives (~5 h zopiclone, ~2.4 h zolpidem¹²) and metabolic clearance mean a lone Z-drug overdose is self-limiting within the same admission.
Severe zopiclone overdoseThe haematological tail
What you seeRarely, cyanosis, a saturation gap or haemolysis.¹
What is happeningThe label's reported methaemoglobinaemia/haemolysis¹ — unexplained, zopiclone-specific, and worth recognising because it is treated quite differently from the sedation.

What the mechanism predicts at the bedside

A lone Z-drug overdose is usually benign and self-limiting. The receptor ceiling and short half-lives mean most patients need airway and supportive care and wake over hours.¹²
The history of co-ingestion is the assessment. The deaths are co-ingestion deaths — the Z-drug is contributory rather than primary⁴ — so what else was swallowed, especially alcohol or an opioid, matters more than the Z-drug dose.
Deep or prolonged coma from a Z-drug alone does not fit. With half-lives of 2–5 hours¹², a patient who stays deeply unconscious should be reassessed for a co-ingestant, an airway complication or another cause.
Flumazenil works, and the labels tell you when not to use it. Both name flumazenil for serious CNS depression but restrict it: the zopiclone label says NOT TO BE USED IN MIXED OVERDOSE OR AS A "DIAGNOSTIC" TEST and notes its ~1-hour half-life¹; the zolpidem label warns it may contribute to the appearance of neurological symptoms (convulsions) and wears off in 40–80 minutes.² See flumazenil.
Do not send them for dialysis. Zolpidem is not dialyzable² and zopiclone has no dialysis role; extracorporeal treatment cannot help.
Activated charcoal for a recent significant ingestion — the zopiclone label suggests it after more than 150 mg in an adult within an hour.¹ See activated charcoal.
In a severe zopiclone overdose, look for a saturation gap. The label predicts methaemoglobinaemia and haemolysis¹; unexplained cyanosis or a low SpO₂ that does not correct with oxygen points to the methaemoglobin syndrome, which is managed on its own terms.
Do not be reassured by 'just a sleeping tablet'. The English data put zopiclone/zolpidem overdose at nine times the fatal toxicity of diazepam³ — safer than the most dangerous benzodiazepines, but not the harmless agents their reputation suggests.

The antidote, from the poison's side

There is a pharmacological antidote — flumazenil — and it is a rare case in this library of the labels recommending an antidote and then spending most of their words restraining its use. That tension is the whole of the antidote story here.

Flumazenil: A competitive antagonist at the benzodiazepine site, so it reverses both Z-drugs. Both labels permit it for serious CNS depression and then restrict it: NOT TO BE USED IN MIXED OVERDOSE OR AS A "DIAGNOSTIC" TEST¹, short-acting (~1 hour; 40–80 minutes¹²) so resedation is expected, and may contribute to… convulsions² — a real hazard where a proconvulsant co-ingestant or benzodiazepine dependence is present. See flumazenil.
Airway and ventilatory support: The mainstay, because the danger is respiratory depression and the half-lives are short. Supporting ventilation through the plateau is usually all a single-agent overdose needs.
Activated charcoal: For a recent significant ingestion — the zopiclone label suggests it after >150 mg in an adult within an hour, or gastric lavage within an hour of a potentially life-threatening overdose.¹ See activated charcoal.
Methylthioninium (methylene blue): Only for the zopiclone-specific complication — the label's reported methaemoglobinaemia¹. It treats a different problem from the sedation and is relevant only if a saturation gap appears; see methaemoglobin inducers.
Dialysis: No role. Zolpidem is not dialyzable²; zopiclone is cleared metabolically. Extracorporeal treatment cannot lower the concentration meaningfully.

Critical appraisal

The 9× and 12× figures are real, specific and diazepam-referenced. Geulayov 2018 compared zopiclone/zolpidem with diazepam in England, 2005–2012, and found fatal-toxicity and case-fatality indices of nine and twelve³. They are quoted with the comparator attached, because against a different benzodiazepine the numbers would differ.
This is not a manufactured 'sources disagree'. Geulayov (vs diazepam, 9×³) and Reith (vs benzodiazepines as a group, not significantly different⁴) genuinely reach different rankings — but the two figures sit under different comparators, not the same one, and the page reconciles them rather than presenting them as a bare contradiction. Diazepam is a low-toxicity benzodiazepine; the class contains more toxic members; the Z-drugs sit between.
Both studies agree the deaths are co-ingestion deaths. Reith found hypnosedatives were contributory factors rather than primary substances⁴ and the sole agent in only one of 200 poisoning deaths; the labels say the same¹. The page's load-bearing claim rests on agreement, not on the disputed ranking.
The methaemoglobinaemia/haemolysis is documented but unexplained, and is badged for its occurrence, not for a mechanism. The zopiclone label lists it among severe-overdose findings¹; that the association occurs in humans is established by the label, but no mechanism is given and zolpidem does not share it. The page deliberately badges only the fact that it happens and manufactures no pathway — and it is not badged Traditional teaching, because that tier requires a citation contesting a received mechanism, and there is no received mechanism here to contest.
Flumazenil's harms are better evidenced than its benefit here. The restrictions the labels place on it¹² reflect the general toxicology of flumazenil in overdose — resedation and seizure risk — rather than Z-drug-specific trials; the page states the label position and cross-links the flumazenil monograph for the reasoning.
No lethal dose is given. The zopiclone label states the fatal dose [is] not known¹; the numbers on this page are relative-toxicity indices, not lethal doses, and risk assessment belongs to TOXBASE and NPIS.
Zaleplon is grouped by class, not by its own data. The page's figures are zopiclone's and zolpidem's; zaleplon is named as the third Z-drug on shared pharmacology and is not separately quantified here.

References

1
Zimovane 7.5 mg film-coated tablets (zopiclone) — Summary of Product Characteristics. electronic medicines compendium, product 2855 (Sanofi). Sections 4.9 (Overdose), 5.1 (Pharmacodynamic properties, mechanism of action) and 5.2 (Pharmacokinetic properties). medicines.org.uk/emc/product/2855
2
Zolpidem Tartrate 10 mg Tablets — Summary of Product Characteristics. electronic medicines compendium, product 3975 (Zentiva). Sections 4.9 (Overdose), 5.1 (Pharmacodynamic properties) and 5.2 (Pharmacokinetic properties). medicines.org.uk/emc/product/3975
3
Geulayov G, Ferrey A, Casey D, et al. Relative toxicity of benzodiazepines and hypnotics commonly used for self-poisoning: an epidemiological study of fatal toxicity and case fatality. Journal of Psychopharmacology 2018 Jun;32(6):654–662. PMID 29442611. England 2005–2012; zopiclone/zolpidem 9× the fatal toxicity and 12× the case fatality of diazepam, with alcohol involvement unlikely to account for it.
4
Reith DM, Fountain J, McDowell R, Tilyard M. Comparison of the fatal toxicity index of zopiclone with benzodiazepines. Journal of Toxicology — Clinical Toxicology 2003;41(7):975–980. PMID 14705844. New Zealand 2001; zopiclone's fatal toxicity not significantly different from benzodiazepines as a group when adjusted for usage, with alprazolam and chlormethiazole the toxic outliers, and hypnosedatives contributory rather than primary in poisoning deaths.

Z-drugs (zopiclone and zolpidem)

At a glance