Z-Drugs (zolpidem, zaleplon, eszopiclone, zopiclone)

Non-benzo GABA-A α1-preferring sedatives marketed 1992–2004 as "cleaner benzos for sleep." They are not. They carry a 2019 FDA Boxed Warning for fatal complex sleep behaviors at any dose after a single tablet, produce anterograde amnesia, build tolerance + dependence + rebound insomnia within weeks, and (Z-hypnotics >28 days/quarter) are associated with 79% higher dementia risk in older adults. Daridorexant and seltorexant achieve sleep by removing wake drive instead of forcing sedation, preserve REM/N3 architecture, do not cause rebound, and protect next-day cognition. SKIP-PERMANENT for any chronic use; OPTIONAL only for an acute prescriber-supervised short course (≤2 weeks) when DORAs are unavailable.

The shared mechanism — GABA-A positive allosteric modulation, BZ1-preferring.

GABA-A receptors are pentameric chloride channels assembled from combinations of α (1–6), β (1–3), and γ (1–3) subunits. The α subunit determines the receptor's pharmacology and physiological role:

• α1 (BZ1): highest density in cortex, thalamus, cerebellum; mediates sedation, anterograde amnesia, and most anticonvulsant effect.
• α2/α3: limbic, spinal cord; mediate anxiolysis and muscle relaxation.
• α5: hippocampus; mediates learning + memory modulation.

Benzodiazepines bind non-selectively across α1/α2/α3/α5 → broad effect (sedation + anxiolysis + muscle relaxation + amnesia).

Z-drugs bind the same allosteric site as benzos (the "benzodiazepine recognition site" between α and γ subunits), but with subunit selectivity:

Zolpidem and zaleplon's α1 selectivity is what defines "Z-drug" pharmacologically — sedation > anxiolysis vs benzos. Eszopiclone and zopiclone are pharmacologically closer to benzos in subunit profile despite the marketing distinction. (Carlson et al., Frontiers in Neuroscience 2020.)

What this does to brain physiology (and why it differs from DORAs):

Z-drugs do not remove a wake input. They impose extra GABAergic sedation on every α1-bearing circuit in the cortex and thalamus simultaneously — including circuits that have nothing to do with sleep. This produces:

• Forced sedation that overrides normal sleep architecture. Multiple PSG comparisons show zolpidem suppresses or alters slow-wave sleep (N3) and REM relative to physiological sleep (Landolt et al.; J Neurosci spindle data).
• Sleep spindle augmentation — zolpidem dramatically increases N2 sleep spindle density. This was initially marketed as a memory benefit (consolidation effect) but the same pharmacology produces anterograde amnesia for everything encoded after dosing, and the consolidation signal has not replicated cleanly.
• Anterograde amnesia. α1 PAM at hippocampal-projecting circuits blocks new memory formation while the drug is on board. This is why "I did what?" cases (sleep-driving, sleep-eating, sleep-shopping, sleep-emails) are amnestic — the user has no memory of the period because none was encoded.
• Falls + ataxia. α1 modulation in cerebellum + thalamocortical circuits produces motor incoordination — the elderly fall mechanism.

Half-life-driven phenotypic differences:

Sources: Hesse et al. Drug Saf 2003; Patat et al.; Drugs.com label data.

The "cleaner benzo" claim — modestly true, fundamentally inadequate.

Z-drugs are "cleaner than benzos" in three narrow senses: (1) less daytime anxiolytic spillover at sleep doses because α2/α3 sparing (zolpidem/zaleplon only), (2) shorter half-life than diazepam-class benzos so less morning hangover, (3) marginally lower respiratory depression risk when used alone (still substantial when combined with opioids/alcohol — see boxed warnings). They are not cleaner for sleep architecture preservation, anterograde amnesia, dependence/tolerance, complex sleep behaviors, or dementia signal. DORAs (daridorexant, seltorexant, lemborexant, suvorexant) achieve all the "cleaner" claims and additionally preserve sleep architecture — which is the point of sleep.