IUPAC: 5-ethyl-5-phenylpyrimidine-2,4,6(1H,3H,5H)-trione · C₁₂H₁₂N₂O₃ · MW 232.24 g/mol · CAS 50-06-6 · ATC N03AA02 · ChEMBL40
Phenobarbital (Luminal, Gardenal). The oldest anticonvulsant still in clinical use — introduced by Bayer in 1912, the prototype barbiturate and the drug against which every later sedative-hypnotic was measured. A positive allosteric modulator of the GABAA receptor at the barbiturate/anaesthetic site in the transmembrane domain, it prolongs channel open duration and, at high concentration, gates the channel directly even without GABA. That single property — direct gating with no ceiling — is why the barbiturates were displaced by the benzodiazepines and why their overdose kills. Names: Luminal, Gardenal, phenobarbitone, "phenobarb," "purple hearts."
Phenobarbital and diazepam both amplify GABAA inhibition, but they do it in mechanistically opposite ways, and the difference is the whole story of barbiturate danger. The benzodiazepine binds an extracellular pocket and increases the frequency of channel openings, but only when GABA is already there. Phenobarbital binds sites in the transmembrane domain (TMD) — the intersubunit "anaesthetic" cavities near the β+/α− and α+/β− interfaces resolved in the cryo-EM structure at right — and increases the mean open duration of each opening. At low concentration it is a positive allosteric modulator riding on ambient GABA. At high concentration it becomes a direct agonist, gating the chloride channel with no GABA present at all. That is the fatal property: a modulator has a ceiling (it can only amplify what GABA is doing); a direct gate does not.
GABAA is a pentameric ligand-gated anion channel. GABA at the two β+/α− orthosteric sites opens the pore; Cl⁻ flows in, hyperpolarizing the neuron and raising firing threshold. This is the substrate both benzodiazepines and barbiturates act upon.
Phenobarbital stabilizes the open state, prolonging the mean channel open time rather than opening more often. Total Cl⁻ charge transfer per GABA event rises. Contrast benzodiazepines: they raise opening frequency and leave open duration essentially unchanged.
Above roughly the low-millimolar range, phenobarbital opens the channel by itself, GABA-independently — a barbiturate acting as an agonist, not just a modulator. Respiratory-drive neurons can be silenced outright. This is the mechanistic reason barbiturate overdose has no self-limiting envelope.
The Kim/Hibbs cryo-EM structures (PDB 6X3W) place phenobarbital in transmembrane intersubunit pockets — the same class of cavities used by etomidate, propofol and pentobarbital — distinct from the extracellular benzodiazepine site. Multiple copies bind; occupancy scales the effect smoothly, without saturation.
Beyond GABAA, phenobarbital blocks AMPA/kainate glutamate receptors at anticonvulsant-relevant concentrations, cutting excitatory drive. Dual action — enhancing inhibition and suppressing excitation — underlies its broad, if crude, antiseizure efficacy.
At higher concentration it dampens voltage-gated Na⁺ currents (ChEMBL IC50 ≈ 10 µM vs brain Na channels) and high-voltage-activated Ca²⁺ channels, contributing to membrane stabilization and to cardiovascular depression in overdose.
Because the effect is not use-dependent on ambient GABA once direct gating begins, phenobarbital has no equivalent of the benzodiazepine's self-limiting inhibition. Where isolated diazepam overdose plateaus, phenobarbital's dose–response marches continuously into deep coma, apnea and cardiovascular collapse. The benzodiazepine antagonist flumazenil is useless here — it competes at the extracellular BZD site, which phenobarbital does not occupy.
Phenobarbital is a weak acid (pKa ≈ 7.3) with modest lipophilicity: well absorbed orally (bioavailability ~80–100%) but with a slow onset — peak plasma levels take hours, and the CNS peak lags further, which is why it is a poor "as-needed" sedative and a good maintenance anticonvulsant. Its defining PK feature is an extraordinarily long half-life (~79–100 h in adults, longer in neonates): dosing once daily is sufficient, but steady state takes 2–3 weeks to reach, and a single overdose lingers for days. The second defining feature is that phenobarbital is a potent, broad inducer of hepatic drug-metabolizing enzymes — it remodels the liver's entire clearance machinery, including its own.
Metabolism cascade: Roughly three-quarters of a dose is hepatically cleared — CYP2C9 (with minor CYP2C19/2E1) aromatic-hydroxylates phenobarbital to p-hydroxyphenobarbital (inactive), which is then glucuronidated/sulfated and excreted. The remaining ~20–25% is eliminated unchanged in urine, and because phenobarbital is a weak acid this route is strongly pH-dependent: alkalinizing the urine (sodium bicarbonate) ionizes the drug, traps it in the tubule, and markedly accelerates clearance — the pharmacologic basis of overdose management.
Enzyme induction is the interaction engine. Phenobarbital is a classic activator of the constitutive androstane receptor (CAR), up-regulating CYP3A4, CYP2C9, CYP2C19, CYP1A2, CYP2B6, UGTs and P-glycoprotein over 1–3 weeks. It therefore accelerates the clearance of a huge list of co-medications — warfarin, direct oral anticoagulants, oral contraceptives (contraceptive failure), corticosteroids, many antiepileptics (lamotrigine, valproate), immunosuppressants (cyclosporine, tacrolimus), antiretrovirals, and more — often to sub-therapeutic levels. It also induces its own metabolism (autoinduction). Stopping phenobarbital then reverses the induction over weeks, and co-drugs that were being cleared fast can rebound to toxic levels. Every start and every stop is a drug-interaction event.
For half a century phenobarbital and its cousins were the sedative, hypnotic and anxiolytic pharmacopoeia. They also filled the morgue. The problem was never efficacy — it was the therapeutic index: the ratio between a dose that calms and a dose that kills is narrow for barbiturates (on the order of a handful-fold), and the curve has no plateau. When the benzodiazepines arrived in the 1960s with comparable anxiolytic and anticonvulsant efficacy but a flat, ceilinged overdose profile, they displaced barbiturates for anxiety and insomnia almost completely. Phenobarbital survived only where its cheapness and long half-life still win: epilepsy (especially neonatal seizures and the WHO essential-medicines role in low-resource settings), status epilepticus, and the management of severe alcohol and barbiturate withdrawal.
Because phenobarbital gates the chloride channel directly at high concentration, escalating dose produces escalating CNS and respiratory depression without limit: sedation → stupor → coma → loss of brainstem respiratory drive → apnea → death. There is no self-limiting envelope and no specific antidote — flumazenil does not bind the barbiturate site. This is the exact property benzodiazepines lack, and the reason they replaced barbiturates.
Two tolerances stack: pharmacodynamic (GABAA systems adapt to sustained potentiation) and metabolic (autoinduction of CYP2C9 speeds clearance). Sedative tolerance grows faster than the lethal-dose threshold, so the gap between "the dose I now need" and "the dose that stops my breathing" narrows with chronic use — a uniquely dangerous tolerance geometry.
Chronic exposure drives compensatory GABAergic down-regulation and glutamatergic up-regulation. Abrupt cessation unmasks a hyperexcitable brain: tremor, anxiety, insomnia, orthostatic hypotension, and — in severe cases — generalized seizures, status epilepticus, delirium and death. Barbiturate withdrawal is, like alcohol and benzodiazepine withdrawal, one of the few withdrawal syndromes that is directly lethal.
Never stop abruptly after regular use. Phenobarbital's own very long half-life makes it a relatively forgiving taper agent (it self-buffers the decline), and it is used to taper people off shorter-acting barbiturates. Withdrawal onset can be delayed several days by that half-life, masking causation. Discontinuation is a supervised medical procedure with seizure precautions — not willpower.
The lesson of phenobarbital is the lesson of the whole barbiturate class: a molecule can be genuinely effective and genuinely essential (it is still on the WHO Model List) while carrying a therapeutic margin so thin, an overdose curve so unforgiving, and an interaction and withdrawal profile so treacherous that a safer mechanism — the ceilinged, GABA-dependent benzodiazepine — rightly took its place for nearly everything except seizures.
Evidence-based, non-judgmental. Phenobarbital is not a benzodiazepine: it has no overdose ceiling, no antidote, a days-long half-life, and it rewires liver metabolism. The two lethal vectors are combination and abrupt discontinuation.
| Target | Potency | Rel. | Action |
|---|---|---|---|
|
GABAA
potentiation of GABA current
|
EC50 ≈ 50–100 µM
electrophysiology
|
PAM | |
|
GABAA
direct gating (GABA-independent)
|
EC50 ≈ 0.5–3 mM
high-dose agonism
|
AGONIST | |
|
NaV (brain)
voltage-gated Na⁺ channel
|
IC50 ≈ 10 µM
[³H]BTX displ. · ChEMBL
|
BLOCK | |
|
AMPA / kainate
glutamate receptor block
|
high µM
functional
|
BLOCK | |
|
CaV (HVA)
high-voltage Ca²⁺ current
|
high µM
secondary
|
weak |