GHB

GHB is a small molecule the body makes on its own. It is a short-chain fatty acid formed from GABA, the brain's main inhibitory neurotransmitter, and it is present in normal brain tissue ^[8]. Taken as a drug it is a depressant: euphoria, loss of inhibition, drowsiness and amnesia at lower exposures, and sedation, respiratory depression, coma and death at higher ones ^[8].

It reaches people in two very different ways. The prescription salt is sodium oxybate, and there is a mixed calcium, magnesium, potassium and sodium version carrying 92 percent less sodium ^[22]. Both treat cataplexy and excessive daytime sleepiness in narcolepsy, and the lower-sodium one is also approved for idiopathic hypersomnia, where it was the first medication ever approved for the condition ^[4]. The street drug is the same molecule without the label, the known concentration or the supervision.

What it does in the brain is now reasonably settled. GHB binds two very different things: high-affinity sites that have been known for decades, and the GABA-B receptor, which it engages only weakly ^[19]. A mouse study separated them. Animals bred without a working GABA-B receptor still carried the high-affinity GHB sites, in the same distribution and number, but GHB no longer produced any of its effects in them ^[18]. So the high-affinity site is real, its identity is still being worked out, and the drug's effects, including the dangerous ones, run through GABA-B ^[8].

The reason overdose is common has two parts and neither is about potency. The relationship between blood concentration and effect is steep, and the processes that carry GHB into and out of the body are all saturable, so exposure rises faster than the amount taken ^[8]. The practical result is that the distance between the effect someone wants and unconsciousness is short, and alcohol shortens it further ^[7].

Two facts carry most of the safety weight. First, GHB kills without help from other drugs: in a series of 226 deaths, 35 percent had no other drug present ^[9]. Second, it kills people who had company; in that same series, help was delayed or never sought in 66 of the 72 cases where somebody had already seen adverse effects, and dozens were left to sleep ^[9]. GHB coma reverses abruptly, over about half an hour, after a deep phase that lasted a median of 90 minutes in one series of unconscious patients, and that pattern is exactly what makes waiting feel reasonable ^[10].

Daily use produces dependence, and stopping is the other dangerous moment. GHB withdrawal is described as life-threatening, it frequently does not respond to benzodiazepines, and it is managed with barbiturates, with baclofen, or with a supervised taper using pharmaceutical GHB ^[13][14][16]. The largest data set, 229 patients across six Dutch centres, achieved detoxification in 85 percent of admissions in about 12.5 days for most patients, with the main complications being hypertension and anxiety; 69 percent relapsed within three months ^[16].

Acute overdose is the common event rather than the rare one, and it looks the same each time: sleepiness deepening into coma, bradycardia, hypotension and respiratory failure ^[25]. There is no antidote and no reversal agent. Management is supportive, which in practice means continuous monitoring of vital signs and protecting the airway until the drug clears, and with that care most people recover fully within about six hours ^[24]. Vomiting while unconscious is part of what makes the coma dangerous rather than merely alarming; six episodes of vomiting occurred in a single monitored laboratory session that combined GHB with alcohol ^[5].

The most important finding on this page is about bystanders. In a series of 226 GHB-associated deaths, 213 died of cardiorespiratory arrest, and 78 of the 226, 35 percent, had no other drug present at all ^[9]. Cardiorespiratory arrest happened before any hospital contact in every one of the 184 cases where a history was available. Of 72 cases where adverse effects had been witnessed beforehand, medical help was delayed or never sought in 66; 40 of those people were left to sleep it off although the GHB use was known, and another 30 were left sleeping and found dead ^[9]. Unconsciousness after GHB is not sleep and is not something to sit beside. Fatality counts drawn from postmortem toxicology deserve one caveat, and that series built the caveat into its own design: GHB is generated in a body after death, so it required far higher concentrations in postmortem samples than in blood drawn from the living before counting a case as GHB-positive ^[9].

Alcohol is the drug most often found alongside GHB in acute intoxication, and the combination is worse than either alone ^[7]. The only controlled human experiment gave sixteen healthy adults GHB and ethanol separately and together under monitoring: oxygen saturation fell with each drug alone and fell furthest with both, and the combination produced hypotension in two participants and vomiting in six ^[5]. The average fall in oxygen saturation was small at the modest amounts used in that laboratory, so the honest reading of it is direction rather than magnitude; nobody has run the equivalent study at recreational amounts, with recreational drinking, or alongside opioids and benzodiazepines, which is where the deaths actually occur. Animal work shows why waiting is not a strategy: ethanol lengthened GHB sleeping time beyond the sum of the two drugs' separate effects while leaving GHB blood exposure essentially unchanged, so the interaction is happening at the receptor and not in the liver ^[6].

Withdrawal is the other way this compound kills, and it does not behave like benzodiazepine withdrawal. The literature calls it life-threatening and states plainly that it does not always respond to benzodiazepines ^[13]. In the founding review of 38 cases, 53 percent deteriorated into delirium, benzodiazepine-refractory cases needed barbiturates or chloral hydrate instead, and one patient died ^[11]. One published case makes the failure concrete: a woman in withdrawal received 225 mg of diazepam with no effect at all, then needed propofol and intubation, and two further days of propofol, alfentanil and dexmedetomidine to stay sedated ^[12]. The mechanism explains it, because GHB acts at GABA-B while benzodiazepines and most anaesthetic agents act at GABA-A, so the standard answer is aimed at the wrong receptor ^[12]; that is also why baclofen, a GABA-B agonist, keeps appearing in these cases ^[13], and why phenobarbital is reached for when benzodiazepines and baclofen together are not enough ^[14]. Risk tracks how often someone dosed rather than how long they used, because the duration of action is short enough that a dependent user is redosing around the clock ^[11]. The commonly quoted five to seven days is not reliable either; one man's withdrawal delirium recurred repeatedly across 56 days, with laboratory proof that he had not used in between ^[15].

The prescription product carries a different and much more mundane set of risks. Nausea, headache, dizziness, anxiety and vomiting were the common adverse events in the idiopathic hypersomnia trial ^[4], and enuresis joined that list in the original narcolepsy trial ^[1]. Sodium oxybate also delivers a real sodium load, large enough that a version carrying 92 percent less sodium was developed specifically to reduce it ^[22]. Whether that load matters clinically is genuinely unsettled: no threshold has been established at which the sodium in sodium oxybate raises cardiovascular risk, and the data collected so far show a very low frequency of cardiovascular side effects and no overall association with cardiovascular risk in narcolepsy ^[21]. GHB is also used to commit sexual assault; two cases facilitated by the pharmaceutical product were confirmed in postmarketing review, and the drug's reputation in that respect is not a myth ^[23][8]. This entry is educational and is not medical advice.

The medical and recreational histories of GHB run alongside each other rather than in sequence. In the 1990s it was sold as a dietary supplement and used as a club drug, and both facts show up in the mortality record: of 226 GHB-associated deaths, sixteen involved products sold as supplements and one involved the pharmaceutical version ^[9]. The date-rape label and the club-drug reputation both date from that decade ^[7].

The medicine arrived from the other direction. A multicentre randomised trial in 136 narcolepsy patients, who were having a median of 21 cataplexy attacks a week, produced dose-related reductions in cataplexy, daytime sleepiness and clinician-rated severity, and that trial is what turned GHB into sodium oxybate ^[1]. A 12-month open-label extension in 118 of those patients found the benefit held, was maximal after about eight weeks, and showed no evidence of tolerance ^[2]. The product reached the market in 2002 and was dispensed through a restricted distribution programme from the start ^[23].

What followed was a series of narrower approvals. Sodium oxybate was approved in Italy and Austria for alcohol use disorder, a use with its own trial record and its own pharmacovigilance base ^[27]. A mixed calcium, magnesium, potassium and sodium oxybate carrying 92 percent less sodium was then developed and tested in a randomised-withdrawal design in narcolepsy with cataplexy ^[3], and in August 2021 that formulation became the first medication ever approved for idiopathic hypersomnia; until that month there was none at all ^[4][22].

Dependence became a public-health story mainly in the Netherlands, where the number of patients in treatment for GHB dependence rose sharply ^[26] and six addiction treatment centres built a national monitoring project around a detoxification protocol using pharmaceutical GHB ^[16].

GHB has three reputations that rarely meet. In sleep medicine it is a legitimate and rather unglamorous medicine: cumulative postmarketing review across roughly 26,000 patients recorded abuse in 0.039 percent, dependence in 0.016 percent, and five diversion incidents across about 600,000 bottles distributed ^[23]. The alcohol-treatment literature reads similarly; pharmacovigilance covering 299,013 patients exposed in Austria and Italy reports transitory dizziness and vertigo as the main adverse events, rare serious events, and no death attributed to the drug, with abuse and dependence described as rare in patients without psychiatric comorbidity or polydrug use ^[27].

In emergency medicine and forensic toxicology the same molecule has the opposite reputation, and it is earned in a different population: unmeasured amounts, unlabelled product, alcohol already on board, and 226 deaths in a single series ^[9]. In addiction medicine it is the drug that made a national monitoring project necessary ^[16].

The recreational defence usually offered, that GHB must be safe because the body already makes it, is the weakest of these positions. GHB is endogenous ^[8], and the natural experiment for chronic endogenous elevation is succinic semialdehyde dehydrogenase deficiency, where the enzyme that clears this pathway is missing, 4-hydroxybutyric acid accumulates, and the phenotype is developmental delay, hypotonia, ataxia, behavioural problems and seizures in about half of patients ^[20]. Being endogenous is a fact about origin, not a safety argument.

One more asymmetry is worth naming plainly. Sodium oxybate is an approved treatment for alcohol withdrawal and for maintaining abstinence in two European countries ^[27] while being dangerous to co-ingest with alcohol ^[5][7]. Both are true at once, and holding them together is most of what it means to understand this drug.