The sedative action of potassium bromide was recognized more than 150 years ago. In adequate doses the compound was found to cause sedation, drowsiness, and sleep. However, the sleep induced by the bromide salt was neither as deep nor as “refreshing” as that produced by chloral hydrate or paraldehyde, and hangover was frequent. In addition, single doses sufficient to cause sedation and sleep were very difficult to retain without vomiting, which led to its displacement by the barbiturates and benzodiazepines for the treatment of insomnia.

The compound was synthesized by Liebig in 1832, and introduced into medicine by Liebreich in 1869 [19]. By that time ethyl alcohol, opium, and cannabis were mainly used for the treatment of insomnia. In small doses the compound induces sedation, whereas greater doses, in the order of 1–2 mg, taken in a quiet environment induce sleep accompanied by dreams. Sleep usually starts after a latency of 60 min and lasts for at least 5 h. Greater amounts of chloral hydrate have been shown to increase the duration of sleep, but also to augment the incidence of hangover. Habituation, tolerance, and addiction were described with some frequency in patients taking the derivative for long periods of time.

The drug was discovered by Wiedenbusch in 1829 and introduced into medicine by Cervello in 1882 [8]. Administration of a dose of 4–8 ml by oral route was reported to induce sleep in 10–15 min which persisted for only 2–3 h, and was accompanied with some frequency by after effects. Paraldehyde was considered an efficacious hypnotic drug but its penetrating odor and burning taste led to its withdrawal for the treatment of insomnia.

Barbital (veronal) , the oldest barbiturate was introduced into medicine by Fisher and von Mering in 1903. About 10 years later Loewe [20] brought into use phenobarbital (luminal). An additional number of derivatives with shorter duration of action were synthesized during the next few years including secobarbital, amobarbital, and pentobarbital. Barbiturates became the drugs of choice for the treatment of insomnia, and this practice persisted until the introduction of the benzodiazepine derivatives in the 1970s. Almost all available barbiturates were shown to be effective for the treatment of “simple insomnia” (presently called primary insomnia) after administration of appropriate oral doses. They differed mainly in speed of onset of action and in duration of action. Sleep usually started with a latency of 20–60 min and, depending on the elimination half-life of the derivative, lasted from 2 to 8 h. In the absence of sleep laboratory studies, it was initially proposed that the sleep induced by barbiturates closely resembled physiological sleep. Notwithstanding this, patients reported that their sleep was dreamless. Tolerance and addiction were shown to develop during the chronic use of barbiturates. Of interest, depending on the dose administered, all the barbiturates used for the treatment of insomnia were known to produce a broad range of effects, extending from sedation to deep coma and death, which can explain the relatively high incidence of suicide attempts with these drugs .

The 1,4-benzodiazepine derivatives were synthesized by Sternbach et al. in the late 1950s [34, 35] . Soon after, Randall et al. characterized the sedative property of chlordiazepoxide and diazepam in laboratory animals [31]. A few years later several other benzodiazepine derivatives were introduced for the selective treatment of insomnia, and became for almost 20 years the most frequently prescribed group of hypnotics, because of their efficacy and relative safety compared with the barbiturates and chloral hydrate. According to their elimination half-life they were divided into short- (midazolam, triazolam), intermediate- (flunitrazepam, temazepam), and long-acting (flurazepam, nitrazepam) derivatives. The evaluation of the effect of benzodiazepine hypnotic drugs on sleep induction and maintenance was carried out mainly in patients with primary chronic insomnia, and was based on sleep laboratory studies and subjective data from clinical trials. Oswald and Priest [29] evaluated for the first time a benzodiazepine hypnotic (nitrazepam) in the sleep laboratory in the year 1965. A few years later, Kales et al. [15] and Monti et al. [25] examined the effects of flurazepam and flunitrazepam, respectively, on sleep variables in patients with a diagnosis of insomnia . It was found that the sleep induced by benzodiazepine hypnotics was characterized by a shortened sleep-onset latency, decreased number of nocturnal awakenings, reduced time spent awake, an increase in stage 2 NREMS, a consistent reduction in slow wave sleep, a dose-dependent suppression of REMS, and an improvement in the subjective quality of sleep compared with no treatment. Thus, from the point of view of sleep architecture, they did not induce a physiological sleep. It was learned also that the short-acting benzodiazepine derivatives were effective predominantly in patients with a prolonged sleep-onset latency, and that rebound insomnia occurred upon their sudden withdrawal. The benzodiazepines, irrespective of their eliminations half-life, showed to be effective hypnotics when administered over a relatively short period of time (4–6 weeks) in patients with chronic insomnia. In addition, dependence was described in patients who were given large doses over a prolonged period of time .

In addition to the benzodiazepine hypnotics, a structural dissimilar group of non-benzodiazepine derivatives has become available for the treatment of insomnia. They include zopiclone, eszopiclone, zolpidem, and zaleplon that belong to the cyclopyrrolone, imidazopyridine, and pyrazolopyrimidine classes, respectively . Together with the benzodiazepine derivatives they act as GABA-A receptor allosteric modulators. Melatonin and the melatonin receptor agonist ramelteon are being used also for the treatment of insomnia.