Benzodiazepine Dependence: Animal Studies


Animal models are available for the production of both psychological and physical dependence.

The most important animal model for the study of psychological dependence is the operant model using an intravenous self-administration technique originally used for opiate studies (see for example references-) . The animal is trained to self-administer the drug solution through an indwelling cannula by pressing a bar which activates the injection pump. The literature, particularly that relating to opiates has been extensively reviewed and is beyond the scope of this paper. These studies have shown that different groups of drugs have different levels as operant reinforcers. Thus opiates, amphetamines and cocaine are highly potent, ethanol and barbiturates moderately so and mescaline and phenothiazines relatively ineffective. Not all animals of the same species respond to reinforcement in the same way but some develop drug intake patterns which, like those of dependent humans, lead to physical illness and gross withdrawal reactions.

In such experiments benzodiazepines have shown negligible evidence of dependence production. Thus Findley, Robinson and Peregrino studied the effect of intravenous administration of chlordiazepoxide at doses of 1 mg/kg each 3 h. When the animals were then given the chance of intravenous self-administration of chlordiazepoxide (1 mg/kg) or secobarbitol (9 mg/kg) they gradually changed over to secobarbitol dependence over a period of 60 days. Yanagita and Takahashi on the other hand found that dependence could result in monkeys after 1 month on either daily intravenous administration of chlordiazepoxide (75-113 mg/kg) diazepam (8-10 mg/kg) or oxazolam (20-60 mg/kg) or their intravenous self-administration. The dosage is, however, enormous (i.e. about 100Q times the pharmacological dose).

Signs of physical dependence can also be produced in various animal species by chronic administration in high doses of all the general central nervous system depressants that have been studied. These include ethanol, the barbiturates, other hypnotics and tranquillizers.

Much of the animal work has been undertaken on ethanol or barbiturates and the picture appears to be essentially the same in all the species tested and with either ethanol or sedatives. Tremor, convulsions and a peculiar behaviour suggestive of the responses to hallucinatory stimuli have been found in monkeys, cats, dogs,, rats and mice. The rate at which the physical signs developed varied with the species for the same drug-, and the daily dosage level at which physical signs could be produced also seems to vary both between species and within species, but the data are too fragmentary to draw firm conclusions.

Studies on chlordiazepoxide by Stolerman, Kumar and Steinberg in rats showed that availability of an aqueous solution of 0.5 mg/ml produced no dependence. Harris, Claghorn and Schoolar forced rats to drink a similar solution by depriving them of other liquid. The animals reduced their liquid intake and when allowed to choose between drugged and pure water subsequently chose the latter, showing no sign of dependence. In a further study in which the drinking of the drugged water was encouraged by only giving food when the water was drunk, the drugged water was taken in preference to pure water after 25 days’ administration. Goterdam allowed rats to self-administer intragastric medazepam by a lever-pressing response in food-deprived animals. The responses of the medazepam group were significantly higher than that of the control.

In another recent study on rats chlordiazepoxide showed similar weight-change patterns during 8 h daytime periods to narcotics but the study is yet to be confirmed. Its relevance to the human is also far from clear since nalorphine, a narcotic antagonist, shows a more pronounced effect than several accepted narcotics and it therefore appears that the test is not a valid one for predicting human dependence of any type.

Another of the well-authenticated methods to assess the ability of drugs to produce dependence in animals is the ability to cross substitute in an animal previously rendered dependent. Deneau and Weiss studied various substances in dogs rendered physiologically dependent on sodium barbital. All the sedative drugs they tested, including other barbiturates, chloral hydrate, paraldehyde, ethanol, meprobamate, glutethimide, methyprylone, methaqualone and chlordiazepoxide all substituted.

Although the development of tolerance to a drug in animals cannot be accurately correlated with dependence production it is usually regarded as an indication of dependence potential. Hoogland et al. showed that tolerance to chlordiazepoxide could be produced experimentally in rats and depended on the induction of liver drug metabolizing enzymes.

It has sometimes been suggested that phenothiazines do not produce dependence but Boyd showed in rats that after administration of high doses for 40 weeks, sudden withdrawal led to hyperkinesia, diarrhoea and 2 deaths. Thus it must be concluded that all these sedative compounds have a dependence potential in animals, but that there are differences in the level of risk as determined by the ratio of dependence production to pharmacologically effective doses and the period of regular administration that is required.

Such animal experiments demonstrate that by chronic administration of high doses of most, if not all the central nervous system depressants that have been studied, it is possible to produce a syndrome in animals which is akin to dependence in the human. So far as can be determined from observations on these animal species, the physical signs of physical dependence to this whole group of drugs has considerable similarity and is also like that seen with similar drugs in humans (). These experiments also give one a broad indication of the degree of dependence liability which exists between different groups of drugs in animals, but there is of course no direct evidence that this has relevance to the human situation where the pattern of intake is different (for example in the psychological dependence in animal studies the intravenous route is the one most commonly used).

However, even if such animal experiments which indicate a dependence liability at high dosage in animals can be regarded as relevant to the human, they do not indicate the potential for the development in the human of dependence at therapeutic dosage levels which is the most important single feature of interest for therapeutic agents.