Testing for dependence liability of stimulants and depressants in animals and man
Physical dependence in animals
Self-administration in animals
PROCEDURES DESIGNED TO COMPARE THE EFFECTIVENESS OF DIFFERENT DRUGS AS REINFORCERS
Behavioural toxicity in animals
CIRCUMSTANCES OF USE
CONSEQUENCES OF EXPOSURE SCHEDULES
BEHAVIOURAL TOXICOLOGY AND ABUSE LIABILITY
APPROPRIATE BEHAVIOURAL TECHNIQUES
Behavioural evaluation in man
PROFILES OF ABUSED DRUGS
Recommendations for further research
Pages: 21 to 31
Creation Date: 1977/01/01
Report by the Committee on Problems of Drug Dependence, Assembly of Life Sciences, National Research Council
In 1966, the Committee on Problems of Drug Dependence of the National Academy of Sciences-National Research Council prepared a document entitled, "Testing for Dependence Liability in Animals and Man." 1 That document was updated in 1972. 2 Since then, not only have the Committee's interest in abuse potential of drugs continued to broaden, but abuse of a wide variety of substances that act on the central nervous system has increased. Before the 1966 report, nearly all published research was concerned with testing morphine-like drugs in physically dependent monkeys and human addicts. However, very significant advances have occurred in developing new techniques for predicting the abuse liability of a broader range of drugs, including stimulants and depressants.
In all probability no single method will have such substantial predictive validity that drug abuse liability can accurately be assessed using one measure. Moreover, the relative weighting attached to various predictors may vary considerably from one drug class to another. Thus, in considering the various predictive tools, one must be constantly wary of single approaches (e.g., using solely self-administration or physical-dependence methods). Finally, in drug development, abuse potential must always be considered in the context of therapeutic efficacy. That is, a unique and critically needed compound should not be discarded solely because early laboratory testing predicts some dependence liability; but a newly developed drug that is shown by several laboratory predictors to have a very high abuse potential would be difficult to justify if it were not absolutely essential to medical practice.
The purpose of this document is to present a brief résumé of the current state of knowledge concerning prediction of abuse liability of central nervous system stimulants and sedative-hypnotics. Predictive procedures in four categories will be considered: testing for physical dependence in animals, self-administration in animals, psychological and behavioural evaluation in man, and behavioural toxicity and other predictive procedures in animals.
Central nervous system stimulants are not known to cause physical dependence in animals. However, central nervous system depressants that have been abused by man generally produce physical dependence in some animal models. These compounds include the barbiturates, ethanol, the nonbarbiturate sedative-hypnotics, and the minor tranquillizers. Testing for physical dependence in animals is important, to identify the pharmacologic profile of the test drug in comparison with the profiles of prototypic drugs that are abused by man, to predict the individual and social harm that might result from the abuse of the test drugs, and to predict the extent to which withdrawal need may be associated with drug-seeking behaviour. As with centrally acting analgesics and their antagonists, there are three principal ways of determining physiologic dependence capacity, the likelihood that a depressant will produce physical dependence: direct development of dependence on chronic administration, by the animal or by an operator; the effect of substitute drug on withdrawal symptoms in a dependent animal; and the effect of a single dose of the drug on withdrawal symptoms.
*Prepared with support from the Drug Enforcement Administration, the Food and Drug Administration, and the National Institute on Drug Abuse, under contract No. 271-76-3322.1
1Committee on Problems of Drug Dependence, Testing for Dependence Liability in Animals and Man, Bulletin on Narcotics, XXII: 1, 11-17, 1970.2
2 Committee on Problems of Drug Dependence, Testing for Dependence Liability in Animals and Man (Revised 1972), Bulletin on Narcotics, XXV: 2, 25-39, 1973.
Several measures of withdrawal-symptom severity are available, so reproducibility has been generally satisfactory. The methods now used were developed with dogs, monkeys, cats, and mice and have been validated with such standard drugs as sodium barbital (dogs and monkeys), pentobarbital (cats), and ethanol (mice). Although most, but not all, depressants that are abused by man produce physical dependence in these animal models, physical dependence may not be as useful as other pharmacologic properties in predicting abuse potential. Nevertheless, determination of physical dependence recommends itself as a relatively rapid, economical, and fairly reliable preliminary screen.
Depressants with abuse potential are widely divergent in chemical structure. It therefore seems appropriate to examine different prototypic drug standards whose abuse potential in man is known, to enhance predictability. These might include secobarbital, pentobarbital, amobarbital, chloral hydrate, glutethimide, methaqualone, diazepam, meprobamate, and ethanol.
Because the type of physical dependence and the severity of withdrawal symptoms depend in large measure on the pharmacodynamic and pharmacokinetic properties (biologic half-life, metabolism, etc.) of a drug, it is desirable to use standards whose pharmacokinetics are well established or can be readily determined, such as hexobarbital, phenobarbital, chlorodiazepoxide, oxazepam, and bromide ion.
A primary pharmacological property of drugs germane to predicting abuse liability is their ability to reward or reinforce self-administration behaviour. There is substantial face validity to the use of drug-taking behaviour in laboratory animals as a method of studying important aspects of drug-seeking and drug-taking behaviour in man.
It must be recognized that the reinforcing action of a drug is only one of the pharmacologic properties that eventually dictate the public-health consequences of its use. Other properties of the drug-such as physical dependence liability, toxicity, and range of therapeutic uses-are also important determinants of abuse liability, as are social factors.
Over the last 15 years, laboratory methods have been developed and perfected for studying the reinforcing properties of drugs in laboratory animals.
Basically, these procedures arrange conditions so that a discrete behavioural response is followed by the administration of the drug. For example, intravenous drug self-administration studies have typically used animals with chronic venous catheters attached to a remotely operable infusion pump. The pump can be made to operate in response to an action, such as a lever-press, and infuse a fixed quantity of drug solution. Systematic changes in the performance required to produce an infusion yield detailed quantitative data on the drug's reinforcing properties. Such techniques have become increasingly refined and have been applied in a variety of species of laboratory animals, including rats, cats, dogs, and non-human primates.
Most compounds that serve as reinforcers in laboratory animals and that are used by man are abused by man. This isomorphism is clearest in the case of opioid analgesics: analgesic agonists with demonstrated abuse liability in man are readily self-administered by both naive and physically dependent animals. Furthermore, narcotic antagonists, such as nalorphine and cyelazocine, are devoid of abuse liability in man and are not positive reinforcers under most conditions in which they have been studied in laboratory animals. The narcotic agonists, morphine, codeine, meperidine, heroin, and propoxyphene, have been shown to serve as positive reinforcers in rhesus monkeys. In the initial studies, the subjects were made physically dependent before they had any opportunity for self-administration; later studies have demonstrated conclusively that a state of physical dependence is not necessary for narcotic agonists to serve as positive reinforcers, although the reinforcing efficacy of a centrally acting analgesic is generally increased if the subjects are physically dependent.
The agonist-antagonists pentazocine and propiram fumarate serve as positive reinforcers in non-dependent monkeys, but not, under most conditions, in monkeys physically dependent on morphine. In contrast, neither the partial agonists nalorphine and cyclazocine nor the antagonist naloxone serve as positive reinforcers in non-dependent monkeys. These findings correlate well with the available data on human abuse, suggesting that self-administration procedures are useful in predicting the abuse liability of the narcotic analgesics and their antagonists in man.
Self-administration procedures have been widely used to study psychomotor stimulant drugs. The data on psychomotor stimulants are consistent in that cocaine, the amphetamines, methylphenidate, phenmetrazine, and diethylpropion serve as reinforcers in a number of species. These compounds have all been reported to be abused by man. Fenfluramine and chlorphentrazine, however, are not reliably self-administered by animals; this corresponds with clinical reports that their abuse potential differs from that of the more typical psychomotor stimulants. In short, present evidence suggests that self-administration procedures are useful in evaluating abuse liability of psychomotor stimulant drugs.
A number of barbiturates and minor tranquillizers serve as reinforcers in animal self-administration procedures; however, they have not been studied extensively and systematic comparisons among drugs have not been carried out. In spite of the limited validation of self-administration procedures for studying the abuse liability of depressants, it seems that all drugs shown to be reinforcers in infrahuman self-administration are also abused by man.
A common approach to the sequential evaluation of the abuse liability of a stimulant or depressant involves the procedures described below.
Animals are trained to self-administer a prototypic drug at a dose previously determined to be effective in producing repeated lever-pressing. Typically, daily experimental sessions last for several hours, during which the drug is delivered once for every n (e.g., 10) responses. After establishment of a stable rate and pattern of drug-reinforced-responding, test drugs may be substituted for the standard drug. The rate and pattern of response generated by the test drug are compared with those observed with the prototypic drug and with saline (vehicle) substitution. A test drug is considered to be a positive reinforcer if it maintains a higher rate of responding than that observed when saline is substituted. It is essential that a broad range of doses be tested, because a dose that is higher or lower than optimal may not maintain a response rate above that observed with saline or vehicle.
The substitution procedure permits evaluation of a range of test doses in a relatively short period. If a drug maintains a response rate significantly above that with saline, initial evidence of abuse potential is thereby established. In interpreting such initial evidence, the environmental and pharmacologic history of the test subject, the schedule of drug delivery, the duration of access to the drug, the species, and the route of administration must be considered; all have been shown to influence the degree to which drugs may serve as reinforcers.
Animals are allowed to self-administer drugs without dose or time limitation every day for several weeks. The initial test dose is based on the results obtained in the substitution procedure, which allows a determination of the amount of drug taken and the pattern of daily intake. The procedure is especially useful for assessing toxicity, tolerance, and physical dependence development at self-regulated doses. Food intake and water intake are measured concurrently. Measures of conditioned and unlearned behaviours can be obtained concurrently.
The use of experimentally naive animals in this procedure has the added advantage of demonstrating that prior drug history is not needed for the test drug to serve as a reinforcer. The use of naive animals, however, is complicated by the fact that learning the drug-reinforced response is influenced by a number of variables. For example, failure to obtain self-administration may reflect inadequate experimental procedures, rather than a lack of reinforcing effect of the drug. The continuous access procedure yields primarily qualitative information and does not readily lend itself to refined quantitative comparison of the reinforcing strength of different drugs.
In general, attempting to quantify the relative reinforcing strengths of drugs within a class is more useful than comparison among drug classes. Two approaches appear to have merit in the prediction of the relative abuse liability of stimulant and depressant drugs: the progressive-ratio procedure and the choice procedure. A third approach, involving second-order schedules of reinforcement, shows great promise, but has not been studied sufficiently to allow statements concerning predictive validity.
Progressive-ratio procedure. This is perhaps the most widely used self-administration assessment procedure. A baseline performance is obtained with a standard drug, the test drug is substituted, and the response requirement systematically increased. For example, the number of responses required for drug reinforcement could be progressively doubled. The response requirement that the animal fails to meet (e.g., 100 responses) is called the breaking point. It has been shown, with this procedure, that the breaking point can be used to rank-order some psychomotor stimulant drugs according to reinforcing strength. The rank order of these drugs correlates highly with their reported dependence liability in man. Dose and duration of action appear to be important variables in the effective use of this procedure.
Choice procedures. This procedure permits .the experimental animal to choose between two drug solutions. Two measures can be derived: per cent choice of one solution compared with the other and rate of response maintained by the two solutions. Animals are trained to choose between a prototypic drug and saline. After initial training, the test drug may be substituted at various doses. To ensure exposure to each of the drug solutions, a period of self-administration of each drug alone may be provided. Animals choose various psychomotor stimulant drugs over saline, higher doses of a given drug over a lower dose, and, finally, one drug over another, regardless of doses. This procedure appears to have considerable promise in the rank-ordering of drugs.
Second-order schedule procedure. Animals are trained to respond on a simple schedule of reinforcement producing a brief stimulus presentation (e.g., a light or tone). Periodically, the brief stimulus is presented concurrently with a drug. The brief stimulus is able to maintain impressive large amounts of behaviour, even when only intermittently coupled with drug delivery. This procedure has not been applied to the prediction of relative abuse liability of drugs. Because such second-order schedules maintain high rates of responding, it may be possible to use this procedure in a manner analogous to the progressive-ratio procedure. This could be done by determining how infrequently drug reinforcement could be given while still maintaining response. A primary advantage of this procedure is that extended samples of behaviour leading to drug reinforcement can be maintained with little or no confounding effect of previous drug administration (e.g., earlier the same day).
Behavioural toxicology usually is considered within the context of environmental and occupational exposure although no logic limits it to those bounds. Indeed, nontherapeutic actions of drugs helped to spur the development of behavioural toxicology, a historical alliance that meshes it with the issues discussed here. Even if therapeutic effects are accompanied by conjoint adverse responses, the latter are tolerated, in view of some subjective equation of risks and benefits. The same effects, if produced by administration outside a therapeutic context, can be conceived of as toxic - they would not be acceptable if they arose from an environmental agent.
Chronic or long-term, as well as acute, behavioural observations seem a compelling necessity, given our current appreciation of long-term consequences of drug exposure. Prolonged amphetamine abuse, for example, is believed by clinicians to lead to persistent behavioural or psychologic changes. Experiments on drug administration in animals indicate that it is possible to produce a long-term residual supersensitivity that predisposes to behavioural and motor disorders as responses to later pharmacologic or environmental manipulations.
Such tests require that the behaviour under observation at least be sensitive to acute levels of the doses that will be studied chronically and others should be sensitive to toxic side effects unrelated to the desired clinical effects. Techniques range from tests of simple locomotor activity to tests using reinforcement schedules, complex behavioural surveys, and examinations of specific sensory and motor systems and learning. For example, punishment procedures, such as those sensitive specifically to barbiturates and minor tranquillizers, would be useful in assessing compounds of that class.
Patterns of abuse are not always easily included in the toxicologic categories of "acute", "subacute", and "chronic". For many circumstances, a more accurate descriptive term is "episodic", which is more congruent with the abuse patterns often associated with amphetamines, cocaine, sedative-hypnotic agents, and solvents. Even opiate use may be more accurately characterized by such a term.
Acute Phase. While a person is experiencing the peak effects of drug administration, the prime toxicologic focus is on functions that are maximally disabled, that is, functions so impaired that the subject's capacity to deal adequately with the environment is compromised. We are most interested in relatively specific functions for such an assessment: learning, sensory intactness, motor co-ordination, and complex behavioural sequences.
Postepisodic phase. The course of recovery of the functions impaired during an episode must be specified. Rates of recovery vary from function to function, and generalizing from one to another is likely to be unprofitable.
A second question concerns the extent to which additional impaired functions intrude into the postepisodic phase that may have played negligible roles in the acute phase. For example, rebound phenomena might well produce a pattern of effects distinctly different from those seen during the acute phase (cf. repeated administration of amphetamine at short intervals). In addition, nonspecific blunting of functional integrity may occur - a slightly out-of-focus behavioural capacity, rather than sharply-defined impairments. Such changes are difficult to assess; techniques for gauging behavioural function in everyday environments are badly needed.
Chronic phenomena. If episodic abuse is considered a common pattern with centrally acting drugs, the term "chronic" takes on further connotations. Instead of referring only to a relatively stable exposure situation, it describes one overlaid with the fluctuating pattern of administration and recovery, and it turns our attention toward the way in which successive episodes and their immediate recovery phases are modified through an extensive history of such episodes. An alternative approach may be chronic exposure to just sublethal doses.
These issues are complicated by two apparently opposing processes. One is the development of tolerance, which is characteristic of many substances of abuse.
The other is the progressive modification of central nervous system mechanisms to sensitize the organism to further administration of the same or different drugs. These two manifestations of prolonged use may be reflections of the same process. Also demanding attention are the possible irreversible neurochemical and neurophysiologic manifestations associated with behavioural abnormalities in chronic stimulant intoxication.
The consequences of long-term chemotherapy in psychiatry exemplify some of the problems relevant to a consideration of stimulants and despressants. A major side effect of neuroleptic therapy is the production of movement disorders. Early in treatment, these closely mimic the symptoms of Parkinson's disease: retardation of movement (akinesia), a peculiar motor restlessness (akathisia), and sometimes a resting tremor. Prolonged use sometimes induces a movement disorder known as tardive or persistent dyskinesia, a syndrome now evoking increasing concern. There is emerging evidence that a variety of chronically administered compounds may induce a residual excessive sensitivity to stimulants.
The development of motor disorders and psychosis, the persistence of delusional ideation, and the residual sensitivity to reactivation of psychosis by moderately high doses may also constitute a series of abuse liabilities in the chronic use of stimulants. Existing evidence from experimental animals indicates that there are evolving chronic behavioural disorders and dyskinesia, including residual effects, and warrants continued investigation with behavioural methods. Although current behavioural surveys are not completely satisfactory, methods for assessing behaviour of experimental animals offer sensitive indexes of the development of behavioural and motor disorders, as well as of the sentitization to drug effects and residual effects.
Behavioural observations focusing on impaired function represent a major avenue of assessing dependence liability. Because conventional assays, even in their chronic form, are not congruent with patterns of drug abuse, and because abuse is such a socially important feature of the use of centrally active drugs, behavioural toxicology in this sense may represent an even more important facet than in its more conventional toxicological framework.
Two pertinent parameters of abuse patterns cannot be predicted from usual preclinical test results: dosages that maintain self-administration, and the associated temporal parameters of self-administration behaviour.
Doses and administration patterns for behavioural toxicology should ideally be chosen on the basis of the parameters found in self-administration studies. These are probably the best preclinical guides available. Although tolerance and dependence studies are also relevant, they possess a less immediate relationship to the core problem.
An additional reason for using self-administration patterns as a guide is the impossibility, as indicated earlier, of separating, in one situation, the reinforcing properties of a substance from its general behavioural effects, such as blunting of stimulus control.
Anticipated advances in behavioural technologies argue against the desirability of making any recommendations that would tend to institutionalize current practices. Any behavioural function can be assessed by a broad variety of procedures and, for maximal validity, must be evaluated in the most appropriate species. Although it is becoming more and more a standard practice, behavioural teratology should not and must not be neglected. Lifetime studies are indispensable for determining many of the most vital features of toxicity and also serve to amplify toxic effects that in another context would remain covert.
Depending on the nature of initial surveys, a variety of sensitive behavioural assessments are available, including sensory capability, motor co-ordination, and activity levels. Acquisition and retrieval of learned responses and maintenance of schedule-controlled behaviour are also important indicators of toxicity. Observation techniques occupy an important role in the assessment of behavioural and motor disorders.
Important factors in the choice of agents for evaluation in man will be the results of animal testing and the proposed therapeutic use. In this regard, there are differences among species in reactivity to psychoactive drugs and in the metabolism and kinetics of the drugs. It is necessary to evaluate the abuse potential of psychoactive agents in man before they are used therapeutically. In general, an agent should be evaluated for dependence potential if it is found during clinical trials to have hypnotic or sedative effects, to modify mood, to intoxicate, to disturb cognitive processes, or to alter the subject's control over his own functioning.
The risk to a human subject from an evaluation of an amphetamine-like or sedative-hypnotic drug during a well-designed study of dependence liability is quite small. Idiosyncratic toxicity can be minimized by choosing subjects who have already used these drugs - whether in benign or substantial patterns of abuse. Their reaction tends to resemble that of a hypothetical population that may be described as "susceptible". A valid test of dependence liability can be obtained with fewer such subjects, because they interpret the effects of a drug in terms of their past experience and are less likely to give "placebo" responses.
Sedative and psychostimulant drugs can become incorporated into patterns of dependence among those in various walks of life, so such drugs should be evaluated in populations that are "normal" in terms of behaviour alterations or impairments that may arise in ordinary life circumstances.
Any such experiments should be subject to peer review, and the customary constraints and safeguards, based both on common sense and ethical considerations, must be observed. Initiation and continuation of a subject participation in a study should be voluntary; all subjects should be informed of the purpose of the study and the associated risks and must have the opportunity to withdraw consent without jeopardy.
Study of the kinetics of depressant drugs is essential in predicting their dependence potential and the type of withdrawal reaction that they will evoke.
The drugs most likely to produce dependence are those whose effects are swift, sure, and short. The more rapid the onset of a clearly discernible effect, the more likely that the drug will be used for intoxicating purposes. Rapid absorption and short duration of action meet these needs.
The development of physical dependence requires a sustained pharmacologic effect over a finite period. Drugs with a high degree of lipid solubility, a large apparent volume of distribution, and a long half-life (either of the parent drug or an active metabolite), generally meet these requirements. The severity of the withdrawal reaction is inversely related to the rate of decline. Conceivably, a drug with a sufficiently large cumulative effect and a slow disappearance rate would produce no Withdrawal symptoms, despite the fact that it might induce physical dependence.
The most frequently abused drugs are those which are rapidly absorbed and which have a moderately long half-life (12-24 hours), e.g., secobarbital, meprobamate, and ethanol. Drugs that are slowly absorbed and long-acting are less frequently abused. Thus, analysis of the kinetics of a new sedative drug during phase 1 or 2 clinical studies might predict potential abuse.
The traditional approach in assessing human drug-dependence potential is to compare the subjective, behavioural, and physiologic effects of an unknown compound with those of known compounds with well-established dependence potential. Profiles of pharmacologic action of prototypic stimulant and depressant drugs appear to be useful models. If a new compound displays a pharmacologic profile similar to that of a drug with established dependence liability, considerable dependence liability of the new compound would be expected.
A number of measures of drug effects are currently used for such purposes. Subjective effects of drugs are assessed with questionnaires, Such as the Addiction Research Center Inventory. From scale scores, drugs can be categorized according to their euphoric or sedative effects. Subjective effects are also assessed through unstructured observations by the subject. Behavioural tasks used in establishing profiles for predicting dependence liability include performance tests (reaction time, tapping time) and measures of perceptual-motor functioning. Operant tasks, which are highly sensitive in assessing drug effects in animals, are being investigated, as are such other phenomena as changes in risk-taking behaviour, social interaction, sexual functioning, appetite, vigilance, and arousal. Physiologic measures - including electroeardiography, electroencephalography, heart rate, blood pressure, temperature, respiration, evoked potentials, pupillary changes, and postrotational nystagmus are also being employed to establish profiles of drug effects.
Another general approach for predicting drug-dependence liability may involve the use of human self-administration for assessing the reinforcing effects of drugs. As with animal models, human self-administration studies appear likely to be useful in predicting dependence liability and with appropriate proper ethical considerations this approach may prove successful as a predictive technique in its own right or as a useful adjunct with other techniques.
Evaluation for dependence liability includes an assessment of the toxicity of a psychoactive drug, the dose at which toxic effects occur (compared with thera-peutic and reinforcing doses), and the degree to which behavioural toxicity may be affected by other psychoactive substances. Among the toxic effects that must be considered are the following:
The degree to which the substance produces ordinarily nonreinforcing side effects (e.g., dysphoria) that will tend to limit self-administration. In general dependence potential will be greater if such untoward side effects occur only a doses well above those which are substantially reinforcing. It must be recognized that some events are unpleasant for most persons, yet positively reinforcing for other persons.
The capacity of the drug to produce aberrations in behaviour associated with stimulation or depression; changes in perception, mood, consciousness, cognitive processes, or judgement; and decrements in motor co-ordination. In general, abuse potential will be greater if a drug's capacity to produce such changes is greater.
The degree to which symptoms occur, when drug use is discontinued, thai tend to reinforce further drug taking (e.g., changes in mood or feeling states, psychic distress, and physical discomfort), whether or not such symptoms are associated with physical dependence itself or with other drug-related phenomena.
The degree to which behavioural toxicity or reduction in motor coordination results from one or two therapeutic doses of a drug alone or in combination with a blood alcohol content of 30 mg per cent or higher.
The validation of methods involving human subjects to predict dependence liability of a drug will initially involve demonstrating the ability to measure both the reinforcing properties of prototypic drugs and the equivalence of the test drug to prototypic drugs. Ultimately, predictions made on the basis of experimental studies can be validated only by the actual occurrence of drug-related problems, when the test drug is generally available. Because drugs are introduced under various circumstances of availability and in various social and cultural settings, it is imperative that monitoring programmes be established, to produce data for validation of predictive methods and to reveal unanticipated drug-related problems related to changes or differences in mores.
Drug interactions and cross-substitutions should be studied as factors is multiple drug use and as contributions to the development of withdrawal manifestations.
Differences and similarities between animals and man, with regard to tolerance physical dependence, and withdrawal, should be investigated.
Objective measures of withdrawal symptoms should be developed.
The relationship of withdrawal characteristics to dependence liability should be studied.
Specific antagonists to the depressant drugs should be developed.
Because there is some diversity of opinion as to the role of physical dependence in determining the abuse potential of depressants (sedatives-hypnotics and minor tranquillizers), more experience must be gained regarding the extent of abuse of widely used, prototypic drugs.
There is considerable need for further research in drug self-administration by nonintravenous routes; in particular, methods for chronic intragastric and oral self-administration of drugs that are not water-soluble require exploration.
With the increasing cost of rhesus and squirrel monkeys, it becomes more important .to develop improved methods for chronic parenteral self-administration of drugs in rodents and less expensives primates. A large body of information on rhesus monkeys has already been gathered; therefore consideration should be given to developing mechanisms for maintaining continuing colonies of these species.
Far greater emphasis should be placed on concurrent or interpolated measures of other behavioural actions of self-administered drugs; this is particularly important in chronic studies in which drugs are continuously available for self-administration.
Despite the considerable advances in our knowledge of the variables that influence drug self-administration, a great many issues remain to be resolved, such as the role of pharmacologic and environmental histories and the significance of the duration of access to drugs and of schedules of drug self-administration.
Care must be taken to distinguish drug-taking from drug-reinforced behaviour i.e., drug self-administration in animals and humans may be induced, not by the drug, but by another inducing condition (e.g., a sweet-flavoured solution for liquid-deprived rats); it is essential that drug self-administration be demonstrated in the absence of other inducing conditions.
Greater emphasis should be placed on validating data on infrahuman self-administration in controlled situations, so that they may be properly compared with data on humans in the same situations.
Research with human subjects should not be limited to long-term hard-core abusers of multiple drugs, but should include subjects with various patterns of drug abuse.
The relations between subjective ratings of drug effects (e.g., the Addiction Research Center Inventory) and objective behavioural measures of drug actions (e.g., operant tasks) should be investigated in human subjects.
Acute, postepisodic, and long-term effects of abused drugs on objectively measurable aspects of natural and conditional responses of animals should be investigated.
Methods for assessing subtle behavioural effects of abused drugs (e.g., on complex discrimination, problem-solving, and learning of complex tasks) should be improved.
Behavioural teratologic effects of administration of high doses of abused drugs to pregnant female animals and their offspring should be investigated.