An Addiction Science Network Resource

Reprinted from J.E. Henningfield, R.E. Johnson, and D.R. Jasinski (1987), Clinical procedures for the assessment of abuse potential. In M.A. Bozarth (Ed.), Methods of assessing the reinforcing properties of abused drugs (pp. 573-590). New York: Springer-Verlag.
Brain Reward System
Back to book TOC
(Search book contents) (Search entire ASNet)


Chapter 27

Clinical Procedures for the Assessment of Abuse Potential

Jack E. Henningfield, Rolley E. Johnson, and Donald R. Jasinski

Addiction Research Center
National Institute on Drug Abuse
PO Box 5180
Baltimore, Maryland 21224

Procedures are described for assessing the abuse liability of drugs and their potential to produce physiological dependence in human volunteers. Abuse liability of a drug is determined by showing that a drug (a) is psychoactive, (b) produces euphoria, and (c) serves as a reinforcer. The strategies currently used to collect such data include objective assessment of euphoria by use of structured questionnaires and the intravenous drug self-administration paradigm. Dependence potential is determined by the ability of a drug to produce withdrawal when chronic administration is terminated (direct addiction procedure) or to block withdrawal when substituted for a drug known to produce physiological dependence (substitution procedure). Physiological dependence is viewed as a concomitant factor in the abuse potential of a drug, but not the cardinal indicator.



Assessment of abuse potential of drugs is presently pursued at several levels of analysis through many subspecialty areas in pharmacology and psychology as evidenced in this volume. The present chapter is a summary of procedures currently used at the National Institute on Drug Abuse’s Addiction Research Center to assess abuse potential using human volunteers as research subjects. The methods are essentially those developed over nearly one half century of clinical research at the Addiction Research Center (Jasinski, 1977) and, more recently, supplemented by the strategies of behavioral pharmacology (Thompson & Johanson, 1980). Studies with opioid drugs using both clinical and behavioral pharmacological procedures have generated a model of substance abuse that is the standard to which other forms of compulsive behavior are compared. The strength of these procedures is that they provide a validated means of quantifying abuse potential, as evidenced by studies involving drugs of previously established liability for abuse (cf. Jasinski, 1977; Jasinski, Henningfield, & Johnson, 1984) and studies of new compounds with suspected potential for abuse (Brady & Lukas, 1984).

Most abused substances share, at least, two measurable properties. First, they serve as reinforcers, as determined from animal and human studies of drug self-administration, and secondly they serve as euphoriants, as determined from human self-report data (Jasinski, 1977; Yanagita, 1980). Additionally, abused drugs may be categorized on the basis of their interoceptive stimulus effects, measured by using psychometric instruments in human subjects (Haertzen & Hickey, this volume) or drug discrimination procedures in animals (Colpaert, this volume; Overton, this volume; Woods, Herling & Young, 1982). Drugs of abuse, notably the opioids and sedatives, produce physiological dependence which may be viewed as a rebound phenomenon that accompanies abrupt abstinence after a period of chronic drug administration (Jaffe, 1980). In this review abuse liability refers to those properties of a drug that are critical to the maintenance of compulsive drug-seeking behavior; dependence potential refers only to the ability of the drug to produce physiological dependence (see also Brady & Lukas, 1984). Currently used procedures for quantifying the subjective effects, reinforcing efficacy, and dependence potential of substances will be summarized.

Dependence Potential Assessment

Procedures for assessing a drug’s dependence potential were developed in pioneering studies by Himmelsbach and his coworkers during the 1930s and 1940s (Martin & Isbell, 1978). The fundamental observation was that abrupt termination of morphine administration to a person who had been receiving daily doses of morphine produced a sequence of signs and symptoms that could be quantified. Himmelsbach observed that the intensity of this abstinence syndrome was related to the duration and magnitude of the dosing regimen and that certain signs only occurred after termination of a high dose regimen. Early onsetting (within 24 hours) signs of withdrawal of low to moderate intensity were assigned one point; later onsetting, more severe signs, were assigned three points. Himmelsbach’s rating scale, in various forms, remains the cornerstone of most studies of opioid withdrawal in human subjects and in animals. An assumption in the studies by Himmelsbach was that the abuse liability of a drug was measured by its potential to produce physiological dependence. However, as will be discussed below, it was subsequently learned that physiological dependence is not the only indicator of abuse liability.

The strategy used to assess opioid withdrawal has been adopted to study withdrawal from sedatives, alcohol, phencyclidine, and, most recently, benzodiazepines. Current adaptations of these methods include characterization of possible withdrawal effects accompanying termination of chronic use of tobacco and psychomotor stimulants.

In the 1960s, Martin, Haertzen, and their colleagues evaluated the subjective concomitants of opioid withdrawal and showed that subjective withdrawal effects did not necessarily covary with physiological withdrawal effects (Jasinski, 1977). These studies showed that termination of cyclazocine administration to cyclazocine-dependent subjects produced significant withdrawal scores but did not increase subjective discomfort or desire to take the drug (Fraser & Isbell, 1960). Two additional examples of the dissociation between physiologic withdrawal and subjective indices are as follows: First is the finding that buprenorphine administration to opioid-dependent subjects can produce physiological signs of withdrawal without subjective discomfort (Jasinski, Haertzen, Henningfield, Johnson, Makzoumi, & Miyasato, 1982). Second is the finding that clonidine administration during morphine withdrawal blocks physiologic signs without a proportional attenuation of subjective discomfort (Jasinski, Johnson, & Kocher, 1985). Clearly, an important area of study remains the investigations of the relations among the physiological, behavioral and subjective phenomena of drug withdrawal.

Another of Himmelsbach’s findings was that administration of either codeine or morphine could reverse the opioid abstinence syndrome at any point in its course. This finding established the principle of cross-dependence and provided the foundation of the substitution strategy to determine which drugs were morphine-like (Himmelsbach & Andrews, 1943). Similar strategies were applied to the measurement and treatment of sedative and alcohol withdrawal in the 1950s; cross-dependence also characterized these drugs.

The procedure initially used by Himmelsbach to study dependence potential was termed the direct addiction procedure. Essentially, volunteers were given gradually increasing doses of the drug and then were maintained for a time (usually a week or more) on a scheduled dosing regimen that maintained relatively high and stable blood levels of the drug. Drug administration was then terminated, and signs and symptoms of withdrawal were measured. Withdrawal effects are rebound phenomena which are generally opposite in direction of those produced as a direct result of drug administration and are presumably due to drug-induced neuroadaptation. For example, opioid-induced constipation may be replaced by diarrhea, and sedative-induced muscle relaxation may be replaced with convulsive activity (Jaffe, 1980). Test compounds can be administered during abstinence to assess their ability to either reverse or attenuate the syndrome. This procedure has been most recently used to evaluate possible withdrawal phenomena following buprenorphine administration to volunteers (Jasinski, Griffith, & Pevnick, 1978).

The most common strategy currently used to evaluate new compounds for efficacy in the treatment of opioid withdrawal is the 24-hour substitution procedure described by Fraser and Isbell (1960). Four doses of morphine (15 mg/dose) are given each day for several days. On test days three of the doses are replaced with either saline or graded doses of the test compound or morphine (as a positive control). Under this dosing regimen saline substitution results in reliable onset of early signs and symptoms of opiate withdrawal. The syndrome is reversed by the return to the morphine dosing regimen at the time of the next normally scheduled dose; however, the syndrome is mild enough that research subjects (moderately paid volunteers) tolerate the procedure and will repeat the experiment, permitting each drug condition to be tested in each subject.

The efficacy of clonidine in the treatment of opioid withdrawal was assessed in a recent study utilizing the 24-hour substitution procedure (Jasinski, Johnson, & Kocher, 1985). As mentioned earlier, when clonidine was substituted for morphine, clonidine reversed the sympathetically mediated signs of withdrawal even more effectively than did morphine (Figure 1, upper frame); however, clonidine did not influence self-reported sickness (Figure 1, lower frame). This finding that physiological signs do not necessarily covary with subjective effects is an important concept in strategies for measuring abuse liability. It also is of clinical relevance, because the subjective effects appear to be more important than the physiological effects in predicting abuse liability.

Abuse Liability Assessment

Abuse liability is assessed by procedures that predict the likelihood that drugs will be taken in the nonlaboratory setting. The earliest strategies were those which determined the qualitative nature of the subjective effects of drugs. Drugs shown to produce euphoria and feelings of well being were generally those which were abused, while drugs producing no such effects were not abused (Jasinski, 1977). A more recently developed strategy is to determine the probability that a drug will be voluntarily taken in a laboratory setting by animal and human subjects. Each strategy provides some unique information; however, there is a close correspondence between the predictions resulting from both. For instance, Griffiths and Balster (1979) showed that opioid drugs which serve as reinforcers for animals closely correspond to those drugs that produce morphine-like subjective effects in human subjects (see also, Griffiths, Bigelow, & Henningfield, 1980; Weeks & Collins, this volume).

Suppression of morphine withdrawal by morphine and by clonidine
Figure 1: Eleven morphine-dependent subjects were given four maintenance doses of morphine each day. On test days these doses were substituted with three doses of morphine, clonidine, or placebo to produce the total daily doses indicated on the abscissa. Mean staff-reported withdrawal scores (including observations such as vomiting and rhinorrhea recorded on the Himmelsbach Scale of opioid withdrawal) are shown in the upper frame. Self-reported sickness scores are shown in the lower frame. Signs (upper frame) and symptoms (lower frame) of withdrawal shown in the figure are mean scores determined 20 to 24 hours following the last maintenance dose of morphine. The dashed line indicates the 95% confidence interval of the mean response to placebo. All doses of morphine and clonidine significantly reduced physiologic signs of withdrawal, whereas only the high dose of morphine reduced subject discomfort. Reprinted with permission from Jasinski, Henningfield, & Johnson, 1984.

Quantification of Subjective Effects of Drugs

Unfortunately, there were no generally accepted methods of quantitating subjective effects of drugs early in the twentieth century. The major scientific advance was by Beecher and his colleagues who adapted quantitative bioassay procedures to assess subjective responses to pain stimuli and to the effects of analgesic drugs on such responses (Beecher, 1959). Beecher also established the importance and utility of the crossover design with inclusion of standard and placebo drug control procedures, the double-blind technique, and randomized experimental design to assess subjective drug effects. These pioneers in the assessment of abuse potential of drugs made the following observations that provide the bedrock of current strategies: (a) Drugs produce reproducible transitory psychoactive states, (b) similar states produced by drugs may be used to define drug classes, (c) individuals reliably discriminate drug induced subjective states, (d) the subjective state characteristic of drugs of abuse is one of euphoria, and (e) the nature of the state is related to the propensity to take the drug. A correlate of the observation that abuse liability was not necessarily correlated with physiologic dependence was the observation that it was the subjective effects of drugs that were critical (Jasinski, 1977). The rationale, and even the fundamental scientific strategy used to study abuse liability in humans, was most elegantly described by Isbell (1948):

Since most people begin the use of drugs and become addicted because the drugs produce effects which they regard as pleasurable, the detection of euphoria is a very important procedure in evaluating abuse liability. The method used is simple: single doses of the drug under test are administered to former drug abusers, and the subjects are unobtrusively watched for a period of 6 hours or more by specially trained observers. For our purposes, euphoria is defined as a series of effects similar to those produced by morphine. These effects are: increased talkativeness, boasting, greater ease in the experimental situation, expression of satisfaction with the effects of the drug, requests for increased doses of the drug, increased motor activity, and, with larger doses, slurring of speech, motor ataxia, and evidence of marked sedation. As many experiments are done as necessary to reach a clear-cut conclusion. The observations are controlled by administering 30 mg of morphine to the same subjects on other occasions. Initially, small subcutaneous doses of the drug under test are used, and if no untoward toxic effects are observed, the dosage is increased progressively in subsequent experiments until evidence of euphoria, roughly equivalent to that produced by 30 mg of morphine is detected or if no evidence of euphoria is detected, the dosage is elevated until further increases would be regarded as dangerous. If euphoria is detected, blind experiments are arranged in which neither the subject nor the observer are aware whether the drug given was morphine or the compound under test. Finally various doses of the drug are administered intravenously.

It is noteworthy that the initial definition of euphoria was judged solely by observed signs, and that the data were exclusively those that could be collected by observers and that the model was based on the effects of morphine.

In recent and ongoing studies at the Addiction Research Center, subjects are tested on a residential research unit, under medical supervision, and formal procedures to protect their rights and welfare (Hickey & Jasinski, note 1). Single doses of drugs are given at sufficient time intervals to eliminate residual drug effects. Drugs are administered according to double-blind procedures. Intervals range from about one hour after short acting drugs such as nicotine to several days after long lasting drugs such as benzodiazepines. Following the drug administration, a variety of physiological, observer-reported (signs or behaviors) and self-reported (symptoms or subjective) effects are assessed at intervals which range from a few seconds to several hours, depending on the time course of drug action.

Typically, the test compound is compared to placebo and to an appropriate positive control (such as morphine in a study of analgesics, or pentobarbital in a study of sedatives). Both the test and control compounds are given at two or more dose levels. A limited number of subjects (usually about 10) participate according to a design whereby every subject is tested under each condition (cross-over or within-subject design). Data are expressed as changes from the pre-drug (baseline) observations and are averaged across subjects. Depending on the temporal pattern, drug effects may be expressed either as peak effect or as the area under the time-action curve for changes in scores.

Self-report measures of subjective drug effects are collected from the subjects using structured questionnaires. The Single-Dose Questionnaire, developed by Fraser and his coworkers (Fraser, Isbell, Martin, Van Horn, & Wolbach, 1961) is among the most elegant psychometric instruments in its simplicity and predictive power. It contains four scales: (a) the first asks whether the drug was felt and thereby determines whether the drug is psychoactive, (b) the second is a 14-item list of substances from which the subject is asked which the administered compound is most like and thereby permits classification (the list includes blank and other), (c) the third is a 14-item list of sensations (including normal and high) that characterizes and quantifies symptoms, and (d) the fourth is a 5-point liking scale which is a measure of euphoria. A similar questionnaire is completed by staff observers.

A variety of drugs have been assessed using the Single-Dose Questionnaire. Figure 2 shows Liking scores obtained following the administration of some of these drugs. As shown in the figure, drugs known to produce widespread compulsive use, such as morphine, d-amphetamine, and pentobarbital, produce dose-related increases in Liking scores. Other substances not known to be abused (e.g., zomepirac, chlorpromazine, or placebo) do not significantly elevate Liking scale scores.

Drug identification responses are also dose-related and a similar effect is also found in animal studies of drug discrimination (Woods, Herling, & Young, 1982). Figure 3 shows that placebo was never identified as drug and that percent of correct drug identification responses for nicotine, nabilone, and diazepam were directly related to drug dose.

Another self-report measure is the Addiction Research Center Inventory (ARCI). The ARCI is a true-false questionnaire with empirically derived scales sensitive to various classes of psychoactive substances (Haertzen, 1974; Haertzen & Hickey, this volume). It contains about 600 items. A 40-item version of the ARCI is used commonly; it contains subsets from three scales: (a) the Morphine-Benzedrine Group (MBG) scale, which reflects feelings of euphoria and well-being; (b) the Pentobarbital-Chlorpromazine-Alcohol Group (PCAG) scale, which reflects sedation and intoxication; and (c) the Lysergic Acid Diethylamide (LSD) specific scale, which may reflect dysphoria and feelings of fear. Scores from the MBG scale usually covary with Liking scale scores, supporting the notion that this scale reflects euphoric drug effects. Dose-related increases in Liking and MBG scale scores are the cardinal subjective effects of abused drugs and define a drug as a euphoriant.

Drug liking scores for various substances
Figure 2: Mean scores on the Liking Scale of the Single Dose Questionnaire from subjects tested at the Addiction Research Center. The number of subjects in each group range from 6 (pentobarbital and chlordiazepoxide) to 13 (d-amphetamine). The high dose of each drug except zomepirac produced significant (p < 0.05) increases in scores above placebo data. The responses are peak responses which occurred after the drug had been given. The time of the peak response ranged from about 1 minute (nicotine) to 5 hours (buprenorphine). Morphine and zomepirac data are from the same group of subjects as are the pentobarbital and chlordiazepoxide data. The P+T point on the pentazocine graph is the score, by the same subjects, to 40 mg pentazocine given in combination with 50 mg tripelennamine (an antihistamine that produced a Liking score of 0.9)—the street combination called "T’s and Blue’s." The M point on the delta-9-THC graph is the score, from the same subjects, obtained after smoking a marijuana cigarette that contained 10 mg (1% by weight) delta-9-THC. Reprinted with permission from Jasinski, Henningfield, & Johnson, 1984.

Dose-response analysis of drug identification responding
Figure 3: Data are summarized from three different studies involving nicotine (Henningfield, Miyasato, & Jasinski, 1985), diazepam (Jasinski, & Johnson, unpublished observations), and nabilone (Boren, Johnson, & Jasinski, unpublished observations). Percent correct drug identifications of the groups tested are graphed as a function of drug dose (nicotine, n = 8; nabilone, n = 10; diazepam, n = 12).

Verification of self-report data is provided by reports from the trained observers. Observer measures of drug-induced changes in overall subject status include nonstructured reports by trained nurse-research staff and structured questionnaires. The most useful questionnaire is the Observer’s Single Dose Questionnaire which is an analogue of the Subject’s Single-Dose Questionnaire (described above). Data obtained from trained and vigilant observers parallel those obtained from the subjects’ self-reports. Such correspondence between self- and observer-reported data confirmed that self-report data were also reliable and could be objectively quantified and confirmed the premise that subjective drug effects could be objectively assayed (Jasinski, 1981). The use of drug-experienced subjects also enhances the orderliness of the data since such persons have learned to accurately discriminate drug-induced states and do not respond to placebo. Certain physiological parameters covary with subjective parameters and can also be used to monitor drug activity. Additionally, the physiological data are used to evaluate possible toxic effects of the drugs. Commonly used physiological measures include pupil diameter, heart rate, blood pressure, respiratory rate, oral and rectal temperature. More specialized measures such as electroencephalographic (EEG) responses and changes in neuroendocrine parameters and brain metabolism are also used in certain studies.

From a historical perspective it is interesting to note the changes in the relative role ascribed to what are now termed physiologic and subjective effects of drugs in the phenomenon of drug abuse. In brief, at the beginning of the twentieth century, drug abuse (narcotic addiction) was thought due largely to physiologic processes including immorality and weakness of character. Himmelsbach and colleagues rejected all such mentalistic (subjective) notions and based their studies on observable physiologic consequences of drug taking and drug abstinence. Later, studies by Isbell, Beecher, Frazier, and others showed that subjective responses could be equally valid, if not more important, markers of abuse liability. Subjective responses to drugs are considered the hallmark indicators of abuse potential, and they are now bioassayed with precision comparable to that in physiologic studies. Furthermore, the subjective responses are known to be direct consequences of the activity of pharmacologic agents of specific physiologic loci (e.g., receptors), thus they are no less "physical" than their physiologic concomitants.

Figure 4 provides an example of the data that may be obtained from some of the measures described above. These profiles of pharmacological activity were determined for orally and subcutaneously administered methadone and for subcutaneously administered morphine. As shown in the figure, the drugs differed primarily in their duration of action along certain parameters.

Quantification of Reinforcing Properties of Drugs

A reinforcer is an event that strengthens and maintains behavior leading to its presentation (positive) or removal (negative). By definition, abused drugs are those which serve as reinforcers and thereby maintain drug-seeking behavior. Whether or not reinforcers are pleasant or aversive is an empirical question and is an issue that is not necessarily addressed by procedures for assessing reinforcement. Reinforcing efficacy is directly evaluated in drug self-administration paradigms in which the performance of a specific behavior results in drug administration.

Procedures for the Assessment of Positively Reinforcing Properties of Drugs

In an early study by Wikler, the self-administration behavior was simply the act of preparing a morphine solution, of applying a tourniquet, and of injecting the drug (Wikler, 1952). In animal studies one can arrange for an arbitrary response, such as a lever press, to produce an injection via an implanted cannula (Yokel, this volume) or to present a single dose of a solution which may be consumed orally (Meisch & Carroll, this volume). Mendelson and Mello, at Harvard Medical School, pioneered the adaptation of the drug self-administration methods used with animals to study orally ingested drugs by human subjects (Mendelson & Mello, 1966, this volume). The strategy has since been adapted at several other laboratories (cf. Griffiths, Bigelow, & Henningfield, 1980). Most recently, the strategy has been added to those used at the Addiction Research Center for abuse liability testing.

Time course for morphine and methadone effects
Figure 4: Time-course of morphine sulfate (20 mg/kg, subcutaneous), methadone hydrochloride (20 mg/kg, subcutaneous), and methadone hydrochloride (20 mg/kg, oral), effects on pupillary diameter, opiate symptoms, MBG scale, and subject’s Liking. Each point represents the mean of 12 subjects. Reprinted with permission from Martin, Jasinski, Haertzen, Kay, Jones, Mansky, and Carpenter, 1973. Copyright 1973 by The American Medical Association. 

At the Addiction Research Center, the intravenous drug self-administration procedure used with animals was explicitly adapted to use with human subjects. Figure 5 shows a schematic of the intravenous drug self-administration paradigm for human subjects. In the initial studies this strategy was utilized to clarify the role of nicotine in cigarette smoking. Pressing a lever produced various levels of nicotine or saline which were delivered via an indwelling forearm catheter to cigarette smokers who were not permitted to smoke during the 3-hour test sessions. The response requirement was 10 lever presses per injection. Injections were about 15 seconds long and were followed by a 1-minute time-out.

In the first seven subjects tested, nicotine maintained orderly patterns of lever-pressing behavior. The patterns resembled those observed when cigarettes are available to cigarette smoking human subjects or when psychomotor stimulants are available for self-injection by animals. As is discussed in more detail in the chapter by Yokel in this volume (see also the chapter by Meisch & Carroll), demonstration of self-administration behavior is not equivalent to showing that a drug is serving as a reinforcer. In brief, procedures that may be used to demonstrate reinforcing efficacy include (a) substituting saline for drug, (b) providing concurrent access to both drugs and saline, and (c) manipulating drug dose to show that the behavior is sensitive to such manipulations. These procedures are essentially those developed previously in animal drug self-administration studies (Pickens & Thompson, 1968; Yanagita, 1976; Young & Herling, 1983). These procedures are used in the human drug self-administration studies at the Addiction Research Center.

The intravenous self-administration paradigm
Figure 5: The intravenous drug self-administration paradigm is shown. Prior to the session the subject is equipped with a catheter in a forearm vein. The catheter line is kept patent by a slow intravenous drip. Two different drug solutions can be concurrently provided with the use of a third pump (washout) to push the balance of a dose given from the left or right pumps into the vein.

Figure 6 shows acquisition of drug self-administration behavior in a subject without a history of drug abuse. After seven sessions nicotine was replaced with saline. Lever-pressing rates and saline injections increased for three sessions, then decreased to rates below those maintained by nicotine. Such data suggest that the drug itself was necessary for maintenance of the behavior. An additional lever, the operation of which had no scheduled consequence, was available for the subject to press during this study. The subject pressed this lever between 10 and 20 times during the first few sessions and then did not press it any more. This simple procedure demonstrated that injections were necessary for maintenance of lever-pressing behavior.

Intravenous nicotine self-administration
Figure 6: Subject KO was a cigarette smoker without a history of drug abuse. He responded under the simple fixed ratio schedule (FR-10) of nicotine or saline injection. Nicotine was available (1.5 mg nicotine/injection) during seven consecutive sessions; then saline was substituted for an additional seven sessions. Number of injections per session are shown on the ordinate. Reprinted with permission from Henningfield & Goldberg, 1983. Copyright 1983 by ANKHO International, Inc.

One of the simplest ways to determine whether or not a drug is serving as a reinforcer is to provide concurrent access to drug and saline and then to alternate the levers which produce either solution. Figure 7 shows the results of this procedure in a subject on the nicotine self-administration study. As shown in the figure, regardless of the lever from which nicotine was obtained, nicotine was consistently preferred to saline. In another study a subject was pretreated with mecamylamine given orally 1 hour before sessions (Henningfield, 1983). As the dose of mecamylamine was increased from 2.5 to 10 mg, nicotine preference to saline decreased. Additionally, postinjection drug Liking scores were inversely related to mecamylamine dose.

Finally, in a study by Goldberg and Henningfield (unpublished observations), manipulation of delivered dose for similarly tested human and squirrel monkey subjects produced orderly and similar (across species) changes in behavior. Number of drug injections was inversely related to unit dose, while there was a slight positive relationship between total session drug intake and dose. This sort of compensatory change in the rate of self-administration behavior is typical of that seen when drugs are serving as reinforcers in animal and human subjects (Griffiths, Bigelow, & Henningfield, 1980)

Responding for intravenous nicotine and saline
Figure 7: Subject had a history of alcoholism and had been previously tested over a range of nicotine doses. In this study he was given concurrent access to both nicotine (1.5 mg/injection) and saline during 3-hour sessions. Ten responses (FR-10) were required to produce an injection and the levers which were associated with nicotine and saline injections reversed each day. Number of injections per session are shown on the ordinate. As an additional pilot procedure, a mecamylamine capsule was given 1 hour before each session. The mecamylamine doses were: 0 mg, Sessions 1, 3, and 8; 2.5 mg, Sessions 2 and 6; 5 mg, Sessions 4 and 5; 10 mg, Session 7. Mecamylamine attenuated self-reported liking of the nicotine injections and appeared to decrease the rate of nicotine-maintained responding; this may simply have been a trend across sessions. Reprinted with permission from Henningfield & Goldberg, 1983. Copyright 1983 by ANKHO International, Inc.

Procedures for the Assessment of Negatively Reinforcing Properties of Drugs

The focus of most abuse potential studies is on the pleasant and the positively reinforcing properties of drugs. However, it is well known that drugs of abuse produce multiple effects which may be both pleasant and aversive. In drug abusing persons, effects considered aversive to non-drug abusers (or effects that were aversive during the drug abuser’s initial experience) may become part of the constellation of desired drug effects. For instance, the nausea produced by opioids may be reported as a "pleasant sick;" inability to concentrate, dizziness, and sedation may also be reported as pleasant in the drug abuser (Goldberg, Shannon, & Spealman, 1981; Jasinski, 1977). These effects of drugs may be quantified by their ability to suppress drug-taking behavior (punishing effects) or by their ability to maintain behavior leading to the avoidance of drug administration (negatively reinforcing effects). Such procedures have been developed using animals as subjects (e.g., Goldberg & Spealman, 1982) and were extended to human studies at the Addiction Research Center (Henningfield & Goldberg, 1983).

A subject who self-administered nicotine at a very low rate was tested in a drug avoidance paradigm. Nicotine injections were programmed to be delivered at 15-minute intervals unless the subject blocked the injections by pressing the lever 10 times. As shown in Figure 8, the subject blocked most injections when nicotine was present and took most injections when saline was present. Additionally, ratings of aversive or "bad" effects produced by injections were directly related to avoidance behavior. Additional manipulations in two other subjects showed that the strength of the avoidance behavior was directly related to the dose of nicotine.

Avoidance of intravenous nicotine injections
Figure 8: Subject PA was tested during 3-hour sessions on a concurrent schedule in which pressing the right lever (FR-10) produced a nicotine injection and pressing the left lever extinguished the left lever light and avoided the next programmed injection (12 injections were programmed at 15-minute intervals). The dose available was administered once, 15 minutes prior to the start of the session. The subject did not press the right lever. The number of programmed injections and the number of responses (left lever) per session are shown for seven consecutive sessions. The negative Visual Analog Scale scores are the number of mm away from the neutral point on a 100-mm line analogue scale; the instructions on the scale were to place a mark on the line that "indicates the strength of any bad or negative effect which you don’t like." The score shown is the mean produced by the first three injections. Reprinted with permission from Henningfield & Goldberg, 1983. Copyright 1983 by ANKHO International, Inc.


The strategies which are described in this chapter may be used to evaluate drugs for abuse liability and dependence potential. The theoretical framework underlying such studies is that quantifiable pharmacological properties of drugs determine abuse liability. This theory is not inconsistent with the observation that social, genetic, personality, and other factors are involved in any given instance of substance abuse. Rather, carefully controlled, systematic studies provide a rational means for differentiating substances in accord with the degree to which their pharmacological effects determine their abuse.

It is apparent that techniques developed in studies with human subjects can be adapted to studies with animals (e.g., drug discrimination testing); the converse is also true (e.g., drug self-administration studies). Furthermore, findings from studies with human subjects generally correspond to those obtained with animal subjects, supporting the validity of the methods as well as the biological basis of substance abuse. This cross validation of human and animal findings supports the hypothesis that substances of abuse share common, biologically based mechanisms of action, assessed as euphoriant and reinforcing properties. The ultimate validity of these models of euphoria and reinforcing efficacy will be the extent that they continue to predict and explain the phenomena of drug abuse. An important area of future research will be to further investigate the relationship between subjective drug effects and the properties of drugs as reinforcers.

As is evidenced by the papers in this volume, there are some apparent discrepancies in the evaluation of abuse liability when the results of different methods are compared to one another. For instance, the results of studies involving place preference indicate that drugs known to serve as reinforcers in drug self-administration paradigm may have aversive properties (e.g., ethanol). These findings, however, may not be discrepant at all but may simply reflect the different functional properties of drugs that are measured by the two procedures. The self-administration procedure is a measure of propensity to take a drug while the place preference procedure may reflect the qualitative nature of certain interoceptive effects of drugs.

The notion that drugs of abuse produce both noxious and pleasurable effects is well known. In animal studies, as well, it has been shown that drugs known to serve as reinforcers under certain conditions serve as negative reinforcers and punishers under other conditions (Goldberg & Spealman, 1982; Spealman, 1983; see also, Wise, de Wit, & Yokel, 1976). Observations made from the human studies of intravenous nicotine self-administration may help to resolve these discrepancies. In the human studies it is concurrently possible to assess the reinforcing properties of the drugs, as evidenced by response rate measures, and to assess the degree of noxious and pleasurable effects produced by the drugs (Henningfield, 1983; Henningfield & Goldberg, 1983). Results from an ongoing study suggest that while a drug may be administered at a monotonic rate across a 3-hour session, the subjective effects may change considerably. Typically, liked effects diminished across successive injections (tolerance) while disliked (toxic) effects increased. Following sessions, it was not unusual for a subject to report that he would have taken more injections than he did except that the pleasurable effects of the drug were counterbalanced by noxious effects (e.g., nausea, burning in the arm, dizziness). These data provide direct evidence that a range of subjective drug effects may occur when a drug is serving as a reinforcer and that simple measurement of the nature of the effect may not be a reliable predictor of propensity to take the drug. Thus, self-administration data may provide a means of quantifying the abuse liability of a drug, while measures of subjective effects (e.g., pleasure inducing or place preference) may be more useful in the descriptive pharmacology of the drug.

The advent of new strategies and the refinement of old strategies for assessing abuse liability and dependence potential have paralleled new concepts underlying the process of drug dependence. The fundamental advance was the finding that drugs of abuse produce orderly changes in mood and feeling (e.g., euphoria), that these effects could be quantified using bioassay-like procedures, and that these changes generally correspond with reinforcing efficacy. Another major conceptual advance was the discovery that physiological dependence was neither necessary nor sufficient to maintain drug-seeking behavior. On the contrary, abuse is maintained by the reinforcing properties of substances, and these properties do not necessarily vary as a function of physiological dependence. Finally, the self-administration paradigm provides a means to quantify the reinforcing properties of substances. Findings from these strategies have blurred the distinction between psychological and physiological dependence—the subjective and reinforcing properties of drugs are clearly a function of physiological alterations in the central nervous system.


Brady, J. V., & Lukas, S. E. (Eds.) (1984). Report on the assessment of abuse liability and dependence potential (National Institute on Drug Abuse Research Monograph 52). Washington, DC: U.S. Government Printing Office.

Fraser, H. F., & Isbell, H. (1960). Human pharmacology and addiction liabilities phenazocine and levophenalcylmorphan. Bulletin on Narcotics, 12, 15-23.

Fraser, H. F., Van Horn, G. D., Martin, W. R., Wolbach, A. B., & Isbell, H., (1961). Methods for evaluating addiction liability. (A) "Attitude" of opiate addicts toward opiate-like drugs. (B) A short term "direct" addiction test. Journal of Pharmacology and Experimental Therapeutics, 133, 371-387.

Goldberg, S. R., Shannon, H. E., & Spealman, R. D. (1981). Psychotropic effects of opioids and opioid antagonists. In F. Hoffmeister & G. Stille (Eds.), Handbook of experimental pharmacology. (Vol. 55, pp. 269-304). Berlin: Springer-Verlag.

Goldberg, S. R., & Spealman, R. D. (1982). Maintenance and suppression of responding by intravenous nicotine injections in squirrel monkeys. Federation Proceedings, 41, 216-220.

Goldberg, S. R., Spealman, R. D., Risner, M. E., & Henningfield, J. E. (1983). Control of behavior by intravenous nicotine injections in laboratory animals. Pharmacology Biochemistry & Behavior, 19, 1011-1020.

Griffiths, R. R., & Balster, R. L. (1970). Opioids: Similarity between evaluations of subjective effects and animal self-administration results. Clinical Pharmacology and Therapeutics, 25, 611-617.

Griffiths, R. R., Bigelow, G. E., & Henningfield, J. E. (1980). Similarities in animal and human drug taking behavior. In N. K. Mello (Ed.), Advances in substance abuse: Behavioral and biological research. (pp. 1-90). Greenwich, CN: JAI Press.

Haertzen, C. A. (1974). An overview of addiction research center inventory scales (ARCI): An appendix and manual of scales (DHEW Publication No. ADM 74-92). Washington, DC: U.S. Government Printing Office.

Henningfield, J. E. (1983). Measurement issues in cigarette smoking research: Basic behavioral and physiological effects and patterns of nicotine self-administration. In J. Grabowski, & C. S. Bell (Eds.), Measurement issues in tobacco research and treatment (National Institute on Drug Abuse Research Monograph 48, pp. 27-38). Washington, DC: U.S. Government Printing Office.

Henningfield, J. E., & Goldberg, S. R. (1983). Control of behavior by intravenous nicotine injections in human subjects. Pharmacology Biochemistry & Behavior, 19, 1021-1026.

Henningfield, J. E., Miyasato, K., & Jasinski, D. R. (1983). Cigarette smokers self-administer intravenous nicotine. Pharmacology Biochemistry & Behavior, 19, 887-890.

Henningfield, J. E., Miyasato, K., & Jasinski, D. R. (1985). Abuse liability and pharmacodynamic characteristics of intravenous and inhaled nicotine. Journal of Pharmacology and Experimental Therapeutics, 234, 1-12.

Hickey, J. E., & Jasinski, D. R. (note 1). Protection of the rights and welfare of human subjects. Manuscript in preparation.

Himmelsbach, C. K. & Andrews, H. L. (1943). Studies on modification of the morphine abstinence syndrome by drugs. Journal of Pharmacology and Experimental Therapeutics, 77, 17-23.

Isbell, H. (1948). Methods and results of studying experimental human addiction to the newer synthetic analgesics. Annals of the New York Academy of Sciences, 51, 108-122.

Jaffe, J. H. (1980). Drug addiction and drug abuse. In A. G. Gilman, L. S. Goodman, & A. Gilman (Eds.), The pharmacological basis of therapeutics (pp. 535-584). New York: Macmillan.

Jasinski, D. R. (1977). Assessment of the abuse potentiality of morphine like drugs (methods used in man). In W. R. Martin (Ed.), Drug addiction I: Morphine, sedative/hypnotic and alcohol dependence (Handbook of Experimental Pharmacology, Vol. 45, pp. 197-258). Berlin: Springer-Verlag.

Jasinski, D. R. (1981). Opiate withdrawal syndrome: Acute and protracted aspects. Annals of the New York Academy of Sciences, 362, 183-186.

Jasinski, D. R., Griffith, J. D., & Pevnick, J. S. (1978). Human pharmacology and abuse potential of the analgesic buprenorphine. Archives of General Psychiatry, 35, 501-516.

Jasinski, D. R., Haertzen, C. R., Henningfield, J. E., Johnson, R. E., Makhzoumi, H. M., & Miyasato, K. (1982). Progress report of the Addiction Research Center. In L. S. Harris (Ed.), Problems of drug dependence, 1981 (National Institute on Drug Abuse Research Monograph 41, pp. 45-52). Washington, DC: U.S. Government Printing Office.

Jasinski, D. R., Henningfield, J. E., & Johnson, R. E. (1984). Abuse liability assessment in human subjects. Trends in Pharmacological Sciences, 5, 196-200.

Jasinski, D. R., Johnson, R. E., & Kocher, T. R. (1985). Clonidine in morphine withdrawal: Differential effects on signs and symptoms. Archives of General Psychiatry, 42, 1063-1066.

Lukas, S. E., & Griffiths, R. R. (1982). Precipitated withdrawal by a benzodiazepine receptor antagonist (RO 15-1788) after 7 days of diazepam. Science, 217, 1161-1163.

Martin, W. R. (1977). Chemotherapy of narcotic addiction. In W. R. Martin (Ed.), Drug addiction I: Morphine, sedative/hypnotic and alcohol dependence (Handbook of Experimental Pharmacology, Vol. 45, pp. 279-318). Berlin: Springer-Verlag.

Martin, W. R., & Isbell, H. (1978). Drug Addiction and the U.S. Public Health Service (DHEW Publication No. 77-434). Washington, DC: U.S. Government Printing Office.

Martin, W. R., Jasinski, D. R., Haertzen, C. A., Kay, D. C., Jones, B. E., Mansky, P. A., & Carpenter, R. W. (1973a). Methadone—a reevaluation. Archives of General Psychiatry, 28, 286-295.

Mendelson, J. H., & Mello, N. K. (1966). Experimental analysis of drinking behavior of chronic alcoholics. Annals of the New York Academy of Science, 133, 828-845.

Spealman, R. D. (1983). Maintenance of behavior by postponement of scheduled injections of nicotine in squirrel monkeys. Journal of Pharmacology and Experimental Therapeutics, 227, 154-159.

Thompson, T., & Johanson, C. E. (Eds.). (1981). Behavioral pharmacology of human drug dependence (National Institute on Drug Abuse Research Monograph 37). Washington, DC: U.S. Government Printing Office.

Thompson, T., & Schuster, C. R. (1968). Behavioral pharmacology. Englewood Cliffs, NJ: Prentice-Hall.

Wikler, A. (1952). A psychodynamic study of a patient during self-regulated readdiction to morphine. Psychiatric Quarterly, 26, 270-293.

Wise, R. A., de Wit, H., & Yokel, R. A. (1976). Both positive reinforcement and conditioned aversion from amphetamine and from apomorphine in rats. Science, 191, 1273-1275.

Woods, J. H., Herling, S., & Young, A. M. (1982). Classification of narcotics on the basis of their reinforcing, discriminative, and antagonist effects in rhesus monkeys. Federation Proceedings, 41, 221-227.

Yanagita, T. (1976). Some methodological problems in assessing dependence-producing properties of drugs in animals. Pharmacological Reviews, 27, 503-509.

Yanagita, T. (1980). Self-administration studies on psychological dependence. Trends in Pharmacological Sciences, 161-164.

©1987 Springer-Verlag (printed version)
©2000-2009 Addiction Science Network (web-enhanced version)

ASNet Home

ASNet Profile

Illicit Drug Index

Virtual Lab Tour

Research Reports

Drug Classification

A Primer on Drug Addiction

Experimental Methods

Treatment Resources

Biological Basis of Addiction

Before Prohibition: Early Psychoactive Medicines

Links to Other Websites

Click here to enter the Addiction Science Network Discussion Forum

Brain Reward System©1999-2009 Addiction Science Network
This page was last revised 08 April 2009 16:59 EDT.
Send comments to:
Report technical problems to: