Drug Preference in Humans: Lorazepam


A drug’s capacity to reinforce behavior in a laboratory setting usually correlates with its dependence potential in the general population. In laboratory tests, diazepam is not an effective positive reinforcer, either in laboratory animals using drug self-administration tests () or in normal human volunteer subjects using a choice test (). The failure to find evidence for a positive reinforcing effect of diazepam in these experimental tests is inconsistent with clinical reports that diazepam is used excessively by some people.

The failure to demonstrate the positive reinforcing efficacy of diazepam in an experimental situation may be due in part to the drug’s long duration of action (half-life = 24 – 48 hours). In animal self-administration studies that test the reinforcing efficacy of drugs, it has been found that benzodiazepines that have shorter durations of action are also more effective reinforcers (). In the present study, human subjects were tested for preference for lorazepam, a benzodiazepine with effects similar to diazepam but with a shorter half-life than diazepam (half-life = 12-15 hours).


Subjects. Twelve normal healthy volunteers, aged 21 to 27 (4 males, 8 females) participated in this study. They were recruited using advertisements in the local student newspaper, notices posted on the University campus, and word-of-mouth referrals. Prior to acceptance, subjects were interviewed to explain the nature of the study and to ascertain their medical, psychiatric and drug use histories. Subjects were accepted if they were considered normal and healthy on the basis of this interview and a subsequent EKG. Most subjects had some experience with psychotropic drugs but none had a history of any type of drug abuse.

Subjects signed a consent form prior to participation which outlined the study in detail and indicated all possible side effects of any drug they might be given. They were informed that they would not be told what drug they ingested at the time, except that it would either be a psychomotor stimulant, minor tranquilizer, or a placebo, and that the dose would be within the daily therapeutic range. Bach subject also agreed not to take other drugs, except their normal amounts of coffee and cigarettes, 12 hours before and 6 hours after taking a capsule. Except for the actual drug ingested, subjects were completely informed of all other procedural details as outlined below.

Procedure: All subjects participated in each of four experiments, presented in counterbalanced order. The procedure for each experiment was identical except for the drugs available, which were as follows:

  • Experiment 1: lorazepam, 0.5 mg versus placebo
  • Experiment 2: lorazepam, 1.0 mg versus placebo
  • Experiment 3: lorazepam, 2.0 mg versus placebo
  • Experiment 4: lorazepam, 1.0 mg. versus diazepam, 5 mg

The doses of lorazepam are within the therapeutic range for the drug’s anxiolytic effect, and the doses of lorazepam and diazepam tested in Expt. 4 are considered therapeutically equipotent ().

Every experiment consisted of three sessions per week over a 3-week period, resulting in a total of nine sessions. During the first four sessions, the subject reported to the experimental room between 9 and 10 a.m. At that time, he/she filled out mood forms (see below) and received a colored capsule for immediate ingestion. Approximately half of the subjects received drug during sessions 1 and 3 and placebo (or diazepam in Expt. 4) during sessions 2 and 4. The order was reversed for the other half. For each subject, each drug was dispensed in a capsule of a consistent and distinctive color in order to facilitate identification. Capsule colors were assigned randomly across subjects to avoid the influence of color preference. Each subject was instructed during the initial four sessions to note the capsule. Colors, and to try to associate each of the two colors with the effects of the substances contained in them. After ingesting the capsule, subjects were free to leave. They took three additional mood forms with them, which they were to fill out 1, 3, and 6 hr later. In addition, subjects filled out a questionnaire at hour 6, indicating whether they liked the drug (from “disliked a lot” to “liked a lot”), what they thought it was (stimulant, tranquilizer and placebo), and whether they had experienced any unusual reactions. During the last five sessions, the procedure was identical in every respect except that the subjects were given a choice of the two colored capsules to ingest.

Subjective Effects. The scales used to assess mood were an experimental version of the Profile of Mood States (POMS) and a shortened version of the Addiction Research Center Inventory (ARCI). Both have been shown to be sensitive to the effects of psychotropic drugs (). The POMS consists of 72 adjectives commonly used to describe momentary mood states. Subjects indicate how they feel at the moment in relation to each of the 72 adjectives on a 5-point scale ranging from “not at all” (0) to “extremely” (4). There are eight clusters of items (subscales) which have been separated empirically using factor analysis (Anxiety, Depression, Anger, Vigor, Fatigue, Confusion, Friendliness, and Elation). The value of each subscale is determined by adding the numbers checked for each adjective in the cluster and dividing the total by the number of adjectives. Two additional subscales, Arousal and Positive Mood, were derived from the other subscales as follows: Arousal = (Anxiety + Vigor) – (Fatigue + Confusion), Positive Mood = Elation – Depression. The ARCI consists of 49 true/false items which have been separated into 5 clusters described as measuring typical drug effects such as stimulant-like (A and BG), euphoric (MBG), sedative (PCAG) and dysphoric (LSD).

For each experiment the ten POMS scores were averaged across sessions for each subject separately for drug and placebo at each of the four time periods. A two-way analysis of variance (drug x hour) was performed separately for each factor. If a significant (P < 0.05) drug x hour interaction was found, further statistical tests were conducted to determine at which hours the scores for the two drugs were significantly different.

Drug Preparation. Drug tablets (Ativan, Wyeth Laboratories; Valium, Hoffman-LaRoche) of the required dose were placed in opaque gelatin capsules (size 00) which then were filled with dextrose powder. Placebo capsules were identical in size and contained dextrose powder alone.

Results. The number of drug choices in each of the four experiments is illustrated in Figure “Mean (and S.E.M.) number of lorazepam choices (out of 5) in comparisons between lorazepam (0.5, 1.0 or 2.0 mg) and placebo, and lorazepam (1.0 mg) and diazepam (5.0 mg). Dashed line indicates level of choice that would be expected by chance; asterisk indicates significant differences from chance”. The mean number of lorazepam choices out of 5 were as follows: 0.5 mg dose: 2.5 (49%); 1.0 mg: 1.6 (32%); and 2.0 mg 0.6 (16%). Only the- proportion of choices for 2.0 mg lorazepam over placebo differed significantly from chance (one-sample t-test, t = 4.2, df = 11, p <.01) but a trend of decreased preference related to dose is apparent. Subjects showed no preference for either drug in the comparison between lorazepam and diazepam (46% lorazepam choice). The POMS and ARCI scores reflected lorazepam’s known anxiolytic and sedative effects. At the 0.5 mg dose only Anxiety scores (POMS) were decreased at hour 6. At 1.0 mg the drug decreased Arousal and increased Fatigue and Confusion on the POMS. At 2.0 mg, Anxiety, Vigor and Arousal were decreased, and Confusion increased. On the ARCI, PCAG scores (sedative effects) were increased and BG (stimulant effects) scores were decreased at 1.0 mg and 2.0 mg. LSD scores were elevated at the 2.0 mg dose. There were no significant differences in the comparison between lorazepam and diazepam. For purposes of illustration, the drug’s effects on one POMS (Arousal) and one ARCI (PCAG) subscale are presented in Fig. “Examples of POMS (Arousal) and ARCI (Sedation) subscales showing subjective effects of lorazepam (solid lines) versus placebo (broken lines) or diazepam (dashed line, last panel) over a 6-hour period. Asterisks indicate significant (p <.05) differences between drug and placebo (Fisher’s LSD post hoc test)” (a and b).

Although the data are not shown here, the results from the post-session questionnaires on liking and drug identification were consistent with the choice and subjective effects results already described. Accuracy in labelling lorazepam as a tranquilizer increased as the dose increased, and mean ratings of drug liking were in agreement with mean drug choices for the three doses, that is, both measures decreased with increasing doses. The degree of concordance between the liking rating on Sessions 1 – 4 and the number of times the drug was subsequently selected was determined by calculating Pearson’s r between these two measures for each of the 3 experiments. Only in the 2.0 mg comparison did this correlation reach significance (r = 0.798, df = 10, p <.01).


These experiments showed that normal volunteer subjects did not prefer lorazepam over placebo in a laboratory test of choice. At the lowest dose tested (0.5 mg), subjects appeared to be indifferent to the drug whereas at the highest dose (2.0 mg) there was a clear preference for the placebo. When therapeutically equipotent doses of lorazepam and diazepam were compared, subjects showed no preference for either drug.

The subjective effects of lorazepam were in general consistent with the drug’s known sedative properties, and, with some exceptions (see below), the same as the effects of diazepam. The most notable and surprising finding from the point of view of the purpose of this study was the relatively long duration of effect of this drug. On each of the subscales where there was a significant drug x hour interaction, the difference between drug and placebo was present 6 hours after drug ingestion (as measured by Fisher’s LSD post-hoc tests). In some cases the drug’s effect was also evident at hours 1 and 3, but in no case did the effects peak before hour 3 or disappear by hour 6. These results are in marked contrast to previous findings with 5 and 10 mg diazepam (), when the drug effects peaked at hour 1 and had largely dissipated by the 6th hour.

While it seems paradoxical that a drug (lorazepam) whose half-life is shorter than another (diazepam) should have a longer duration of effects, closer examination of the pharmacokinetic properties of these two drugs provides an explanation. After oral administration, peak plasma concentrations of lorazepam are reached at about 2 hours () whereas for diazepam peak levels are reached 30 minutes after administration (). In both cases, the peak clinical effects of the drugs correspond to the peak plasma levels. In addition, lorazepam crosses the blood-brain barrier less readily because of its lower lipophilicity. The drugs also differ in the extent to which they are distributed to tissue. Diazepam is more readily absorbed by peripheral tissue, decreasing plasma levels somewhat but delaying total elimination time. Thus, plasma levels of lorazepam remain higher after peak levels are reached, but are then also more rapidly eliminated.

Lorazepam’s effects on mood are very similar to the effects of diazepam. In the experiment comparing equipotent doses of the two drugs (1 mg lorazepam versus 5 mg diazepam), there were no differences on the POMS or ARCI subscales. Furthermore, the results of the three dose comparisons with placebo can be compared to previous findings with diazepam versus placebo. Both drugs produced dose-dependent increases in Fatigue and Confusion, and decreases in Vigor and Arousal (). While direct comparisons cannot be made across experiments, casual inspection of the data indicate that the direction and magnitude of the effects of the two drugs are similar, although the time course of effects is notably different, as described above. Anxiety scores on the POMS were decreased after lorazepam, an effect reported with diazepam in one study () but not in another (). The drugs appear to have similar effects on the PCAG and BG subscales of the ARCI, but the increased LSD scores observed after the highest dose of lorazepam have never been reported with diazepam.

It is notable that decreased Anxiety was the only measurable mood effect after 0.5 mg lorazepam, and that it occurred in the absence of any measurable sedative-like effects. This dissociation of effects has implications both for the drug’s usefulness in treatment, and for our understanding of the mechanisms of action of anxiolytic drugs.

In conclusion, these experiments provide data on the reinforcing property of another benzodiazepine, lorazepam, in normal volunteer subjects. The results with lorazepam are in agreement with previous findings using diazepam, both in the drug’s sedative-like effects on mood, and the notable absence of a preference for the drug over placebo. The absence of preference for these drugs suggests that in this subject population their dependence potential is low. The hypothesis that a shorter-acting benzodiazepine is more reinforcing could not be tested, due to the unexpectedly long duration of lorazepam’s effects.


Selections from the book: “Problems of Drug Dependence, 1983. Proceedings of the 45th Annual Scientific Meeting, the Committee on Problems of Drug Dependence, Inc.” Louis S. Harris, Ph.D., ed. A collection of papers which together record a year’s advances in drug abuse research; also includes reports on tests of new compounds for efficacy and dependence liability. National Institute on Drug Abuse Research Monograph 49, March 1984.