Preferences for ethanol and diazepam in anxious individuals: an evaluation of the self-medication hypothesis.

The self-medication hypothesis of addictive disorders postulates that individuals with psychiatric symptoms use drugs to alleviate their symptoms. Although commonly cited to explain the etiology of substance abuse, self-medication has not been experimentally validated. This study evaluated one version of the self-medication hypothesis by formulating it into a testable hypothesis: are highly anxious volunteers more likely to self-administer anxiolytic drugs than non-anxious controls. Anxious (ANX, n = 22) and control (CTL, n = 23) subjects participated in two double-blind placebo-controlled experiments, one testing ethanol (0.8 g/kg) and the other testing diazepam (20 mg). Subjects sampled and then chose between ethanol and placebo in one experiment, and diazepam and placebo in the other. The main dependent measures were choice of drug over placebo and subjective responses to the drugs. Ethanol decreased self-reported anxiety in ANX subjects, but ANX subjects did not choose ethanol more often than CTL subjects. Diazepam did not measurably reduce anxiety, but ANX subjects nevertheless chose diazepam more often than did CTL subjects. Thus, there were some differences in drug responses between the ANX and CTL subjects, and the study provided limited support for the self-medication hypothesis. However, drug choice was not directly related to anxiolytic drug effects with either ethanol or diazepam. The procedure may be used to test other formulations of the self-medication hypothesis (e.g., examining other psychiatric risk factors).