Michael Feinberg, Jean-Paul Pegeron, and Meir Steiner
NIDA Research Monograph Series 43, pp. 245-250, 1982.

Resurgence of interest in the association between narcotics and mental illness suggests that a careful reading of the older clinical literature may point the way toward fruitful experiments which will shed light on the pathophysiology of some mental illnesses and may suggest radically new treatments. Comfort (1977) has briefly reviewed the literature written in English.

We have reviewed the clinical literature on the treatment of melancholia with morphine or opium, covering the years from 1850 to 1960, to address several questions:

  1. Is morphine effective in relieving the symptoms of endogenous depression?

  2. If so, is this merely nonspecific sedation, or is there some more specific treatment of symptoms?

  3. Does the relief of symptoms persist after the administration of morphine is stopped?

  4. How do depressed patients respond to morphine? What dose is needed for relief of symptoms, or for sedation? What is the incidence of tolerance and dependence?

Nearly every author cited concluded that morphine provided symptomatic relief in melancholia, and some suggested that morphine cured the patient. Kielholz (1959) compared opium with imipramine and concluded that opium produced symptomatic relief but that the symptoms might return if treatment were stopped. (It is interesting to note that this represents the current clinical wisdom about imipramine: the symptoms may return if administration of the drug is stopped too early.) Tigges (1864) was the only author who found opium useless in the treatment of melancholia. He described the treatment of 39 patients, and provided a good deal of information about each patient's history, symptoms, and treatment. He concluded that the poor response may have been caused by the patients' having chronic, rather than acute, melancholia. It is equally likely that the lack of response was due in part to the relatively low doses of opium used. Some older textbooks of psychiatry recommend opium as a sedative-hypnotic in endogenous depression (Bucknill & Tuke, 1858; Cramer et al., 1907, Maudsley, 1867), while others found it useless (Henderson and Gillespie, 1933). The latter did not comment on the doses used, and the former only occasionally provided guidelines for its use.

Nearly all of the patients described were psychotic. Morphine had an antipsychotic effect out of proportion to its effect on other symptoms, such as depressed mood. Kielholz (1959), Mickle (1874), and Marce (1857) were quite specific about this, and Ziehen (1889) found that the presence of hallucinations predicted a good response to opium. These reports are even more interesting in light of the more recent finding that psychotically depressed patients may not respond to treatment with tricyclic antidepressants (Glassman et al., 1975).

Some of the papers cited describe psychotic patients who probably did not have endogenous depression (Engelken, 1851; Knecht, 1872; Voisin, 1881; 1891); in these cases the antipsychotic effect was clearly separate from an antidepressant (or antimelancholic) effect. The antipsychotic effect of morphine, found in a wide variety of patients, is difficult to reconcile with the current theory that psychosis may be due to a functional excess of endorphins and be relieved by narcotic antagonists (Bloom et al., 1976; Berger et al., 1981). Gold et al. (1977) have suggested that morphine may have an antipsychotic effect, based on the morphine-induced increase in plasma prolactin.

Voisin also described an increased sensitivity to morphine, which occurred suddenly when patients became well. He could find no reason for this, or for the wide range of sensitivity to morphine in his patients; he said that he'd asked Claude Bernard, who was similarly puzzled.


The available evidence clearly suggests that morphine produces symptomatic relief in melancholic patients, and that this relief is out of proportion to any sedative effects. These effects seen to be dose related, as authors using lower doses (<60 mg of morphine/day) reported fewer responses to treatment than did those using higher doses. It is likely that rather than being a cure, this is symptomatic relief similar to that provided by antidepressant drugs, although most of the authors do not give enough information to allow a clear decision. This finding suggests that abnormal functioning of an endorphin "system" may be responsible for some of the symptom of melancholia. It is tempting, and may be fruitful, to speculate that this abnormal functioning may underlie the illness itself. It is also possible that morphine, an artificial endorphin, acts to remedy symptoms caused by abnormal function elsewhere in the brain.

The evidence suggests that we should mount a double-blind placebo-controlled study of morphine or other centrally acting endorphin agonist in patients with melancholia. Such a study should probably include a narcotic antagonist and one or more of the partial agonists. The drugs used should be selected for their differential effects on the several postulated endorphin receptors in the brain.

Michael Feinberg, M.D., Ph.D.; Meir Steiner, M.D., Ph.D.;
Jean-Paul Pegeron, M.D. Mental Health Research Institute and
Adult Psychiatric Service, Dept. of Psychiatry, University of
Michigan, Ann Arbor, Michigan 48109

Morphine: structure
Opiates as antidepressants
Is morphine a smart drug?
Opioids, mood and cognition
Prozac and the opioid pathways
Is morphine an antidepressant?
Depression, opioids and the HPA
Buprenorphine as an antidepressant
Opiates for treatment-resistant depression
Methadone and morphine as antidepressants?
Endogenous opioids mediate basal hedonic tone
Antidepressive response and depression subtypes
Use of oxycodone and oxymorphone to treat major depression
Kappa agonists, antagonists and partial agonists for mood disorders

Go To Good Drug Guide
BLTC Research
Designer Drugs
Utopian Pharmacology
The Hedonistic Imperative
When Is It Best To Take Crack Cocaine?