Hormonal Treatments for Women With Schizophrenia

Emil Kraepelin proposed links between dementia praecox (schizophrenia) and hormones in 1892. Other early researchers, such as Hoskins, studied endocrine changes in people with schizophrenia at postmortem. The discovery of insulin further stimulated interest in the interaction between behavior and metabolism. Between 1940 and 1970, there was considerable interest in the psychoendocrinology of schizophrenia. Mason (1975) demonstrated the mimicking of psychotic symptoms by administering high doses of steroid hormones. Brambilla and Penati (1978) reviewed the evidence for malfunctioning of the adrenal, pituitary and thyroid glands and of the gonads in people with schizophrenia. Endocrinopathies were only rarely present, thus suggesting that endocrine abnormalities in these patients are not the cause of schizophrenia. As the role of neurotransmitters in the regulation of pituitary hormone release via hypothalamic hormones becomes clearer, neuroendocrine studies in people with schizophrenia aim to use the pituitary gland as the "window to the brain." This involves using probes to modify the secretion of anterior pituitary hormones to detect abnormalities in tubero-infundibular pituitary function reflecting similar abnormalities in the mesolimbic system. Lieberman and Koreen (1993) concluded in a large review that the hypothalamic-pituitary-thyroid axis, in general, has not been shown to manifest disturbances in schizophrenia. Nevertheless, studies of growth hormone have tended to show a response to a wide variety of pharmacological probes (e.g., apomorphine [Uprima], tryptophan, fenfluramine [Pondimin], baclofen [Liofen, Lioresal], clonidine [Catapres], methylphenidate [Ritalin] and bromocriptine [Parlodel]), suggestive of enhanced noradrenergic activity in patients with schizophrenia compared with controls. A clear omission from the neuroendocrine studies described to date is the study of hormonal profiles in women with schizophrenia. Researchers have often excluded women because of a putative inability to control the variable fluctuating monthly hormonal cycles. In considering the relationship between gonadotrophin and other gonadal steroids and mental state in patients with schizophrenia, there are very few studies examining the female patient. Riecher-Rossler and colleagues (1994) studied 32 acutely psychotic women and found a significant excess of admissions in the premenstrual and menstruation phases with an inverse relationship between serum estradiol and severity of psychotic symptoms.

Estrogen Protection

Pioneering researchers in the area of gender differences in the onset, treatment and outcome of schizophrenia have proposed that, in women, estrogen may confer protection against the early onset of severe schizophrenia (Lewine, 1988; Seeman and Lang, 1990). These researchers have suggested that women are vulnerable to relapses of schizophrenia, or their first episode of illness, in the perimenopausal period when estrogen production diminishes. As articulated by Hafner and colleagues (1998), the hypothesis also encompasses early estrogen effects on the developing brain such that a structural effect of estrogen acting during brain maturation causes the delay of first onset schizophrenia in females. From puberty, this putative structural effect is reinforced by a functional effect. Fading estrogen secretion around menopause causes women predisposed for schizophrenia (but who until then had been protected by estrogens) to fall ill with late-onset schizophrenia. Essentially, the estrogen protection hypothesis is based on three broad lines of work: epidemiological studies, biological or basic science studies in animal models, and clinical studies (Hafner et al., 1998).

Epidemiology

A gender difference in the age at first admission, with women being older than men, was noted by Kraepelin (1913-1915). The finding was largely ignored until the late 1980s when Angermeyer and Kuhn (1988) confirmed the age difference at first admission in 50 out of 53 international studies. It is now widely accepted that males presenting with their first episode of schizophrenia are usually between the ages of 16 and 19 years, while women are up to five to 10 years older at first presentation.

Animal Models

Behrens and colleagues (1992) showed that estrogen has a complex action on the dopamine system in animals, but were able to conclude from a series of studies that estrogen downregulates dopamine transmission. This finding has been replicated in animals (Clopton and Gordon, 1986; Ferretti et al., 1992; Perry et al., 1981). More recently, Fink and colleagues (1999, 1998) have shown that estrogen induces a significant increase in 5-HT_2A receptors and the serotonin transporter (SERT) in regions of the rat forebrain that are concerned with mental state mood, cognition, memory, emotion and neuroendocrine control in humans. The precise mechanism of the estrogen serotonin interaction is not yet clear. The net effect of the action of estrogen on 5-HT_2A receptors and SERT dopamine in general could be thought to mimic so-called atypical or newer antipsychotic medications. The significance of these findings for schizophrenia suggests that in fact the effect of estrogen in the brain on key neurotransmitter systems involved in the production of psychotic symptoms accounts for the estrogen protection hypothesis.

Clinical Studies

Clinically, there have been case reports of psychotic symptoms occurring cyclically in women with schizophrenia during the late luteal phase when estrogen levels are at their lowest (Endo et al., 1978; Riecher-Rossler et al., 1992). Seeman (1983) reported that neuroleptic dose requirements in women who are psychotic rise as they approach menopause and described premenopausal women as "already neuroleptized by their antidopaminergic estrogens." The psychiatric literature contains case reports of women suffering from psychotic symptoms around the time of menopause (Berlin et al., 1982). There are also reports of improvement in psychotic symptoms with the addition of a combined estrogen-progesterone oral contraceptive (Dennerstein et al.,1983; Felthous et al., 1980). Following these clues, it appears that estrogen has inherent protective aspects that may prove to be clinically important in considering hormone treatments for women with schizophrenia. As a clinical trial, my colleagues and I conducted an open-label pilot study in which 11 women of childbearing age with schizophrenia were given 0.02 mg oral ethynyl estradiol as an adjunct to antipsychotic drug treatment for eight weeks. Their progress was compared to a similar group who received antipsychotic drugs only (Kulkarni et al., 1996). The group receiving estrogen made a significantly more rapid recovery from acute psychosis symptoms. Subsequently, we conducted a dose-finding study for the optimal use of estradiol in women with DSM-IV schizophrenia (Kulkarni et al., 2001). In this three-arm, double-blind, placebo-controlled, 28-day study, 12 women received 50 mcg transdermal estradiol plus antipsychotic medication; 12 women received 100 mcg transdermal estradiol plus antipsychotic medication; and 12 women received transdermal placebo plus antipsychotic medication. The main finding was that the women who received 100 mcg estradiol adjunct made a more rapid and greater recovery in their psychotic symptoms (positive, negative and general symptoms as measured by the Positive and Negative Syndrome Scale [PANSS]). To confirm this finding, we are continuing a two-arm, 28-day, double-blind, placebo-controlled study comparing symptom resolution in women with schizophrenia receiving transdermal 100 mcg estradiol plus antipsychotic medication, with women receiving transdermal placebo plus antipsychotic medication. This is an ongoing study, and interim data are included below.

Study Method

The sample size reported here is 36 (18 women in each group). Women were excluded if they were taking any synthetic steroids (including the oral contraceptive) or illicit drugs; they were also excluded if they were pregnant, lactating or postmenopausal. All patients gave written informed consent, and the Alfred Hospital Ethics Committee approved the study. Each subject was enrolled for 28 days (one menstrual cycle) and received baseline psychopathology and hormone assessment, followed by assessments at days 4, 7, 14, 21 and 28. Psychopathology was measured with PANSS. Hormone assays for serum estrogen, progesterone, prolactin, luteinizing hormone, follicle stimulating hormone and testosterone were performed.

Results

There were no statistically significant differences between the women receiving the 100 mcg transdermal estradiol adjunct and the women who received the placebo adjunct, in terms of age, menstrual cycle phase, race, illness duration, diagnosis or antipsychotic drug dose. The mean baseline PANSS was 86.3±12.3 for the estrogen group and 73.3±12.3 for the placebo group. The difference is statistically significant (p