12 Schizophrenia in women

Sex and gender differences in schizophrenic disorders are one of the most exciting areas of psychiatry, both for clinicians and researchers. For researchers, these differences could be a window to explore the aetiology or at least some pathogenetic mechanisms of these disorders as sex differences—especially regarding age of onset—are one of the most stable and well-established findings in schizophrenic psychoses. Not only biological differences but also differences regarding psychosocial factors influencing the outbreak and the course of these disorders are certainly of great importance in this context.

More knowledge about these mechanisms could also be helpful for developing new therapeutic strategies. For clinicians such new therapeutic strategies would be most welcome. A better understanding of the differences between men and women and their specific needs would also allow to provide specific services tailored to their different life circumstances.

Gender differences in schizophrenia have been described for a long time. In particular, differences in age at onset have been found in many studies (Häfner et al. 1989, 1993a, b; Jablensky et al. 1992). Less consistent are the findings concerning differences in symptomatology and the course of the disease.

Inconsistencies might be partly due to methodological problems in earlier studies (Riecher-Rössler and Rössler 1998). Thus, populations examined were frequently neither representative nor restricted to first-contact patients or first admissions. Often the studies did not rely on direct investigation or standardized assessment.

Few studies on this topic fulfil the essential methodological standards. One which does is the DOSMD Study (Determinants of Outcome of Severe Mental Disorders) by the World Health Organization (WHO), in which 1379 first-contact patients in 12 centres all over the world were examined (Jablensky et al. 1992). Another one is the Finnish First Contact Study by Salokangas and Stengard (1990), who examined 227 first-contact patients. In our own studies on gender differences (Häfner et al. 1989, 1993a, b; Riecher-Rössler et al. 1992) we examined a representative population of 267 first-time hospitalized patients directly and with standardized instruments retrospectively. We also examined 1169 first-time hospitalized patients using data from the Danish Case Register. In these studies many of the mentioned gender differences could be confirmed.

Concerning the difference in age of onset observed in many centres of the WHO studies (Jablensky et al. 1992) as well as in the Danish and the Mannheim Case Register (Häfner et al. 1989), the age of women at first contact or first admission was higher than that of men. The difference was on average between three and a half and nearly six years. In our direct investigation of the first onset of the disease we could demonstrate that this gender difference is genuine and is already evident at the first onset of signs and symptoms of the disease; there is not just a delay in help-seeking or hospital admission among women (Häfner et al. 1993a, b).

Searching for the explanation to this gender difference, a number of psychosocial factors have been suggested. However, none of them has so far been empirically proven (for a review see Riecher-Rössler and Häfner 2000).

The DOSMD study by the WHO shows that the observed sex difference in the age of onset is a phenomenon found in almost all cultures investigated all over the world (Jablensky et al. 1992). This finding probably also speaks against psychosocial causes of the gender difference.

At the same time we were able to show that the cumulative lifetime risk for men and women is identical (Häfner et al. 1989, 1993a, b). Men just fall ill earlier than women. This shows that gender has no stable influence on the outbreak of the disease, but that this influence is modulated by age. Men show their peak of first admissions in their early twenties, women only in their late twenties. Furthermore, there is a second, smaller peak of onsets in women after the age of 45 (Häfner et al. 1989, 1993a, b).

This distribution, together with the lack of psychosocial explanations for the gender difference in age of onset, directed our attention towards the oestrogen hypothesis (Riecher-Rössler and Häfner 1993a, b). According to this hypothesis, women are protected against schizophrenia between puberty and menopause to some extent by their relatively high physiological oestrogen production during this phase. Then, around age 45, several years before menopause sets in, oestrogen production begins to fall. Women lose the protection oestrogens potentially offer, and this may account for the second peak of illness onset in women after age 45.

The oestrogen hypothesis is not only based on epidemiological data, but also on increasing evidence from clinical and basic research, which shows that oestrogens have significant effects on the mental state of women, including women suffering from schizophrenia (see later).

As to the symptom-related gender differences noted in symptoms, it has often been reported that negative symptoms occur more frequently in men and affective symptoms more often in women. More recent studies have not consistently confirmed these findings. Addington et al. (1996) have shown that gender differences in negative and affective symptoms disappear when the sample is restricted to narrowly defined schizophrenia. Also, in our own direct investigation we did not find many gender differences, especially with respect to psychotic symptoms. The differences we found were more related to illness behaviour. For example, men showed more self-neglect or drug abuse, while women tended to be overadaptive (Häfner et al. 1993a, b).

Comorbidity, especially the greater prevalence of substance abuse in men, might have contributed to the gender differences found in the studies, which did not control for this. It may also be that negative symptoms have been reported more often in men, since men on average are prescribed higher doses of neuroleptics, and negative symptom rating scales do not always adequately distinguish between negative symptomatology and the neuroleptic side effects (Seeman 1996a).

Women have often been reported to have a more favourable course and a better psychosocial ‘outcome’ than men (Riecher-Rössler and Rössler 1998). They were said to have fewer and shorter hospital stays, better social adjustment, and a better living situation than men, whereas the symptom-related course seems to be similar for both genders. Women’s mortality also seems to be lower, mainly due to their lower suicide rate.

However, looking at more recent and more reliable studies, it seems important to differentiate carefully, i.e. between narrowly or broadly defined schizophrenia, young or older women, etc.

Thus, e.g. in our study based on the Danish Case Register cohort, the frequency and duration of hospitalizations over a 10-year period were the same for both genders when the analyses were restricted to narrowly defined schizophrenia (Maurer 1995). Similar results were found in the Nottingham subsample of the DOSMD study (Harrison et al. 1996).

Retterstøl (1991) showed that a more favourable course of disease is only observed in women up to menopause where after this the course deteriorates. Also Ciompi and Müller (1976) found that the symptoms in schizophrenic women are relatively mild at a young age, but become worse as age advances. In men, this pattern was reversed. Seeman (1983) demonstrated that only young women show a better response to neuroleptics as compared to men of the same age group.

There are several reasons for these contradictory results. Firstly, gender differences might be partly due to artefacts, e.g. results of gender-specific diagnostic habits. As a consequence of this, some differences disappear if one looks only at narrowly defined core-schizophrenia.

Secondly, the remaining ‘true’ effects of gender are—at least in part—age-dependent: on the one hand there is obviously a direct influence of actual age on the symptom-related course of disease. In view of the aforementioned results, this could be due to with the protective effect of oestrogens in younger women up to menopause.

The better overall social course of women on the other hand might also be a consequence of the higher age of onset and the accordingly better social integration of the female patients to start with. The first symptoms in men usually occur around the age of 20, but in women not until the mid 20s; thus, at illness onset women have usually already achieved much more stability in their various social roles as compared to men. Very often, women have already completed education, have a stable job and partnership, might be married, and so on. This better social integration significantly improves the prognosis—at least for the social course of the illness.

Finally, there are indications that women receive different and potentially better treatment than men do. Salokangas and Stengard (1990) were able to show that psychiatric teams primarily recommend psychotherapy for women, but rehabilitation concentrating on working capacity and basic social skills for men. Moreover, women actually used psychotherapy more often than men did.

A further reason for the better course in women is probably their better compliance and possibly also a lower susceptibility to expressed emotions. Other potentially contributing psychosocial factors have not yet been sufficiently examined (Riecher-Rössler and Rössler 1998; Riecher-Rössler 2000).

Summing up the findings on epidemiological and clinical gender differences, the lifetime risk for schizophrenia seems to be the same in men as in women. Women fall ill later and tend to have a better course of illness despite only few differences in symptoms to begin with. This applies especially to young women. After age 40, women not only fall ill twice as frequently as men, their symptoms and the course of their disease also seem to be worse.

The explanations for these gender differences are probably biological as well as psychosocial. It seems that the female sex hormone estradiol has a certain protective effect in young women (see later). The tendency for the course to be better in women certainly also has to do with the later age of onset, which is associated with better social integration. Other contributing factors could be better care offered to women and women’s better compliance.

Male and female patients with schizophrenia have the same pattern of structural brain abnormalities, but male patients appear to manifest greater severity, especially with regard to ventricular enlargement (Nopoulos et al. 1997). Sex as well as schizophrenic psychosis might also influence structural lateralization of the brain. Results remain relatively mixed, particularly at the regional level (Beaton 1997; Shapleske et al. 1999). Nonetheless, empirical evidence supporting the presence of greater structural asymmetries in men compared to women are well replicated in perisylvian, temporal, and orbital brain regions (Wada et al. 1975; Witelson and Kigar 1992; Bryant et al. 1999; Good et al. 2001; Knaus et al. 2004). Other structural imaging findings, however, fail to support the presence of sex differences in right-frontal and left-occipital width asymmetries or in hemispheric length asymmetries (Chui and Damasio 1980; Narr et al. 2007). Recent structural and functional magnetic resonance imaging (MRI) findings suggest disrupted sexual brain dimorphisms in schizophrenia are associated with sex-specific language deficits (Sommer et al. 2003) and left hippocampal abnormalities (Bühlmann et al. 2009), in particular, contribute to language dysfunction among men (Walder et al. 2007). In functional MRI studies, sex differences in cerebral activations in schizophrenia patients deviate from what has been observed in the general population. In men, exposure to and experience of negative affect evoked significantly greater activations in the thalamus, cerebellum, temporal, occipital, and posterior cingulate cortex, while women exhibited greater activations in the left middle frontal gyrus (Mendrek et al. 2007). Advances in neuroimaging techniques allow for a detailed investigation of morphometric and functional variability in the brain. It should be emphasized that sex differences in brain structure and function refer to average differences between men and women and that differences between individuals within each sex are much greater than the average difference between sexes (Resnick and Driscoll 2008).

The presence of neuropsychological impairment in patients suffering from schizophrenia is well established matter of fact. Deficits affect the areas of memory, attentional processes, and executive functions. Language, motor, and visuospatial abilities are less consistently impaired (for a review see Gopal and Variend 2005). However, gender differences with regard to neuropsychological capabilities in schizophrenia are an issue of controversy. Some research groups underscore inferior neuropsychological performance in men (e.g. Goldstein et al. 1998), others in women (e.g. Perlick et al. 1992), and finally there is a substantial number of reports which cannot provide evidence for gender differences at all (e.g. Goldberg et al. 1995).

In this context it is important to note that the impairment of neuropsychological functions depends on a wide variety of conditions, which might differ between men and women; i.e. the age of onset (Hoff et al. 1996), the severity of negative symptoms (O’Leary et al. 2000), or the overall symptom severity. Neuropsychological performance in women may fluctuate with their monthly cycling oestrogen levels, complicating the isolation of gender effects even more (Hoff et al. 2001). Contrary to women, strong laterality effects have been reported for male patients indicating a more pronounced impairment in language skills as compared to visuospatial capabilities (Ragland et al. 1999).

More recent reports emphasize conceptual differences of gender and sex in terms of a more psychosocially founded and multidimensional term of gender on the one hand versus a biologically reduced and dichotomous term of sex on the other hand. Lewine et al. (2006) report results that clearly indicate stronger gender than sex effects. While sex yielded two main effects in the domains of language and visual perception in favour of men, gender was associated with five main effects (language, verbal and spatial memory, visual perception, and motor speed) favouring the ‘feminine’ over the ‘masculine’ individuals. In this study gender was defined as ‘gender role’, i.e. the overt behaviour displayed in society to establish a position associated with the evaluation of one’s gender.

The same controversy holds true considering the course of the illness as a determining factor for gender differences in neurocognitive functions. There is only scarce evidence for gender differences in the first episode of psychosis. One of the few studies shows that men outperform women in tests tackling visuospatial abilities while women perform superior in verbal learning and memory (Albus et al. 1997). However, Hoff et al. (1998) could not find any gender differences in first-episode patients of psychosis after controlling for symptom severity.

There is hardly any evidence concerning individuals at-risk for or in a prodromal state of psychosis. Weiser et al. (2000) retrospectively analysed military draft board data and reported on individuals who later developed schizophrenia. In these prodromal individuals women performed poorer than men in a small set of verbal as well as non-verbal tests. Currently, there is no prospective study regarding individuals in a clinically defined prodromal state of psychosis.

These findings on gender differences clearly indicate that we need a more gender-sensitive attitude in our diagnostic and therapeutic approaches.

As regards biological aspects, consideration needs to be given to the large field of gender-specific pharmacotherapy in general and especially in pregnancy and during nursing. This broad topic, however, exceeds the scope of this chapter. Other important biological influences are those of gonadal dysfunction and oestrogens, which will be discussed in the next section. And just as important as considering these gender-specific biological influences is a gender-sensitive approach in the psychosocial treatment of schizophrenia, which means taking into account the influence of gender roles, sexuality, or motherhood.

Recent research increasingly points to the importance of oestrogens and the hypothalamic–pituitary–gonadal axis in schizophrenic psychoses.

On the one hand there are reports of gonadal dysfunction and states of oestrogen deficiency in women with schizophrenia; on the other hand there is mounting evidence from clinical as well as from epidemiological and basic research that estradiol, the main component of oestrogens, exerts protective effects in schizophrenia.

Important findings from basic research were the identification of oestrogen receptors in the limbic system of the brain, and the observation that the effects of oestrogens in rodents are, in some respects, similar to those of neuroleptics. Furthermore, it was shown that oestrogens can modulate the sensitivity and number of dopamine receptors. It was therefore hypothesized that oestrogens exert their antipsychotic effects in a manner similar to that of traditional neuroleptics, at least partly by blockade of dopaminergic transmission (for review see Riecher-Rössler and Häfner 1993a, b).

We now know that oestrogens, and especially 17-beta-estradiol (the natural oestrogen that is most active in the brain), produce many other neuroprotective and psychoprotective effects. For example, they appear to improve cerebral blood flow and glucose metabolism, promote neuronal sprouting and myelination, enhance synaptic density and plasticity, facilitate neuronal connectivity, act as antioxidants, and inhibit neuronal cell death. They have also been shown to exert profound effects on brain differentiation during development, particularly during late gestation and during the early postnatal period, and are important in normal maintenance of brain function during ageing (for reviews see Cyr et al. 2002; Goldstein et al. 2002; Oesterlund 2002; Vedder and Behl 2005).

In a well-controlled MRI study, Goldstein et al. (2002) showed that normal patterns of sexual brain dimorphism (brain regions found to be structurally different between normal men and women) are disrupted in schizophrenia, especially in the cortex. Apart from later ‘activational’ effects of circulating hormones (e.g. during puberty), those investigators suggested that these early ‘organizational’ effects of gonadal hormones that occur during the developmental period (which is probably critical for at least some forms of schizophrenia) could be partly responsible for that finding.

The mechanisms of action of oestrogens are now known not only to depend on the classical genomic pathway but also to involve non-genomic, rapid interactions, which explains the differing latency of effects. They clearly modulate the dopaminergic and other neurotransmitter systems that are believed to be relevant to schizophrenia, such as the serotonergic and glutamatergic system, but also the noradrenergic and cholinergic system (for reviews see Garcia-Segura et al. 2001; Stahl 2001a, 2001b; Cyr et al. 2002; McEwen 2002; Oesterlund 2002;). Recently it has even been suggested that 17-beta-estradiol in the brain might rather be regarded as a neurotransmitter itself than as a hormone (Balthazart and Ball 2006).

There are at least two subtypes of oestrogen receptors, namely oestrogen receptor-a and oestrogen receptor-b, which are transcribed from two distinct genes (Oesterlund 2002). Autopsy studies showed that oestrogen receptor-a messenger RNA is expressed in discrete areas of the human brain such as amygdala, hypothalamus, cerebral cortex, and hippocampus; these areas are associated with neuroendocrine function, as well as with emotion, memory, and cognition (Oesterlund et al. 2000).

Epidemiological studies into sex differences in schizophrenic disorders suggest that the physiologically high estradiol production in young fertile women contributes to the later age of onset of schizophrenia in women as compared with men, to the second peak of onset in women around the menopause, and to the better course of the disease in young women (Häfner et al. 1993a, b; Riecher-Rössler et al. 1997). A number of risk factors appear to counteract the protective effect of oestrogens. Thus, the sex difference in age of onset diminishes in the subgroup of cases with a genetic risk and in patients with perinatal complications (Könneke et al. 2000; Häfner 2005).

Recent results regarding age of menarche further support the hypothesis that physiological oestrogens play a protective role against development of the disease. We could demonstrate a significantly later age of menarche in a representative group of first admitted women with schizophrenia as compared with a healthy control group (Riecher-Rössler 2002). Seeman and coworkers (Cohen et al. 1999; Hayeems and Seeman 2005) found that later menarche was associated with an earlier onset of the illness, an association that was independent of factors such as family history and obstetric complications.

Clinically, psychotic symptomatology has often been found to correlate with the oestrogenic state of women (for reviews see Riecher-Rössler and Häfner 1993a, b; Seeman 1996a). For example, during high oestrogen phases such as pregnancy, chronic psychoses appear to improve, whereas there is an excess of psychoses after delivery.

Psychosis associated with oestrogen withdrawal due to conditions other than the puerperium was recently reviewed by Mahé and Dumaine (2001). Those investigators reported cases of premenstrual psychosis; postabortion psychosis; and psychoses associated with removal of hydatiform mole, cessation of oral contraceptives, clomiphene and tamoxifen administration (both oestrogen receptor antagonists), and gonadorelin agonist administration (which blocks pituitary stimulation of endogenous oestrogen secretion). Psychotic episodes were acute, short, and with a wide range of psychotic, but also affective, symptomatology. Recurrences were often reported when oestrogen withdrawal recurred, and puerperal psychosis was frequent in the history of the patients who were affected.

Psychotic symptoms in schizophrenic patients have also often been shown to deteriorate premenstrually or perimenstrually (i.e. in the low oestrogen phase of the cycle; for review see Riecher-Rössler and Häfner 1993a, b; Seeman 1996b; Riecher-Rössler 2002). Most studies, however, did not correlate symptomatology with estradiol serum levels directly. This was also true for a more recent study (Choi et al. 2001), which found only behavioural, affective, and somatic symptoms in schizophrenia (not psychotic ones) to be associated with menstrual cycle phase. In fact, an elevated number of admissions during the perimenstrual period has also been identified in other disorders (Althaus et al. 2000), and exacerbation of many psychiatric symptoms (not only psychotic ones) during the perimenstrual period was observed in schizophrenia patients (Riecher-Rössler et al. 1994; Harris 1997). In theory, this lack of specificity is to be expected because of the multiple effects of oestrogens on mental functioning.

Rather than examining cyclic fluctuations, Hoff et al. (2001) assessed the relation between average oestrogen levels from four consecutive weeks on the one hand and psychopathology and cognitive function on the other hand in 22 female inpatients with chronic schizophrenia, aged 22–63 years. There was no significant association between oestrogen levels and psychopathology, but higher average oestrogen levels were strongly associated with better cognitive abilities. However, this finding may partly be due to the effects of ageing.

Studies conducted in cases of late-onset schizophrenia demonstrated the importance of age. We showed not only that there are twice as many women as men with onset of schizophrenia beyond the age of 40 years, but also that they suffered from unexpectedly severe disease in terms of symptomatology and course (Riecher-Rössler et al. 1997; Riecher-Rössler 2002). One explanation for this could again be the loss of oestrogens just before and during menopause. In accordance with those findings are the results of long-term studies in schizophrenia, which have shown that the course of schizophrenia in women tends to deteriorate during the menopause and thereafter (for review see Riecher-Rössler and Häfner 1993a, b; Riecher-Rössler and Rössler 1998).

Intervention studies have also been conducted over long periods, mainly with positive results (Korhonen et al. 1995; Lindamer et al. 1997) (for a review see Riecher-Rössler and Häfner 1993a, b). In a systematic trial conducted in 1996, Kulkarni et al. (1996) found that schizophrenic women receiving estradiol as an adjunct to neuroleptic treatment exhibited more rapid improvement in psychotic symptoms than did women receiving neuroleptics only. In a double-blind, 28-day, placebo-controlled study (Kulkarni et al. 2001), 12 women were administered transdermal 17-beta-estradiol (patches) 50 mg/24 hours, another 12 women received 100 mg/24 hours, and the third group received placebo. The 100-mg group experienced greater improvement than either the 50-mg or the placebo groups; striking improvements were observed with respect to key psychotic symptoms. Louza et al. (2004) could not find a respective effect, which they themselves discuss to be possibly due to then having used conjugated oestrogens rather than 17-beta-estradiol, which is known to be active in the brain.

Hormone replacement therapy in postmenopausal women also appears to exert a positive effect. Lindamer et al. (2001) reported on a community sample of perimenopausal and postmenopausal women with schizophrenia. Of the women studied, 24 were on hormone replacement therapy for gynaecological reasons and 28 had never received such therapy. The users of hormone replacement therapy needed a relatively lower average dose of antipsychotic medication and suffered from less severe negative symptomatology.

Ahokas et al. (2000) demonstrated positive effects of oestrogen substitution in women with postpartum psychosis. In those women who exhibited sustained oestrogen deficiency states, the substitution of 17-beta-estradiol, without any further medication, yielded a dramatic antipsychotic effect within one week. However, the proportion of schizophrenia-like psychoses in the sample was not given.

Regarding the therapeutic effect of oestrogens, it must be noted that both, the numerous direct effects on the brain as well as indirect effects may play a role. For example, oestrogens may also increase blood levels of antipsychotic drugs via their actions on liver metabolism (Yonkers et al. 1992).

Several studies have recently confirmed earlier findings of disturbed gonadal function and hypo-oestrogenism in schizophrenic women (Riecher-Rössler and Häfner 1993a, b; Riecher-Rössler et al. 1994; Kulkarni et al. 1996; Riecher-Rössler et al. 1998; Choi et al. 2001; Hoff et al. 2001; Huber et al. 2001; Bergemann et al. 2002; Canuso et al. 2002; Smith et al. 2002; Zhang-Wong and Seeman 2002). They describe menstrual irregularities and reduced blood levels of estradiol, progesterone, and gonadotropins (follicle-stimulating hormone, luteinizing hormone) throughout the menstrual cycle, as well as an ovulation in the majority of women with schizophrenia. Reduced fertility was also reported.

There appear to be multiple reasons for these disturbances. Partly, they are probably a consequence of stress and/or neuroleptic-induced hyperprolactinaemia, which is known to suppress gonadal function (Maguire 2002). However, these are probably not the only causes, because other psychiatric disorders accompanied by similar ‘stress’ do not show the same disturbances or at least not to the same degree (Riecher-Rössler et al. 1998; Huber et al. 2001), and hypo-oestrogenism was observed long before the introduction of neuroleptics. Furthermore, recent findings are not unequivocal. Smith et al. (2002) found the dose of typical neuroleptics to correlate with prolactin levels especially in women and prolactin to correlate inversely with estradiol serum levels. In contrast to these findings, Huber et al. (2001) were unable to identify a significant association of prolactin and estradiol in 43 women with acute psychosis, 14 women with other diagnoses, and nine healthy control women. Nevertheless, women with schizophrenia had significantly lower estradiol serum levels than did the control women. Women with other psychiatric diagnoses fell in between the psychotic and the healthy group with regard to estradiol and prolactin levels. Also, in 16 premenopausal women with schizophrenia and schizoaffective disorders, Canuso et al. (2002) found a high rate of ovarian dysfunction and estradiol levels below normal, irrespective of medication type or prolactin status. Interestingly, Warner et al. (2001) found prolactin levels in unmedicated schizophrenic patients to be even lower than in control individuals. Those investigators suggested that this was due to a disordered dopaminergic system because dopamine tonically inhibits prolactin.

Taken together, these results imply that the hypothalamic–pituitary–gonadal axis is disturbed in many women with schizophrenia, and that the reasons for this are far from clear, yet. An interesting research question in this context is whether gonadal dysfunction with oestrogen deficiency could even be part of the underlying pathogenetic process, at least in a subgroup of women (Riecher-Rössler 2002).

Further research into the impact of gonadal function and oestrogens on schizophrenia is warranted because new diagnostic and therapeutic strategies could emerge that would benefit the many women worldwide who suffer from this disorder.

Because there is growing evidence that many even younger women with schizophrenia are in a state of oestrogen deficiency, in the future oestrogens and the gonadal axis should be considered more seriously in the treatment of women with schizophrenia. Psychiatric history taking should always include questions regarding menstrual irregularities, amenorrhoea, and galactorrhoea. Also, prolactin and oestrogen serum levels should be tested, if necessary. Gonadal dysfunction and hypo-oestrogenic states can often be found even in menstruating women (Riecher-Rössler et al. 1994; Riecher-Rössler et al. 1998; Smith et al. 2002). In addition, hyperprolactinaemia is clearly underdiagnosed (Maguire 2002). Some authors have therefore suggested routine laboratory tests (Smith et al. 2002).

Most neuroleptics can cause hyperprolactinaemia and can—especially if they are taken over a number of years—theoretically induce ‘iatrogenic early menopause’ via suppression of physiological estradiol production. The attendant risks include both short-term effects, such as hot flushes and sexual dysfunction, and long-term consequences, including osteoporosis and potentially cardiovascular disease or cognitive deterioration (Maguire 2002; Oesterlund 2002). In schizophrenic patients, these risks are further increased by additional risk factors such as smoking, poor diet, and reduced exercise (Smith et al. 2002).

Furthermore, menopausal complaints may lead to compliance problems. In the case of hyperprolactinaemia with secondary oestrogen deficiency, prolactin-sparing neuroleptics, e.g. clozapine, quetiapine, aripiprazol, or maybe olanzapine (Maguire 2002), should therefore be preferred. If a switch to these neuroleptics is not possible for clinical reasons or if hypo-oestrogenism persists despite switching, then oestrogens should be substituted. Issues regarding contraception must be taken into account in such cases because, when switching to prolactin-sparing neuroleptics, the menstrual cycle often normalizes and fertility is regained, with high risk for unplanned pregnancy (Neumann and Frasch 2001).

First trials of oestrogens in schizophrenia indicate that estradiol could be used as an adjunct to neuroleptic medication. However, further replication of these findings in larger controlled studies by different groups are needed before recommendations for broad clinical application can be made.

In women who suffer from frequent perimenstrual psychotic relapses, ‘cycle modulated’ neuroleptic therapy or, if contraception is needed at the same time, continuous use of oral contraceptives without hormone-free intervals may be strategies worthy of research (Braendle et al. 2001; Riecher-Rössler 2002).

Even more promising could be hormonal replacement with oestrogens in women with schizophrenia in peri- and postmenopause, since oestrogens in other disorders such as depression have proven to be especially helpful when they are used to restore hormonal balance.

In any case, oestrogen replacement therapy ameliorates perimenopausal complaints, which can act as stressors and theoretically provoke relapses. Hormone replacement therapy has also been recommended for other reasons, such as for prophylaxis of osteoporosis, for a delay of the age-dependent cognitive decline, and possibly also for Alzheimer’s disease (for a review see Riecher-Rössler and de Geyter 2007). Schizophrenia may be an additional indication to be studied.

Research should also be conducted concerning the best mode of hormone replacement therapy for psychiatric patients. Progestogens are usually added to oestrogens in order to prevent endometrial cancer, but they can antagonize the positive effects of oestrogens with respect to mental state (Braendle et al. 2001; Cyr et al. 2002). Furthermore, other risks associated with hormone therapy, such as breast cancer and cardiovascular disease for certain combinations, must be considered (Writing Group for the Women’s Health Initiative Investigators 2002), although these risks have been overestimated lately due to uncritical interpretations of studies such as the Women’s Health Initiative Study (Pines et al. 2007; Riecher-Rössler and de Geyter 2007; Rossouw et al. 2007).

Nevertheless, alternatives to conventional hormone replacement therapy, compounds with more specific and potent oestrogenic activity in the brain as opposed to other tissues, should be sought (Halbreich 2002; Riecher-Rössler 2002). Such compounds would both minimize the side effects of hormonal therapy and permit new therapeutic strategies in men. Possible candidates are the selective oestrogen receptor modulators, which have agonist or antagonist properties that depend on the target tissue. However, the effects of the available selective oestrogen receptor modulators on the brain remain to be clarified. Raloxifene, for example, appears to exert its main effects on the bone, although recent data suggest that it also acts on different brain receptors (Cyr et al. 2002). Also, the synthetic steroid tibolone appears to cause less endometrial proliferation, but its effects on the central nervous system are still not clear, apart from the fact that it appears to have an androgenic effect and increases beta-endorphin levels, with improvement in mood and libido (Davis 2002). Further studies on the brain-specific effects of selective oestrogen receptor modulators and other oestrogenic compounds (e.g. phyto-oestrogens, xeno-oestrogens, and dihydroepiandrosterone) are urgently needed.

In summary, hopes are emerging that oestrogens, as neuroprotective and psychoprotective adjunctive therapies, may complement the traditional drug therapies in schizophrenia in the future. However, it must be emphasized that most strategies are still being researched. In particular, results from larger controlled clinical trials are needed before oestrogens may be recommended as adjunct therapy in younger women without proven oestrogen deficiency. In contrast, other strategies should already be part of standard clinical care (Grigoriadis and Seeman 2002). These include examination of the gonadal axis, with therapeutic consequences, if indicated. Regarding oestrogen replacement therapy, it must be stressed that the decision must always be made on the basis of an individual risk–benefit assessment (NAMS 2000; Writing Group for the Women’s Health Initiative Investigators 2002) and in close cooperation with a gynaecologist.

For future research, many questions remain unresolved, not only regarding new therapeutic strategies and compounds but also regarding the poorly understood disturbances of oestrogens and the hypothalamic–pituitary–gonadal axis in women with schizophrenia. Further research in this area may even contribute to our understanding of the pathogenesis of this disease, at least in a subgroup of women.

Gender-sensitive therapy also means being sensitive to gender differences in symptom perception and illness concept, in coping, illness, and help-seeking behaviour (Riecher-Rössler 2000). The influence of current social roles (partnership, motherhood, professional roles, etc.), social status, social stress, and social support has to be taken into account just as much as the influence of gender-specific socialization and ‘gendered’ role behaviour.

It also means that we should be more aware of the practical needs of our patients. Thus, day-treatment or weekend home-leaves during a hospital stay for a housewife and mother may be a much greater stress than for a man, who may be free of any duties at home.

Also, significantly more women than men with schizophrenia have children, which is often associated with specific burdens and stressors (see later).

Also the later age of onset in women means that the disease has different effects on them than on men. As a consequence of this, women frequently have to cope with losses in many forms, such as in relationships with partners or children, or in the professional sphere. Thus, whereas for men the goal is often to attain certain roles for the very first time, for women, the focus in therapy often has to be the maintenance or re-establishment of certain roles.

Most women with schizophrenia are sexually active (Miller 1997), although sexual desire and orgasmic function can be impaired not only by the illness itself, but also by side effects of medication, especially anti-psychotic-induced hyper-prolactinaemia (see earlier).

Women with schizophrenia often suffer from sexual victimization, experience sexual abuse, or feel pressured to sexual activities which place them at risk for sexually transmitted diseases (for a review see Seeman and Fitzgerald 2000). When hospitalized, women often feel threatened by the aggressive or sexually assaultive behaviour of male patients. Doctors should be aware of these problems, ask their female patients actively regarding problems in this area, and offer specific help in form of counselling and arranging practical help. Cooperation with gynaecologists, social workers, specialized institutions for women such as women’s shelters and counselling agencies is often required. Establishing special ‘women’s areas’ in hospitals can be helpful.

In recent times, more women with schizophrenia are married and more often they have children. In a review, Mary Seeman (2004) concludes that over 50% of all female patients with schizophrenia become mothers, a percentage close to the general population. But pregnancy and motherhood can be difficult for women with schizophrenia.

Many women with schizophrenia desiring motherhood can not fulfil this wish without difficulties. Due to the disease or neuroleptics they can be infertile, often they do not have a suitable and reliable partner. These aspects should always be inquired when caring for women with schizophrenia and they should receive the necessary counselling. If a child is desired, the woman’s general situation should be considered, if possible together with her partner. What is the expected course of the disease? What would the effect of the additional burden of caring for a child be? What psychosocial support is there for mother and child? How much support can be offered? What are the father’s wishes and his situation? How is his psychosocial stability? If, after carefully considering these aspects, motherhood seems possible, the question of whether the course of the disease permits a temporary pause from the neuroleptic medication needs to be addressed, the patient should have been stable and without relapse for at least one to two years. It is advisable to taper out the medication slowly and the contraception should not be stopped before the patient is medication free. Otherwise a teratogenic effect of the neuroleptics during the first three months of pregnancy cannot be definitely excluded.

Unwanted or at least unplanned pregnancies are not rare. In these cases the pregnancy is often noticed very late and a doctor is consulted to discuss a termination, or confronted with the question whether the medication could have led to deformations of the child. These women need to be advised very carefully and their own wishes, possible ambivalence, and their capacity of judgement should be taken into account. Medication during the first trimester alone is not a reason for interrupting a pregnancy, as on the whole the risk of teratogenicity is low. However, the patient should have access to good prenatal care with specialized ultrasound monitoring etc. if she wants to bring the pregnancy to term.

For a woman affected with schizophrenia, motherhood can be a considerable strain, not only because of the care the child needs. The fear that the child could be taken away if she cannot cope has also to be taken into consideration. Partners who may be able to give support usually do not live with the mothers or are unknown (for a review see Seeman 2004). So, some of these fears are well founded, they can be very stressful for these women, and can maintain pathological processes of the disease as well as delaying remission. Because of these fears, many women do not seek help, which causes a vicious circle. Here psychiatric institutions which offer low threshold, specialized care are urgently needed. Parents with psychiatric diseases, especially single mothers, often do not accept the usual offers of care, such as day centres for patients with psychiatric diseases, as they do not want to be separated from their children and these can not be cared for in such institutions. Also most of the institutions are not suited for the care of young children. Here a specialized setting and specific experience with the cooperation of the various professions such as psychiatrists, psychiatric nurses, social workers, midwives, pedagogues, etc., is necessary. Specialized parent/child programmes can specifically train those skills which individuals with a psychiatric disease often do not have spontaneously or sufficiently, the so called ‘parenting skills’ (Hofecker Fallahpour and Riecher-Rössler 2002).

Specialized institutions with these specific settings and experience are being founded, especially in Anglo-American countries (for reviews see Oyserman et al. 1994; Seeman 2004). One of their main aims besides treatment and rehabilitation is the training of parenting. The parents’ skills are trained and improved in a special ‘parenting rehabilitation’ which includes an initial assessment followed by a long-term programme with parenting classes, support groups for parents and children, therapeutic kindergartens, etc. (Oyserman et al. 1994; Mowbray et al. 2001; Seeman 2004).

Schizophrenic disorders show a later age of onset in women and a slightly better course especially in young women. As to pathogenesis there is some evidence that the age difference might be at least partly due to the female sex hormone estradiol being a protective factor. Differences in course might also have to do with this biological factor, but at the same time with the psychosocial advantages of a higher age of onset and other psychosocial factors. Thus, gender differences in schizophrenia are obviously determined by biological as well as psychosocial factors and the complex interplay of both.

For the clinician this means that therapy has to be based on a multilevel approach that integrates biological, psychological, and social factors and takes into account the specific needs of women, including their needs in specific life circumstances such as motherhood.

For research, gender differences in schizophrenic psychoses could be a valuable paradigm regarding the interplay between biological and psychosocial factors. Understanding gender differences could help understanding the complex pathogenetic mechanisms of these disorders and give valuable clues to new therapies.