Mary V. Seeman, Department of Psychiatry, University of Toronto, Centre for Addiction and Mental Health, 250 College St., Toronto, Ontario, Canada M5T 1R8. E-mail: email@example.com
Seeman MV. Menstrual exacerbation of schizophrenia symptoms.
Objective: To better understand premenstrual exacerbations of schizophrenia in women and weigh treatment options.
Method: A PubMed literature search was conducted, using the search terms ‘schizophrenia’, ‘psychosis’, ‘menstrual exacerbation’, ‘hormones’ and assessing relevance to premenstrual exacerbation of schizophrenia symptoms.
Results: Exacerbations are usually distinguishable from periodic or menstrual psychosis, a relatively rare condition. Controversy continues about whether low estrogen periods of the month lead to an increase in schizophrenia symptoms among women of reproductive age or whether some women suffer from both schizophrenia and premenstrual dysphoric disorder (PMDD). No treatment trials of specific interventions have been conducted so that physicians must decide on a case-by-case basis whether to raise antipsychotic doses premenstrually, try estrogens or estrogen/progesterone combinations or selective estrogen receptor modulators, or target PMDD symptoms.
Conclusion: Clinicians need to be aware of premenstrual symptom aggravation in a large minority of women with schizophrenia. Treatment strategies will depend on the nature of the symptoms that are exacerbated. Optimal treatment needs to be adjusted to the individual woman.
• Clinicians treating women with schizophrenia need to inquire about menstrual exacerbation of symptoms.
• The nature of the symptoms that are premenstrually exacerbated will determine the correct intervention.
• Raising the dose of prolactin-raising antipsychotics can be counterproductive because estrogen levels can fall and ovulation can be inhibited.
• There are inherent safety risks to treatment with estrogens, but selective estrogen receptor modulators (SERMs) may prove to be effective.
• SERMs, like oral estrogens, can affect CYP enzymes that metabolize individual antipsychotics.
The following question was recently posted on an online psychiatric forum:
‘… Anyone know anything about the relationship between menstrual cycles and schizophrenia?’ (1). Thinking about this issue led to this review, whose aims are to address the following questions: When symptoms in women with schizophrenia become more severe perimenstrually, is it because hormonal fluxes aggravate psychotic symptoms or because the stress of premenstrual symptoms (anxiety, depression, and somatic problems) is superimposed on the pre-existing psychotic condition? What is the best treatment for menstrual exacerbations?
Many women patients diagnosed with schizophrenia perceive that their symptoms are worse prior to their menstrual periods. The following is an example from an anonymous online post:
..the week before [sometimes one in (sic) a half or two] I start my period I become much more paranoid, I have spacey feelings - like I’m far away looking in - much more often, I feel very anxious and angry, and I hear things others don’t much more often. Is this normal for someone who is schitzo (sic)? (2)
Hormonal effects in schizophrenia
The estrogen hypothesis (3, 4) of schizophrenia theorizes that endogenous estrogen shields women, to some extent, from the worst effects of the illness. Estrogens may also protect men (5), and progesterone, too, may be protective (6, 7). This explains why the menstrual cycle, during which these two hormones fluctuate, may induce changes in the severity of schizophrenia symptoms. Please see Fig. 1 for a diagram of monthly hormonal fluxes in reproductive age women. One recent theory is that dopamine receptors develop increased sensitivity to estrogens or progestins during the luteal phase of the menstrual cycle (8). Positron emission tomography studies of dopamine D2 receptors over menstrual phases have been conducted in monkeys and have shown that receptivity or sensitivity of these receptors in the striatum is almost 12% higher during the luteal phase than during the follicular phase (9). This could have direct bearing on schizophrenia symptoms caused by overactivity of dopamine pathways (10). Functional magnetic resonance imaging studies during different menstrual phases have suggested that estrogen also attenuates arousal pathways in women and modulates the stress response (11, 12). Estrogen could, thus, attenuate stress-induced and psychological trauma-induced psychotic symptoms (13). As well as dampening stress, estrogen may promote healing. Many animal studies have demonstrated that the endogenous estrogen, 17β-estradiol, is protective of the brain during injury (14). This may be pertinent to psychotic symptoms secondary to developmental injury (15).
Aims of the study
The aim of this study was to try to untangle premenstrual exacerbations of schizophrenia from periodic psychosis on the one hand and comorbidity of schizophrenia and premenstrual dysphoric disorder on the other. Treatment would be more effective if causes could be specified.
Material and methods
The present clinical overview looks at studies found by introducing the search terms: ‘schizophrenia’, ‘psychosis’, ‘menstrual exacerbation’, and ‘hormones’ into PubMed. This means that papers published in non-peer-reviewed journals were excluded. There were no language restrictions. All articles (original studies or review papers) were screened for pertinence to the issue of menstrual exacerbation of schizophrenia symptoms.
Early work on menses and symptom severity in schizophrenia
Work from the 1990s and early 2000s suggested that symptom severity fluctuated in women with schizophrenia, at least in some women with schizophrenia, over the menstrual month. The suggestion was that symptoms were more severe during low estrogen phases of the cycle (16–18), although, of the early papers, only Riecher-Rössler et al. (18) actually measured estrogen levels. Anxiety symptoms and depressive symptoms appeared to loom larger than psychotic symptoms, but these symptoms may have been accentuated because trial participants were being treated with antipsychotic drugs throughout the study period. In the work by Riecher-Rossler and colleagues (18), higher estradiol levels were closely associated with less severe positive symptoms of psychosis. A 4th study also found symptom changes over the menstrual cycle in women with schizophrenia, but the changes were not particular to any one phase of the cycle; the symptoms affected were, in the main, mood symptoms (19). The author of this study concluded that, in some women, premenstrual symptoms are superimposed, or added to, psychotic symptoms rather than being symptoms of the original illness aggravated by hormones. Rather than seeing this phenomenon as a periodic flare-up of psychotic symptoms, perhaps treatable with intermittently higher doses of antipsychotic medication, it was conceptualized as comorbidity – psychosis plus premenstrual dysphoria – optimally to be addressed as two separate conditions.
In a later study designed to investigate the effects of serum levels of estrogen, progesterone, testosterone, and cortisol on both psychopathology and neuropsychological functioning in a group of 37 schizophrenia patients (17 women and 20 men), high estrogen levels again correlated with low positive symptom scores (20). Thompson et al., however, found no difference in Positive and Negative Symptom Scale scores between the follicular and luteal phases of the menstrual cycle in 29 premenopausal women with psychosis, four of whom were unmedicated (21). Contradictory findings are not surprising because sample sizes in all these studies were relatively small. Moreover, the study of Thompson et al. included not only women diagnosed with schizophrenia but also women with related psychoses. The menstrual cycle phase was verified by hormone assays, which is the strength of this study. However, testing sessions were held only once in the follicular phase (days 1–14) and once in the luteal phase (days 14–28), 14 days apart, leading to a broad range of estradiol and progesterone levels. The correlations between symptoms and hormone levels were not reported. As in some of the previous studies, the largest effect sizes were for total and general symptoms and not for the primary positive/negative symptoms of schizophrenia.
Time of the month for women has been used as a proxy for estrogen level in many early studies, the participants often carrying a variety of diagnoses. What could be assumed to be low estrogen times of the month were positively associated with time of psychiatric admission (22, 23), suggesting a cause and effect. Gattaz et al. studied the hospital course of 65 women diagnosed with schizophrenia. Those admitted during phases of the menstrual cycle when estrogen levels were low required lower effective doses of antipsychotic medication compared with the group admitted during phases of high estrogen (P < 0.05). The authors concluded that, in all probability, progressively rising levels of estrogen were responsible for the superior therapeutic response of these women (24). In a replication study, 51 female patients with schizophrenia were again divided into two groups according to their reported menstrual cycle phase at admission. Approximately half were admitted in what would be their low estrogen phase, and on discharge, they were being prescribed lower daily antipsychotic doses than the other half (P < 0.001), but only if they were being treated with first-generation antipsychotics. The picture was less clear for those prescribed second-generation antipsychotic drugs (25). Sample sizes were still relatively small. Bergemann and colleagues investigated two larger samples of premenopausal women with schizophrenia (n = 115 and n = 170). In both, they found a significant increase in admissions during the perimenstrual or low estrogen phase – 3 days before and 3 days after the first day of the menses (26). The conclusion from these studies was that low estrogen times of the menstrual cycle precipitated hospital admission in women with psychotic illness (27).
A condition called menstrual psychosis may or may not be distinct from menstrual exacerbation of schizophrenia. The name refers to psychotic episodes recurring during each premenstrual/menstrual period, but remitting during the rest of the month. Brockington reviewed this category thoroughly in 2005 (28). Because menstrual psychosis is conceptualized as a relatively rare condition, the literature describing it is based entirely on case reports. Episodes of menstrual psychosis are the most prevalent around the time of menarche and the postpartum period, which has suggested an association with anovulatory cycles. The clinical picture resembles that of puerperal psychosis, and there is overlap between the two in that the same women sometimes suffer from both. Brockington gives the following description of menstrual psychosis: i) acute onset, against a background of normality; ii) brief duration, with full recovery; iii) psychotic features: confusion, stupor and mutism, delusions, hallucinations, or a manic syndrome; and iv) a circa-mensual (approximately monthly) periodicity, in rhythm with the menstrual cycle. Exacerbation of chronic mental illness is excluded from the definition of menstrual psychosis, but Brockington points out that phasic psychosis can metamorphose into continuous illness. Some cases are premenstrual – they start during the second half of the cycle and end with abrupt recovery at the onset of menstrual bleeding. There are catamenial cases where the psychosis begins with the onset of menstrual flow. There are also paramenstrual psychoses that recur on a monthly basis and start either before, during, or immediately after the end of menses. Mid-cycle psychoses that occur around the time of ovulation have been described in the literature, but relatively rarely. Also mentioned are epochal psychoses, a term used to describe bipolar psychoses where switches from mania to depression and back occur in synchrony with menstrual cycles. Brockington describes possible treatments under three headings: i) hormones, ii) suppression of the menstrual cycle, and iii) miscellaneous therapeutic agents. No clinical trials have been conducted in menstrual psychosis; all reports are of individual cases or small case series (28). Effective treatment for menstrual psychosis could also prove useful for menstrual exacerbations of schizophrenia.
Psychosis and estrogen withdrawal
From a systematic review of the literature for the period 1960–2000, Mahe and Dumaine identified 26 reports of an association between a psychotic disorder and a phase of estrogen withdrawal (29). This association may be exaggerated in schizophrenia because antipsychotic drugs used to treat schizophrenia raise prolactin levels and, thereby, lower estrogen production (30). Work in the area of schizophrenia and estrogen up to 2004, including relevant preclinical animal studies, has been reviewed by Lindamer et al. (31). Several limitations to these earlier studies have been pointed out. The main issues are that clinical samples were generally small and menstrual cycle phases were sometimes ascertained by self-report. When hormone levels were done, they did not necessarily correlate with symptom scores.
Schizophrenia and menstruation – more recent studies
Bergemann et al.’s 2007 study of changes in symptoms across the menstrual cycle included 125 premenopausal women with acute schizophrenia, all being treated with antipsychotics. Hormone levels were assessed in the follicular, peri-ovulatory, and luteal phases of the menstrual cycle. Significant improvement in psychotic, but not depressive, symptoms (as measured by scores on the Positive and Negative Syndrome Scale and the Brief Psychiatric Rating Scale) was observed during the luteal (high estrogen) phase, compared with the other menstrual phases (32). Unexpectedly, no difference was found between the follicular and the peri-ovulatory phase, suggesting the probable contribution of progesterone to symptom improvement.
Rubin et al. (2010) also found that positive symptoms improved in 23 women with chronic stable schizophrenia during the mid-luteal (days 20–22; high estrogen/progesterone) vs. the early follicular phase (days 2–4; low estrogen/progesterone) (P < 0.01). Peripheral estrogen/progesterone levels did not account for the phase changes, but the authors point out that peripheral measures are an imperfect index of CNS hormone levels. In addition, this sample included a high proportion of Afro-Americans in a relatively chronic and stable stage of illness so that their psychopathology scores were low (33). The aim of the Rubin et al.’s study was to investigate the effects of menstrual phase and hormone levels on clinical symptom severity in women with chronic schizophrenia with a focus on estrogen, progesterone, and also oxytocin, which may have antipsychotic properties (34). While oxytocin levels did not fluctuate across menstrual phases in the Rubin et al.’s study (33), higher oxytocin levels in this sample were associated with less severe positive symptoms and less severe overall psychopathology (P < 0.01).
Menstrual effects in illness other than schizophrenia
Menstrual exacerbations have been reported in other psychiatric illnesses (35–39), in pain syndromes (40–42), and in a variety of medical and neurological illnesses (43). This strongly suggests that illness exacerbation in the perimenstrual period is not restricted to schizophrenia or to psychiatric illnesses. It also suggests that some women, and not others, are susceptible to a hormonal effect that can exacerbate whatever illness the woman suffers (43, 44).
Mechanisms of menstrual exacerbations of illness in schizophrenia
Hormone effects on brain morphology. Reflective of functional change, the human brain may change in size over the menstrual cycle (45). Using MR volumetry in a small sample, Hagemann et al. found a significant gray matter volume peak of about 13.5 ml in women, corresponding to the time of ovulation. This peak did not correlate with estradiol levels for the whole group of eight women examined, but, once a statistical outlier was removed from the analysis, a correlation became evident. There was an associated CSF volume loss after ovulation, which, while not statistically significant in itself, correlated with progesterone levels. The eight men who were also examined did not show any significant brain volume alterations. Although the sample was small, the investigators conclude that their findings support a neuroprotective role of estrogens in the human brain analogous to that previously found in rodent brain.
Hormone effects on symptoms. The effects of estrogens on mental health and particularly on schizophrenia have recently been thoroughly reviewed (46, 47). Markham concludes, after her extensive review of the literature: ‘In summary, it is clear that changing estrogen levels in adult women – whether this is across the natural cycle or in the context of hormone therapy – can influence schizophrenia symptom severity’ (47).
What is true for symptoms may also be true for cognitions. Ko et al. (48) studied 35 relatively stable women with chronic schizophrenia and divided them into two subgroups (low estradiol group and normal estradiol group). Patients in the low estradiol group had a significantly lower performance in most of the cognitive function tests administered to them than patients in the normal estradiol group.
From an epidemiological view-point, it has been shown repeatedly that late onset schizophrenia, starting at an age when brain estrogen levels are lower in women than they are in men because testosterone is partly converted to estrogen (See Fig. 2 on lifetime distribution of estrogen and testosterone), is an illness mainly of women (49). Moreover, postpartum psychosis can be conceptualized as an estrogen withdrawal phenomenon (50). Early menarche (early rise in estrogen level) was thought to delay the onset of schizophrenia in girls (51), but this has been disproven in several studies, including the recent cohort studies of Welham et al. (52) and Boden et al. (53).
Given clinical and epidemiological indications that estrogen withdrawal appears to initiate or exacerbate symptoms, Kulkarni (54) began using estrogen as a potential therapeutic agent for schizophrenia. Several randomized double-blind placebo-controlled trials by this group and an open-label study have shown that adjunctive estradiol delivered by transdermal patch is significantly associated with a reduction of symptoms (55). Akhonzadeh et al. (56) confirmed that oral ethinyl estradiol added to haloperidol for 8 weeks was superior to haloperidol alone for general psychopathology symptoms as well as for positive and negative symptoms of schizophrenia. Louza et al. (57) used oral conjugated estrogens (0.625 mg/day) added to 5 mg daily haloperidol for 4 weeks and were unable to show a statistically significant effect. Bergemann et al. (58) were also unable to show a difference between estradiol and placebo added to standard antipsychotic medication. A 2005 Cochrane review (59) on estrogen for the treatment of schizophrenia concluded that adjunctive estrogen with or without progesterone did not appear to offer convincing advantages over placebo. That review was published before the 2008 paper by the Kulkarni group (60). After recruiting from both in-patient acute hospital wards and out-patient clinics of two Australian general hospitals, 102 reproductive age women diagnosed with schizophrenia (acute or chronic) received 100 μg of transdermal estradiol or transdermal placebo for 4 weeks. The transdermal route precluded the possibility that effects were being mediated by altered blood levels of concomitant antipsychotics. The intervention group showed significantly reduced positive (P < 0.05) and general psychopathology symptoms (P < 0.05) relative to women receiving antipsychotic medication alone (60). Discrepancies across studies can be explained by pre-, peri-, and postmenopausal status differences among participants and the specifics of estrogen treatment, as well as its duration. Estradiol adjunctive treatment appears to be superior to conjugated estrogens (Wroolie et al., 61). It is still not known how different estrogens and their metabolites, at various doses and schedules, with or without progestins, interact with estrogen receptors in different tissues to produce variable results (46).
Effect of menses on treatment
Today, most women suffering from schizophrenia are being treated with antipsychotic drugs. It has been known for several decades that menstrual cycle hormonal changes affect drug absorption, distribution, metabolism, and excretion (62). The expression of some, though not all, metabolizing enzymes is influenced by estrogen and progesterone levels and may therefore affect the clearance of some antipsychotic drugs. Their levels in the bloodstream would, thus, differ at different phases of the menstrual cycle, and dose adjustments may be needed over the menstrual month for optimal management of symptoms. Pharmacotherapy for women with schizophrenia has been reviewed by Usall et al. (63).
Assessment of menstrual exacerbations
Because most women of reproductive age, regardless of diagnosis, usually have some complaints relative to their menstrual cycle, it is important for physicians to routinely ask all their women patients about their menstrual experience. Effective treatment depends on hearing what the patient is experiencing (64). Suggesting that women keep a menstrual diary has proven effective in the assessment of premenstrual migraine (65) and could prove equally useful in the assessment of premenstrual exacerbation of symptoms in schizophrenia. It is important to keep in mind that symptom expression may, temporarily, increase in response to daily self-monitoring (66).
Treatment of Menstrual Exacerbations in Schizophrenia
Estrogen/progesterone. Although studies cited above suggest that estrogen augmentation can be beneficial in schizophrenia, treatment with estrogen is controversial and is not currently approved except on an experimental basis. There is a recent case report of combination oral contraceptives successfully preventing the recurrence of psychotic symptoms in premenstrual psychosis (67). Using low dose birth control pills is a safe initial intervention. Estrogen supplementation in the form of daily oral conjugated estrogens (Premarin 0.625) was tried for 3 months on four women with schizophrenia who experienced premenstrual exacerbation of symptoms. Attenuation of the monthly rise in symptoms was achieved for two of four cases (68). If effective, estrogens can be given transdermally via a patch or by subcutaneous implant. Important to remember is that estrogen and progesterone given orally interact with metabolizing enzymes so that their use can, for instance, boost clozapine and chlorpromazine blood levels – an effect that may, in itself, be helpful for the exacerbation of psychotic symptoms (69, 70).
Selective estrogen receptor modulators. Selective estrogen receptor modulators (SERMs) that are safe for both breast and uterus and that pass through the blood brain barrier and mimic estrogen’s action in the brain (71) can also be tried. One SERM, raloxifene, has shown some promise as an adjunctive treatment for postmenopausal women with psychosis even though only a small per cent of the oral dose of this compound enters the brain (72, 73). Kulkarni (74) has conducted a dose-finding study for the use of raloxifene in postmenopausal women in schizophrenia. Usall et al. (75) treated 33 postmenopausal women with schizophrenia with either adjunctive raloxifene 60 mg/day or placebo for 12 weeks. Raloxifene added to regular antipsychotic treatment significantly reduced negative, positive, and general psychopathology symptoms when compared to placebo. As with estrogen, and depending on the SERM and on the antipsychotic drug, antipsychotic blood levels may be raised by the interaction (76, 77), thus making the attribution of efficacy more difficult.
Suppression of ovulation
Another way of treating menstrual exacerbations of symptoms is to interrupt the menstrual cycle. Suppression of ovulation using gonadotropin-releasing hormone (GnRH) agonists works well for premenstrual dysphoric disorder, but ‘add-back’ hormone therapy with estrogen or estrogen/progesterone is required. GnRH agonists have been successfully combined with tibolone, a synthetic estrogen, progesterone, and androgen receptor agonist (78). There have been no trials of these interventions for premenstrual exacerbations of schizophrenia. An easier way of suppressing ovulation in schizophrenia is to treat with antipsychotic drugs that raise prolactin levels sufficiently to block ovulation. This may have been done commonly for premenstrual exacerbations, although perhaps not deliberately, when first-generation antipsychotics were the gold standard for the treatment of schizophrenia. High prolactin levels, however, lead to an unwanted drop in estrogen level, which might prove counterproductive in the longer run.
Intermittent raising of antipsychotic dose
Intermittently raising antipsychotic doses premenstrually to prevent or treat increasing symptoms is yet another strategy. Hsiao and Liu describe the successful intermittent use of the antipsychotic, sulpiride, on premenstrually symptomatic days, (200 and 400 mg) in two women with periodic psychoses (79). Sulpiride, however, is a potent releaser of prolactin (80) and is not the best drug to use for this purpose as hyperprolactinemia will, as mentioned earlier, lower estrogen levels.
Treating exacerbation as premenstrual dysphoric disorder
A fifth strategy is to treat the increase in symptoms separately from schizophrenia, as one would treat premenstrual dysphoric disorder (PMDD) in any woman who presented with recurrent symptoms at that time of the month. A recent review summarizes the current therapeutic management of PMDD, recommending diet, exercise, and selective serotonin receptor inhibitors (78). This is especially indicated when the symptoms seen in the premenstrual phase are mainly mood symptoms. Adding SSRIs intermittently in the late luteal phase or daily throughout the menstrual cycle is a standard treatment of PMDD, but there have been no trials evaluating this approach in schizophrenia patients. SSRIs have shown an ability to alleviate depressive symptoms in schizophrenia patients over 40 years of age (81), but recent schizophrenia PORT guidelines indicate that there is insufficient evidence to make recommendations for or against treating depression in schizophrenia with SSRIs. (82).
It is difficult to glean from the literature how frequently schizophrenia symptoms are exacerbated during the premenstrual period in women. Hsiao et al. (83) set out to establish frequencies of both PMDD and premenstrual exacerbation of a number of psychiatric disorders in a Chinese population. They reported that 52% of 50 women with schizophrenia had PMDD, while 20% experienced premenstrual exacerbation. Dias et al. (35) reported that over 65% of 191 women with a related illness, bipolar disorder, showed premenstrual exacerbation of their symptoms. Shivakumar et al. (84), however, found only 5/41 bipolar women with a mania score that was higher in the luteal phase than in the follicular phase and 8/41 whose depression score was higher in the luteal phase, e.g., 32% with exacerbation. Because it is difficult to distinguish psychosis with PMDD from psychosis with premenstrual exacerbation, from periodic menstrual psychosis, a conservative estimate would be that a large minority of women suffering from schizophrenia probably experience an increase in symptoms during the premenstrual period. It is not clear what makes some women more susceptible than others to such symptom increases. Perhaps, as in menstrual psychosis, these are women who are also vulnerable to postpartum psychosis because of increased sensitivity to hormonal fluxes. Deuchar and Brockington (85) postulate that women with anovulatory menstrual cycles, with high levels of relatively unopposed estrogens that prime the central nervous system prior to a premenstrual estrogen cascade, may be at special risk. This is analogous to what occurs in puerperal psychosis where an estrogen cascade follows a lengthy period of sustained high brain estrogen. If this is the case, antipsychotic medication that raises prolactin levels and inhibits ovulation should be avoided.
It is important for clinicians to ask women patients with schizophrenia, which, if any, of their symptoms worsen prior to menses. If it is depression and anxiety, then the right intervention will likely be diet, exercise, and selective serotonin receptor inhibitors. If the symptoms that worsen premenstrually are psychotic symptoms, boosting the antipsychotic dose of prolactin-sparing drugs for 3–5 days prior to menstruation may be effective. Treatment with estrogen, estrogen/progesterone combinations, or raloxifene can be tried, keeping in mind that antipsychotic levels may be affected unless the medication bypasses the liver. The take-home message of this review is that health care providers treating women with schizophrenia need to be alert to menstrual exacerbation of symptoms because the phenomenon remains understudied.