Mary V. Seeman, MD

Dr. I, a 48-year-old university professor, is brought to the ER by her husband because she has developed an irrational fear of being chased by Nazis. The fears have become increasingly bizarre, her husband reports. She believes her Nazi persecutors are bandaging their arms and using wheelchairs to pretend to be disabled. When out with her husband, Dr. I points to people in wheelchairs, convinced they are after her, will kill her, and are incensed because she left Germany—her country of birth. Her hus­band brought her to the ER when she started to hear her persecutors addressing her in German at night.

Psychoses of unknown cause usually begin in late ado­lescence or early adulthood. Less frequently the onset occurs in later adulthood (age ≥40). Late-onset psycho­sis is much more prevalent in women than in men for reasons that are imperfectly understood.

When you are evaluating a midlife woman with first onset of psychosis, don’t assume an illness of unknown cause (bipolar disorder or schizophrenia) until after you have done a comprehensive search for triggers of her psychotic symptoms. After age 40, women are more likely than men to develop psychosis because of gender-specific medical and psychological precipitants.

Predisposing factors for psychosis
Psychosis is an emergent quality of structural and chemical changes in the brain. As such, it can be ex­pected to surface during:
   • brain reorganization or transition (adolescence, senescence, brain trauma, stroke, starvation, inflammation, or brain tumor)
   • change in brain chemistry (flux in go­nadal, thyroid, or adrenal hormone levels; electrolyte imbalance; fever; exposure to chemical substances; immune response).

Psychological stress impacting the brain via stress hormones also can predispose a person to psychosis.

Because some individuals are more prone than others to develop psycho­sis during brain alteration, chemical and structural changes in the brain are as­sumed to interact with genetic propensi­ties to influence gene expression. Once a psychotic event has occurred, it is thought to sensitize the brain so that subsequent events emerge more readily.¹

Schizophrenia—though not the only ill­ness in which psychosis plays a role—is a prototype for psychotic illness, and sev­eral reported sex differences in this dis­order are worth noting.² The incidence of schizophrenia is approximately the same in both sexes, but women show a later age of onset—a paradox in that the brain de­velops at a faster pace in females and theo­retically should reach the threshold for the first appearance of schizophrenia earlier. Women also require lower doses of anti­psychotic medication to recover from an acute psychotic episode and to maintain remission, at least before menopause.^3,4 Both of these differences can be explained as an effect of estrogen on a) gene expres­sion⁵ and b) liver enzymes that metabolize antipsychotics.⁶

The estrogen hypothesis. Women show a tendency toward premenstrual and post­partum exacerbation of symptoms when estrogen levels are relatively low. These clinical observations, confirmed by some but not all studies, have led to the hypoth­esis that estrogens are neuroprotective⁷ and also protect against psychosis.⁸

Estrogen withdrawal in specific brain cells may release a cascade of events that over time can increase the severity of psy­chotic and cognitive symptoms. The reason for suspecting such effects is based on what we know about estrogenic effects on neurotransmitter, cognitive, and stress-induction pathways, and—more fundamentally—on neuronal growth and atrophy.

According to the estrogen hypothesis, women are—to some degree—protected against schizophrenia by their relatively high gonadal estrogen production between puberty and and menopause. Women lose this protection with the onset of perimeno­pausal estrogen fluctuation and decline, accounting for their second peak of illness onset after age 45.

Epidemiologic studies showing a second peak of schizophrenia onset in women (but not men) around the age of menopause sup­port this hypothesis.^9,10 Longitudinal out­comes for schizophrenia—which are better in women than in men during late adolescence or early adulthood11—gradually even out after the first 15 years of illness, suggest­ing that women’s advantage is lost at a time approximating menopause (Box 1).

The question, then, becomes: Is it only because of estrogen loss after age 40 that women become more prone to develop a psychotic illness? Other differences be­tween the sexes that may play roles include immune function, low iron stores, sleep suf­ficiency, thyroid function, exposure to toxic substances (including therapeutic drugs), societal pressures to be slim while aging (Table), and the experience of stress.¹²

CASE CONTINUED
Exhausted and confused
Dr. I is a well-groomed, handsome woman, but she hardly speaks when interviewed, looking frightened and somewhat bewildered. She has never had a mental health problem, nor has anyone in her family. She agrees to stay in the hospital but is not sure why. She has slept no more than 1 or 2 hours in the last several days.

Her early history is unremarkable. She did well in school. After earning a PhD at the Univer­sity of Leipzig, she and her husband immigrated to Canada. Both are university professors. They never decided not to have children, but children hadn’t come. Her menstrual periods stopped 2 years before admission. The question about children is the only 1 that elicits emotion in Dr. I. When I ask about it, tears come to her eyes as she shakes her head.

Her husband reports that she has not been eating well and has, in the last year, started to drink more alcohol than usual—3 to 4 drinks of whiskey a night. She does not smoke cigarettes, and her health generally is good. She uses no medications. Her hus­band describes their marital relationship as very close, although it has become strained in recent weeks because of her unreasonable fears. He admits that their work is always stressful; competition is fierce, with more and more deadlines and less and less leisure time. The couple has few friends and no hobbies.

Late-onset psychosis symptoms
In late-onset psychosis (after age 45), men appear to suffer substantially milder symp­toms and spend less time hospitalized than women.¹³ Women with late-onset schizophrenia have more severe positive symptoms than men and fewer negative symptoms.^14,15 Overall, patients with late-onset schizophrenia have a lower prevalence of looseness of associations and negative symptoms than those with earlier onset.^16,17

In addition, individuals with schizo­phrenia who become ill in middle age have been reported to:
   • show better neuropsychological per­formance (particularly in learning and ab­straction/cognitive flexibility) than those with early onset
   • possibly have larger thalamic volumes
   • respond to lower antipsychotic doses.¹⁸

Auditory and visual hallucinations frequently are observed in patients with comorbid late-onset schizophrenia and auditory and visual impairment.¹⁶ Palmer et al¹⁸ reported no difference in family his­tory of schizophrenia between early and late onset, but this is controversial. Con­vert et al¹⁶ note that most studies reveal a lower lifetime risk of schizophrenia in first-degree relatives of patients with late-onset than early-onset schizophrenia.

CASE CONTINUED
Medical workup
Dr. I’s physical exam is unremarkable. Her thy­roid is not enlarged; there are no breast lumps. On mental status exam, her mood is flat. She is preoccupied with fears of the Nazis. Routine blood tests show slight anemia; fasting glu­cose levels are within normal range.

I give Dr. I zopiclone, 7.5 mg, to help her sleep. The next day she keeps to herself, eats very little, and appears disinterested in her surroundings. Nursing staff report that she often seems frightened. Dr. I asks to use the ward phone to call Germany but is told that she cannot make long distance calls from that phone. This seems to disturb her.

Differential diagnosis
Sensory impairment, substance abuse, and metabolic changes have been implicated in the appearance of psychosis in later life. More specific to women than men, howev­er, are medical and psychiatric precipitants. These include autoimmune disease (and its treatment) and psychiatric disorders, as well as thyroid dysfunction, self-induced starvation (anorexia nervosa) and diet aids, substance use and abuse, insomnia, and iron deficiency (Table).

Autoimmune disease and treatment. Nearly 80% of patients with autoimmune disease are women, and these disorders (as well as their treatment) can manifest as psychosis. Corticosteroids have a well-documented history of triggering psy­chotic symptoms, which are twice as likely in women than in men. The incidence of severe psychosis while taking oral predni­sone ranges from 1.6% to 50% and averages 5.7%. The average daily dose of corticoste­roids for patients who develop psychosis is 59.5 mg/d.

Corticosteroid creams absorbed through skin as well as inhaled and intranasal cor­ticosteroids in their more potent formula­tions can have systemic effects, including psychosis. Nonsteroidal anti-inflammatory drugs such as ibuprofen also can trigger psychosis.¹⁹

Psychiatric disorders. Posttraumatic stress disorder with psychotic symptoms may overlap with categories such as psy­chogenic psychoses, hysterical psychoses, nonaffective remitting psychoses, acute brief psychoses, reactive psychoses, acute and transient psychoses, and bouffées dé­lirantes (in France, the name for transient psychotic reactions).²⁰ Consider these female-predominant conditions in the differential diagnosis, along with micro­psychotic episodes in borderline personal­ity disorder, in which the predominance of women is 3:1.

Medical treatment for depression and anxiety also can lead to psychotic symp­toms through individual susceptibility to the action of specific drugs or through withdrawal effects.

Clinical assessment
Question all women presenting with psy­chosis about eating habits and diet pills, and check for hypokalemia and hypocal­cemia to rule out starvation effects and reactions to stimulants. Also ask about in­halants, and examine for anemia and thy­roid dysfunction. Consider all medications as having the potential to trigger psychotic symptoms.

A family history of illness is important, with a focus on autoimmune disorder and its treatment. A thorough psychiatric history is crucial and needs to include assess­ment of sleep, mood, and relationships with attachment figures. Do not assume ill­nesses of unknown cause (bipolar disorder or schizophrenia) until after a comprehen­sive search for precipitants of psychotic symptoms.

CASE CONTINUED
Guilty feelings
To address her delusions, I start Dr. I on ris­peridone, 2 mg at bedtime. She goes home for the weekend, and her husband reports that she slept throughout the visit. When she returns, she spends a lot of time in bed but is more communicative.

When I ask Dr. I whether she has called Germany, she says she called her recently widowed father. Dr. I begins to cry when talk­ing of her mother, and tells the nurse she feels guilty for not visiting for the last few years. When her mother died 6 months ago, Dr. I had not seen her in 4 years.

Her fears remit with risperidone, main­tained at 2 mg/d, but Dr. I remains depressed and responds slowly to treatment with cital­opram, 20 mg/d, and supportive therapy. Her final diagnosis is mood disorder with psy­chotic features.

Treatment
When treating women with late-onset psy­chosis, remove all potential triggers and address underlying illness. Cognitive ther­apy targeting specific symptoms is useful; antipsychotics probably will be necessary. Age-related physiologic changes make older persons more sensitive to the thera­peutic and toxic effects of antipsychotics.

Estrogen therapy? Women suffering from schizophrenia show significantly lower es­trogen levels than the general population of women, and they experience first-onset or recurrence of a psychotic episode signif­icantly more often in low estrogen phases of the cycle. Estrogens have therefore been postulated to constitute a protective factor against psychosis, which means perimeno­pause is an at-risk period.²¹ Although evi­dence is limited, preliminary studies have found beneficial effects from short-term, off-label use of estrogen therapy in women with psychotic illness (Box 2).

Because continuous use of estrogen plus progestin has been associated with an in­creased risk of adverse effects,²² off-label use of selective estrogen receptor modula­tors (SERMs) also is being investigated in women with schizophrenia. SERMs act as tissue-specific estrogen agonists and an­tagonists because they can either inhibit or enhance estrogen-induced activation of estrogen response element-containing genes.²³

Wong et al²⁴ used a crossover design to compare the SERM raloxifene with place­bo as adjunctive treatment for 6 postmeno­pausal women with schizophrenia. Each woman received 8 weeks of raloxifene, 60 mg/d, and 8 weeks of placebo. Three be­gan with placebo and 3 with raloxifene.

Verbal memory was measured weekly with the California Verbal Learning Test, using 5 memory trials, free and cued short-delay recall, and long-delay recall. At baseline, the participants had lower scores than older adults in the general popula­tion. Eight weeks of placebo improved scores somewhat, suggesting a practice effect. Eight weeks of raloxifene improved cognitive scores to a level similar to that of schizophrenia-free subjects. After 16 weeks, however, cognitive scores in the 2 groups were indistinguishable.

At present I do not recommend estrogen for women with late-onset schizophrenia because the risk is too high and raloxifene does not enter the brain sufficiently to be a valuable cognitive enhancer. Novel SERMs with more specific efficacy for improving cognitive function may prove useful in the future,²⁵ however, as may phytoestro­gens. Adjunctive hormone modulation is a promising area of gender-specific treat­ment for serious mental illness.²⁶

CASE CONCLUSION
Gradually improving
Dr. I’s depression was triggered by her moth­er’s death and regrets about not visiting and not being a mother. The content of her de­lusions was related to her guilt about not having returned to Germany; the delusions were probably triggered by depression, al­cohol intake, her relative hypoestrogenic state, stress at work, lack of social supports, and dependence on her husband.

Over the next few years, Dr. I is maintained on a low dose of risperidone (reduced from 2 mg/d to 1 mg/d) and citalopram (reduced from 20 mg/d to 10 mg/d). She becomes increas­ingly engaged in supportive dynamic therapy, and her symptoms gradually improve.

BOTTOM LINE
Psychosis onset in midlife is mostly a female phenomenon because a perimenopausal estrogen decline increases women’s susceptibility. Seek specific triggers such as medical illness or response to a drug before assuming an illness of unknown cause such as bipolar disorder or schizophrenia. Cognitive therapy targeting specific symptoms is useful; antipsychotics probably will be necessary.

Related Resources
• Women and psychosis: A guide for women and their families. Centre for Addiction and Mental Health. University of Toronto. www.camh.net/About_Addiction_Mental_ Health/Mental_Health_Information/Women_Psychosis.
• Seeman MV. Women and psychosis. www.medscape.com/ viewarticle/408912.
• Chattopadhyay S. Estrogen and schizophrenia: Any link? The Internet Journal of Mental Health. 2004;2(1). www.ispub. com/journal/the_internet_journal_of_mental_health.html.

Drug Brand Names
Citalopram • Celexa                         Prednisone • Deltasone,
Estradiol • Estrace,                                    Orasone, others
   Estrofem, others                           Raloxifene • Evista
Estradiol transdermal •                     Risperidone • Risperdal
   Estraderm , Climara, others
Methylphenidate • Concerta,
   Ritalin, others

Disclosure
Dr. Seeman reports no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Dr. I, a 48-year-old university professor, is brought to the ER by her husband because she has developed an irrational fear of being chased by Nazis. The fears have become increasingly bizarre, her husband reports. She believes her Nazi persecutors are bandaging their arms and using wheelchairs to pretend to be disabled. When out with her husband, Dr. I points to people in wheelchairs, convinced they are after her, will kill her, and are incensed because she left Germany—her country of birth. Her hus­band brought her to the ER when she started to hear her persecutors addressing her in German at night.

Psychoses of unknown cause usually begin in late ado­lescence or early adulthood. Less frequently the onset occurs in later adulthood (age ≥40). Late-onset psycho­sis is much more prevalent in women than in men for reasons that are imperfectly understood.

When you are evaluating a midlife woman with first onset of psychosis, don’t assume an illness of unknown cause (bipolar disorder or schizophrenia) until after you have done a comprehensive search for triggers of her psychotic symptoms. After age 40, women are more likely than men to develop psychosis because of gender-specific medical and psychological precipitants.

Predisposing factors for psychosis
Psychosis is an emergent quality of structural and chemical changes in the brain. As such, it can be ex­pected to surface during:
   • brain reorganization or transition (adolescence, senescence, brain trauma, stroke, starvation, inflammation, or brain tumor)
   • change in brain chemistry (flux in go­nadal, thyroid, or adrenal hormone levels; electrolyte imbalance; fever; exposure to chemical substances; immune response).

Psychological stress impacting the brain via stress hormones also can predispose a person to psychosis.

Because some individuals are more prone than others to develop psycho­sis during brain alteration, chemical and structural changes in the brain are as­sumed to interact with genetic propensi­ties to influence gene expression. Once a psychotic event has occurred, it is thought to sensitize the brain so that subsequent events emerge more readily.¹

Schizophrenia—though not the only ill­ness in which psychosis plays a role—is a prototype for psychotic illness, and sev­eral reported sex differences in this dis­order are worth noting.² The incidence of schizophrenia is approximately the same in both sexes, but women show a later age of onset—a paradox in that the brain de­velops at a faster pace in females and theo­retically should reach the threshold for the first appearance of schizophrenia earlier. Women also require lower doses of anti­psychotic medication to recover from an acute psychotic episode and to maintain remission, at least before menopause.^3,4 Both of these differences can be explained as an effect of estrogen on a) gene expres­sion⁵ and b) liver enzymes that metabolize antipsychotics.⁶

The estrogen hypothesis. Women show a tendency toward premenstrual and post­partum exacerbation of symptoms when estrogen levels are relatively low. These clinical observations, confirmed by some but not all studies, have led to the hypoth­esis that estrogens are neuroprotective⁷ and also protect against psychosis.⁸

Estrogen withdrawal in specific brain cells may release a cascade of events that over time can increase the severity of psy­chotic and cognitive symptoms. The reason for suspecting such effects is based on what we know about estrogenic effects on neurotransmitter, cognitive, and stress-induction pathways, and—more fundamentally—on neuronal growth and atrophy.

According to the estrogen hypothesis, women are—to some degree—protected against schizophrenia by their relatively high gonadal estrogen production between puberty and and menopause. Women lose this protection with the onset of perimeno­pausal estrogen fluctuation and decline, accounting for their second peak of illness onset after age 45.

Epidemiologic studies showing a second peak of schizophrenia onset in women (but not men) around the age of menopause sup­port this hypothesis.^9,10 Longitudinal out­comes for schizophrenia—which are better in women than in men during late adolescence or early adulthood11—gradually even out after the first 15 years of illness, suggest­ing that women’s advantage is lost at a time approximating menopause (Box 1).

The question, then, becomes: Is it only because of estrogen loss after age 40 that women become more prone to develop a psychotic illness? Other differences be­tween the sexes that may play roles include immune function, low iron stores, sleep suf­ficiency, thyroid function, exposure to toxic substances (including therapeutic drugs), societal pressures to be slim while aging (Table), and the experience of stress.¹²

CASE CONTINUED
Exhausted and confused
Dr. I is a well-groomed, handsome woman, but she hardly speaks when interviewed, looking frightened and somewhat bewildered. She has never had a mental health problem, nor has anyone in her family. She agrees to stay in the hospital but is not sure why. She has slept no more than 1 or 2 hours in the last several days.

Her early history is unremarkable. She did well in school. After earning a PhD at the Univer­sity of Leipzig, she and her husband immigrated to Canada. Both are university professors. They never decided not to have children, but children hadn’t come. Her menstrual periods stopped 2 years before admission. The question about children is the only 1 that elicits emotion in Dr. I. When I ask about it, tears come to her eyes as she shakes her head.

Her husband reports that she has not been eating well and has, in the last year, started to drink more alcohol than usual—3 to 4 drinks of whiskey a night. She does not smoke cigarettes, and her health generally is good. She uses no medications. Her hus­band describes their marital relationship as very close, although it has become strained in recent weeks because of her unreasonable fears. He admits that their work is always stressful; competition is fierce, with more and more deadlines and less and less leisure time. The couple has few friends and no hobbies.

Late-onset psychosis symptoms
In late-onset psychosis (after age 45), men appear to suffer substantially milder symp­toms and spend less time hospitalized than women.¹³ Women with late-onset schizophrenia have more severe positive symptoms than men and fewer negative symptoms.^14,15 Overall, patients with late-onset schizophrenia have a lower prevalence of looseness of associations and negative symptoms than those with earlier onset.^16,17

In addition, individuals with schizo­phrenia who become ill in middle age have been reported to:
   • show better neuropsychological per­formance (particularly in learning and ab­straction/cognitive flexibility) than those with early onset
   • possibly have larger thalamic volumes
   • respond to lower antipsychotic doses.¹⁸

Auditory and visual hallucinations frequently are observed in patients with comorbid late-onset schizophrenia and auditory and visual impairment.¹⁶ Palmer et al¹⁸ reported no difference in family his­tory of schizophrenia between early and late onset, but this is controversial. Con­vert et al¹⁶ note that most studies reveal a lower lifetime risk of schizophrenia in first-degree relatives of patients with late-onset than early-onset schizophrenia.

CASE CONTINUED
Medical workup
Dr. I’s physical exam is unremarkable. Her thy­roid is not enlarged; there are no breast lumps. On mental status exam, her mood is flat. She is preoccupied with fears of the Nazis. Routine blood tests show slight anemia; fasting glu­cose levels are within normal range.

I give Dr. I zopiclone, 7.5 mg, to help her sleep. The next day she keeps to herself, eats very little, and appears disinterested in her surroundings. Nursing staff report that she often seems frightened. Dr. I asks to use the ward phone to call Germany but is told that she cannot make long distance calls from that phone. This seems to disturb her.

Differential diagnosis
Sensory impairment, substance abuse, and metabolic changes have been implicated in the appearance of psychosis in later life. More specific to women than men, howev­er, are medical and psychiatric precipitants. These include autoimmune disease (and its treatment) and psychiatric disorders, as well as thyroid dysfunction, self-induced starvation (anorexia nervosa) and diet aids, substance use and abuse, insomnia, and iron deficiency (Table).

Autoimmune disease and treatment. Nearly 80% of patients with autoimmune disease are women, and these disorders (as well as their treatment) can manifest as psychosis. Corticosteroids have a well-documented history of triggering psy­chotic symptoms, which are twice as likely in women than in men. The incidence of severe psychosis while taking oral predni­sone ranges from 1.6% to 50% and averages 5.7%. The average daily dose of corticoste­roids for patients who develop psychosis is 59.5 mg/d.

Corticosteroid creams absorbed through skin as well as inhaled and intranasal cor­ticosteroids in their more potent formula­tions can have systemic effects, including psychosis. Nonsteroidal anti-inflammatory drugs such as ibuprofen also can trigger psychosis.¹⁹

Psychiatric disorders. Posttraumatic stress disorder with psychotic symptoms may overlap with categories such as psy­chogenic psychoses, hysterical psychoses, nonaffective remitting psychoses, acute brief psychoses, reactive psychoses, acute and transient psychoses, and bouffées dé­lirantes (in France, the name for transient psychotic reactions).²⁰ Consider these female-predominant conditions in the differential diagnosis, along with micro­psychotic episodes in borderline personal­ity disorder, in which the predominance of women is 3:1.

Medical treatment for depression and anxiety also can lead to psychotic symp­toms through individual susceptibility to the action of specific drugs or through withdrawal effects.

Clinical assessment
Question all women presenting with psy­chosis about eating habits and diet pills, and check for hypokalemia and hypocal­cemia to rule out starvation effects and reactions to stimulants. Also ask about in­halants, and examine for anemia and thy­roid dysfunction. Consider all medications as having the potential to trigger psychotic symptoms.

A family history of illness is important, with a focus on autoimmune disorder and its treatment. A thorough psychiatric history is crucial and needs to include assess­ment of sleep, mood, and relationships with attachment figures. Do not assume ill­nesses of unknown cause (bipolar disorder or schizophrenia) until after a comprehen­sive search for precipitants of psychotic symptoms.

CASE CONTINUED
Guilty feelings
To address her delusions, I start Dr. I on ris­peridone, 2 mg at bedtime. She goes home for the weekend, and her husband reports that she slept throughout the visit. When she returns, she spends a lot of time in bed but is more communicative.

When I ask Dr. I whether she has called Germany, she says she called her recently widowed father. Dr. I begins to cry when talk­ing of her mother, and tells the nurse she feels guilty for not visiting for the last few years. When her mother died 6 months ago, Dr. I had not seen her in 4 years.

Her fears remit with risperidone, main­tained at 2 mg/d, but Dr. I remains depressed and responds slowly to treatment with cital­opram, 20 mg/d, and supportive therapy. Her final diagnosis is mood disorder with psy­chotic features.

Treatment
When treating women with late-onset psy­chosis, remove all potential triggers and address underlying illness. Cognitive ther­apy targeting specific symptoms is useful; antipsychotics probably will be necessary. Age-related physiologic changes make older persons more sensitive to the thera­peutic and toxic effects of antipsychotics.

Estrogen therapy? Women suffering from schizophrenia show significantly lower es­trogen levels than the general population of women, and they experience first-onset or recurrence of a psychotic episode signif­icantly more often in low estrogen phases of the cycle. Estrogens have therefore been postulated to constitute a protective factor against psychosis, which means perimeno­pause is an at-risk period.²¹ Although evi­dence is limited, preliminary studies have found beneficial effects from short-term, off-label use of estrogen therapy in women with psychotic illness (Box 2).

Because continuous use of estrogen plus progestin has been associated with an in­creased risk of adverse effects,²² off-label use of selective estrogen receptor modula­tors (SERMs) also is being investigated in women with schizophrenia. SERMs act as tissue-specific estrogen agonists and an­tagonists because they can either inhibit or enhance estrogen-induced activation of estrogen response element-containing genes.²³

Wong et al²⁴ used a crossover design to compare the SERM raloxifene with place­bo as adjunctive treatment for 6 postmeno­pausal women with schizophrenia. Each woman received 8 weeks of raloxifene, 60 mg/d, and 8 weeks of placebo. Three be­gan with placebo and 3 with raloxifene.

Verbal memory was measured weekly with the California Verbal Learning Test, using 5 memory trials, free and cued short-delay recall, and long-delay recall. At baseline, the participants had lower scores than older adults in the general popula­tion. Eight weeks of placebo improved scores somewhat, suggesting a practice effect. Eight weeks of raloxifene improved cognitive scores to a level similar to that of schizophrenia-free subjects. After 16 weeks, however, cognitive scores in the 2 groups were indistinguishable.

At present I do not recommend estrogen for women with late-onset schizophrenia because the risk is too high and raloxifene does not enter the brain sufficiently to be a valuable cognitive enhancer. Novel SERMs with more specific efficacy for improving cognitive function may prove useful in the future,²⁵ however, as may phytoestro­gens. Adjunctive hormone modulation is a promising area of gender-specific treat­ment for serious mental illness.²⁶

CASE CONCLUSION
Gradually improving
Dr. I’s depression was triggered by her moth­er’s death and regrets about not visiting and not being a mother. The content of her de­lusions was related to her guilt about not having returned to Germany; the delusions were probably triggered by depression, al­cohol intake, her relative hypoestrogenic state, stress at work, lack of social supports, and dependence on her husband.

Over the next few years, Dr. I is maintained on a low dose of risperidone (reduced from 2 mg/d to 1 mg/d) and citalopram (reduced from 20 mg/d to 10 mg/d). She becomes increas­ingly engaged in supportive dynamic therapy, and her symptoms gradually improve.

BOTTOM LINE
Psychosis onset in midlife is mostly a female phenomenon because a perimenopausal estrogen decline increases women’s susceptibility. Seek specific triggers such as medical illness or response to a drug before assuming an illness of unknown cause such as bipolar disorder or schizophrenia. Cognitive therapy targeting specific symptoms is useful; antipsychotics probably will be necessary.

Related Resources
• Women and psychosis: A guide for women and their families. Centre for Addiction and Mental Health. University of Toronto. www.camh.net/About_Addiction_Mental_ Health/Mental_Health_Information/Women_Psychosis.
• Seeman MV. Women and psychosis. www.medscape.com/ viewarticle/408912.
• Chattopadhyay S. Estrogen and schizophrenia: Any link? The Internet Journal of Mental Health. 2004;2(1). www.ispub. com/journal/the_internet_journal_of_mental_health.html.

Drug Brand Names
Citalopram • Celexa                         Prednisone • Deltasone,
Estradiol • Estrace,                                    Orasone, others
   Estrofem, others                           Raloxifene • Evista
Estradiol transdermal •                     Risperidone • Risperdal
   Estraderm , Climara, others
Methylphenidate • Concerta,
   Ritalin, others

Disclosure
Dr. Seeman reports no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Dr. I, a 48-year-old university professor, is brought to the ER by her husband because she has developed an irrational fear of being chased by Nazis. The fears have become increasingly bizarre, her husband reports. She believes her Nazi persecutors are bandaging their arms and using wheelchairs to pretend to be disabled. When out with her husband, Dr. I points to people in wheelchairs, convinced they are after her, will kill her, and are incensed because she left Germany—her country of birth. Her hus­band brought her to the ER when she started to hear her persecutors addressing her in German at night.

Psychoses of unknown cause usually begin in late ado­lescence or early adulthood. Less frequently the onset occurs in later adulthood (age ≥40). Late-onset psycho­sis is much more prevalent in women than in men for reasons that are imperfectly understood.

When you are evaluating a midlife woman with first onset of psychosis, don’t assume an illness of unknown cause (bipolar disorder or schizophrenia) until after you have done a comprehensive search for triggers of her psychotic symptoms. After age 40, women are more likely than men to develop psychosis because of gender-specific medical and psychological precipitants.

Predisposing factors for psychosis
Psychosis is an emergent quality of structural and chemical changes in the brain. As such, it can be ex­pected to surface during:
   • brain reorganization or transition (adolescence, senescence, brain trauma, stroke, starvation, inflammation, or brain tumor)
   • change in brain chemistry (flux in go­nadal, thyroid, or adrenal hormone levels; electrolyte imbalance; fever; exposure to chemical substances; immune response).

Psychological stress impacting the brain via stress hormones also can predispose a person to psychosis.

Because some individuals are more prone than others to develop psycho­sis during brain alteration, chemical and structural changes in the brain are as­sumed to interact with genetic propensi­ties to influence gene expression. Once a psychotic event has occurred, it is thought to sensitize the brain so that subsequent events emerge more readily.¹

Schizophrenia—though not the only ill­ness in which psychosis plays a role—is a prototype for psychotic illness, and sev­eral reported sex differences in this dis­order are worth noting.² The incidence of schizophrenia is approximately the same in both sexes, but women show a later age of onset—a paradox in that the brain de­velops at a faster pace in females and theo­retically should reach the threshold for the first appearance of schizophrenia earlier. Women also require lower doses of anti­psychotic medication to recover from an acute psychotic episode and to maintain remission, at least before menopause.^3,4 Both of these differences can be explained as an effect of estrogen on a) gene expres­sion⁵ and b) liver enzymes that metabolize antipsychotics.⁶

The estrogen hypothesis. Women show a tendency toward premenstrual and post­partum exacerbation of symptoms when estrogen levels are relatively low. These clinical observations, confirmed by some but not all studies, have led to the hypoth­esis that estrogens are neuroprotective⁷ and also protect against psychosis.⁸

Estrogen withdrawal in specific brain cells may release a cascade of events that over time can increase the severity of psy­chotic and cognitive symptoms. The reason for suspecting such effects is based on what we know about estrogenic effects on neurotransmitter, cognitive, and stress-induction pathways, and—more fundamentally—on neuronal growth and atrophy.

According to the estrogen hypothesis, women are—to some degree—protected against schizophrenia by their relatively high gonadal estrogen production between puberty and and menopause. Women lose this protection with the onset of perimeno­pausal estrogen fluctuation and decline, accounting for their second peak of illness onset after age 45.

Epidemiologic studies showing a second peak of schizophrenia onset in women (but not men) around the age of menopause sup­port this hypothesis.^9,10 Longitudinal out­comes for schizophrenia—which are better in women than in men during late adolescence or early adulthood11—gradually even out after the first 15 years of illness, suggest­ing that women’s advantage is lost at a time approximating menopause (Box 1).

The question, then, becomes: Is it only because of estrogen loss after age 40 that women become more prone to develop a psychotic illness? Other differences be­tween the sexes that may play roles include immune function, low iron stores, sleep suf­ficiency, thyroid function, exposure to toxic substances (including therapeutic drugs), societal pressures to be slim while aging (Table), and the experience of stress.¹²

CASE CONTINUED
Exhausted and confused
Dr. I is a well-groomed, handsome woman, but she hardly speaks when interviewed, looking frightened and somewhat bewildered. She has never had a mental health problem, nor has anyone in her family. She agrees to stay in the hospital but is not sure why. She has slept no more than 1 or 2 hours in the last several days.

Her early history is unremarkable. She did well in school. After earning a PhD at the Univer­sity of Leipzig, she and her husband immigrated to Canada. Both are university professors. They never decided not to have children, but children hadn’t come. Her menstrual periods stopped 2 years before admission. The question about children is the only 1 that elicits emotion in Dr. I. When I ask about it, tears come to her eyes as she shakes her head.

Her husband reports that she has not been eating well and has, in the last year, started to drink more alcohol than usual—3 to 4 drinks of whiskey a night. She does not smoke cigarettes, and her health generally is good. She uses no medications. Her hus­band describes their marital relationship as very close, although it has become strained in recent weeks because of her unreasonable fears. He admits that their work is always stressful; competition is fierce, with more and more deadlines and less and less leisure time. The couple has few friends and no hobbies.

Late-onset psychosis symptoms
In late-onset psychosis (after age 45), men appear to suffer substantially milder symp­toms and spend less time hospitalized than women.¹³ Women with late-onset schizophrenia have more severe positive symptoms than men and fewer negative symptoms.^14,15 Overall, patients with late-onset schizophrenia have a lower prevalence of looseness of associations and negative symptoms than those with earlier onset.^16,17

In addition, individuals with schizo­phrenia who become ill in middle age have been reported to:
   • show better neuropsychological per­formance (particularly in learning and ab­straction/cognitive flexibility) than those with early onset
   • possibly have larger thalamic volumes
   • respond to lower antipsychotic doses.¹⁸

Auditory and visual hallucinations frequently are observed in patients with comorbid late-onset schizophrenia and auditory and visual impairment.¹⁶ Palmer et al¹⁸ reported no difference in family his­tory of schizophrenia between early and late onset, but this is controversial. Con­vert et al¹⁶ note that most studies reveal a lower lifetime risk of schizophrenia in first-degree relatives of patients with late-onset than early-onset schizophrenia.

CASE CONTINUED
Medical workup
Dr. I’s physical exam is unremarkable. Her thy­roid is not enlarged; there are no breast lumps. On mental status exam, her mood is flat. She is preoccupied with fears of the Nazis. Routine blood tests show slight anemia; fasting glu­cose levels are within normal range.

I give Dr. I zopiclone, 7.5 mg, to help her sleep. The next day she keeps to herself, eats very little, and appears disinterested in her surroundings. Nursing staff report that she often seems frightened. Dr. I asks to use the ward phone to call Germany but is told that she cannot make long distance calls from that phone. This seems to disturb her.

Differential diagnosis
Sensory impairment, substance abuse, and metabolic changes have been implicated in the appearance of psychosis in later life. More specific to women than men, howev­er, are medical and psychiatric precipitants. These include autoimmune disease (and its treatment) and psychiatric disorders, as well as thyroid dysfunction, self-induced starvation (anorexia nervosa) and diet aids, substance use and abuse, insomnia, and iron deficiency (Table).

Autoimmune disease and treatment. Nearly 80% of patients with autoimmune disease are women, and these disorders (as well as their treatment) can manifest as psychosis. Corticosteroids have a well-documented history of triggering psy­chotic symptoms, which are twice as likely in women than in men. The incidence of severe psychosis while taking oral predni­sone ranges from 1.6% to 50% and averages 5.7%. The average daily dose of corticoste­roids for patients who develop psychosis is 59.5 mg/d.

Corticosteroid creams absorbed through skin as well as inhaled and intranasal cor­ticosteroids in their more potent formula­tions can have systemic effects, including psychosis. Nonsteroidal anti-inflammatory drugs such as ibuprofen also can trigger psychosis.¹⁹

Psychiatric disorders. Posttraumatic stress disorder with psychotic symptoms may overlap with categories such as psy­chogenic psychoses, hysterical psychoses, nonaffective remitting psychoses, acute brief psychoses, reactive psychoses, acute and transient psychoses, and bouffées dé­lirantes (in France, the name for transient psychotic reactions).²⁰ Consider these female-predominant conditions in the differential diagnosis, along with micro­psychotic episodes in borderline personal­ity disorder, in which the predominance of women is 3:1.

Medical treatment for depression and anxiety also can lead to psychotic symp­toms through individual susceptibility to the action of specific drugs or through withdrawal effects.

Clinical assessment
Question all women presenting with psy­chosis about eating habits and diet pills, and check for hypokalemia and hypocal­cemia to rule out starvation effects and reactions to stimulants. Also ask about in­halants, and examine for anemia and thy­roid dysfunction. Consider all medications as having the potential to trigger psychotic symptoms.

A family history of illness is important, with a focus on autoimmune disorder and its treatment. A thorough psychiatric history is crucial and needs to include assess­ment of sleep, mood, and relationships with attachment figures. Do not assume ill­nesses of unknown cause (bipolar disorder or schizophrenia) until after a comprehen­sive search for precipitants of psychotic symptoms.

CASE CONTINUED
Guilty feelings
To address her delusions, I start Dr. I on ris­peridone, 2 mg at bedtime. She goes home for the weekend, and her husband reports that she slept throughout the visit. When she returns, she spends a lot of time in bed but is more communicative.

When I ask Dr. I whether she has called Germany, she says she called her recently widowed father. Dr. I begins to cry when talk­ing of her mother, and tells the nurse she feels guilty for not visiting for the last few years. When her mother died 6 months ago, Dr. I had not seen her in 4 years.

Her fears remit with risperidone, main­tained at 2 mg/d, but Dr. I remains depressed and responds slowly to treatment with cital­opram, 20 mg/d, and supportive therapy. Her final diagnosis is mood disorder with psy­chotic features.

Treatment
When treating women with late-onset psy­chosis, remove all potential triggers and address underlying illness. Cognitive ther­apy targeting specific symptoms is useful; antipsychotics probably will be necessary. Age-related physiologic changes make older persons more sensitive to the thera­peutic and toxic effects of antipsychotics.

Estrogen therapy? Women suffering from schizophrenia show significantly lower es­trogen levels than the general population of women, and they experience first-onset or recurrence of a psychotic episode signif­icantly more often in low estrogen phases of the cycle. Estrogens have therefore been postulated to constitute a protective factor against psychosis, which means perimeno­pause is an at-risk period.²¹ Although evi­dence is limited, preliminary studies have found beneficial effects from short-term, off-label use of estrogen therapy in women with psychotic illness (Box 2).

Because continuous use of estrogen plus progestin has been associated with an in­creased risk of adverse effects,²² off-label use of selective estrogen receptor modula­tors (SERMs) also is being investigated in women with schizophrenia. SERMs act as tissue-specific estrogen agonists and an­tagonists because they can either inhibit or enhance estrogen-induced activation of estrogen response element-containing genes.²³

Wong et al²⁴ used a crossover design to compare the SERM raloxifene with place­bo as adjunctive treatment for 6 postmeno­pausal women with schizophrenia. Each woman received 8 weeks of raloxifene, 60 mg/d, and 8 weeks of placebo. Three be­gan with placebo and 3 with raloxifene.

Verbal memory was measured weekly with the California Verbal Learning Test, using 5 memory trials, free and cued short-delay recall, and long-delay recall. At baseline, the participants had lower scores than older adults in the general popula­tion. Eight weeks of placebo improved scores somewhat, suggesting a practice effect. Eight weeks of raloxifene improved cognitive scores to a level similar to that of schizophrenia-free subjects. After 16 weeks, however, cognitive scores in the 2 groups were indistinguishable.

At present I do not recommend estrogen for women with late-onset schizophrenia because the risk is too high and raloxifene does not enter the brain sufficiently to be a valuable cognitive enhancer. Novel SERMs with more specific efficacy for improving cognitive function may prove useful in the future,²⁵ however, as may phytoestro­gens. Adjunctive hormone modulation is a promising area of gender-specific treat­ment for serious mental illness.²⁶

CASE CONCLUSION
Gradually improving
Dr. I’s depression was triggered by her moth­er’s death and regrets about not visiting and not being a mother. The content of her de­lusions was related to her guilt about not having returned to Germany; the delusions were probably triggered by depression, al­cohol intake, her relative hypoestrogenic state, stress at work, lack of social supports, and dependence on her husband.

Over the next few years, Dr. I is maintained on a low dose of risperidone (reduced from 2 mg/d to 1 mg/d) and citalopram (reduced from 20 mg/d to 10 mg/d). She becomes increas­ingly engaged in supportive dynamic therapy, and her symptoms gradually improve.

BOTTOM LINE
Psychosis onset in midlife is mostly a female phenomenon because a perimenopausal estrogen decline increases women’s susceptibility. Seek specific triggers such as medical illness or response to a drug before assuming an illness of unknown cause such as bipolar disorder or schizophrenia. Cognitive therapy targeting specific symptoms is useful; antipsychotics probably will be necessary.

Related Resources
• Women and psychosis: A guide for women and their families. Centre for Addiction and Mental Health. University of Toronto. www.camh.net/About_Addiction_Mental_ Health/Mental_Health_Information/Women_Psychosis.
• Seeman MV. Women and psychosis. www.medscape.com/ viewarticle/408912.
• Chattopadhyay S. Estrogen and schizophrenia: Any link? The Internet Journal of Mental Health. 2004;2(1). www.ispub. com/journal/the_internet_journal_of_mental_health.html.

Drug Brand Names
Citalopram • Celexa                         Prednisone • Deltasone,
Estradiol • Estrace,                                    Orasone, others
   Estrofem, others                           Raloxifene • Evista
Estradiol transdermal •                     Risperidone • Risperdal
   Estraderm , Climara, others
Methylphenidate • Concerta,
   Ritalin, others

Disclosure
Dr. Seeman reports no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.