Mar 6, 1997 - Objective: To characterize the epidemiology of schizophrenia. Method: Narrative literature ... The uneven distribution of prevalence rates is a result of social selection: an ...... In: Tsuang MT, Simpson JC, editors. Handbook of ...
IN REVIEW
Epidemiology of Schizophrenia 1
Heinz Häfner , Wolfram an der Heiden
2
Objective: To characterize the epidemiology of schizophrenia. Method: Narrative literature review. Results: Each year 1 in 10 000 adults (12 to 60 years of age) develops schizophrenia. Based on a restrictive and precise definition of the diagnosis and using standardized assessment methods and large, representative populations, the incidence rates appear stable across countries and cultures and over time, at least for the last 50 years. Schizophrenic patients are not born into ecological and social disadvantage. The uneven distribution of prevalence rates is a result of social selection: an early onset leads to social stagnation, a late onset to descent from a higher social status. The main age range of risk for schizophrenia is 20 to 35 years. It is still unclear whether schizophrenia-like late-onset psychoses (for example, late paraphrenia) after age 60 should be classified as schizophrenia either psychopathologically or etiologically. In 75% of cases, first admission is preceded by a prodromal phase with a mean length of 5 years and a psychotic prephase of one year’s duration. On average, women fall ill 3 to 4 years later than men and show a second peak of onset around menopause. Consequently, late-onset schizophrenias are more frequent and more severe in women than in men. The sex difference in age of onset is smaller in cases with a high genetic load and greater in cases with a low genetic load. Type of onset and core symptoms do not differ between the sexes. The most pronounced sex difference is the socially negative illness behaviour of young men. Conclusions: Among the factors determining social course and outcome are level of social development at onset, the disorder itself (for example, genetic liability, severity of symptoms, and functional deficits), general biological factors (for example, estrogen), and sex- and age-specific illness behaviour.
(Can J Psychiatry 1997;42:139–151) Key Words: schizophrenia, epidemiology, incidence, prevalence, ecology, symptomatology, age at onset, gender differences
O
ne hundred years of research efforts have failed to produce biological indicators specific to a diagnosis of schizophrenia. Studies on schizophrenia morbidity, therefore, continue to rely on descriptive case definitions, and the quality of their results depends on the diagnostic definitions used.
comparable findings. The standards include a precise, operational case definition, the collection of comprehensive data on all “cases” from a sufficiently large, statistically welldocumented population, symptom assessment by transculturally standardized instruments, and computerized diagnosis independent of the interviewer. The World Health Organization (WHO) “Determinants of Outcome” study (2) was the first to satisfy nearly all of these requirements at each study site in the determination of the incidence of schizophrenia (12 centres, 10 countries). Recently, Thornicroft and Johnson (3) have added a further methodological standard that calls for the assessment of met and unmet needs for care in psychiatric surveys.
In a historical analysis, Dohrenwend (1) distinguished 3 periods of epidemiological studies in psychiatry. According to his methodological standards, only the last period, beginning around 1980, has produced valid and transnationally 1
Director, Schizophrenia Research Unit, Central Institute of Mental Health, Mannheim, Germany. 2 Senior Scientist, Schizophrenia Research Unit, Central Institute of Mental Health, Mannheim, Germany. Address for correspondence: Dr H Häfner, Durchwahl 17 03-726, J 5, Mannheim 68159 Germany
Ecological Epidemiology With their ecological Chicago study, Faris and Dunham (4) began an important chapter in the epidemiology of schizophrenia. The authors found the highest first-admission rates
Can J Psychiatry, Vol 42, March 1997
139
140
The Canadian Journal of Psychiatry
for schizophrenia in the inner city, characterized by social disintegration and low-standard housing. The lowest incidence was on the outskirts of the city, in high-standard residential areas. In the wake of their study, similar zonal distribution patterns were reported from several American cities, but also from Oslo, Norway (5), Nottingham, United Kingdom (6), and Mannheim, Germany (7). In Mannheim, the incidence of schizophrenia was reassessed 10 to 15 years (8) and 22 to 23 years (9) later. Parallelling the city’s stable socioecological structure, the ecological distribution pattern of first-admission rates for schizophrenia proved stable. Besides the uneven ecological distribution of the rates, individuals with schizophrenia at first admission are usually found to be personally and socially disadvantaged as well. Dohrenwend and Dohrenwend (10) found the highest rates in the lowest social class in 5 out of 7 first-admission studies, and Eaton (11) found the same in 15 out of 17 studies. In a later metaanalysis of studies on the topic, Eaton and others (12) calculated a ratio of 3 to 1 for the risk of schizophrenia between the lowest and the highest of 3 social classes. Social Drift versus Social Causation Although time and cause were unclear, the social gradient provided the impetus for a causal link between severe social disadvantage and schizophrenia onset (10,13,14). The socialdrift hypothesis—social decline after illness onset—merely explained the uneven social distribution of prevalence rates. Since first admission or first contact, at that time, was taken as the operational definition of the onset of the disorder, any social consequences of the emerging disorder were not considered. Retrospectively studying individual ecological mobility before first admission, Dauncey and others (15) found that the ecological disadvantage observed in schizophrenic patients at first admission had emerged, on average, 5 years before. The ecological selection is brought about by a tendency of individuals later admitted with a diagnosis of schizophrenia to remain during the premorbid period in decaying housing areas, while their healthy and socially more successful peers move on from these areas into better neighbourhoods. Similar conclusions were drawn by Wiersma and others (16) using Dutch data. On the basis of the national Norwegian case register, Ödegaard (17) early pointed out that people with schizophrenia tend to cluster in rapid-turnover and low-status jobs prior to their first admission. He implicated premorbid schizoid personality traits in the explanation. It seems that at least a part of the social disadvantage involved in schizophrenia emerges before first admission. The question of it already being present in the patients’ parents was investigated by Goldberg and Morrison (18) in a methodologically sound intergenerational study in London. They demonstrated, for schizophrenic men, that the lack of upward social mobility emerges after birth. In another study, conducted by Turner and Wagenfeld (19) in New York, the fathers of schizophrenic men showed slightly lower social statuses compared with controls, but this result is probably
Vol 42, No 2
attributable to the selection of a socially disadvantaged study population, that is, first admissions to state mental hospitals. Meanwhile, 3 large studies of birth cohorts from Great Britain (20,21) and Northern Finland (22) followed into the age of risk for schizophrenia have demonstrated that schizophrenic patients are not socially disadvantaged at birth. Dohrenwend and others (23) tested the alternative hypothesis of social causation versus social selection in a methodologically sophisticated study in Israel. They found compelling evidence for the social selection hypothesis in schizophrenia, in contrast to the risk for major depressive disorder in women or substance abuse in men. Hence social causation does not appear to play an essential etiological role in schizophrenia, as is also indicated by the fairly similar morbidity risk in all the populations investigated (2). Prevalence Prevalence is a product of incidence and length of illness. Length of illness is determined by various factors, for example, variation in the life expectancies of the populations studied, excess mortality after disease onset, and varying proportions of symptom-free cases due to differences in treatment, so that comparisons of prevalence rates across studies are difficult. Prevalence rates are primarily used as an indicator of the morbidity of a given population and its need for care. Table 1 lists selected prevalence studies of schizophrenia and their results according to the following criteria: 1) studies published in 1980 or later, 2) population studies or sufficiently representative utilization studies (for example, case registers). For the reasons stated above, the marked variation in the total rates cannot be interpreted as reflecting differences in the morbidity of the populations under study. The point versus period prevalence rates differ only minutely which, as pointed out by Eaton and others (24), is accounted for by the primarily chronic course of the illness involving only slight annual excess mortality. The dependence of the study results on the diagnostic definition is reflected in the fact that only 30% of a representative sample of first-admission cases (N = 276) assigned to a broad clinical diagnosis of schizophrenia according to ICD-9 (295, 297, 298.3/4) (the ABC schizophrenia study sample) qualified for a DSM-III diagnosis of schizophrenia (25). The lifetime prevalence rate of 1.5% reported from the Epidemiologic Catchment Area study (26) is probably too high. A possible reason is that the assessments were conducted by lay persons using instruments of low discriminative power (27). The most precise estimates of the lifetime risk so far, though only up to the fifth decade of life, have been reported from the British birth cohort study (28). The directly assessed cumulative prevalence until age 43 was 0.63. Assuming that the share of late-onset schizophrenias in the total schizophrenic morbidity is fairly low (15%), this figure presumably represents a fairly close estimate of the true
March 1997
Epidemiology of Schizophrenia
141
Table 1. The prevalence of schizophreniaa Authors
Period
Babigian (98)
USA
1 year
4.1
Bamrah and others (99)
UK
1 year
7.5
No
Bebbington and others (100)
UK
Point
10.9
Yes
Ben-Tovim and Cushnie (101)
Botswana
1 year
5.3
No
Canada
Lifetime
0.3
USA USA Puerto Rico
6 months 6 months 6 months
6.0 to 11.0 3.0 to 6.0 15.0
No No No
China Germany
1 year Point
1.9 to 4.7 3.9
Yes No
Croatia UK
3 months 1 year
1.5 to 5.1 6.8
No No
Bland and others (102) Blazer and others (103) Burnam and others (104) Canino and others (105) Cheung (106) Dilling and Weyerer (107) Folnegovic and Folnegovic-Smalc (108) Freeman and Alpert (109) Halldin (110)
Age correction Yes
No
Sweden
1 year
6.0
No
Netherlands
Point
7.3
Yes
Hwu and others (112)
Taiwan
Lifetime
3.1
No
Lee and others (25,113) Lehtinen and others (114,115)
Korea Finland
Lifetime Point
Hodiamont and others (111)
Lin and others (116)
3.1 to 5.4 1.3
No No
China
Point
4.2
Yes
Mavreas and Bebbington (117)
UK
Point
13.0
No
Murphy and Taumoepeau (118) Myers and others (119) Nandi and others (120)
Tonga USA India
1 year 6 months Point
0.9 to 1.3 6.0 to 11.0 2.2
No No Yes
Shen and others (121,122)
China
Point
1.8
Sikanerty and Eaton (123)
Ghana
Point
1.1
No
Bulgaria
Point
2.8
Yes
Temkov (124)
Yes
Vasquez-Barquero (125)
Spain
Point
5.6
No
Walsh (126)
Ireland
Point
9.8
No
4.0
No
Weissman and Myers (127) Widerlöv and others (128) Zimmerman-Tansella and others (129) a
Prevalence per 1000 population
Country
USA
Point
Sweden Italy
Lifetime 1 year
0.7 to 5.0 1.3
No No
Abridged version of a compilation by Warner and de Girolamo (36), supplemented by own review of literature.
cumulative lifetime prevalence as based on the ICD-9 diagnosis 295. The prevalence rates for old age probably need to be adjusted. It is still unclear how many of the delusional disorders of the old and the very old can be subsumed under the diagnosis of schizophrenia and, thus, be considered part of the prevalence and incidence of schizophrenia in these age groups (29). The reason is that attempts at a precise empirical distinction at the psychopathological level have failed. Incidence As stated, incidence rates are the key indicator of morbidity risk. Their distribution patterns across geographic, cultural and social units, family membership, age, and sex also provide some external criteria for testing the validity of a diagnosis (30) and for formulating etiological and pathogenetic hypotheses. Since the onset of schizophrenia is a comparatively rare event, the accurate assessment of incidence figures is faced
with considerable practical and methodological problems. This is especially the case when age distributions are studied or risks for certain age groups are estimated. To ensure accuracy, the following preconditions should be fulfilled: 1) the population at risk must be sufficiently large, geographically stable, and statistically well documented to provide exact data on age, sex, and social status for the denominator in the calculation; 2) the study sample must be representative, that is, comprise all the cases fulfilling the diagnostic criteria during the study period; 3) the study sample must be large enough to cover the whole age-of-risk range for schizophrenia; 4) the cases should be entered in the study as soon after onset as possible in order to keep recall deficits, postonset deaths, and other distorting effects to a minimum; 5) onset must be defined and its occurrence determined precisely. The common administrative definition of onset, first admission to hospital or first psychiatric contact, is flawed by uncertainty about the duration of the time period of illness preceding that event. The more precise, but still artificial, definition of a “diagnostic onset” (the time point when a set of diagnostic
142
The Canadian Journal of Psychiatry
criteria is fulfilled for the first time), proposed by Pogue-Geile (unpublished observation), is an acceptable basis for rough estimates of the incidence. If, however, the intention is to determine the time point or age of onset precisely, both the prodromal phase and the development of illness before the diagnostic criteria are met must be taken into account. An ideal approach to studying incidence would be a prospective birth cohort design with cross-sectional assessments at short intervals. A prospective birth cohort design is adopted in studies of the offspring of schizophrenic mothers, but for cost reasons, the cross-sectional assessments are carried out at lengthy intervals. Moreover, the incidence rates calculated for these high-risk groups do not provide any information on incidence in the population at large. Birth cohorts of general populations were studied retrospectively by Fremming (31) in Denmark and for over 70 years by Helgason (32) in Iceland, but the number of 34 or 36 cases of schizophrenia is too low for calculating age-specific risks. The British birth cohort study (28) provided population-based first-admission rates for schizophrenia until age 43 years, but did not produce data on the age distribution of onset. Only in a minority of cases is the onset of schizophrenia marked by conspicuous symptoms. In three-quarters of the cases, the first psychotic episode is preceded by a prodromal phase of several years’ duration, during which nonspecific symptoms predominate. The psychotic prephase, too, frequently evolves unnoticed by others, at least for some time (33). This means that schizophrenia is currently diagnosable only after it has started to produce psychotic symptoms and is perceived by the patient himself or herself or by others. Its onset and all associated variables, such as age distribution of the morbidity risk, must therefore be determined retrospectively by suitable methods. The problems involved and possible solutions have been discussed by Maurer and Häfner (34) in the context of the standardized interview developed to address them: Instrument for the Retrospective Assessment of the Onset of Schizophrenia (IRAOS) (35). The interview was designed to produce data on prodromal signs, symptoms, functional impairment, social disability, and objective social development from 3 sources and on the basis of a time matrix, which enables accurate timing by means of anchor events. Population studies are impractical because of the low morbidity risk. A 2-stage design should be used, therefore, that is, one which identifies all incidence cases of nonaffective psychosis from all the clinical records of a large, defined population over the course of one or 2 years. For this method to be accurate, however, the lifetime likelihood of schizophrenic patients coming into contact with the mental health services of the catchment area must be close to 100%. This percentage, depending on the availability and cost of mental health services, is bound to decline with time because of the growing proportion of patients with schizophrenia treated by general practitioners. An additional search for incidence cases not in contact with mental health services can improve
Vol 42, No 2
epidemiological validity, especially in developing countries (2). At the second stage, standardized interviews are administered to the index population obtained through the initial screening. The operational accuracy of the diagnosis and all the other design variables ensured, onset is then assessed retrospectively. Comparative Epidemiology Table 2 is based on a review by Warner and de Girolamo (36) but restricted to studies published after 1980. Rates calculated with populations not corrected for age are so indicated. The incidence rates range from a minimum of 0.07 to a maximum of 7.1. Tempting though it would be to explain these differences as secondary to regional heterogeneity of environmental or genetic factors, the studies should first be compared for their methodological standards. In the WHO “Determinants of Outcome” study (2), all patients were assessed for psychopathology with a semistructured interview, the Present State Examination (PSE) (37). The PSE is supplemented by a computer algorithm, known as CATEGO, that can be used at several levels of classification. The patients are allocated to 1 of 9 descriptive CATEGO classes, depending on the prevailing symptomatology as, for example, S+, representing central schizophrenic conditions (thought intrusion and delusions of control), P (paranoid symptomatology), or O (other psychoses). The results from 7 countries are illustrated by data for 8 study sites, depicted in Figure 1 in sections II and III. For comparison, section I gives the rates from 8 selected national studies. The transnational variability of the incidence rates decreases as the stringency of studies increases. The national studies show the greatest variation in methodology, for example, diagnostic definitions and case findings. In the intermediate section, annual incidence rates for a broadly defined diagnosis of schizophrenia (ICD 295 or CATEGO S, P, and O), based on the same sample as in section III, show less but still considerable variation, whereas in section III, based on a restrictive and highly reliable diagnosis of nuclear schizophrenia (CATEGO S), no significant differences can be seen, the annual rates varying around 10/100 000. Incidence rates of the same size in all populations and cultures studied render a substantial role of cultural, social, or climatic factors in the morbidity risk for schizophrenia inconceivable. Zubin (38) tried to explain this unusual epidemiological pattern by assuming that schizophrenia is the common final pathway arising from a great number of causes. Another common disorder with an almost identical age-corrected incidence throughout the world is dementia in old age (39,40), which shows a clear-cut causal heterogeneity, but largely identical symptoms. Age at Onset In the majority of studies (age of risk beyond 60 years unconsidered), the mean age at first contact or first admission
March 1997
Epidemiology of Schizophrenia
143
Table 2. The incidence of schizophreniaa Country
Diagnosis
Incidence per 1000 population per year
Age correction
Babigian (98)
USA
Clinical diagnosis
0.94
No
Bamrah and others (99)
UK
ICD
0.19
No
Bates and van Dam (130)
Canada
Clinical diagnosis
0.10
Yes
Der (76) Dilling and others (131)
UK Germany
Clinical diagnosis Clinical diagnosis
0.09 to 0.16
No Yes
Eagles and Whalley (75) Folnegovic and others (132)
Scotland Croatia
ICD ICD
0.12 to 0.20 0.27
Yes No
Netherlands
ICD
0.11
No
Authors
Giel and others (133)
0.48
Häfner (1996, unpublished observations)
Germany
ICD
0.16
No
Hagnell and others (134)
Sweden
Clinical diagnosis / PSE-CATEGO
0.24
Yes
Jablensky and others (2) Jablensky and others (2) Jablensky and others (2) Jablensky and others (2) Jablensky and others (2) Jablensky and others (2) Jablensky and others (2) Krasik and Semin (135) Krupinski (1983, unpublished observations)
Denmark India USA USSR Japan UK Ireland USSR Australia
Clinical diagnosis / PSE-CATEGO Clinical diagnosis / PSE-CATEGO Clinical diagnosis / PSE-CATEGO Clinical diagnosis / PSE-CATEGO Clinical diagnosis / PSE-CATEGO Clinical diagnosis / PSE-CATEGO Clinical diagnosis / PSE-CATEGO Clinical diagnosis —
0.07 to 0.18 0.09 to 0.35 0.09 to 0.16 0.12 to 0.28 0.10 to 0.21 0.14 to 0.24 0.09 to 0.22 0.08 to 0.23
No No No Yes No No No Yes Yes
Munk-Jørgensen (42) Ninuallain (136)
Denmark Ireland
Clinical diagnosis Clinical diagnosis / ICD / PSE-CATEGO
0.07 to 0.11
0.18 0.37
No No
Eaton (137)
India
—
0.58
No
Shen (121,122)
China
Clinical diagnosis
0.11
Yes
Tien and Eaton (138)
USA
DSM
1.0 to 7.1
No
a
Abridged version of a compilation by Warner and de Girolamo (36), supplemented by own review of literature.
lies between 25 and 35 years for men and women (41,42). In the ABC schizophrenia study (43), mean age at onset according to different definitions as assessed by the IRAOS interview (35) was 24.0 years for the first sign of the disorder, 25.5 years for the first negative symptom, 29.0 years for the first positive symptom, 30.1 years for a first peak of positive symptoms (climax of the first psychotic episode), and 30.3 years for the first admission to hospital.
women. On the basis of first admissions with a clinical diagnosis of schizophrenia to other British registers and to the Dutch national case register, van Os and others (45) recently found an increase from about 10/100 000 in the age group 60
Late-Onset Schizophrenia and Late Paranoid Psychoses The question whether first episodes of schizophrenia also occur beyond age 60 is still unclarified. A great obstacle to epidemiological studies of nonaffective functional psychoses in old age is not only the diagnosis but also the fact that neither psychiatric nor general medical services nor population studies guarantee access to all the cases with paranoid or schizophrenic symptoms. This is why the results from recent studies continue to vary considerably. Harris and Jeste’s (44) review of European studies on late-onset schizophrenia, as well as national Danish case-register data (41), showed a decrease in first-admission rates after age 60 based on a clinical diagnosis of schizophrenia. In contrast, applying the more restrictive DSM-III-R diagnosis of schizophrenia, Castle and Murray (29), who rated Camberwell register data for 1965 to 1984, demonstrated a considerable increase in old age, particularly in
Figure 1. Incidence rates from selected national studies and the WHO “Determinants of Outcome” study per 100 000 population, aged 15 to 54 (both sexes), and for a “broad” and “restrictive” definition of schizophrenia.
144
The Canadian Journal of Psychiatry
Vol 42, No 2
Table 3. Studies on the psychopathology of late-onset schizophrenia Author Fish 1960 (47)
Kay and Roth 1961 (139)
Sample All 41 “paranoid states” > 60 years admitted in 1957, including 9 late-onset and 7 early-onset schizophrenia (cut-off 60 years) 99 admissions > 60 years diagnosed with “late paraphrenia”
Type of study/assessment Case notes, partly personal examination
Case records and clinical assessment
Harris and others 1988 (140)
5 cases aged 56 to 67 years with “late-onset schizophrenia” (first and readmissions)
Case study
Pearlson and others 1989 (51)
54 late-onset (> 45 years) versus 54 early-onset (young) versus 22 early-onset (elderly); first and readmissions
Rating of case records
Howard and others 1993 (53)
134 late-onset (> 45 years) versus 336 early-onset cases; first admissions
Rating of case records
Mayer and others 1993 (52)
130 first admissions 40 to 63 years versus 126 first admissions 18 to 23 years
Evaluation of systematic clinical assessment
Yassa and Suranyi-Cadotte 1993 (141)
20 late-onset schizophrenia (> 45 years) versus 7 “paraphrenia” versus 13 “late-onset paranoia” (first and readmissions)
Clinical assessment
Jeste and others 1995 (142)
25 late-onset schizophrenia versus 39 early-onset schizophrenia versus 35 normals
Clinical and neuropsychological assessment
Symptomatology in late-onset schizophrenia No specific schizophrenia syndrome
Risk factors for late-onset schizophrenia Paranoid personality traits
Auditory hallucinations and “many typically schizophrenic symptoms”; systematic and fantastic, usually persecutory delusions; rare incoherence of thought Bizarre (persecutory) delusion, auditory hallucinations, visual hallucinations, paranoid subtype More visual, tactile, olfactory hallucinations, more diversity of hallucinations, more persecutory delusions; less affective flattening, fewer thought disorders Persecutory and organized delusions; third person running commentary; auditory hallucinations More depression; more symptoms of autonomic nervous system, less psychosocial impairment Late-onset schizophrenia: bizarre delusions, auditory hallucinations; paraphrenia: nonbizarre delusions; late-onset paranoia: nonbizarre delusions, organic conditions Paranoid subtype, better work adjustment
Female gender, deafness, personality disorder, few relatives
—
Sensory impairment, schizoid premorbid personality, female gender, living alone
Female gender
Female gender
Female gender (all diagnoses); no increased sensory deprivation
—
to 74 to 19/100 000 in the group 75 and over in England and Wales and to 25/100 000 for those aged 90 and over in the Netherlands (Figure 2).
age at the symptom level in sufficiently large study populations did the methods adopted produce a clear-cut empirical demarcation (44,51,52).
There are, however, ongoing controversies whether early-onset and late-onset schizophrenia are different or similar disorders and which delusional disorders of the elderly should be included with late-onset schizophrenia and which of them classified separately as, for example, late paraphrenia, paranoid psychoses, or merely persistent delusional syndromes. Behind the question of a valid diagnosis is the issue of the underlying etiologies or, at least, of the similarities or differences of symptoms and course (46–50).
The main risk factors for paranoid disorders in old age are cognitive and sensory impairment (44,53), whereas in early-onset schizophrenia, signs of delayed development, including cognitive impairment, predominate. It is therefore reasonable to assume that in early-onset and very late-onset schizophrenia, respectively, disordered or delayed brain development and degenerative processes of the brain, in both instances ranging from mild to moderate in severity, are the key risk factors. The reaction pattern of schizophrenia or paranoid psychosis as a common final pathway shows clearcut psychopathological similarities but also some age-related differences in both periods of life (54).
Studies focusing on the differentiation of delusional syndromes in later life are reviewed in Table 3. Final conclusions are difficult to draw because of methodological limitations: all of these studies employed small convenience or treatment samples (44,47). Often first-episode cases as well as readmissions were included or no information was given on this issue. In none of the recent studies attempting to distinguish between primary paranoid disorder and schizophrenia in old
Sex Differences in Age at First-Ever Onset and Lifetime Risk As early as 1909, Kraepelin observed that women with dementia praecox, when hospitalized for the first time, were several years older than men. In their review of 53 studies,
March 1997
Epidemiology of Schizophrenia
145
25 years, steadily fell. Female onsets were slower to increase, reached a lower peak somewhat later in life and, after decreasing, peaked again, but less markedly in the age group 40 to 45 years. In first-admission rates, a peak would be expected to occur about 5 years later, which we were able to demonstrate on first-admission rates from the Danish national case register (Häfner and others 1996, unpublished observations). The distribution pattern published by Castle and others (57), based on Camberwell case-register data over 20 years, also displayed a second peak of female onsets just prior to menopause.
Figure 2. Annual first-contact rates of nonaffective psychosis at age 60+ (based on administrative data).
Loranger (58), Seeman (59), Lewine (60), and Seeman and Lang (61) have previously suggested that the sex difference in age at onset might be accounted for by a protective effect of estrogen. Animal experiments have shown that acute and long-term estrogen applications actually reduce dopaminergic behaviour by attenuating the sensitivity of central D2 receptors (62,63). Estrogen effects on serotonin receptors at the neurochemical and the molecular level (gene expression) have been reported by Sumner and Fink (64) and on glutamate receptors by Woolley and McEwen (65). In a controlled clinical study of schizophrenic and depressive women with normal menstrual cycles, Riecher-Rössler and others (66) showed that, when estrogen levels were at their peak, schizophrenic and nonspecific symptom measures were significantly reduced but depressive symptoms were unaffected in both diagnostic groups of women. The sex difference in age of onset, however, disappears in cases with a high genetic load, as shown by DeLisi and others (67), Leboyer and others (68), and Albus and Maier (69), whereas in the definitely nonfamilial cases of Albus and Maier’s sample, the onset-age difference grew to 5.7 years. Evidently, the protective effect of estrogen is the more likely to prevail, the lower the genetic liability. In addition, patients of both sexes with a high genetic load seem to develop schizophrenia at a relatively young age (69).
Figure 3. Mean age values at 5 definitions of onset until first admission. First-episode sample of schizophrenia, broad definition (N = 232).
Angermeyer and Kühn (55) found that women at first admission and frequently also at onset, which has been defined and ascertained in different ways, were older than men. Figure 3, taken from the ABC schizophrenia study and based on a representative sample of 232 first-episode cases, shows a significant age difference of about 3 to 4 years for all definitions of onset. The pooled data of the WHO “Determinants of Outcome” study revealed a mean age difference of 3.4 years (56). A comparison of the distribution patterns of onset for men and women on the basis of ABC study data showed that onset for males started to increase somewhat earlier and more steeply and, after a pronounced peak in the age group 15 to
The lifetime risk for schizophrenia until age 60 appears to be the same for men and women. In earlier studies in the United States, higher lifetime incidence rates for men predominated, in part because of the age cut-off of 45 years in the DSM-III diagnosis, in part because of the samples recruited from public mental hospitals with higher proportions of male patients. The use of criteria representing more stringently defined schizophrenia (for example, duration criterion in DSM-III and DSM-IV) seems to yield a significantly more disparate male-to-female ratio (70). Hambrecht and others (56) have recently discussed the methodological pitfalls that explain the predominance of men in some studies, for example, an underrepresentation of primarily female late-onset cases. A cumulative depiction of the incidence rates for men and women in 5-year age bands until age 59 from the ABC study shows that the cumulative incidence, a good indicator of the lifetime risk as based on an ICD-9 295 diagnosis of schizophrenia is practically the same for both sexes at 13.2/100 000 for men and 13.1/100 000 for women. Men clearly consume
146
The Canadian Journal of Psychiatry
Table 4. Symptomatology at first admission: CATEGO syndrome scores and total score by 3 age groups (N = 232) Syndrome DAH
12 to 20
Onset agea in years 21 to 35
36 to 59
11.2
10.3
10.1
ANOVA P < 0.7
BSO
8.7
7.8
7.4
P < 0.4
SNR
9.6
6.7
6.6
P < 0.003
NSN
16.3
15.5
13.3
P < 0.08
Total score
45.9
40.4
37.7
P < 0.03
a
Onset defined by first psychotic symptom.
their lifetime risk more rapidly than women do, but from age 30 to 35 on, the rates for women approximate and finally meet those for men. Temporal Trends Considering the even distribution of morbidity across countries and cultures (see Figure 1), great variation over historical time is unlikely. This is a difficult assumption to verify, however, because it requires the registration and diagnosis on a fairly identical basis of all cases from a large, statistically well-prepared population over a long period controlling for demographic changes. The belief that schizophrenia did not exist before the 18th century (71) is extremely unlikely to be true, and so is the conviction that its incidence has increased with the bourgeoning population in the 19th and 20th centuries (71–74). Decreases of 40% to 50% in the first-admission rates for schizophrenia from the mid-1960s to about the mid-1980s were recently reported by Munk-Jørgensen (42) from Denmark, Eagles and Whalley (75) from Scotland, and Der (76) from England and Wales. Der’s interpretation of the findings as true morbidity trends triggered a critical debate in The Lancet and the European Archives of Psychiatry and Clinical Neuroscience. The main point raised was that the number of available hospital beds and first admissions for a few other diagnoses, such as affective psychoses and neurotic disorder, also fell by approximately the same amount in these countries over the same period (77,79). Strömgren (78) and Kendell and others (80) also pointed out that the diagnosis of schizophrenia had become more restrictive during this period. Warner and de Girolamo (36), referring to repeated, crosssectional investigations with the same methodology at some study sites over the period in question, found stable, age-corrected rates. Only 3 longitudinal studies fulfilled the requirements mentioned for the investigation of long-term morbidity trends in schizophrenia (17,81,82). They were conducted on case-register data for Norway, Iceland, and the state of Victoria, Australia. The longest period was covered in the Australian study, in which a retrospective application of operationalized diagnoses to random samples of case records was used. All of these studies found fairly stable rates, indicating that temporal changes in schizophrenia morbidity currently or in the recent past must be considered unlikely.
Vol 42, No 2
Clinical Epidemiology Clinical epidemiology aims at generally valid conclusions from clinical samples. The precondition is that study samples are representative of the disorder in question or at least of a defined subtype. The diagnostic definition should not include any course criteria, such as the 6-month criterion in DSM-III and DSM-IV, which is bound to lead to a biased selection in the sample and thus to distorted results. Some epidemiological aspects of a disorder with an irregular, episodic course also require comparisons at an identical stage of illness to avoid confounds from different duration of illness (83). This requirement has been fulfilled only recently in primarily first-episode studies (43,84). Symptomatology of the First Episode across the Life Cycle Early-onset schizophrenia (under age 21) and very earlyonset schizophrenia (under age 14) are characterized by disorganization and a high frequency of poorly differentiated and nonspecific symptoms, severe conduct disorders, and functional impairment: the more prominent these symptoms, the younger the patients are at onset (85,86). The sex distribution of early-onset schizophrenia shows a slight predominance of males in most studies, but the opposite has been demonstrated in other studies. The studies, mostly based on first-contact samples at child psychiatric institutions, however, usually do not include admissions after age 18. For this reason, some young male cases of schizophrenia may be missed because their asocial behaviour brings them into contact with social and youth guidance services as well as juvenile courts before reference to mental health services (87). Testing the conclusion drawn from several studies—that early-onset schizophrenias are more severe than late-onset schizophrenias—investigators from the ABC schizophrenia study, on the basis of PSE and CATEGO syndrome scores (37), found that the PSE total scores were significantly lower for late-onset patients, men and women taken together. As shown in Table 4, the result was accounted for to a lesser extent by the specific symptom scores DAH (delusional and hallucinatory syndromes) and BSO (behaviour and speech), and primarily by the 2 nonspecific scores SNR (specific neurotic syndromes) and NSN (nonspecific neurotic syndromes). The implication of this finding is that the greater severity of early-onset schizophrenia is primarily attributable to a higher frequency of symptoms not specific to psychosis. A comparison of the symptom scores in early- and lateonset schizophrenia (< 21 years versus ≥ 40 years) by sex yielded different age trends for men and women. On 4 out of 8 symptom dimensions, late-onset men scored significantly lower than early-onset men. For late-onset women, not a single indicator was significantly more favourable and one, the SANS global score, was more unfavourable in late-onset women (Table 5). This means that the milder symptomatology in late-onset schizophrenia as a whole was accounted for by men alone, while the total scores for late-onset women did not decrease, and the SANS score, measuring negative symptoms, even
March 1997
Epidemiology of Schizophrenia
147
Table 5. Sex differences in symptom scores at time of first psychotic episode—early versus late onseta Men Symptomatology DAH
Women Late (n = 9)
12.1
Wilcoxon 0.02a
↓5.7
8.6
0.29
7.3
SNR
10.7
0.11
NSN
18.9
0.03b
Total score
50.3
0.02b
↓31.8
SANS
9.3
0.29
PIRS
10.7
0.26
3.0
0.06c
↓1.8
BSO
DAS-M
Early (n = 28)
Early (n = 21)
Late (n = 24)
10.0
Wilcoxon 0.95
8.9
0.44
7.9
7.3
8.2
0.42
7.1
↓11.4
13.0
0.58
13.8
40.0
0.80
39.2
6.6
6.7
0.08c
↑9.5
8.4
9.8
0.73
10.5
1.9
0.61
1.8
10.5
a
Age at first psychotic symptom < 21 years versus ≥ 40 years. P ≤ 0.1. c P ≤ 0.05. Arrows mark the direction of the age difference. b
increased in comparison with early-onset cases. In keeping with the estrogen hypothesis (58–60,62,88), we assume that this finding resulted from the waning protective effect of estrogen—responsible for the delay in onset and the milder early symptomatology in women. In the premenopausal period, this would explain the higher frequency and severity of schizophrenias developed by women at this age. The protective effect of estrogen lacking, men develop relatively severe cases at a young age, but relatively milder cases with growing age of onset.
By using syndrome scores derived from the systematic Arbeitsgemeinschaft für Methodik und Dokumentation in der Psychiatrie (AMDP) symptom check list (93), the main items of which are identical with the PSE, we compared a large, consecutive, first-admission sample of schizophrenic patients (n = 1109) from the Mannheim-based CIMH in the years 1978 to 1992 by 5-year age bands across the total age range (> 15 to ≥ 75). Neither significant sex differences nor age trends were found in any of the schizophrenic and affective syndrome scores.
Age and Sex Differences in Illness Behaviour
At the level of single symptoms, the majority showed only little variation with age. Only very few symptoms showed highly significant and pronounced age differences: as illustrated in Figure 4, based on 5-year age groups from 15 to 19 years until 75 to 79 years, all the percentages of cases presenting systematic and paranoid delusions increased linearly with age across the total age-of-onset range. Systematic delusions showed a sixfold increase from 7% in the youngest age group to 44% in the oldest group. Goodness of fit tests (linear logit model) demonstrated highly significant linear increases or decreases in frequency in logits with age.
A series of studies has consistently reported a higher frequency of conduct disorders, especially asocial and criminal behaviours, and of substance abuse in schizophrenic men (89). The WHO “Determinants of Outcome” study (2) also found more first admissions triggered by acts of violence in men than in women (22.2% versus 13.9%) (56). In the ABC schizophrenia study, comparisons by sex, age not taken into account, did not reveal any significant differences in the core symptoms. As stated by Goldberg and Gold (90) with respect to neuropsychological test results in firstepisode patients, the disorder itself appears to be fairly similar in men and women. The most pronounced sex difference emerged in behavioural items, such as self-neglect, reduced interest in a job, social withdrawal, and deficits of communication, which were all significantly more frequent in men (Table 6). The socially negative items in men showed the highest frequency before age 25 and the lowest frequency after age 35. Considering that population studies have shown conduct disorders, antisocial and violent behaviour, and substance abuse to be significantly overrepresented in men from puberty to early adulthood (91), it is more plausible to classify this sex difference in schizophrenia as illness behaviour rather than as a direct expression of the illness. This presumes that the socially negative behaviour of young males, also reflected in a deficient compliance with care provision, has an unfavourable impact on outcome (92).
Unlike the delusional symptoms, the key symptoms of “incoherence of thought” and “disordered sense of self,” indicators of the disorganizational syndrome dimension particularly frequent in early-onset schizophrenia (85,87), decreased linearly over the entire age range. These 2 trends are probably attributable to developmental factors and, again, are not expressions of the basic psychophysiological process of the disease. This means that the level of cognitive and personality development at the onset of the disorder is reflected in symptoms (85,94,95). As the increase in paranoid and systematic delusions indicates, the cognitive and coping abilities of personality at more mature stages of personality development become more differentiated and increasingly stable, thus reducing the disorganizing effect of the psychosis on overall mental functioning. Like the socially negative illness behaviour of young males, the age trends in the leading positive symptoms—
148
The Canadian Journal of Psychiatry
Vol 42, No 2
Table 6. Gender differences in symptomatology at first admission and during early course (ABC Schizophrenia First-Episode Sample, N = 232) Symptom
Males (%)
Females (%)
P(χ2)a
Cross-sectional at first admission (PSE, PIRS, SANS, DAS) Behaviours: Overadaptiveness/conformity
5.8
14.9
0.029
38.1 81.9 76.4 68.3 84.7 58.1 36.5 48.0
18.0 54.1 50.7 31.3 62.7 39.5 19.5 31.3
0.001 0.001 0.001 0.002 0.002 0.005 0.005 0.012
79.6
92.7
0.003
39.8 31.5 74.1
21.0 16.9 55.6
0.002 0.009 0.003
Behaviours more frequent in men: Self-neglect Deficits of free-time activities Deficits of communication Reduced interest in a job Social disability (overall estimate) Loss of interests Deficits in personal hygiene Social inattentiveness Cumulative prevalence during early course (IRAOS) Behaviours more frequent in women: Restlessness Behaviours more frequent in men Drug abuse Alcohol abuse Reduced activities in free time a
The table gives results only for those items (out of the total of 303 items of PSE, PIRS, SANS, DAS, and IRAOS) in which (for alpha-correction) a significant gender difference was validated with at least a statistical trend (P < 0.10) in both randomly split subsamples.
Eugen Bleuler (96) regarded them as “secondary” symptoms—but not in the negative symptoms, illustrate the extent to which personality development, together with genetic and probably also environmentally determined brain processes, influences the clinical expression and probably also the course of the disorder. This issue was recently raised by Galdos and van Os (97) with reference to the fact that fully elaborated positive symptoms, such as paranoid and systematic delusions, require a certain degree of cognitive maturity and education, for example, knowledge of the systems included as explanatory principles in the delusions. The interaction of disease and development is likely to become one of the key topics for future schizophrenia research and to help build a bridge between genetic, epidemiological, and personality research.
morbidity risk, and the distinction of empirical subtypes of schizophrenia. The knowledge available allows some fairly reliable conclusions about the epidemiology of schizophrenia, however: the average age-corrected morbidity risk for schizophrenia of a restrictive definition is 0.1/1000 population per year, with
Discussion The epidemiology of schizophrenia is still faced with a series of unsolved methodological problems, making conclusive interpretations and meaningful comparisons difficult. Among these problems are the definition of the diagnosis by conventional criteria, the variability of illness courses despite the similarity of psychopathological symptoms, and the heterogeneity of the study populations. The difficulties resulting from nonrepresentative study populations or from a failure to check the findings against alternative interpretations are reflected in the controversy produced by the reports of decreasing first-admission rates for schizophrenia from the mid1960s to the 1980s in some regions. Equally controversial issues are the similarity or difference of age at onset in men and women, the association between social class and the
Figure 4. Psychotic symptoms with significant age trends.
March 1997
Epidemiology of Schizophrenia
a range from 0.07 to 0.14/1000. It is unlikely that cultural, social, or ecological factors play a crucial part in the etiology of schizophrenia.
7.
Schizophrenia is primarily a disorder of adolescents and young adults. The mean age at first contact with mental health services is 25 to 35 years for both sexes together. In threequarters of first-episode cases, the first psychotic symptom is preceded by a prodromal period of 6 years on average. Negative and nonspecific symptoms emerge, on average, 5 years earlier than positive symptoms. Both symptom categories show an exponential accumulation until the climax of the first episode and disappear altogether or at least partly after that. Irrespective of how onset is defined, women fall ill 3 to 4 years later than men. The age distribution of onsets for women shows a second peak in those aged 40 to 45 years. Despite the difference in age of onset, the core symptoms and almost all the other actual disease variables do not differ between the sexes. A pronounced difference is the predominance of socially negative behaviour and substance abuse in young male patients, but this finding very likely reflects ageand sex-specific illness behaviour and is not a direct expression of the disorder.
9.
Overall, the onset of schizophrenia is determined by various factors—genetic and general biological (for example, estrogen)—and these same factors, interacting with age- and sex-specific behaviour and level of development at onset, influence the clinical expression and social course.
8.
10. 11. 12.
13. 14. 15. 16. 17. 18. 19. 20.
21.
22. 23. 24.
Clinical Implications • Incidence rates of schizophrenia appear stable across countries,
25.
cultures, and over time.
• Given a narrow definition of schizophrenia, the annual incidence rates vary around 1 in 10 000.
• On average, women fall ill 3 to 4 years later than men, probably
26. 27.
because of a protective effect of estrogen. 28.
Limitations
29.
• Comparisons among studies are difficult because of differential
30.
standards in methodology.
• It is still unclear whether schizophrenia-like late-onset psychoses should be classified as schizophrenia.
31. 32. 33.
References
34. 35.
1.
2.
3. 4. 5. 6.
Dohrenwend BP. The problem of validity in field studies of psychological disorders. In: Robins LN, Barrett JE, editors. The validity of psychiatric diagnosis. New York: Raven Press; 1989. p 35–55. Jablensky A, Sartorius N, Ernberg G, Anker M, Korten A, Cooper JE, and others. Schizophrenia: manifestations, incidence and course in different cultures: a World Health Organization ten-country study. Psychol Med 1992;20:1–27. Thornicroft G, Johnson S. True versus treated prevalence of psychosis—the Prism Case Identification Study. European Psychiatry 1996;11:185. Faris REL, Dunham HW. Mental disorders in urban areas: an ecological study of schizophrenia and other psychosis. Chicago: University of Chicago Press; 1939. Sundby P, Nyhus P. Major and minor psychiatric disorders in males in Oslo: an epidemiological study. Acta Psychiatr Scand 1963;39:519–49. Giggs J, Cooper JE. Ecological structure and the distribution of schizophrenia and affective psychosis in Nottingham. Br J Psychiatry 1987;151:627–33.
36. 37. 38. 39. 40.
149
Häfner H, Reimann H, Immich H, Martini H. Inzidenz seelischer erkrankungen in Mannheim 1965. Social Psychiatry 1969;4:125–35. Weyerer S, Häfner H. The stability of ecological distribution of the incidence of treated mental disorders in the city of Mannheim. Soc Psychiatry Psychiatr Epidemiol 1989;24:57–62. Löffler W, Häfner H. Die ökologische verteilung schizophrener ersterkrankungen in zwei deutschen grossstädten (Mannheim und Heidelberg). Fundamenta Psychiatrica 1994;8:103–15. Dohrenwend BP, Dohrenwend BS. Social status and psychological disorder: a causal inquiry. New York: J Wiley; 1969. Eaton WW. Residence, social class, and schizophrenia. J Health Soc Behav 1974;15:289. Eaton WW, Day R, Kramer MS. The use of epidemiology for risk factor research in schizophrenia: an overview and methodological critique. In: Tsuang MT, Simpson JL, editors. Handbook of schizophrenia, Volume 3, Nosology, epidemiology and genetics. Amsterdam: Elsevier Science; 1988. p 169–204. Kohn M. Social class and schizophrenia: a critical review and reformulation. Schizophr Bull 1973;7:60–79. Hollingshead A, Redlich F. Social class and mental illness. New York: J Wiley; 1958. Dauncey K, Giggs J, Baker H, Harrison G. Schizophrenia in Nottingham: lifelong residential mobility of a cohort. Br J Psychiatry 1993;163:613–9. Wiersma D, Giel R, de Jong A, Slooff CJ. Social class and schizophrenia in a Dutch cohort. Psychol Med 1983;13:141–50. Ödegaard O. Epidemiology der psychosen. Nervenarzt 1971;42:569–75. Goldberg EM, Morrison SL. Schizophrenia and social class. Br J Psychiatry 1963;109:785–802. Turner RJ, Wagenfeld MO. Occupational mobility and schizophrenia. American Sociological Review 1967;32:104–13. Done DJ, Crow TJ, Johnstone EC, Sacker A. Childhood antecedents of schizophrenia and affective illness: social adjustment at ages 7 and 11. BMJ 1994;309:699–703. Crow TJ, Done DJ, Sacker A. Birth cohort study of the antecedents of psychosis: ontogeny as witness to phylogenetic origins. In: Häfner H, Gattaz WF, editors. Search for the causes of schizophrenia, vol.III. Berlin: Springer-Verlag; 1995. p 3–20. Jones PB. Does schizophrenia result from pregnancy, delivery and perinatal complications? a 28-year study in the 1966 North Finland birth cohort [abstract]. European Psychiatry 1996;11(4 Suppl):243S. Dohrenwend BP, Levav I, Shrout PE, Schwartz SL, Naveh G, Link BG, and others. Socioeconomic status and psychiatric disorders: the causation-selection issue. Science 1992;255:946–52. Eaton WW, Mortensen PB, Herrman H, Freeman HE, Bilker W, Burgess P, and others. Long-term course of hospitalization for schizophrenia, part I: risk for rehospitalization. Schizophr Bull 1992;18:217–27. Lee CK, Kwak YS, Yamamoto J, Rhee H, Kim YS, Han JH, and others. Psychiatric epidemiology in Korea, part II: urban and rural differences. J Nerv Ment Dis 1990;178:247–52. Eaton WW, Kessler LG. Epidemiologic field methods in psychiatry: the NIMH Epidemiologic Catchment Area Program. New York: Academic Press; 1985. Anthony J, Folstein MF, Romanoski A. Comparison of the lay Diagnostic Interview Schedule and a standardized psychiatric diagnosis: experience in Eastern Baltimore. Arch Gen Psychiatry 1985;42:667–75. Jones PB, Rodgers B, Murray RM, Marmot M. Child development risk factors for adult schizophrenia in the British 1946 birth cohort. Lancet 1994;344:1398–402. Castle D, Murray RM. The epidemiology of late-onset schizophrenia. Schizophr Bull 1993;19:691–700. Häfner H, Maurer K. The contribution of epidemiology to the study of diagnosis. European Psychiatry 1994;9:3–12. Fremming KH. The expectations of mental infirmity in a sample of the Danish population. London: Cassell; 1996. Helgason T. Epidemiology of mental disorders in Iceland. Acta Psychiatr Scand 1964;40 (173 Suppl):1S–258S. Hambrecht M, Häfner H, Löffler W. Beginning schizophrenia observed by significant others. Soc Psychiatry Psychiatr Epidemiol 1994;29:53–60. Maurer K, Häfner H. Methodological aspects of onset assessment in schizophrenia. Schizophr Res 1995;15:265–76. Häfner H, Riecher-Rössler A, Hambrecht M, Maurer K, Meissner S, Schmidtke A, and others. IRAOS: an instrument for the assessment of onset and early course of schizophrenia. Schizophr Res 1992;6:209–23. Warner R, de Girolamo G. Schizophrenia. Geneva: World Health Organization; 1995. Wing JK, Cooper JE, Sartorius N. Measurement and classification of psychiatric symptoms. London: Cambridge University Press; 1974. Zubin J. Epidemiology and course of schizophrenia: discussion. In: Häfner H, Gattaz WF, Janzarik W, editors. Search for the causes of schizophrenia, vol I. Berlin: Springer-Verlag; 1987. p 114–9. Jorm AF. Cross-national comparisons of the occurrence of Alzheimer’s and vascular dementias. Eur Arch Psychiatry Neurol Sci 1991;240:218–22. Haass C, Lemare C, Capell A, Citron M, Seubert T, Schenk D, and others. The Swedish mutation causes early-onset Alzheimer’s disease by beta-secretare clevage within the secretory pathway. Natural Medicine 1995;1:1291–6.
150
41. 42. 43.
44. 45. 46. 47. 48. 49. 50.
51. 52. 53. 54.
55. 56.
57. 58. 59. 60. 61. 62.
63.
64. 65. 66. 67. 68. 69. 70. 71. 72. 73.
The Canadian Journal of Psychiatry
Häfner H, Riecher A, Maurer K, Löffler W, Munk-Jørgensen P, Strömgren E. How does gender influence age at first hospitalization for schizophrenia? a transnational case register study. Psychol Med 1989;19:903–18. Munk-Jørgensen P. Schizophrenia in Denmark: incidence and utilization of psychiatric institutions. Acta Psychiatr Scand 1986;73:172–80. Häfner H, Maurer K, Löffler W, Bustamante S, an der Heiden W, Riecher-Rössler A, and others. Onset and early course of schizophrenia. In: Häfner H, Gattaz WF, editors. Search for the causes of schizophrenia, vol.III. Berlin: Springer-Verlag; 1995. p 43–66. Harris MJ, Jeste DV. Late-onset schizophrenia: an overview. Schizophr Bull 1988;14:39–55. van Os J, Howard R, Takei N, Murray RM. Increasing age is a risk factor for psychosis in the elderly. Soc Psychiatry Psychiatr Epidemiol 1995;30:161–4. Roth M, Morrissey J. Problems in the diagnosis and classification of mental disorders in old age. Journal of Mental Science 1952;98:66–80. Fish F. Senile schizophrenia. Journal of Mental Science 1960;106:938–46. Grahame PS. Schizophrenia in old age (late paraphrenia). Br J Psychiatry 1984;145:493–5. Holden NL. Late paraphrenia or the paraphrenias? a descriptive study with a 10-year follow-up. Br J Psychiatry 1987;150:635–9. Almeida OP, Howard R, Förstl H, Levy R. Late onset paranoid disorders, part I: coming to terms with late paraphrenia. In: Chiu E, Ames D, editors. Functional psychiatric disorders of the elderly. Cambridge: Cambridge University Press; 1994. p 303–16. Pearlson GD, Kreger L, Rabins PV, Chase GA, Cohen BM, Wirth JB, and others. A chart review study of late-onset and early-onset schizophrenia. Am J Psychiatry 1989;146:1568–74. Mayer C, Kelterborn G, Naber D. Age of onset in schizophrenia: relations to psychopathology and gender. Br J Psychiatry 1993;162:665–71. Howard R, Castle D, Wessely S, Murray RM. A comparative study of 470 cases of early-onset and late-onset schizophrenia. Br J Psychiatry 1993;163:352–7. Häfner H, Maurer K. Epidemiology of positive and negative symptoms in schizophrenia. In: Shriqui CL, Nasrallah HA, editors. Contemporary issues in the treatment of schizophrenia. Washington (DC): American Psychiatric Press; 1993. p 125–54. Angermeyer MC, Kühn L. Gender differences in age at onset of schizophrenia: an overview. Eur Arch Psychiatry Neurol Sci 1988;237:351–64. Hambrecht M, Maurer K, Häfner H, Sartorius N. Transnational stability of gender differences in schizophrenia? an analysis based on the WHO study on determinants of outcome of severe mental disorders. Eur Arch Psychiatry Clin Neurosci 1992;242:6–12. Castle D, Wessely S, Der G, Murray RM. The incidence of operationally defined schizophrenia in Camberwell, 1965-84. Br J Psychiatry 1991;159:790–4. Loranger AW. Sex differences in age of onset of schizophrenia. Arch Gen Psychiatry 1984;41:157–61. Seeman MV. Interaction of sex, age, and neuroleptic dose. Compr Psychiatry 1983;24:125–8. Lewine RRJ. Gender and schizophrenia. In: Tsuang MT, Simpson JC, editors. Handbook of schizophrenia. Volume III, Nosology, epidemiology and genetics of schizophrenia. Amsterdam: Elsevier Science; 1988. p 379–97. Seeman MV, Lang M. The role of estrogens in schizophrenia gender differences. Schizophr Bull 1990;16:185–94. Häfner H, Behrens S, De Vry J, Gattaz WF. Oestradiol enhances the vulnerability threshold for schizophrenia in women by an early effect on dopaminergic transmission—evidence from an epidemiological study and from animal experiments. Eur Arch Psychiatry Clin Neurosci 1991;241:65–8. Gattaz WF, Behrens S, De Vry J, Häfner H. Östradiol hemmt dopamin-vermittelte verhaltensweisen bei ratten: ein tiermodell zur untersuchung der geschlechtsspezifischen unterschiede bei der schizophrenie. Fortschr Neurol Psychiatr 1992;1:8–16. Sumner BEH, Fink G. Oestradiol-17ß in its positive feedback mode significantly increases 5-HT2a receptor density in the frontal, cingulate and piriform cortex of the female rat. J Physiol 1995;483:52. Woolley CS, McEwen BS. Estradiol regulates hippocampal dendritic spine density via an N-methyl-D-aspartate receptor-dependent mechanism. J Neurosci 1994;14:7680–7. Riecher-Rössler A, Häfner H, Dütsch-Strobel A, Oster M, Stumbaum M, van Gülick-Bailer M, and others. Further evidence for a specific role of estradiol in schizophrenia? Biol Psychiatry 1994;36:492–5. DeLisi LE, Bass N, Boccio A, Shields G, Morganti C, Vita A. Age of onset in familial schizophrenia. Arch Gen Psychiatry 1994;51:334–5. Leboyer M, Filteau M-J, Jay M, Campion D, Rochet T, d’Amato T, and others. No gender effect on age at onset in familial schizophrenia [letter]. Am J Psychiatry 1992;149:1409. Albus M, Maier W. Lack of gender differences in age at onset in familial schizophrenia. Schizophr Res 1995;18:51–7. Lewine RRJ, Burbach D, Meltzer HY. Effect of diagnostic criteria on the ratio of male to female schizophrenic patients. Am J Psychiatry 1984;141:84–7. Torrey EF. Schizophrenia and civilization. New York: Jason Aronson; 1996. Astrup C. Nervöse erkrankungen und soziale verhältnisse. Berlin: Volk und Gesundheit; 1956. Eaton WW. A formal theory of selection for schizophrenia. American Journal of Sociology 1980;86:149–58.
74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94.
95. 96. 97. 98. 99. 100. 101. 102. 103. 104.
105. 106. 107.
Vol 42, No 2
Hare E. Was insanity on the increase? Br J Psychiatry 1983;142:439–45. Eagles JM, Whalley LJ. Decline in the diagnosis of schizophrenia among first admissions to Scottish mental hospitals from 1969-78. Br J Psychiatry 1985;146:151–4. Der G. Is schizophrenia disappearing? [letter]. Lancet 1990;335:513–6. Angst J. Is schizophrenia disappearing? Eur Arch Psychiatry Clin Neurosci 1991;240:373. Strömgren E. Changes in the incidence of schizophrenia? Br J Psychiatry 1987;150:1–7. Häfner H, Gattaz WF. Reply: is schizophrenia disappearing? Eur Arch Psychiatry Clin Neurosci 1991;240:374–6. Kendell RE, Malcolm DE, Adams W. Is the incidence of schizophrenia falling? Schizophr Res 1992;6:100. Helgason T. Expectancy of schizophrenia in Iceland during the twentieth century. In: Glel R, editor. Changing the course and outcome of mental disorder. World Psychiatric Association meeting; Groningen, The Netherlands; 1993 Sept 1–3. Krupinski J, Alexander L. Patterns of psychiatric morbidity in Victoria, Australia, in relation to changes in diagnostic criteria 1848-1978. Social Psychiatry 1983;18:61–7. Häfner H. The epidemiology of onset and early course of schizophrenia. In: Häfner H, Wolpert EM, editors. New research in psychiatry. Toronto: Hogrefe & Huber; 1996. p 33–61. Loebel AD, Lieberman JA, Alvir JMJ, Mayerhoff D, Geisler SH, Szymanski SR. Duration of psychosis and outcome in first-episode schizophrenia. Am J Psychiatry 1992;149:1183–8. Remschmidt H, Schulz E, Martin M, Warnke A, Trott G-E. Childhood-onset schizophrenia: history of the concept and recent studies. Schizophr Bull 1994;20:727–45. Asarnow RF, Asamen J, Granholm E, Sherman T, Watkins JM, Williams ME. Cognitive/neuropsychological studies of children with a schizophrenic disorder. Schizophr Bull 1994;20:647–69. Häfner H, Nowotny B. Epidemiology of early-onset schizophrenia. Eur Arch Psychiatry Clin Neurosci 1995;245:80–92. Riecher-Rössler A, Häfner H, Stumbaum M, Maurer K, Schmidt R. Can estradiol modulate schizophrenic symptomatology? Schizophr Bull 1994;20:203–14. Häfner H, Böker W. Crimes of violence by mentally abnormal offenders. London: Cambridge University Press; 1982. Goldberg TE, Gold JM. Neurocognitive deficits in schizophrenia. In: Hirsch SR, Weinberger DR, editors. Schizophrenia. London: Blackwell Science; 1995. p 146–62. Choquet M, Ledoux S. Epidémiologie et adolescence. Rhône-Poulenc rorer specia 1994;27:287–309. Häfner H, Fätkenheuer B, Nowotny B, an der Heiden W. New perspectives in the epidemiology of schizophrenia. Psychopathology 1995;28:26–40. Gebhardt R, Pietzker A, Strauss A, Stoeckel M, Langer C, Freudenthal K. Skalenbildung im AMDP-System. Archiv für Psychiatrie und Nervenkrankheiten 1983;233:223–45. Asarnow RF, Caplan R, Asarnow JR. Neurobehavioral studies of schizophrenic children: a developmental perspective on schizophrenic disorders. In: Häfner H, Gattaz WF, editors. Search for the causes of schizophrenia, vol.III. Berlin: Springer-Verlag; 1995. p 87–113. Blanz B, Schmidt MH, Detzner U, Lay B. Is there a sex-specific difference in onset age of schizophrenia that started before age of 18? European Child and Adolescent Psychiatry 1996;3:267–76. Bleuler E. Dementia praecox oder die gruppe der schizophrenien. Leipzig-Wien: Deuticke; 1911. Galdos PM, van Os J. Gender, psychopathology, and development: from puberty to early adulthood. Schizophr Res 1995;14:105–12. Babigian HM. Schizophrenia: epidemiology. In: Kaplan HI, Freedman AM, Sadock BJ, editors. Comprehensive textbook of psychiatry. Baltimore: Williams & Wilkins; 1980. p 860–6. Bamrah JS, Freeman HL, Goldberg DP. Epidemiology in Salford, 1974-84: changes in an urban community over ten years. Br J Psychiatry 1991;159:802–10. Bebbington P, Hurry J, Sturt E, Wing JK. Epidemiological study of mental disorders in Camberwell. Psychol Med 1981;11:561–79. Ben-Tovim DI, Cushnie JM. The prevalence of schizophrenia in a remote area of Botswana. Br J Psychiatry 1986;148:576–80. Bland RC, Orn H, Newman SC. Lifetime prevalence of psychiatric disorders in Edmonton. Acta Psychiatr Scand 1988;338:24–32. Blazer D, George LK, Landerman R, Pennybacker M, Melville ML, Woodbury M, and others. Psychiatric disorders: a rural/urban comparison. Arch Gen Psychiatry 1985;42:651–6. Burnam MA, Hough RL, Escobar JI, Karno M, Timbers DM, Teller CA, and others. Six-month prevalence of specific psychiatric disorders among Mexican Americans and non-Hispanic whites in Los Angeles. Arch Gen Psychiatry 1987;44:687–94. Canino GJ, Bird HR, Shrout PE, Rubio-Stipec M, Bravo M, Martinez R, and others. The prevalence of specific psychiatric disorders in Puerto Rico. Arch Gen Psychiatry 1987;44:727–35. Cheung P. Adult psychiatric epidemiology in China in the 80s. Culture, Medicine and Psychiatry 1991;15:479–96. Dilling H, Weyerer S. Prevalence of mental disorders in the small-town, rural region of Traunstein (Upper Bavaria). Acta Psychiatr Scand 1984;69:60–79.
March 1997
Epidemiology of Schizophrenia
108. Folnegovic Z, Folnegovic-Smalc V. Schizophrenia in Croatia: interregional differences in prevalence and a comment on constant incidence. J Epidemiol Community Health 1992;46:248–55. 109. Freeman HL, Alpert M. Prevalence of schizophrenia in an urban population. Br J Psychiatry 1986;149:603–11. 110. Halldin T. Prevalence of mental disorder in an urban population in central Sweden. Acta Psychiatr Scand 1984;69:503–18. 111. Hodiamont P, Peer N, Sybern N. Epidemiological aspects of psychiatric disorder in a Dutch health area. Psychol Med 1987;17:495–505. 112. Hwu HG, Yeh E-K, Chang L-Y. Prevalence of psychiatric disorders in Taiwan defined by the Chinese diagnostic interview schedule. Acta Psychiatr Scand 1989;79:136–47. 113. Lee CK, Kwak YS, Yamamoto J, Rhee H, Kim YS, Han JH, and others. Psychiatric epidemiology in Korea, part I: gender and age differences in Seoul. J Nerv Ment Dis 1990;178:242–6. 114. Lehtinen V, Joukamaa M, Lahtela K, Raitasalo R, Jyrkinen E, Maatela J, and others. Prevalence of mental disorders among adults in Finland: basis results from the Mini Finland Health Survey. Acta Psychiatr Scand 1990;81:418–25. 115. Lehtinen V. The prevalence of PSE-CATEGO disorders in a Finnish adult population cohort. Soc Psychiatry Psychiatr Epidemiol 1990;25:187–92. 116. Lin KM. Overview of mental disorders in Chinese cultures: review of epidemiological and clinical studies. In: Kleinman A, Lin TY, editors. Normal and abnormal behaviour in Chinese cultures. Dordrecht: Reidel; 1981. p 237–72. 117. Mavreas VG, Bebbington P. Psychiatric morbidity in London’s Greek-Cypriot immigrant community. Social Psychiatry 1987;22:150–9. 118. Murphy HB, Taumoepeau BM. Traditionalism and mental health in the South Pacific: a re-examination of an old hypothesis. Psychol Med 1980;10:471–82. 119. Myers JK, Weissman MM, Tischler GL, Holzer CE, Leaf P, Orvaschel H, and others. Six month prevalence of psychiatric disorders in three sites. Arch Gen Psychiatry 1984;41:959–67. 120. Nandi DN, Mukherjee S, Boral GC, Banerjee G, Ghosh A, Sarkas S, and others. Socio-economic status and mental morbidity in central tribes and castes in India: a cross-cultural study. Br J Psychiatry 1980;136:73–85. 121. Shen Y. Investigations of mental disorders in Beijing suburban district. Chinese Medical Journal 1981;94:153–6. 122. Shen Y. A survey of mental disorders in a suburb of Beijing. International Journal of Mental Health 1988;16:75–80. 123. Sikanerty T, Eaton WW. Prevalence of schizophrenia in the Labadi district of Ghana. Acta Psychiatr Scand 1984;69:156–61. 124. Temkov I. Use of reported prevalence data in cross-national comparisons of psychiatric morbidity. Social Psychiatry 1980;3:111–7. 125. Vasquez-Barquero JL. A community mental health survey in Cantabria: a general description of morbidity. Psychol Med 1987;17:227–41.
151
126. Walsh D. The treated prevalence of mental illness in the Republic of Ireland: the three county case register study. Psychol Med 1980;10:465–70. 127. Weissman MM, Myers JK. Psychiatric disorders in a U.S. community: the application of research diagnostic criteria to a resurveyed community sample. Acta Psychiatr Scand 1980;62:99–111. 128. Widerlöv B, Borgå P, Cullberg J, Stefansson CG, Lindqvist G. Epidemiology of long-term functional psychosis in three different areas in Stockholm County. Acta Psychiatr Scand 1989;80:40–6. 129. Zimmerman-Tansella C. Bringing into action the psychiatric reform in SouthVerona: a five year experience. Acta Psychiatr Scand 1985;71:71–86. 130. Bates CE, van Dam CH. Low incidence of schizophrenia in British Columbia coastal Indians. J Epidemiol Community Health 1984;38:127–30. 131. Dilling H, Weyerer S, Fichter M. The Upper Bavarian studies. Acta Psychiatr Scand 1989;348:113–40. 132. Folnegovic Z, Folnegovic-Smalc V, Kulcar Z. The incidence of schizophrenia in Croatia. Br J Psychiatry 1990;156:363–5. 133. Giel R, Sauer HC, Slooff CJ, Wiersma D. Epidemiological observations on schizophrenia and disability in the Netherlands. Tijdschrift voor Psychiatrie 1980;11-12:710–22. 134. Hagnell O, Essen-Möller E, Lanke J. The incidence of mental illness over a quarter of a century: the Lundby longitudinal study of mental illness in a total population based on 42,000 observation years. Stockholm: Almquist & Wiksell International; 1990. 135. Krasik ED, Semin IR. Epidemiological aspects of first admissions of schizophrenic patients. Zhurnal Nevropatologij Psikijatrii 1980;80:1354–9. 136. Ninuallain M, O’Hare A, Walsh D. Incidence of schizophrenia in Ireland. Psychol Med 1987;17:943–8. 137. Eaton WW. Update on the epidemiology of schizophrenia. Epidemiol Rev 1991;13:320–8. 138. Tien AY, Eaton WW. Psychopathologic precursors and sociodemographic risk factors for the schizophrenia syndrome. Arch Gen Psychiatry 1992;49:37–46. 139. Kay DWK, Roth M. Environmental and hereditary factors in the schizophrenia of old age (“late paraphrenia”) and their bearing on the general problem of causation in schizophrenia. Journal of Mental Science 1961;107:649–86. 140. Harris MJ, Cullum CM, Jeste DV. Clinical presentation of late-onset schizophrenia. J Clin Psychiatry 1988;49:356–60. 141. Yassa R, Suranyi-Cadotte B. Clinical characteristics of late-onset schizophrenia and delusional disorder. Schizohpr Bull 1993;19:701–7. 142. Jeste DV, Harris MJ, Krull A, Kuck J, McAdams LA, Heaton RK. Clinical and neuropsychological characteristics of patients with late-onset schizophrenia. Am J Psychiatry 1995;152:722–30.
Résumé Objectif : Caractériser l’épidémiologie de la schizophrénie. Méthode : Analyse documentaire narrative. Résultats : Chaque année, un adulte sur 10 000 (âgé entre 12 et 60 ans) est atteint de schizophrénie. Lorsqu’on utilise une définition restrictive et précise du diagnostic, des méthodes d’évaluation normalisées et des populations larges et représentatives, on constate que les taux d’incidence semblent être demeurés stables, d’un pays et d’une culture à l’autre ainsi qu’au fil des ans, et ce depuis au moins 50 ans. Les malades schizophrènes ne viennent pas au monde socialement défavorisés, ni défavorisés quant à leur milieu. La répartition inégale des taux de prévalence est le résultat de la sélection sociale : l’apparition précoce de la maladie mène à une stagnation sociale alors qu’une apparition tardive entraîne une diminution du statut social. La principale zone d’âge, pour ce qui est des risques de schizophrénie, se situe entre 20 et 35 ans. On ne sait pas encore si les psychoses d’allure schizophrénique à début tardif (comme les paraphrénies tardives), qui se déclarent après l’âge de 60 ans, devraient être classées comme des schizophrénies, que ce soit au plan psychopathologique ou étiologique. Dans 75 % des cas, la première hospitalisation est précédée d’une phase prodromique d’une durée moyenne de 5 ans et d’une préphase psychotique d’un an. Chez les femmes, la maladie se manifeste généralement 3 ou 4 ans plus tard que chez les hommes et une deuxième période d’apparition élevée coïncide à peu près avec la ménopause. Les schizophrénies à début tardif sont donc plus fréquentes et plus graves chez les femmes que chez les hommes. Les différences entre les sexes quant à l’âge d’apparition de la maladie sont moindres lorsque le fardeau génétique est élevé et elles sont plus grandes lorsque le fardeau génétique est faible. Le mode d’apparition et les principaux symptômes de la maladie sont similaires chez les deux sexes. Les différences les plus marquées entre les sexes se situent au niveau du comportement socialement négatif dû à la maladie, qui est observé chez les jeunes hommes. Conclusions : Parmi les facteurs qui déterminent l’évolution sociale et l’issue figurent le niveau de développement social au moment de l’apparition de la maladie, le trouble lui-même (par exemple le fardeau génétique, la gravité des symptômes et les déficits fonctionnels), les facteurs biologiques généraux (par exemple, l’oestrogène) et les comportements pathologiques spécifiques au sexe et à l’âge.
