Acta psychiat. scand. (1979) 60, 483-503 Department of Clinical Psychiatry (Head Prof. Dr. Helmchen), Free University of Berlin, West Germany

Fertility and sibship size in a psychiatric patient population A COMPARISON WITH NATIONAL CENSUS DATA

H.P. VOGEL Fertility and sibship size of 2,518 psychiatric inpatients during 1968-1975 were compared with national census data and examined according to psychiatric diagnoses, psychiatric diseases in firstdegree relatives, early psychic disturbances, duration of disease, and school achievement. Fertility is markedly reduced in all diagnostic subgroups, though particularly in schizophrenia. While psychic disturbances before the age of 15, as an index of a severe disturbance of personality, reduced fertility even further, no significant correlation was found with the duration of the patients’ actual disease. Less than ordinary schooling, but also higher qualifications led to a further reduction in fertility. This is particularly so in schizophrenia. Schizophrenics have their children at a later stage of their reproductive career. The psychological and biological consequences of this fact for the transmission of schizophrenia are discussed in detail. No change of fertility relative to the normal population was detected during the observed 8 years. There is suggestive evidence that patients stem from families smaller than expected from the national data. The inherent methodological problems are discussed. The results do not favour the hypothesis of a balanced polymorphism as a mechanism which could explain the constant incidence of psychoses in spite of the severe selection pressure against them.

Key words: Psychic diseases - fertility - sibship size - national census data

- population genetics.

It is we11 known that the fertility of psychiatric patients - particularly of schizophrenics - is markedly reduced (Essen-Moller (1959), Slater et al. (1971), Mac Sorley (1964), Price et al. (1971), Stevens (1969)). As genetic factors play an important role in the causation of these altogether common diseases (Slater & Cowie (1971)) the question arises why they still occur at such a high and - as far as one can see - constant incidence, although there is a severe selection pressure against them. Fresh mutations alone certainly are not sufficient to guarantee a constant incidence, as many calculations show that a mutation rate in the would be necessary, while the human gene-mutability is believed range of to be in the order of lov5per locus (Kishimoto (1957), Bock (1953), Haldane (1949)). Huxley et ul. (1964) were the first to propose a genetic polymorphism on the basis of a heterozygote advantage in order to solve this problem. The best known example for such a mechanism is sickle cell anaemia, where the homo0001-6!30X/79/100483-21$02.50/0 @ 1979 Munksgaard, Copenhagen

484 zygote is clinically ill, while the heterozygote has an advantage with respect to a relative immunity towards malaria (Allison (1964)). In order to elucidate such a system reasonably reliable data of the psychiatric patients and their families should be collected with respect to: 1) Fertility, 2) mortality up to the end of the reproductive period, and 3) length of the generation cycle. The advantage of the heterozygote therefore could be found in an increased fertility, a decreased mortality, and a shortening of a generation cycle. All three items can have psychosocial as well as biological determinants. One could think of favourable psychological attributes which lead to a higher differential fertility and/or physiological advantages. As the death-rate between 1 year and the main part of the reproductive age is not high, a selective advantage should work predominantly between conception and the first year of life. One could think of a modified tolerance towards maternal-foetal blood group incompatibilities or certain other immunological properties which increase the chance of survival. Some immunological studies in schizophrenics demonstrate increased levels of IgA, IgG and IgM (Amkraut et al. (1973)), but the subject is still controversial (Hussar et al. (1971)). Erlenmeyer-Kimling (1968) found that the first year mortality in the offspring of schizophrenics was significantly lower than in the general U.S. population. Carter & Watts (1971) demonstrated a diminished incidence of various infections among schizophrenic’s relatives and a decreased rate of accidents. A decreased postneonatal death-rate and a decreased accident-proneness are both mainly socially determined figures which show a familial clustering far greater than that of genetically determined diseases (Newcombe (1966), Kincaid (1965), Vogel & Knox (1975)). This could in part be due to a family style of great carefulness. Perhaps mothers in families with an increased risk of schizophrenia are more protective than normal mothers, which would be a benefit to the offspring. Only the exaggeration into overprotectiveness has disabling consequences on the personality of the young, which by some authors is assumed to be important in the development of frank schizophrenia (Kuttner & Lorincz (1966)). The discrepancy between an insufficient mutation rate and a constant incidence of the common psychiatric diseases, in spite of the reduced fertility, is particularly large if one suggests a monogenic mode of transmission as Slater et al. (1971) do for the schizophrenic psychosis. Under Slater’s conditions (gene frequency 0.015; penetrance 0.263) a heterozygote advantage of approximately 10 % is needed to keep the balance (Moran (1965b)). If one proposes a polygenic modus (as e.g. Gottesman & Shields (1967) do for schizophrenia) several gene-loci would be responsible for the disease, each of which would add the average mutability rate of 10-5 per locus. But according to Mayr (1966) even six different loci would still not be sufficient to explain the genetic balance. Erlenmeyer-Kimling discusses yet another possible explanation, i.e. genotypic heterogeneity. If e.g. schizophrenia is not viewed as a genetic entity but as a heterogeneous collection of genotypes which produce similar phenotypes, again mutation plays a more important role. If it were true, it would be a similar situation as with the phenotypical condition “mental deficiency”, where very many heritable lesions and

485

chromosomal defects have been detected by now (Erlenmeyer-Kimling Q Paradowski (1966)). It was the aim of this investigation to examine the patients’ and their parents’ fertility (i.e. the patients’ sibship size) because one parent at least must be a carrier. More detailed knowledge about the fertility pattern of patient-parents, subgroups of patient-parents, and subgroups of patients would help to clarify the still open question of the mode of transmission in schizophrenia, in affective psychoses, and to some extent also in neuroses (Miner (1973)). METHODS In the Psychiatric Clinic of the Free University of Berlin the AMP-documentation-system (Angst et aE. (1969), Scharfetrer (1971)) has been used regularly for every newly admitted inpatient since 1967. All patients from 1968 to 1975 who were admitted for the first time to this hospital were included in this study. About 10 % could not be evaluated due to coding errors in the identification characteristics. 2,518 cases were eventually processed. The following items were extracted for further analysis: Age, sex, first and second psychiatric ICD-diagnosis, number of sibs, birth rank within the sibship, number of children, family status, age at first marriage (if married), school achievement, duration of the actual psychiatric disease (in years, if known), psychiatric disturbances before and after the 15th year of age (only coded, yes, no, or questionable), psychiatric diseases in fist-order relatives and in more distant relatives (if known, it is coded whether it is the same as the patient’s disease or another, or whether the type of the disease cannot be identified; number of affected relatives is not coded), halfsibs, twins, and year of admission. Although very interesting, the effects of birth rank order were not examined, as the number of observed patients is too small to render reliable data, due to the great methodological difficulties inherent in this problem (Newcornbe (1964), Hare & Price (1969, 1970, 1974), Hartog (1974)).

As in a routine documentation not all data are expected to be complete, a special code for “missing values” was provided. Missing and most probably also incorrect values occur, but there is no reason to believe that the frequency of either differs among the different sets of data, thereby introducing any serious bias. The two ICD-diagnoses were used to form another diagnostic code. A hierarchical scheme was used. If one of the two diagnoses (if the patient had two diagnoses) was psychotic or not psychotic syndrome in connection with a somatic disease (ICD 290-294, 309-3 15) the patient was labelled “organic disease”. If one of the two diagnoses was a schizophrenia-like syndrome (ICD 295, 297, 298.3) and the other not “organic disease” the patient was labelled “schizophrenia”. If one of the two diagnoses was an affective psychosis (ICD 296, 298.0) and the other neither an “organic disease” nor a “schizophrenia” the patient was labelled “affective psychosis”. If one of the two diagnoses was a psychoreactive syndrome (ICD 300-308) and the other none of the three already

486 mentioned syndromes, the patient was labelled “neurosis”. School achievement was coded as follows: Less than ordinary school or ordinary school but with more than one repetition of a school year: 1 Ordinary school or “middle school” but with more than one repetition of a school year: 2 “Middle school” or school leading to university entry, but with more than one repetition of a school year: 3 School leading to university entry: 4 Marital status, number of children, number of sibs, and age at first marriage were data which allowed a comparison with census data. National statistical data were used rather than those referring only to Berlin for the following reason (apart from being easier to obtain): the actual fertility of a certain cohort in a certain year is not of importance in this study, but the cumulative fertility rate, i.e. mean family size at a certain time. As there has always been a large amount of migration into and out of Berlin, an unknown number of patients will have spent a considerable amount of time during their reproductive period outside of Berlin. Even if the fertility in Berlin is slightly lower than the national average, it would not substantially alter the results (e.g. 1974 marital fertility in Berlin: 47.2, in the Federal Republic of Germany: 49.9 live births per 1,000 married women aged 1545!). Three measurements of fertility were used: 1. Age-specific cumulative fertility, i.e. total number of children to a woman (married or unmarried) of a certain age. 2. Age-specific cumulative marital fertility, i.e. total number of children to a married woman of a certain age. 3. Cumulative marital fertility by duration of marriage, i.e. total number of children to a married woman being married a certain number of years. Only the last measurement gives some idea also of male marital fertility, as husband and wife are always married for the same length of time. German birth statistics do not collect any paternal information. As birth statistics were interrupted by the events of World War I1 the agespecific cumulative fertility is based on the 1950-census and the consecutive birth statistics. The other two measurements rely on the 1970-census and the consecutive birth statistics. Table 1 A and B demonstrate national data in 5-year steps concerning women up to the age of 44 and a marriage duration up to 19 years. Data for the years 1968 to 1974 are calculated separately. Data for 1975 were not yet available at the time, therefore patient data of 1975 are also compared with 1974; the difference, however, is not expected to be great. Table 1 C and D give the 1970 data of women above 44 years and being married for longer than 19 years. All patients in this age and marriage duration range had to be compared with the 1970-national data as the German statistical yearbooks publish age-specific and marriage duration-specific data only up to

0.617 0.950 1.485 1.949 2.183 2.086

0.118 0.720 1.720 1.913 2.059 1.906 0.649 0.937 1.547 1.935 2.178 2.084

1970

0.118 0.710 1.351 1.900 2.104 1.933 0.670 0.910 1.427 1.918 2.171 2.079

All Married women women

1971

0.118 0.694 1.342 1.880 2.133 1.975

0.666 0.877 1.394 1.900 2.162 2.176

Married women women

AU

0.8% 1.699 2.044 2.054

1968

0.886 1.687 2.041 2.054

1969

0.871 1.651 2.035 2.053

1970 0.847 1.634 2.028 2.048

1971 0.813 1.624 2.016 2.045

1972 0.769 1.638 2.003 2.040

1973

1972

0.723 1.672 1.988 2.035

1974

0.112 0.661 1.327 1.847 2.143 2.026

0.692 0.832 1.359 1.879 2.151 2.073

All Married women women

20-24 25-24 30-34 above 34

(Y-N

2.065 2.045 2.046 2.297

D) National marital cumulative fertility rates by duration of marriage 1970 for durations above 19 years Duration of marriage

C ) National age-specific cumulative fertility rates 1970 f o r women above age 44 Agsgroup All Married women women (years) 45-49 1.910 1.988 50-54 1.931 1.991 55-59 1.963 2.011 60-64 1.951 2.017 65-69 1.871 1.965 70-74 1.878 1.982 75-79 1.987 2.103 above 79 2.168 2.280

0-4 5- 9 10-14 15-19

bWlrS)

Duration ofmarriage

B) National marital cumulative fertility rates by duration o f marriage 1968-1974

0.117 0.716 1.402 1.917 2.003 1.864

Married women women

15-19 20-24 25-29 30-34 35-39 40-44

1969

All

Married women women

All

Agegroup (Yeas)

1968

A) National age-specific cumulative fertility rates 1968-1974

Table 1

1973

0.102 0.623 1.293 1.801 2.129 2.100

0.701 0.770 1.322 1.855 2.137 2.069

All Married women women

1974

0.094 0.590 1.265 1.753 2.100 2.158

0.657 0.708 1.261 1.819 2.121 2.065

All Married women women

$ 4

the age of 44 and up to 19 years of marriage. The basic data leading to Table 1 are partly unpublished, but kindly provided by the Statistisches Bundesamt. The main part stems from the statistical year books and the quoted special publication of the Statistisches Bundesamt. As can be seen from Table 1, there is a considerable decline in fertility over the observed period of time, particularly in younger women. Therefore, for every female patient and every male married patient, fertility indices were calculated: observed number of childredexpected number of children (derived from the patient’s age or duration of marriage, and year of admission). From the 1970national census the Statistisches Bundesamt has published the mean number of children born to married couples according to the year of contraction of marriage. A frequency distribution of the raised family sizes is published as well. Table 2 gives this information. In order to compare the distribution of sibship sizes of the patient population with the quoted census data, the year of marriage of patient-parents had to be assumed thus: Mean birth intervals correlate negatively with sibship size. The following numerical values are published for an urban community in Britain (Stewart & Barber (1963)) (sibship size 2: 4.1; 3: 3.4; 4: 3.1; 5: 2.7; 6: 2.5; 7: 2.6; 8: 2.3; 9: 2.2; 10 and more: 1.8; always years). They were used in the study. From the patient’s age at admission, his ranking place in his sibship, the total number of sibs and the respective birth interval, one can calculate to which 5year cohort his parents probably belonged. As the assumed birth intervals were speculative, the model was also calculated with different birth intervals from 1.5 to 3 years. The results did not differ considerably from those obtained by using the intervals found by Stewart & Barber (1963).

The frequency distribution of the patients’ sibship sizes according to the assumed year of concentration of their parents’ marriage is given in Table 5 (see Results). These data can only be compared with the census data (Table 2) after further modification and, even then, only with great caution, for the following reasons: 1. If one determines family sizes from samples taken through members of the sibships one gets an over-representation of larger families (e.g., there is always only one person saying that he comes from a one-child family, but 10 persons saying that they come from one 10-children family). Greenwood & Yule (1914) were the first who proposed overcoming this problem by dividing the number of persons in each birth rank by the appropriate birth rank. After repercentaging, this distribution gives a better picture of the family size distribution of the patient-parents. More recently it was Bytheway (1974) who called attention to this statistical trap associated with family size. This transformation has been done with the patient data. In Table 6 this information is presented in a reduced form in order to be easier compared with the census data (Table 2) by combining all birth ranks of four and more. 2. The census data (Table 2) only regard legitimate births to German women,

489 Table 2. Distribution of raised family sizes by year of contraction of the parents’ marriage (legitimate live-born children to German parents; national census data) Year marriage contracted

Single child

Two children

Three children

%

%

%

1899 and beifore 1900-1904 1905-1909 1910-1912 1913-1918 1919-1921 1922-1925 1926-1930 1931-1935 1936-1940 1941-1945 1946-1950 1951-1955

9.86 12.69 16.50 19.80 23.01 27.52 28.68 27.83 26.13 28.51 29.00 29.86 28.44

12.71 18.10 22.30 24.95 27.92 28.94 29.53 29.88 32.54 35.57 36.27 34.79 35.32

13.47 17.04 18.51 19.23 19.39 18.15 17.62 18.31 19.95 19.93 19.33 18.94 19.73

Four and more children %

Total no. childred 100 couples

63.97 52.21 42.70 36.06 29.67 25.27 24.18 23.98 21.38 15.99 15.54 16.42 16.40

489.9 393.0 334.8 293.6 251.6 233.7 221.7 222.6 217.9 204.6 204.5 206.5 205.3

while among the patients there are illegitimate children as well as a few foreigners. The census data as well as the patient data do not take into account the rather complex relationship between social class and fertility (e.g. Piepmeier & Adkins (1973), Stokes (1973)). The census data (as they are recorded in Table 2) are distorted by “maternal” mortality. At the 1970-census only women who were still alive answered the questionnaire and for many women this was several decades after childbearing. Differences in mortality between women with only one child and those with many children cannot be excluded. The distribution of sibship size (as it is recorded in Table 5 ) is distorted by “infant” mortality. Differences in mortality between single children and those from larger sibships are probable and certainly cannot be excluded. These two different mortalities as well as hypothetical differences concerning mortality between the normal population and the families of psychiatric patients are factors which lead to further inaccuracies. There is some likelihood that older sibs who died early were not reported by the patients, as they might not have played any important role in their lives. The frequency distributions and already mentioned indices have undergone further analysis. The data were processed on the central computer of the university (CDC Cyber 71). The necessary statistical calculations were performed using the SPSS-system (Nie et al. (1975)).

RESULTS General findings concerning fertility Table 3 gives some details in order to characterize the patient-population. The University Hospital is not obliged to admit every patient. Therefore this sample 31

490

Table 3. Some characteristics of the patient-population Sex ~

~~

n

%

1.442 1.076

57.3 42.7

~~

Male Female

2.518

Age

n

%

< 20 20-24 25-29 3&34 35-39 4044 >44 Missing

244 343 408 399 225 159 728 12 2.518

9.7 13.6 16.2 15.8 8.9 6.3 28.9 0.5

~~

Diagnostic classification “Organic diseases” Schizophrenia Affective psychosis Neurosis

~

712 708 282 816

28.3 28.1 11.2 32.4

Duration of psychiatric disease since its 6rst manifestation Years n % Unknown

9

655 256 813 339 455

26.0 10.2 32.3 13.5 18.0

is not representative for all psychiatric patients. With respect to schizophrenia the selection is probably a favourable one as there is no department for chronic patients attached and the number of patients having already had long-term hospitalization somewhere else is negligible. As neurotics are generally treated as outpatients, neurotic inpatients are probably a selection of particularly severe cases. The usual duration of hospitalization is in the range of 4 to 12 weeks. Fig. 1 shows the proportion of never married patients, Fig. 2 the proportion of divorced patients by sex, age, and diagnosis. All diagnostic groups - not only the schizophrenic patients - reveal marked ditferences to the normal population. With respect to the celibacy rate schizophrenics deviate most; with respect to the divorce rate neurotic patients deviate most from normality (if one only views the columns with a sufficiently large number of patients). Table 4 shows age-specific cumulative fertility in percent of normal values for the total psychiatric population, and for the four diagnostic subgroups separately. For all subgroups the fertility is markedly reduced. With respect to the

491 general fertility, which also takes the celibacy rate into account, schizophrenics show particularly low values. The two indices of marital fertility are higher and there are no significant differences between the diagnostic subgroups (t-test). An increased age at first marriage may be a reason for the reduced marital fertility in some cases but certainly not in all, as the mean age at first marriage in male patients is 26 years, in females 24 years. There are no differences between the two indices of marital fertility, although the index by duration of marriage also takes the male patients into account. This suggests a similar pattern in both sexes. If one considers only female patients above the age of 44, i.e. after completion of their reproductive period, very similar results are obtained. Further diagnostic sub-classification (acc. to the figure behind the decimal point of the ICD-number) did not yield substantial results. Statistical problems arose due to the great differences in the number of cases to be compared and due to a great inhome geneity of variances. The fertility indices were not significantly different among the different sub-classifications. Two facts, however, should be mentioned: 1. Schizophrenia simplex (ICD 295.0) and hebephrenic schizophrenia (ICD 295.1) showed a particularly decreased fertility. Larson & Nyman (1973) published similar findings. 2. Obsessional neurotics showed a particularly decreased fertility as compared with the other neurotic disturbances (see Hare et al. (1972)). In the normal population completed family size has a variance larger than the mean indicating the negative binomial distribution rather than a Poisson distribution. The same is true for the patient population (female above 44 in order to have completed fertility). Only in schizophrenics is the ratio of mean to variance almost one. The significance of such differences for the survival of genes is beyond the scope of this paper (see e.g. Kojima & Kelleher (1962), Crow & Morton (1955), Imaizumi et al. (1970)).

Findings by specified parameters B y psychiatric diseases in first-degree relatives. Fig. 3 gives the age-specific fertility indices for schizophrenic and affective psychoses. In schizophrenics fertility is even more reduced if first-grade relatives are affected as well, the opposite pattern is found in affective psychoses. Marital fertility shows similar, though not significant trends. No relationship was found between sibship size and psychiatric diseases in relatives, except for affective psychoses. Patients with affected first-degree relatives come from larger families than normal (125 % of normal, P C 0.04). However, it is almost impossible to interpret this fact, if the number of alTected relatives is not known, as the probability of having affected relatives increases with the number of sibs and the patients’ age (the older the patient, the longer also his relatives have lived at risk of falling ill). By psychic disturbances before the age o f 15. Fig. 4 gives the age-specific fertility indices for schizophrenics in relation to whether psychic disturbances before the age of 15 are known (questionable or certain). If this is so, fertility is markedly reduced. The same tendency is found in “organic diseases”, but not in affective psychoses and neurotic disorders. Patients with early psychic disturbances come from larger families than those without (P < 0.05).

0

5

N* 18 2 51 29 AGE-GROUP 15-19vrs

0

Ly

0 02

y1

p 10 =

4

u

Ly

u

a

0

J

5

10

40 5 51 10 20-24yrs

25-29vrs

54 B 16 23

S

91 19 98 52

5

3 0 - 34yrs

51 10 53 25

S

98 19 82 50

S

I ( c e n i u ~d a m 1970 I

Normal population

35-39yrs

29 10 33 17

42 8 34 49

40-44vrs

21 12 23 15

20 5 2 6 3 4

N' : P 1OOp.c. ltotal NO/ age - group I

N : Neurosis 0 : "Organic d i r e a d '

A : Affective Psvchosts

S : Schizophrenia

.

above44yrs

93 120 58 151

S

FEMALE

25 52 62 160

MALE

a

10

20

AGE-GROUP 15 - 1 9 ~ s

w p.

20

- 24yrs 25

- 29yrs

30

-

34yrs

N' : Z loop c. (total No/

0 : "Organic diseases"

A : Affective Psychosis N : Neurosis

Normal population I census dala 1970 I S : Schizophrenia

. I

35

- 39yrs

40

- 44vrr

A

above 4 4 y r r

25 52 62160

S

MALE

494 Table 4. Fertility relative to the normal population Age-specific cumulative fertility Total patient population

Mean SEM n

Statistical significance vs. normal population (i.e. 100 %)

P

Fertility and sibship size in a psychiatric patient population. A comparison with national census data.

Acta psychiat. scand. (1979) 60, 483-503 Department of Clinical Psychiatry (Head Prof. Dr. Helmchen), Free University of Berlin, West Germany Fertili...
1MB Sizes 0 Downloads 0 Views