39
Psychiatry Research, 38:39-50
Elsevier
Familial Thyroid Disease and Delayed Language Development in First Admission Patients With Schizophrenia Lynn E. DeLisi, Angela M. Boccio, Henry Rliordan, Anne L. Hoff, Arlene Dorfman, Joyce McClelland, Maureen Kushner, Olga Van Eyl, and Neal Oden Received February 15, 1991; revised version received May 21, 1991; accepted June 4, 1991. Abstract. One hundred consecutive first admission patients with a DSM-III-R diagnosis of schizophrenia, schizoaffective disorder, or schizophreniform disorder were compared with 100 randomlly selected community controls. Childhood histories of physical, medical, and perinatal trauma, as well as physical and cognitive development, were examined by structured interviews with all available mothers of patients and controls. The prevalence of specific psychiatric disorders and several medical illnesses among first degree and more distant relatives was determined by family history questionnaires. The patient group did not have an excess of childhood head injuries, serious infections, or perinatal/ birth complications compared wit’h controls. With social class level taken into account, it was found that the acquisition of reading skills occurred significantly later in patients than controls. Family histories of schizophrenia and thyroid disorders were significantly more frequent among patients than controls. These data fail to indicate any childhood physical or medical environmental trauma that could lead to an increased risk for schizophrenia, although patients were substance abusers to a greater extent than controls. This study also confirms the already known contribution of familial factors and suggests an association of the inheritance of thyroid disorders with schizophrenia. Delayed development of reading skills suggests that precursers of illness may appear early in life before psychosis is evident. Key Words. Schizophreniform premorbid history.
disorder,
schizoaffective
disorders,
family history,
Although many studies suggest a genetic etiology for schizophrenia, the disorder evades the usual Mendelian characteristics of’ inheritance and has a relatively high discordance rate among monozygotic twins. Thus, the development of schizophrenia may depend on the coexistence of other risk factors in a genetically vulnerable individual. Some factors previously associated with psychosis have been obstetric and other perinatal complications (McNeil and Kaij, 1979; Parnas, 1986; Lewis,
Lynn E. DeLisi, M.D., Angela M. Boccio, MS., Henry Riordan, Ph.D., Anne L. Hoff, Ph.D., Arlene Dorfman, M.S.W., Joyce McClelland, R.N., Maureen Kushner, M.S.W., Olga Van Eyl, M.D., and Neal Oden, M.D., are in the Department of Psychiatry, SUNY Stony Brook, Stony Brook, NY, and the Stony Brook Unit at the Kings Park Psychiatric Center, Kings Park, NY. (Reprint requests to Dr. L.E. DeLisi, Dept. of Psychiatry, HSC, T-10, SUNY Stony Brook, Stony Brook, NY 11794, USA.) 0165-1781/91/%03.50
@ 1991 Elsevier Scientific
Publishers
Ireland
Ltd.
40
1987; Murray and Lewis, 1987), viral infections (i.e., influenza) prenatally or postnatally (DeLisi and Crow, 1986), and serious head trauma (DeLisi et al., 1986; Wilcox and Nasrallah, 1987). The genetic etiology of schizophrenia may also be linked to the gene or genes for other neuropsychiatric disorders or inherited medical conditions that are occasionally characterized by concurrent psychotic symptoms, such as Huntington’s disease, endocrine disorders, and several mental retardation syndromes. The presence of such a disorder among family members might serve as an additional risk factor for schizophrenia; in addition, detecting such a link might provide clues for future searches for a genomic locus for schizophrenia. Not only the development of schizophrenia, but also its onset, may be determined by the presence of premorbid stresses or brain insults (e.g., head injury or substance abuse). The precise onset of schizophrenic illness, however, is difficult to define. Whether early signs of abnormality, other than classical psychotic symptoms, are an indication of onset of the pathological process is controversial. Prospective studies of individuals at high genetic risk for schizophrenia have found that some childhood characteristics (e.g., delayed neuromotor development and attentional deficits; Watt et al., 1984; reviewed in Parnas and Mednick, 1991) consistently predict later development of schizophrenia. Early signs of neuromotor and cognitive difficulties may indicate that schizophrenia develops slowly from birth into adulthood. This report assesses risk factors in a large group of patients and controls that may be associated with an increased vulnerability for schizophrenia. It also explores whether those factors affect the timing of onset of overt psychotic symptomatology.
Methods One hundred patients consecutively admitted to an inpatient psychiatric evaluation program who were being seen for the first time for a schizophrenia-like psychosis were compared with 100 community controls. Patients were selected from approximately 3,500 first admissions screened from September 1987 through December 1990. They showed hallucinations, delusions, formal thought disorder, grossly bizarre behavior, or catatonia; had never been treated before for a psychiatric illness; were not classically manic; did not have a nonpsychotic depression; and did not have substance abuse that could account for the symptoms directly leading to hospitalization. All patients were admitted to an evaluation unit at the Kings Park Psychiatric Center, a New York State hospital that serves a county catchment area of approximately 1 million people. The patients were evaluated by structured clinical interviews (the Schedule for Affective Disorders and Schizophrenia, SADS; Spitzer and Endicott, 1978) history from relatives and clinical observation during the course of hospitalization. For the patient group, a DSM-III-R diagnosis (American Psychiatric Association, 1987) was established taking into account all relevant information from interviews, hospital records and multiple family members. Sixty-three patients were diagnosed as schizophreniform, 10 as subchronic schizophrenia, 8 as chronic schizophrenia, 16 as schizoaffective psychosis, and 3 as psychosis not otherwise specified. Severity of illness was estimated by a total score on the Brief Psychiatric Rating Scale obtained by a trained rater on admission and the amount of time spent in the hospital during the index episode. Control subjects were recruited from the lobby of a general public hospital (University Hospital, SUNY, Stony Brook) that serves the same catchment area as the psychiatric hospital. These persons were approached by a research interviewer at random and asked to participate in a short interview study. An attempt was made to obtain a group of control
41 subjects that was closely matched to the patients for age, sex, and social class. These individuals were interviewed by a structured medical questionnaire and asked for permission to contact their relatives. All available mothers of controls were then interviewed with the same questionnaire that was used for relatives of patients. Controls were eliminated if they were present in the hospital lobby for a psychiatric clinic appointment. Otherwise, the selection of controls was strictly random. These persons did not volunteer to be studied, but rather were approached if they were present in the lobby and appeared to be of similar age to the patient sample. The subjects were not screened for major psychiatric illness and thus represent a general random population. Very few persons refused to answer questions or to allow researchers to contact a relative, although the number of refusals was not recorded. The mean age of the patients was 28.11 (SD = 6.85) and that of the controls was 28.05 (SD = 7.35). There were 64 males and 36 females in the patient group, and 62 males and 38 females in the control group. Structured questionnaires focusing on birth complications, early childhood development, and significant medical or physical trauma were used in interviews with the subject’s mother or closest living relative if the mother was deceased (DeLisi, 1991). A structured family history questionnaire was also used to investigate several inherited metabolic, neurological, and psychiatric illnesses among first, second, and third degree relatives. If a positive history of a major psychiatric illness was solicited, further details about symptoms were ascertained to establish whether the relative met criteria for a D&W-III-R diagnosis of schizophrenia or affective disorder. Information was collected on birth/perinatal complications, childhood illnesses or head injuries, timing of developmental motor or language landmarks (age in months at first steps, first words, first sentence, and first words read), and history of substance abuse (by DSMZZZR criteria). Perinatal complications were scored using a previously published method (DeLisi et al., 1987, 1988). The following were considered complications and received 1 point each: bleeding, hypertension, toxemia, serious injury, and thyroid disease in mother during pregnancy, at least 2 weeks premature or postmature birth, breech delivery, cord around neck, convulsions, jaundice, use of oxygen or incubator, and infection or other perinatal illness. Any patient or control subject who admitted present or past substance abuse was further interviewed using the drug and alcohol section of the SADS to determine if he or she met DSM-III-R criteria for drug or alcohol abuse. For the patients, age of onset was defined as the age at which the patient or a family member first noticed positive psychotic symptoms. The socioeconomic status (SES) of the probands and their parents was estimated using the Hollingshead-Redlich Two-Factor Index of Social Position, which takes into account both educational and occupational level reached (reviewed in Myers and Bean, 1968). The parental SES level was considered to be the highest of either father or mother. Fifty-six of the 100 patients had magnetic resonance imaging (MRI) scans performed according to a previously described protocol (DeLisi et al., 199la, 19916). The volumes of the left and right lateral ventricles and the temporal lobes were measured on 5-mm coronal slices with a 2-mm gap. A trend for larger ventricles was present in this patient group compared with controls, while no temporal lobe differences were found. These measurements were used in correlations with the developmental variables examined in the present study. Data were analyzed using the Complete Statistical System software version B640 (01988, StatSoft, Inc.). Demographic and developmental variables were analyzed by independent t tests (2-tailed) to assess differences between groups. Because there were sex differences within groups and because parental SES differed between groups, two-way analyses of covariance (ANCOVAs, sex X group), with parental SES as the covariate, were conducted on the developmental variables (e.g., age first walked, age first spoke in sentences; see Table 2). Pearson (r) correlations were also computed between perinatal complications and developmental variables. Prevalences of family history of psychiatric and medical disorders, as well as potential clinical risk factors, were analyzed by ~2 statistics. Each family was treated as one unit; a family was considered positive if one or more relatives had illness. No additional weight was
42
given in the analysis to families with more than one ill relative. The data were further examined by path analysis to construct a pictorial representation of a variance matrix for all the variables studied (Li, 1975) and to see whether they contributed to the onset of pathology (i.e., delayed language development). A path diagram specifies the exact nature of a causal structure, and the strengths of the relationships are derived from the underlying variance matrix of the variables. Stepwise regression @ to add = 0.05, p to drop = 0.10) was used within the two sexes separately to specify the causal relations (single-headed arrows) in the diagram. For each regression, two different starting configurations were investigated, and variables were used in the final path diagrams if they were chosen from either starting configuration. After causal connections were stipulated, the strengths of connections were calculated and appear in the path diagram as path coefficients. All path coefficients were calculated by standardized regression through the origin. To obtain the most accurate estimates possible, the maximum number of (nonmissing) cases was used for each regression, so that the number of cases varied from one node to the other across the diagram. Also depicted in the path diagram are all observed correlations (double-headed arrows) with significance values < 0.10.
Results Perinatal Complications and Early Developmental Landmarks. Schizophrenic patients were described by their relatives as speaking their first sentences at a later age (2 = -2.29,~ < 0.05) and learning to read at a later age (t = -4.24,~ < 0.001) than controls (see Table 1). There were similar trends for age at first words and age when first walked, with schizophrenic patients reaching these milestones later (NS). Patients and controls did not differ in reported perinatal complications. A trend was present for left lateral ventricular volume to be positively correlated with the age at which subjects first read (r = 0.29, p = 0.06), although no correlations were significant for other developmental variables and brain morphological measurements. Severity of illness, as measured by the total BPRS score and time spent in hospital at the index admission, also did not correlate with any developmental measures. Two-way ANCOVAs (sex X group; Table 2) with parental social class as a covariate performed on the developmental dependent variables (see Table 1) revealed a main effect of sex on perinatal complications, with females having fewer complications than males (F= 3.81, p < 0.05) across both groups. As above, there was also a large main effect for group on the age at which subjects first read, with schizophrenic patients reading later than controls (F= 12.04, p < 0.001). No other main effects or sex X group interactions emerged. As can be seen from Table 3, a history of birth complications was positively correlated with age at first spoken sentence (r = 0.31, p < 0.05) within the schizophrenic group only, indicating that perinatal complications may adversely affect language development in schizophrenic patients. Premorbid
Childhood
and Later Environmental
in Table 4, schizophrenic patients p < 0.039) and drugs (~2 = 14.54,~
were
more
likely
Risk Factors. As can be seen to abuse
alcohol
h2
= 4.23,
< 0.0001) than controls. However, schizophrenic patients were no more likely than controls to have histories of head injury, childhood infections, hyperactivity, or learning disabilities.
43 Table 1. Characteristics of subjects investigated Controls
First-episode patients Mean
SD
n
Mean
SD
n
28.18
6.83
100
27.97
7.39
100
Males (M)
26.45
5.74
64
29.69
7.31
62
2.662
Females
31.25
7.59
36
25.18
6.74
38
-3.242
(F)
t = 3.643
M vs. F
t 0.06
t =-2.52*
Age at onset of psychotic symptoms (yr) Males Females
23.85
6.5
63
30.6
7.18
35
t = 4.623
M vs. F
3.77
1.66
90
3.19
1.49
90
-2.372
Males
3.57
1.74
58
3.28
1.46
57
-0.96
Females
4.12
1.77
32
3.03
1.57
33
-2.643
Parental socioeconomic
t = -1.44
M vs. F Perinatal
status
complications
t = -0.76
0.61
1.24
80
0.57
1.02
95
-0.26
Males
0.80
1.43
54
0.67
1.20
60
-0.52
Females
0.23
0.54
26
0.40
0.60
35
t = 1.952 1st walked
(mo)
1.51
t= 1.22
13.23
5.15
73
12.05
3.63
89
-1.69’
Males
13.58
5.47
50
12.31
4.41
55
-1.31
Females
12.46
4.39
23
11.63
1.78
34
-0.98
t = -0.86
M vs. F 1st words (mo)
t = -0.85
13.17
8.59
66
10.98
5.49
62
-1.70’
Males
13.42
9.61
46
11.71
5.96
35
-0.92
Females
12.57
5.76
20
10.04
4.75
27
-1.65’
t = -0.37
M vs. F 1st sentence
(mo)
t= -1.19
20.86
10.99
52
16.22
8.59
45
-2.292
Males
21.54
11.99
36
17.28
10.67
25
-1.47
Females
19.31
8.44
16
14.90
4.88
20
-1.96’
M vs. F
t = -0.92
t = -0.67 60.13
14.70
71
50.94
11.79
79
-4.243
Males
62.36
14.53
47
51.46
11.52
47
-4.033
Females
55.75
14.32
24
50.15
12.34
32
-1.57
1st read (mo)
M vs. F
t = -1.82’
t = -0.48
l.p
Psychiatry Research, 38:39-50
Elsevier
Familial Thyroid Disease and Delayed Language Development in First Admission Patients With Schizophrenia Lynn E. DeLisi, Angela M. Boccio, Henry Rliordan, Anne L. Hoff, Arlene Dorfman, Joyce McClelland, Maureen Kushner, Olga Van Eyl, and Neal Oden Received February 15, 1991; revised version received May 21, 1991; accepted June 4, 1991. Abstract. One hundred consecutive first admission patients with a DSM-III-R diagnosis of schizophrenia, schizoaffective disorder, or schizophreniform disorder were compared with 100 randomlly selected community controls. Childhood histories of physical, medical, and perinatal trauma, as well as physical and cognitive development, were examined by structured interviews with all available mothers of patients and controls. The prevalence of specific psychiatric disorders and several medical illnesses among first degree and more distant relatives was determined by family history questionnaires. The patient group did not have an excess of childhood head injuries, serious infections, or perinatal/ birth complications compared wit’h controls. With social class level taken into account, it was found that the acquisition of reading skills occurred significantly later in patients than controls. Family histories of schizophrenia and thyroid disorders were significantly more frequent among patients than controls. These data fail to indicate any childhood physical or medical environmental trauma that could lead to an increased risk for schizophrenia, although patients were substance abusers to a greater extent than controls. This study also confirms the already known contribution of familial factors and suggests an association of the inheritance of thyroid disorders with schizophrenia. Delayed development of reading skills suggests that precursers of illness may appear early in life before psychosis is evident. Key Words. Schizophreniform premorbid history.
disorder,
schizoaffective
disorders,
family history,
Although many studies suggest a genetic etiology for schizophrenia, the disorder evades the usual Mendelian characteristics of’ inheritance and has a relatively high discordance rate among monozygotic twins. Thus, the development of schizophrenia may depend on the coexistence of other risk factors in a genetically vulnerable individual. Some factors previously associated with psychosis have been obstetric and other perinatal complications (McNeil and Kaij, 1979; Parnas, 1986; Lewis,
Lynn E. DeLisi, M.D., Angela M. Boccio, MS., Henry Riordan, Ph.D., Anne L. Hoff, Ph.D., Arlene Dorfman, M.S.W., Joyce McClelland, R.N., Maureen Kushner, M.S.W., Olga Van Eyl, M.D., and Neal Oden, M.D., are in the Department of Psychiatry, SUNY Stony Brook, Stony Brook, NY, and the Stony Brook Unit at the Kings Park Psychiatric Center, Kings Park, NY. (Reprint requests to Dr. L.E. DeLisi, Dept. of Psychiatry, HSC, T-10, SUNY Stony Brook, Stony Brook, NY 11794, USA.) 0165-1781/91/%03.50
@ 1991 Elsevier Scientific
Publishers
Ireland
Ltd.
40
1987; Murray and Lewis, 1987), viral infections (i.e., influenza) prenatally or postnatally (DeLisi and Crow, 1986), and serious head trauma (DeLisi et al., 1986; Wilcox and Nasrallah, 1987). The genetic etiology of schizophrenia may also be linked to the gene or genes for other neuropsychiatric disorders or inherited medical conditions that are occasionally characterized by concurrent psychotic symptoms, such as Huntington’s disease, endocrine disorders, and several mental retardation syndromes. The presence of such a disorder among family members might serve as an additional risk factor for schizophrenia; in addition, detecting such a link might provide clues for future searches for a genomic locus for schizophrenia. Not only the development of schizophrenia, but also its onset, may be determined by the presence of premorbid stresses or brain insults (e.g., head injury or substance abuse). The precise onset of schizophrenic illness, however, is difficult to define. Whether early signs of abnormality, other than classical psychotic symptoms, are an indication of onset of the pathological process is controversial. Prospective studies of individuals at high genetic risk for schizophrenia have found that some childhood characteristics (e.g., delayed neuromotor development and attentional deficits; Watt et al., 1984; reviewed in Parnas and Mednick, 1991) consistently predict later development of schizophrenia. Early signs of neuromotor and cognitive difficulties may indicate that schizophrenia develops slowly from birth into adulthood. This report assesses risk factors in a large group of patients and controls that may be associated with an increased vulnerability for schizophrenia. It also explores whether those factors affect the timing of onset of overt psychotic symptomatology.
Methods One hundred patients consecutively admitted to an inpatient psychiatric evaluation program who were being seen for the first time for a schizophrenia-like psychosis were compared with 100 community controls. Patients were selected from approximately 3,500 first admissions screened from September 1987 through December 1990. They showed hallucinations, delusions, formal thought disorder, grossly bizarre behavior, or catatonia; had never been treated before for a psychiatric illness; were not classically manic; did not have a nonpsychotic depression; and did not have substance abuse that could account for the symptoms directly leading to hospitalization. All patients were admitted to an evaluation unit at the Kings Park Psychiatric Center, a New York State hospital that serves a county catchment area of approximately 1 million people. The patients were evaluated by structured clinical interviews (the Schedule for Affective Disorders and Schizophrenia, SADS; Spitzer and Endicott, 1978) history from relatives and clinical observation during the course of hospitalization. For the patient group, a DSM-III-R diagnosis (American Psychiatric Association, 1987) was established taking into account all relevant information from interviews, hospital records and multiple family members. Sixty-three patients were diagnosed as schizophreniform, 10 as subchronic schizophrenia, 8 as chronic schizophrenia, 16 as schizoaffective psychosis, and 3 as psychosis not otherwise specified. Severity of illness was estimated by a total score on the Brief Psychiatric Rating Scale obtained by a trained rater on admission and the amount of time spent in the hospital during the index episode. Control subjects were recruited from the lobby of a general public hospital (University Hospital, SUNY, Stony Brook) that serves the same catchment area as the psychiatric hospital. These persons were approached by a research interviewer at random and asked to participate in a short interview study. An attempt was made to obtain a group of control
41 subjects that was closely matched to the patients for age, sex, and social class. These individuals were interviewed by a structured medical questionnaire and asked for permission to contact their relatives. All available mothers of controls were then interviewed with the same questionnaire that was used for relatives of patients. Controls were eliminated if they were present in the hospital lobby for a psychiatric clinic appointment. Otherwise, the selection of controls was strictly random. These persons did not volunteer to be studied, but rather were approached if they were present in the lobby and appeared to be of similar age to the patient sample. The subjects were not screened for major psychiatric illness and thus represent a general random population. Very few persons refused to answer questions or to allow researchers to contact a relative, although the number of refusals was not recorded. The mean age of the patients was 28.11 (SD = 6.85) and that of the controls was 28.05 (SD = 7.35). There were 64 males and 36 females in the patient group, and 62 males and 38 females in the control group. Structured questionnaires focusing on birth complications, early childhood development, and significant medical or physical trauma were used in interviews with the subject’s mother or closest living relative if the mother was deceased (DeLisi, 1991). A structured family history questionnaire was also used to investigate several inherited metabolic, neurological, and psychiatric illnesses among first, second, and third degree relatives. If a positive history of a major psychiatric illness was solicited, further details about symptoms were ascertained to establish whether the relative met criteria for a D&W-III-R diagnosis of schizophrenia or affective disorder. Information was collected on birth/perinatal complications, childhood illnesses or head injuries, timing of developmental motor or language landmarks (age in months at first steps, first words, first sentence, and first words read), and history of substance abuse (by DSMZZZR criteria). Perinatal complications were scored using a previously published method (DeLisi et al., 1987, 1988). The following were considered complications and received 1 point each: bleeding, hypertension, toxemia, serious injury, and thyroid disease in mother during pregnancy, at least 2 weeks premature or postmature birth, breech delivery, cord around neck, convulsions, jaundice, use of oxygen or incubator, and infection or other perinatal illness. Any patient or control subject who admitted present or past substance abuse was further interviewed using the drug and alcohol section of the SADS to determine if he or she met DSM-III-R criteria for drug or alcohol abuse. For the patients, age of onset was defined as the age at which the patient or a family member first noticed positive psychotic symptoms. The socioeconomic status (SES) of the probands and their parents was estimated using the Hollingshead-Redlich Two-Factor Index of Social Position, which takes into account both educational and occupational level reached (reviewed in Myers and Bean, 1968). The parental SES level was considered to be the highest of either father or mother. Fifty-six of the 100 patients had magnetic resonance imaging (MRI) scans performed according to a previously described protocol (DeLisi et al., 199la, 19916). The volumes of the left and right lateral ventricles and the temporal lobes were measured on 5-mm coronal slices with a 2-mm gap. A trend for larger ventricles was present in this patient group compared with controls, while no temporal lobe differences were found. These measurements were used in correlations with the developmental variables examined in the present study. Data were analyzed using the Complete Statistical System software version B640 (01988, StatSoft, Inc.). Demographic and developmental variables were analyzed by independent t tests (2-tailed) to assess differences between groups. Because there were sex differences within groups and because parental SES differed between groups, two-way analyses of covariance (ANCOVAs, sex X group), with parental SES as the covariate, were conducted on the developmental variables (e.g., age first walked, age first spoke in sentences; see Table 2). Pearson (r) correlations were also computed between perinatal complications and developmental variables. Prevalences of family history of psychiatric and medical disorders, as well as potential clinical risk factors, were analyzed by ~2 statistics. Each family was treated as one unit; a family was considered positive if one or more relatives had illness. No additional weight was
42
given in the analysis to families with more than one ill relative. The data were further examined by path analysis to construct a pictorial representation of a variance matrix for all the variables studied (Li, 1975) and to see whether they contributed to the onset of pathology (i.e., delayed language development). A path diagram specifies the exact nature of a causal structure, and the strengths of the relationships are derived from the underlying variance matrix of the variables. Stepwise regression @ to add = 0.05, p to drop = 0.10) was used within the two sexes separately to specify the causal relations (single-headed arrows) in the diagram. For each regression, two different starting configurations were investigated, and variables were used in the final path diagrams if they were chosen from either starting configuration. After causal connections were stipulated, the strengths of connections were calculated and appear in the path diagram as path coefficients. All path coefficients were calculated by standardized regression through the origin. To obtain the most accurate estimates possible, the maximum number of (nonmissing) cases was used for each regression, so that the number of cases varied from one node to the other across the diagram. Also depicted in the path diagram are all observed correlations (double-headed arrows) with significance values < 0.10.
Results Perinatal Complications and Early Developmental Landmarks. Schizophrenic patients were described by their relatives as speaking their first sentences at a later age (2 = -2.29,~ < 0.05) and learning to read at a later age (t = -4.24,~ < 0.001) than controls (see Table 1). There were similar trends for age at first words and age when first walked, with schizophrenic patients reaching these milestones later (NS). Patients and controls did not differ in reported perinatal complications. A trend was present for left lateral ventricular volume to be positively correlated with the age at which subjects first read (r = 0.29, p = 0.06), although no correlations were significant for other developmental variables and brain morphological measurements. Severity of illness, as measured by the total BPRS score and time spent in hospital at the index admission, also did not correlate with any developmental measures. Two-way ANCOVAs (sex X group; Table 2) with parental social class as a covariate performed on the developmental dependent variables (see Table 1) revealed a main effect of sex on perinatal complications, with females having fewer complications than males (F= 3.81, p < 0.05) across both groups. As above, there was also a large main effect for group on the age at which subjects first read, with schizophrenic patients reading later than controls (F= 12.04, p < 0.001). No other main effects or sex X group interactions emerged. As can be seen from Table 3, a history of birth complications was positively correlated with age at first spoken sentence (r = 0.31, p < 0.05) within the schizophrenic group only, indicating that perinatal complications may adversely affect language development in schizophrenic patients. Premorbid
Childhood
and Later Environmental
in Table 4, schizophrenic patients p < 0.039) and drugs (~2 = 14.54,~
were
more
likely
Risk Factors. As can be seen to abuse
alcohol
h2
= 4.23,
< 0.0001) than controls. However, schizophrenic patients were no more likely than controls to have histories of head injury, childhood infections, hyperactivity, or learning disabilities.
43 Table 1. Characteristics of subjects investigated Controls
First-episode patients Mean
SD
n
Mean
SD
n
28.18
6.83
100
27.97
7.39
100
Males (M)
26.45
5.74
64
29.69
7.31
62
2.662
Females
31.25
7.59
36
25.18
6.74
38
-3.242
(F)
t = 3.643
M vs. F
t 0.06
t =-2.52*
Age at onset of psychotic symptoms (yr) Males Females
23.85
6.5
63
30.6
7.18
35
t = 4.623
M vs. F
3.77
1.66
90
3.19
1.49
90
-2.372
Males
3.57
1.74
58
3.28
1.46
57
-0.96
Females
4.12
1.77
32
3.03
1.57
33
-2.643
Parental socioeconomic
t = -1.44
M vs. F Perinatal
status
complications
t = -0.76
0.61
1.24
80
0.57
1.02
95
-0.26
Males
0.80
1.43
54
0.67
1.20
60
-0.52
Females
0.23
0.54
26
0.40
0.60
35
t = 1.952 1st walked
(mo)
1.51
t= 1.22
13.23
5.15
73
12.05
3.63
89
-1.69’
Males
13.58
5.47
50
12.31
4.41
55
-1.31
Females
12.46
4.39
23
11.63
1.78
34
-0.98
t = -0.86
M vs. F 1st words (mo)
t = -0.85
13.17
8.59
66
10.98
5.49
62
-1.70’
Males
13.42
9.61
46
11.71
5.96
35
-0.92
Females
12.57
5.76
20
10.04
4.75
27
-1.65’
t = -0.37
M vs. F 1st sentence
(mo)
t= -1.19
20.86
10.99
52
16.22
8.59
45
-2.292
Males
21.54
11.99
36
17.28
10.67
25
-1.47
Females
19.31
8.44
16
14.90
4.88
20
-1.96’
M vs. F
t = -0.92
t = -0.67 60.13
14.70
71
50.94
11.79
79
-4.243
Males
62.36
14.53
47
51.46
11.52
47
-4.033
Females
55.75
14.32
24
50.15
12.34
32
-1.57
1st read (mo)
M vs. F
t = -1.82’
t = -0.48
l.p
