Examining the validity of DSM-III-R schizoaffective disorder and its putative subtypes in the Roscommon Family Study.

OBJECTIVE: The authors sought to assess whether the DSM-III-R category of schizoaffective disorder differs meaningfully from schizophrenia and affective illness in clinical features, outcome, and familial psychopathology. In addition, the authors evaluated the validity of two proposed subtyping systems for schizoaffective disorder: 1) bipolar versus depressive (based on presence or absence of a full manic syndrome in the past) and 2) good versus poor interepisode recovery. METHOD: In the epidemiologically based Roscommon Family Study, index probands with diagnoses of schizophrenia or affective illness were selected from a case registry. Personal interviews were conducted with 88% of traceable, living probands and 86% of traceable, living first-degree relatives. RESULTS: Probands with schizoaffective disorder differed significantly from both those with schizophrenia and those with affective illness on lifetime psychotic symptoms as well as on outcome and negative symptoms assessed as follow-up. Relatives of probands with schizoaffective disorder had significantly higher rates of affective illness than relatives of schizophrenic probands and significantly higher rates of schizophrenia than relatives of probands with affective illness. Probands with bipolar and depressive schizoaffective disorder did not differ substantially with respect to psychotic symptoms, negative symptoms, outcome, or family history. Schizoaffective disorder probands with good interepisode recovery had fewer negative symptoms and a better outcome than those with poor recovery, but there were no significant differences in family history. Both the epidemiologic and family data are consistent with the hypothesis that schizoaffective disorder results from the co-occurrence of a high liability to both schizophrenia and affective illness. CONCLUSIONS: On the basis of the validators examined, DSM-III-R criteria for schizoaffective disorder define a syndrome that differs meaningfully from both schizophrenia and affective illness. The division of schizoaffective disorder into bipolar and depressive subtypes was, however, not validated. The separation of schizoaffective disorder into subtypes based on level of interepisode recovery defined subtypes that differed clinically but not with respect to familial psychopathology.

Schizotypal symptoms and signs in the Roscommon Family Study. Their factor structure and familial relationship with psychotic and affective disorders.

BACKGROUND: Although schizotypal personality disorder aggregates in relatives of schizophrenic probands, the criteria for this disorder may not be optimal either in describing the dimensions of schizotypal phenomena or in identifying those with a high familial liability to schizophrenia. METHODS: In the Roscommon Family Study, an epidemiologically based family study of major psychiatric disorders conducted in the west of Ireland, we examined 25 individual schizotypal symptoms and signs, assessed by structured personal interview, in 1544 first-degree relatives (without chronic psychosis or mental retardation) of five proband groups: schizophrenia; other nonaffective psychoses; psychotic affective illness; nonpsychotic affective illness; and matched, unscreened controls. RESULTS: We obtained seven meaningful schizotypal factors: negative schizotypy, positive schizotypy, borderline symptoms, social dysfunction, avoidant symptoms, odd speech, and suspicious behavior. Taken individually, all of these factors, except borderline symptoms, significantly discriminated relatives of schizophrenic probands from relatives of controls; in descending order of the odds ratios, they were odd speech, social dysfunction, suspicious behavior, negative schizotypy, avoidant symptoms, and positive schizotypy. In a multivariate analysis, four of these factors remained significant: odd speech, negative symptoms, social dysfunction, and avoidant symptoms. These schizotypal factors differed in their specificity. Three of the four most predictive schizotypal factors also significantly discriminated relatives of probands with other nonaffective psychoses from relatives of controls. CONCLUSION: "Schizotypy" is a complex, multidimensional clinical construct, whose various dimensions differ widely both in the degree and specificity with which they reflect the familial liability to schizophrenia. Subpsychotic thought disorder; negative schizotypal signs, such as poor rapport and odd behavior; deficient occupational functioning; and social isolation/avoidance best characterized relatives of schizophrenic probands compared with relatives of matched controls.

Hamilton Anxiety Rating Scale Interview guide: joint interview and test-retest methods for interrater reliability.

The Hamilton Anxiety Rating Scale (HARS) is the most widely used semistructured assessment scale in treatment outcome studies of anxiety. Interrater reliability coefficients for the HARS have been previously reported. However, differences in the way clinicians assess symptom severity may reduce reliability. A structured interview guide--The Hamilton Anxiety Rating Scale Interview Guide (HARS-IG)--was developed to standardize clinical probe questions and to minimize interrater variance. Joint-interview and test-retest methods of interrater reliability assessment were used in a group of 30 inpatients. Intraclass coefficient calculations revealed improved interrater agreement with the HARS-IG versus the HARS. The findings of this study demonstrate that the HARS-IG is a more reliable assessment instrument than the semistructured HARS and that it meets established standards of reliability assessment.

Independent diagnoses of adoptees and relatives as defined by DSM-III in the provincial and national samples of the Danish Adoption Study of Schizophrenia.

BACKGROUND: This report describes the independent application of DSM-III criteria to the adoptees and relatives in the Provincial sample of the Danish Adoption Study of Schizophrenia of Kety and colleagues. We report these results and combine them with those reported previously for the Copenhagen sample to form the National sample. METHODS: Personal interviews and institutional record summaries of adoptees and biological and adoptive relatives were "blindly" diagnosed using DSM-III criteria. "Schizophrenia spectrum" was a priori defined as schizophrenia; schizoaffective disorder, mainly schizophrenic subtype; and schizotypal and paranoid personality disorders. RESULTS AND CONCLUSION: In the Provincial sample, the prevalence of "spectrum" disorders was significantly greater in biological relatives of schizophrenia spectrum vs control adoptees. The results were also consistent with the genetic transmission of individual diagnoses within the spectrum. When combined into the National sample, the results provided strong evidence for (1) the genetic transmission of DSM-III schizophrenia; (2) a genetic relationship between DSM-III schizophrenia, mainly schizophrenic schizoaffective disorder, and schizotypal personality disorder; and (3) the absence of a significant genetic relationship between the schizophrenia spectrum and either psychotic nonspectrum disorders, major depression, or anxiety disorders. We found no evidence for the familial environment transmission of schizophrenia spectrum disorders. These results are consistent with the findings reported by Kety and coworkers from their diagnostic review.

Outcome and family study of the subtypes of schizophrenia in the west of Ireland.

OBJECTIVE: The authors sought to clarify differences in outcome and familial psychopathology among the classical subtypes of schizophrenia. METHOD: In the epidemiologically based Roscommon Family Study, personal interviews were conducted with 88% of traceable living probands (N = 415) an average of 16 years after illness onset and with 86% of traceable living first-degree relatives (N = 1,753). Probands meeting the DSM-III-R criteria for schizophrenia were subtyped by DSM-III-R and ICD-9. RESULTS: By both diagnostic systems, age at onset differed significantly across subtypes, being earliest in the subjects with the hebephrenic and catatonic subtypes and latest in the paranoid subjects. The probands with the paranoid subtype had substantially better outcome, especially in occupational functioning and capacity for self-care. The DSM-III-R criteria for paranoid schizophrenia were considerably more successful than the ICD-9 criteria in selecting good-outcome cases. Neither the risk for schizophrenia nor the risk for schizophrenia spectrum disorders significantly differed in relatives as a function of the proband subtype. The subtypes of schizophrenia did not "breed true" within families. CONCLUSIONS: Paranoid schizophrenia, especially when narrowly defined, as in DSM-III-R, has a substantially better outcome than other subtypes. From a familial perspective, 1) paranoid schizophrenia is not a milder form of schizophrenia and 2) catatonic schizophrenia is probably closely related to typical schizophrenia. The subtypes of schizophrenia are not, from a familial perspective, etiologically distinct syndromes.

Clinical heterogeneity in schizophrenia and the pattern of psychopathology in relatives: results from an epidemiologically based family study.

Individuals with schizophrenia vary widely in their symptoms, course of illness and outcome. Family background is the strongest known risk factor for schizophrenia. We know little of the relationship between clinical variability in this disorder and the level of familial vulnerability to schizophrenia and other major mental disorders. Therefore, in schizophrenic probands meeting DSM-III-R criteria (n = 126) from the epidemiologically based Roscommon Family Study, we systematically assessed 9 major symptoms, course, global outcome, Schedule for Negative Symptoms and the Levels of Functioning Scale. These clinical characteristics were related t the risk of mental disorders in first-degree relatives assessed by personal interview or hospital records (n = 354) utilizing both the "familial/sporadic" and the Cox proportional hazard models. Using either statistical method, no consistent and significant relationship was found between any of our measures of symptoms, course or outcome and the risk for schizophrenia or schizophrenia spectrum disorders in relatives. Similarly, no relationship was found between these clinical measures and the risk for affective illness, alcoholism or anxiety disorders. Our results are not consistent with previously articulated hypotheses that negative symptoms or poor outcome in schizophrenia reflect a high familial liability to illness. While familial factors contribute substantially to an individual's vulnerability to schizophrenia, our results suggest that once an individual is affected, these same factors do not strongly influence either the kinds of symptoms displayed or the course and outcome of the illness.

An epidemiologic, clinical, and family study of simple schizophrenia in County Roscommon, Ireland.

OBJECTIVE: The authors sought to estimate the prevalence of simple schizophrenia, compare the clinical presentations and courses of simple and "typical" schizophrenia, and examine psychopathology in first-degree relatives of probands with simple schizophrenia, probands with typical schizophrenia, and community comparison subjects. METHOD: The authors followed up all individuals with a recorded diagnosis of schizophrenia (N = 285) and 75% of those with a diagnosis of severe affective illness (N = 99) from the Roscommon County Case Register, which includes all individuals seeking psychiatric care in a rural county in western Ireland. The authors interviewed all available first-degree relatives of these groups and of matched unscreened community comparison subjects. RESULTS: Eleven cases of simple schizophrenia were diagnosed in the probands, for an estimated population prevalence and morbid risk in County Roscommon of 5.3 (SE = 1.6) and 6.2 (SE = 1.9) per 10,000, respectively. Individuals with typical schizophrenia (N = 126) had more marked delusions, hallucinations, and positive thought disorder; individuals with simple schizophrenia had more pronounced negative thought disorder and a more chronic course. Neither social/occupational functioning nor negative symptoms differed between the two groups. The risks for schizophrenia and all nonaffective psychoses were greater in the relatives of the probands with simple schizophrenia than in the relatives of the community comparison subjects. CONCLUSIONS: In this sample, simple schizophrenia was relatively rare, was rather debilitating, and resembled typical schizophrenia in presentation and course except for the absence of positive psychotic symptoms. From a familial perspective, simple schizophrenia appears to be related to typical schizophrenia.

The Roscommon Family Study. IV. Affective illness, anxiety disorders, and alcoholism in relatives.

OBJECTIVES: This report seeks to evaluate the specificity of the familial liability to schizophrenia by examining in the relatives of the various proband groups the risk for affective illness (AI), anxiety disorders, and alcoholism. DESIGN: A case-controlled epidemiologic family study using DSM-III-R criteria. PARTICIPANTS: Three hundred eighty-four index probands from a psychiatric case register, 150 unselected control probands from an electoral register and 2043 of their living and traceable relatives, of whom 1753 were personally interviewed. RESULTS: In personally interviewed relatives of schizophrenic probands, the lifetime risk for all AI (24.9% +/- 3.8%) or just bipolar AI (1.2% +/- 0.7%) was very similar to that found in interviewed relatives of controls (22.8% +/- 4.0% and 1.4% +/- 0.7%, respectively). However, the risk for all AI (49.7% +/- 12.9%) or bipolar AI (4.8% +/- 3.2%) was substantially increased in relatives of schizoaffective probands. A substantially higher proportion of relatives of schizophrenic vs control probands who had AI demonstrated psychotic--and specially mood-incongruent psychotic--symptoms when affectively ill. Neither the risk for anxiety disorders nor that for alcoholism was increased in relatives of schizophrenic vs control probands. CONCLUSIONS: The familial liability to schizophrenia possesses some specificity and does not substantially increase the risk to AI, anxiety disorders, or alcoholism. Even when narrowly defined, schizoaffective disorder has a substantial familial link to classic AI. The familial liability to schizophrenia predisposes to psychosis, and especially mood-incongruent psychosis, when affectively ill. Finally, these results do not support the hypothesis that, from a familial perspective, schizophrenia and AI are on a single etiologic continuum.

The Roscommon Family Study. III. Schizophrenia-related personality disorders in relatives.

OBJECTIVES: We sought to clarify the familial relationship between five putative schizophrenia-related personality disorders (schizotypal [SPD], paranoid, schizoid, avoidant, and borderline) and schizophrenia, other nonaffective psychoses, and affective illness. DESIGN: A case-controlled epidemiologic family study using DSM-III-R criteria. PARTICIPANTS: Five hundred thirty-four probands selected from a psychiatric case register or electoral register, of whom 415 were personally interviewed, and 2043 living and traceable relatives, of whom 1753 were personally interviewed. RESULTS: Compared with relatives of unscreened controls, relatives of probands with schizophrenia had a highly significantly increased prevalence of SPD, and modest, but significant, increased prevalences of paranoid, schizoid, and avoidant personality disorders. Borderline personality disorder was rare, with a modest clustering of cases in relatives of affective disorder probands. The prevalence of SPD was also significantly elevated in relatives of probands with SPD and with other nonaffective psychoses but not in relatives of probands with psychotic or nonpsychotic affective illness. In contrast to the pattern seen for schizophrenia, the prevalence rate of SPD was substantially greater in parents than in siblings of schizophrenic probands. CONCLUSIONS: Schizotypal personality disorder has a strong familial relationship with schizophrenia. Paranoid, schizoid, and avoidant, but not borderline, personality disorders may have a significant familial relationship with schizophrenia. Schizotypal personality disorder also reflects the familial liability to other psychotic disorders but probably not to affective illness. Fitness effects may substantially influence the pattern of schizophrenia-related personality disorders in relatives.

The Roscommon Family Study. II. The risk of nonschizophrenic nonaffective psychoses in relatives.

OBJECTIVE: We sought to clarify the familial relationship between the nonschizophrenic, nonaffective psychoses (schizoaffective disorder [SAD], schizophreniform disorder, delusional disorder, and atypical psychosis) and schizophrenia and affective illness (AI). DESIGN: A case-controlled epidemiologic family study using DSM-III-R criteria. RESULTS: Compared with relatives of unscreened controls, the risk of nonschizophrenic, nonaffective psychoses was significantly elevated in relatives of probands with schizophrenia, SAD, schizotypal personality disorder, and psychotic AI. No significant elevation in risk to these disorders was seen in relatives of probands with nonpsychotic AI. The risk for SAD alone was significantly increased in relatives of probands with psychotic or bipolar AI. CONCLUSIONS: The nonschizophrenic, nonaffective psychoses have a significant familial relationship with both schizophrenia and schizotypical personality disorder. Schizoaffective disorder, as defined by DSM-III-R, shares familial etiologic factors with at least some forms of AI.

The Roscommon Family Study. I. Methods, diagnosis of probands, and risk of schizophrenia in relatives.

OBJECTIVES: We sought to examine, in a rural county in the West of Ireland, the degree of familial relationship between schizophrenia and other nonaffective psychoses and affective illness (AI). DESIGN: A case-controlled epidemiologic family study using DSM-III-R criteria. PARTICIPANTS: This study included three proband groups: (1) all cases with a clinical diagnosis of schizophrenia from the Roscommon County Case Register born from 1930 onward (n = 285); (2) a random sample of cases from the register with a clinical diagnosis of severe AI (n = 99); and (3) a matched, random sample of Roscommon residents ascertained from the electoral register (n = 150). Face-to-face structured interviews were conducted with 86% of traceable, living relatives (n = 1, 753) and 88% of traceable, living probands (n = 415). RESULTS: In interviewed relatives, the lifetime risks (+/- SE) for schizophrenia, as a function of the "blind" proband diagnosis, were as follows: schizophrenia, 6.5% +/- 1.6%; schizoaffective disorder, 6.8% +/- 2.5%; schizotypal personality disorder, 6.9% +/- 3.9%; other nonaffective psychoses, 5.1% +/- 2.4%; psychotic AI, 2.8% +/- 1.2%; nonpsychotic AI, 0.6% +/- 0.6%; and control, 0.5% +/- 0.3%. Individuals with schizophrenia reproduced at a rate about one quarter that of controls and the risk for schizophrenia in parents of probands was much less than that found in siblings. CONCLUSIONS: These results support the following hypotheses: (1) in the West of Ireland, as in other populations, schizophrenia is a strongly familial disorder; (2) schizophrenia shares a familial predisposition with a spectrum of clinical syndromes that includes schizoaffective disorder, other nonaffective psychoses, schizotypal personality disorder, and probably psychotic AI, but not nonpsychotic AI; and (3) the diminished reproductive rates associated with schizophrenia have a large impact on the pattern of risk of illness in relatives.

A family study of the subtypes of schizophrenia.

The authors conducted a blind family study based on probands diagnosed as schizophrenic by DSM-III and subtyped by four diagnostic systems. The pattern of psychopathology in relatives did not differ as a function of the probands' subtype diagnosis. Relatives of paranoid and nonparanoid schizophrenic probands had similar risks for schizophrenia and affective illness. Resemblance in subtype diagnosis between probands and their affected relatives did not exceed chance expectation. The etiologic role of familial factors appears to be similar in each of the traditional subtypes of schizophrenia. No evidence was found for familial factors specific for individual subtypes.

A DSM-III family study of the nonschizophrenic psychotic disorders.

The authors conducted a blind DSM-III family study based on probands diagnosed from long-term follow-up information as having schizophreniform disorder, schizoaffective disorder, or psychotic affective illness. The pattern of psychopathology in relatives of schizophreniform probands closely resembled that found previously in relatives of schizophrenic probands. Relatives of schizoaffective probands had an excess risk for schizophrenia, other psychoses, and bipolar illness. The pattern of illness found in relatives of the probands meeting Research Diagnostic Criteria for mainly schizophrenic schizoaffective disorder appeared indistinguishable from that of relatives of schizophrenic probands. Relatives of probands with psychotic affective disorder had an excess risk for schizophrenia and for unipolar and bipolar affective disorder.

Reliability and concordance in the subtyping of schizophrenia.

The authors examined the reliability, frequency, concordance, and demographic characteristics of subtypes of schizophrenia in patients from the Iowa 500 study as defined by four major diagnostic systems: DSM-III, Research Diagnostic Criteria (RDC), ICD-9, and the Tsuang-Winokur criteria. Reliability was higher in diagnostic systems with operationalized than in those with unoperationalized criteria and consistently higher for the paranoid subtype. The frequency of individual subtypes varied widely for the different systems. Concordance for subtype diagnoses between systems ranged from quite high to quite low. Demographic characteristics of the individual subtypes were similar according to all systems.

Psychiatric illness in first-degree relatives of schizophrenic and surgical control patients. A family study using DSM-III criteria.

This report examines the risk for psychiatric illness in 723 first-degree relatives of schizophrenics and 1,056 first-degree relatives of matched surgical control patients. Diagnoses in patients and relatives were made "blind" to one another, using DSM-III criteria. Information on relatives was obtained from personal interview and/or hospital records. Results were analyzed using two levels of diagnostic certainty and with or without relatives on whom only hospital records were obtained. In all analyses, the risk for schizophrenia was significantly greater (at least 18-fold) in the relatives of schizophrenics v controls. Evidence was also found for an increased risk in relatives of schizophrenics for schizoaffective disorder, paranoid disorder, and atypical psychosis but not for unipolar disorder, anxiety disorder, or alcoholism. As defined by DSM-III, schizophrenia is a familial disorder; however, the increased risk for psychotic illness in relatives of schizophrenics does not appear to be confined to schizophrenia alone.

Subtype stability in schizophrenia.

The authors examine the long-term stability of the subtypes of schizophrenia defined by four diagnostic systems. When all patients were considered, agreement between subtype assigned at index and follow-up was modest. This agreement increased considerably when only patients diagnosed as paranoid, hebephrenic, or catatonic at both index and follow-up were considered. As for individual subtypes, stability was highest for paranoid schizophrenia, intermediate for hebephrenia, and virtually absent for undifferentiated schizophrenia. The stability of paranoid schizophrenia was greatest when onset occurred after age 30. As length of follow-up increased, a larger proportion of patients were diagnosed as undifferentiated or residual.

DSM-III schizophrenia: is there evidence for familial transmission.

To resolve some recent controversies as to whether there is any evidence of familial transmission in DSM-III schizophrenia, a blind study of first degree relatives of 332 DSM-III schizophrenia was conducted. A comparison group of first degree relatives of 318 surgical patients who were free of psychiatric symptoms were also studied. The morbidity risks of schizophrenia and related illnesses were estimated among these relatives. The risk of DSM-III schizophrenia in the relatives of schizophrenics was 3.7 + 0.7 which was significantly higher (p less than .01) than among the relatives of surgical controls (0.2% + 0.2) with a relative risk of 18.5 (3.7/0.2). When broad criteria of schizophrenia were used by including all schizophrenia-related illnesses, the risks were approximately 9.0% and 1.0% respectively in the families of schizophrenics and controls with a relative risk of 9.0. These results will be discussed in the light of relevant literature. It is concluded that there is an evidence for familial transmission in DSM-III schizophrenia.

An independent analysis of the Danish Adoption Study of Schizophrenia. VI. The relationship between psychiatric disorders as defined by DSM-III in the relatives and adoptees.

In this report, modified DSM-III criteria were applied to all the available interviews with adoptees from the greater Copenhagen sample of the Danish Adoption Study of Schizophrenia. In the adoptees, reasonable agreement was found between our DSM-III diagnoses and the original diagnoses using global DSM-II-based criteria by Kety et al for their categories of chronic and acute, but not borderline, schizophrenia. Comparing DSM-III-based diagnoses in adoptees and relatives, schizophrenia, schizotypal personality disorder, and paranoid personality disorder were all significantly more common in the biologic relatives of schizophrenic v screened control adoptees. These three diagnoses, which together form a tentative "schizophrenia spectrum," were also significantly concentrated in the biologic relatives of adoptees with schizoaffective disorder, mainly schizophrenic subtype, and schizotypal personality disorder, but not in biologic relatives of adoptees with schizophreniform disorder or atypical psychosis.

Outcome of schizophrenic subtypes defined by four diagnostic systems.

This report examines the short- and long-term outcome of the subtypes of schizophrenia as defined by four diagnostic systems: DSM-III, Research Diagnostic Criteria, Ninth Revision of the International Classification of Diseases, and the Tsuang-Winokur (T-W) criteria. Patients were from the Iowa 500 study and met Washington University criteria for schizophrenia. Subtype diagnosis was based on extensive chart material reviewed by investigators blind to outcome. Short-term outcome, based on chart information, and long-term outcome, based on individual field follow-up, were both better for paranoid than for nonparanoid schizophrenia, the difference being greatest using the T-W criteria. The difference in outcome between paranoid and nonparanoid schizophrenia was greater at long-term than at short-term follow-up, and greater using residential and occupational than psychiatric outcome criteria. Outcome did not differ for the two common forms of nonparanoid schizophrenia: hebephrenic and undifferentiated. The subtyping of schizophrenia has important predictive validity, which was greatest using the T-W criteria.