Characteristics and heterogeneity of schizoaffective disorder compared with unipolar depression and schizophrenia - a systematic literature review and meta-analysis.

BACKGROUND: Comparisons of illness characteristics between patients with schizoaffective disorder (SAD) patients and unipolar depression (UD) are rare, even though UD is one of the most important differential diagnoses of SAD. Also, the variability of illness characteristics (heterogeneity) has not been compared. We compared illness characteristics and their heterogeneity among SAD, UD, and - as another important differential diagnosis - schizophrenia (S). METHODS: In order to reduce sampling bias we systematically searched for studies simultaneously comparing samples of patients with SAD, UD, and S. Using random effects and Mantel-Haenszel models we estimated and compared demographic, illness course and psychopathology parameters, using pooled standard deviations as a measurement of heterogeneity. RESULTS: Out of 155 articles found by an earlier meta-analysis, 765 screened in Medline, 2738 screened in EMBASE, and 855 screened in PsycINFO we selected 24 studies, covering 3714 patients diagnosed according to RDC, DSM-III, DSM-IIIR, DSM-IV, or ICD-10. In almost all key characteristics, samples with schizoaffective disorders fell between unipolar depression and schizophrenia, with a tendency towards schizophrenia. On average, UD patients were significantly older at illness onset (33.0 years, SAD: 25.2, S: 23.4), more often women (59% vs. 57% vs. 39%) and more often married (53% vs. 39% vs. 27%). Their psychopathology was also less severe, as measured by BPRS, GAS, and HAMD. In demographic and clinical variables heterogeneity was roughly 5% larger in UD than in SAD, and samples of patients with schizophrenia had the lowest pooled heterogeneity. A similar picture emerged in a sensitivity analysis with coefficient of variation as the measurement of heterogeneity. LIMITATIONS: Relative to bipolar disorder there are fewer studies including unipolar patients. No studies based on DSM-5 could be included. CONCLUSIONS: Regarding unipolar affective disorder this study confirms what we have shown for bipolar disorders in earlier studies: schizoaffective disorder falls between schizophrenia and affective disorders, and there are relevant quantitative differences in key illness characteristics, which supports the validity of the schizoaffective disorder concept. Contrary to our expectations heterogeneity is not larger in SAD than in UD and not substantially higher than in S. Lower reliability of the diagnosis of SAD therefore cannot be ascribed to higher variability of illness characteristics in SAD.

Schizoaffective disorder diagnosed according to different diagnostic criteria--systematic literature search and meta-analysis of key clinical characteristics and heterogeneity.

BACKGROUND: Schizoaffective disorder is viewed as a heterogeneous diagnosis among psychotic illnesses. Different diagnostic systems differ in their definition with DSM (-IIIR, -IV, and -V) providing a narrower definition than RDC and ICD-10. It is unclear whether this difference is reflected in patient samples diagnosed according to different diagnostic systems. METHODS: Exploratory study based on a systematic review of studies of schizoaffective disorder samples diagnosed by either RDC and ICD-10 (group of "broad criteria") or DSM-IIIR and -IV ("narrow criteria"); comparison (by Mann-Whitney-U-tests) of key characteristics, such as age, number of hospitalizations, or scores in psychometric tests, between more broadly and more narrowly defined schizoaffective disorder samples using standard deviations as a measurement of heterogeneity as well as weighted means and percentages. To reduce selection bias only studies including schizoaffective patient samples together with affective disorder and schizophrenia samples were selected. RESULTS: 55 studies were included, 14 employing RDC, 4 ICD-10, 20 DSM-IIIR, and 17 DSM-IV. Thirteen characteristics were compared: patients diagnosed according to broader criteria had fewer previous hospitalizations (2.2 vs. 5.4) and were both less often male (42 vs. 51%) and married (21 vs. 40%). Heterogeneity was similar in both groups but slightly higher in RDC and ICD-10 samples than in DSM-IIIR and -IV-samples: +4% regarding demographic and clinical course data and +13% regarding psychometric tests (pooled SD). LIMITATIONS: Secular trends and different designs may have confounded the results and limit generalizability. Some comparisons were underpowered. CONCLUSIONS: Differences in diagnostic criteria are reflected in key characteristics of samples. The association of larger heterogeneity with wider diagnostic criteria supports employing standard deviations as a measurement of heterogeneity.

Characteristics of patients diagnosed with schizoaffective disorder compared with schizophrenia and bipolar disorder.

OBJECTIVES: Information on basic demographic and clinical characteristics of schizoaffective disorder is sparse and subject to sampling bias and low diagnostic reliability. In the present study we aimed to: (i) estimate the demographic and clinical descriptors in schizoaffective disorder patients and (ii) compare the findings with those with schizophrenia and bipolar disorder. METHODS: To minimize sampling bias and low reliability, we systematically reviewed studies that simultaneously compared schizoaffective, schizophrenia, and bipolar disorder patients. We estimated demographic, clinical, and psychometric characteristics based on weighted pooling, and compared disorders by meta-analysis. We also estimated whether schizoaffective disorder is closer to schizophrenia or to bipolar disorder. RESULTS: We identified 50 studies that included 18312 patients. Most characteristics of the 2684 schizoaffective disorder patients fell between those of 4814 diagnosed with bipolar disorder and 10814 with schizophrenia. However, the schizoaffective group had the highest proportion of women (52%), had the youngest age at illness onset (23.3 ± 3.8 years), and had the highest standardized ratings of psychosis and depression. Differences in pooled parameters between schizoaffective versus schizophrenia and versus bipolar disorder subjects were similar. Values for patients with schizoaffective disorders mostly were intermediate between schizophrenia and bipolar disorder. However, the majority of studies showed schizoaffective patients to be more like schizophrenia than bipolar disorder patients in seven out of nine demographic and clinical categories as well as in five out of eight psychometric measures. These results remained similar when we restricted the analyses to studies with psychotic bipolar disorder patients only or to studies using the Diagnostic and Statistical Manual of Mental Disorders (DSM)-IIIR and DSM-IV only. CONCLUSIONS: The present study provided estimates of important characteristics of schizoaffective disorder - as balanced as possible in summarizing the findings from observational studies as unbiased as possible. The results did not support the hypothesis that schizoaffective disorder is primarily an affective disorder. The stronger resemblance of schizoaffective disorder to schizophrenia than to bipolar disorder needs further investigation.