The cross-sectional GRAS sample: a comprehensive phenotypical data collection of schizophrenic patients

Abstract

Background: Schizophrenia is the collective term for an exclusively clinically diagnosed, heterogeneous group of mental disorders with still obscure biological roots. Based on the assumption that valuable information about relevant genetic and environmental disease mechanisms can be obtained by association studies on patient cohorts of ≥ 1000 patients, if performed on detailed clinical datasets and quantifiable biological readouts, we generated a new schizophrenia data base, the GRAS (Göttingen Research Association for Schizophrenia) data collection. GRAS is the necessary ground to study genetic causes of the schizophrenic phenotype in a 'phenotype-based genetic association study' (PGAS). This approach is different from and complementary to the genome-wide association studies (GWAS) on schizophrenia.

Methods: For this purpose, 1085 patients were recruited between 2005 and 2010 by an invariable team of traveling investigators in a cross-sectional field study that comprised 23 German psychiatric hospitals. Additionally, chart records and discharge letters of all patients were collected.

Results: The corresponding dataset extracted and presented in form of an overview here, comprises biographic information, disease history, medication including side effects, and results of comprehensive cross-sectional psychopathological, neuropsychological, and neurological examinations. With >3000 data points per schizophrenic subject, this data base of living patients, who are also accessible for follow-up studies, provides a wide-ranging and standardized phenotype characterization of as yet unprecedented detail.

Conclusions: The GRAS data base will serve as prerequisite for PGAS, a novel approach to better understanding 'the schizophrenias' through exploring the contribution of genetic variation to the schizophrenic phenotypes.

Figure 1

Schizophrenia is a complex multigenetic disease. Schizophrenia may be seen as the result of a multifaceted interplay between multiple causative factors, including several genetic markers and a variety of different environmental risks. Cases with a critical genetic load may not need additional external/environmental co-factors, whilst in others, the interaction of a certain genetic predisposition with environmental co-factors is required for disease onset (modified from [84]).

Figure 2

Collaborating centers and patient numbers. Map of Germany displaying the locations of all 23 collaborating centers that were visited by an invariable team of traveling investigators. The table next to the map provides numbers of patients examined in each center. Some centers were visited more than once.

Figure 3

Patient recruitment and flow: (a) Recruitment efficiency 2005 - 2008. Cumulative numbers of recruited patients per quarter of the year are shown in bar graphs. Note that steady-state recruitment is ongoing. (b) Patient flow. Of 1085 patients examined, the diagnosis of schizophrenia or schizoaffective disorder could not be confirmed for 48. Instead, alternative diagnoses had to be given.

Figure 4

Development of the GRAS data bank. Raw data, brought to Göttingen by the traveling team of examiners, were only entered into the data base after careful and comprehensive data re-checking, also based on patient charts and discharge letters. During the whole process, continuous calibration sessions and repeated re-checking of the entered data took place.

Figure 5

Phenotype overview. Various different domains covered by the GRAS data collection are displayed. These domains will also deliver the basis for further sophistication of phenotypical readouts.

Figure 6

Clinical intercorrelation pattern. Correlations between measures of the clinical picture derived from different approaches: Patient self-ratings, clinical rater judgement and 'objective data'. Thickness of the lines represents the strength of correlation between two measures; only significant correlations are displayed. Note the strong internal consistency expressed by a Cronbach's alpha of .753.

Figure 7

Cognitive intercorrelation pattern. Shown are all neuropsychological tests performed, together with their respective cognitive domain. Thickness of the lines represents the strength of correlation between two tests; only significant correlations are displayed. Tests for higher cognitive functions are labelled in orange; tests for basic (mainly basic cognition/fine motor dependent) functions in grey. Measures of higher cognitive functions as well as measures of basic cognition/fine motor functions show powerful internal consistency (Cronbach's alpha of .819 and .801 respectively).

Figure 8

Extrapyramidal intercorrelation pattern. Shown are correlations between different neurological tests for measuring extrapyramidal symptoms. Thickness of the lines represents the strength of correlation between two tests; only significant correlations are displayed. Cronbach's alpha of .675 shows that these measures have a decent internal consistency.