We investigated risk factors for subclinical symptoms of psychosis, and focused on two psychosis dimensions previously identified in the Zurich Study, namely “schizophrenia nuclear symptoms” and “schizotypal signs”. We examined the data from 9814 Swiss conscripts from 2003. The psychosis symptom dimensions were derived from the Symptom-Checklist-90-R (SCL-90-R), and were regressed on a broad range of known risk factors for psychosis. Risk factors typically assigned to schizophrenia and other psychotic disorders – cannabis use, childhood adversity, reading and writing difficulties, attention deficit hyperactivity disorder (ADHD), psychiatric disorders and addiction in parents and the extended family – are relevant also at subclinical levels. Our analyses suggested that specific risk factors may be assigned to distinct psychosis dimensions, as previously determined in an analysis from the Zurich Study. If there are different pathways to psychosis characterized by specific symptom dimensions and risk factors, they mostly co-exist and interact at different symptom load levels.
In recent years, epidemiologic research on risk factors for psychotic disorders has experienced a new development. Incidence of psychotic disorders has been shown to vary not only by sex, but also by a series of trivial or ubiquitous factors such as urban or rural upbringing, ethnicity, and migration (Van Os, 2004). Other risk factors in connection with psychotic disorders have also come into focus, including childhood adversity (Bebbington et al., 2004 and Bak et al., 2005) and substance use, and in particular, cannabis use (Arseneault et al., 2002, van Os et al., 2002, Fergusson et al., 2003, Zammit and Lewis, 2004, Fergusson et al., 2005, Henquet et al., 2005 and Moore et al., 2007). The challenges here derive from the fact that these are all common risk factors with serious leverage effects despite presumably low risk enhancement.
An additional challenge has emerged from the continuum concept of psychotic disorders (Eaton et al., 1991, Kendler et al., 1996, van Os et al., 2000, Johns and van Os, 2001, Johns et al., 2004 and Hanssen et al., 2005): population surveys have commonly yielded prevalence rates of psychotic symptoms which are distinctly higher than clinician-assessed psychotic disorders, which means that most people who experience psychotic symptoms during their life do not develop a psychotic disorder (Van Os et al., 2009). As demonstrated from the longitudinal data of the Zurich Study, there are subgroups of people with persistent enhanced symptom-load, though at subclinical levels. In addition, there are subgroups with a decline of these levels. Persistent high or moderate symptom frequencies have a serious impact on the lives of afflicted individuals. Common consequences include conflicts with partners and other loved ones, problems at the workplace, unemployment, financial problems, and lastly, legal problems (Rössler et al., 2007).
From a methodological perspective, research focusing on low or moderate subclinical symptom load levels has several advantages as compared to research based on clinical psychotic disorders:
–
fewer comorbid disorders, and thus less biased effects;
–
less interference with antipsychotic and other drugs, which may modify symptoms and outcomes;
–
use of larger samples, which permits a better assessment of less frequent risk factors and pathogenetic mechanisms, as well as better statistical modeling opportunities;
–
and easier accessibility to participants, thus enabling population based studies.
In this study we examined moderate psychotic symptoms reported by nearly 10,000 young Swiss conscripts in 2003. In particular, we focused on broadly discussed risk factors for psychotic disorders such as substance use, childhood adversity, and history of mental disorders in parents and relatives (Maki et al., 2005 and Wicks et al., 2005). Firstly, we hypothesized that we would find similar risk factors, at low and moderate symptom load levels, to those that we had identified in a previous study (Rössler et al., 2007) at moderate and elevated symptom load levels. The reasoning is similar such as in most other psychiatric and somatic diseases: a better understanding of initial stages and subclinical symptoms is a clue to improving pathogenetic models. In fact, a great part of pathological processes and associated risk factors seems to take effect both above and below the diagnostic thresholds. Secondly, we aimed to differentiate the risk factors according to two psychosis dimensions; one representing thought alienation and hallucinations (“schizophrenia nuclear symptoms” dimension corresponding to the Schneiderian first rank symptoms), and the other comprising social and interpersonal deficiencies, ideas of reference, suspiciousness and paranoid ideation (representing a “schizotypal signs” dimension). According to the results of our previous study (Rössler et al., 2007) and theoretical considerations (Andreasen, 2000) we hypothesized that different risk factors and etiologies may lead to heterogeneous psychotic syndromes.
