Is the P300 wave an endophenotype for schizophrenia? A meta-analysis and a family study
Introduction
Following recent replications of neuregulin and dysbindin as candidate genes for schizophrenia (Schwab et al., 2003, Stefansson et al., 2003), the search for susceptibility genes for psychosis is regaining momentum. Endophenotypes, also known as intermediate or alternative phenotypes, are anatomical, physiological, or biochemical variables that correlate with the main trait of interest (schizophrenia) but have a more straightforward genetic basis than clinical syndromes. Endophenotypes could improve the definition of psychotic syndromes and help to identify clinically unaffected gene carriers thus increasing the power of linkage and association studies (Gottesman and Gould, 2003, Wickham and Murray, 1997).
The value of the P300 wave in schizophrenia research has received further support from three meta-analyses of case-control studies, which confirmed that patients with schizophrenia have severe P300 amplitude reductions as well as moderate latency delays (Bramon et al., 2004a, Bramon et al., 2004b, Jeon and Polich, 2001, Jeon and Polich, 2003). However, whether similar anomalies are also present in their non-psychotic relatives remains unclear; this is a crucial question if the P300 is to be used as a psychosis endophenotype in molecular research.
Throughout the last 20 years, several studies have explored the P300 wave in non-psychotic relatives of schizophrenic patients with somewhat conflicting conclusions. The P300 amplitude was reported to be reduced in relatives and therefore a promising endophenotype (Frangou et al., 1997, Kidogami et al., 1991, Roxborough et al., 1993, Schreiber et al., 1992) as was its latency (Blackwood et al., 1991, Frangou et al., 1997). However, evidence on the contrary has also emerged (Blackwood et al., 1991, Blackwood et al., 2001, Kidogami et al., 1991, Winterer et al., 2003). Some studies may have been underpowered; in addition, P300 deviances may only be present in a proportion of relatives, and finally, potential publication biases may also have contributed to the existing confusion. Meta-analysis can overcome some of these problems by maximizing statistical power and assessing the extent of publication bias.
The overwhelming majority of the literature has examined the P300 wave in specific midline or temporal electrodes. The P300 at the FZ site, as well as other measures of frontal cognition have been put forward as endophenotypes for schizophrenia (Egan et al., 2001, Gallinat et al., 2003, Tsai et al., 2003). Finally, in addition to being altered in patients and unaffected gene carriers, endophenotypes must be heritable traits or at least show a substantial familial influence (Gottesman and Gould, 2003, Wickham and Murray, 1997).
This paper aims to: (i) Conduct a meta-analytic review of P300 studies exploring non-psychotic relatives of patients with schizophrenia. (ii) Examine the P300 wave in a new sample of families containing members with schizophrenia. (iii) Ascertain the degree of familial influence on the P300.
Section snippets
Meta-analysis of family studies
We conducted a meta-analysis of the literature comparing the P300 wave between non-psychotic relatives of patients with schizophrenia and controls. A search combining key words ‘P300’ and ‘relatives’ or ‘families’ and ‘schizophrenia’ was conducted in both Medline and in Science Citation Index from January 1983 to December 2003. Studies were included if they were peer-reviewed articles using a two-tone frequency oddball task. If means and variances were not reported, we requested these from the
Meta-analysis
We identified 11 studies suitable for analysis with 472 relatives and 513 controls. Another four studies exploring relatives of patients were published but did not fulfil inclusion criteria for analysis (Friedman et al., 1986, Friedman et al., 1988, Saitoh et al., 1984, Steinhauer et al., 1991). There was evidence of significant between-study heterogeneity both for the P300 amplitude (P < 0.001) and P300 latency (P = 0.02), thus supporting the use of random effects meta-analysis methods (Table 1
Is the P300 wave under familial influence?
We have observed a regional variation in familiality (the degree of similarity in P300 performance between family members compared to unrelated participants). In frontal areas (FZ), there seems to be virtually no familial contribution to the P300 waveform. Instead, as we examine central–parietal regions particularly PZ, familiality effects explain 31% and 30% of the variance in amplitude and latency, respectively (both significant). Studies of healthy twins show a strong genetic influence on
Acknowledgments
This study was supported by The Wellcome Trust and by the Schizophrenia Research Fund. E. Bramon and C. McDonald are Wellcome Trust Fellows. We thank the National Schizophrenia Fellowship (Re-think) for help with recruitment. Our special thanks go to all the families who volunteered to take part in this research.
References (53)
- et al.
Utility of WAIS-R short forms in schizophrenia
Schizophr. Res.
(1997) - et al.
Schizophrenia and affective disorders — cosegregation with a translocation at chromosome 1q42 that directly disrupts brain-expressed genes: clinical and P300 findings in a family
Am. J. Hum. Genet.
(2001) - et al.
Mismatch negativity in schizophrenia: a family study
Schizophr. Res.
(2004) - et al.
Meta-analysis of the P300 and P50 waveforms in schizophrenia
Schizophr. Res.
(2004) - et al.
Normal cerebral asymmetry in familial and non-familial schizophrenic probands and their unaffected relatives
Schizophr. Res.
(2004) - et al.
P300 delay and attenuation in schizophrenia: reversal by neuroleptic medication
Biol. Psychiatry
(1998) - et al.
Meta-analysis in clinical-trials
Control. Clin. Trials
(1986) - et al.
The Maudsley family study: 2. Endogenous event-related potentials in familial schizophrenia
Schizophr. Res.
(1997) - et al.
Event-related potentials in children at risk for schizophrenia during 2 versions of the continuous performance-test
Psychiatry Res.
(1986) - et al.
Auditory event-related potentials in children at risk for schizophrenia — the complete initial sample
Psychiatry Res.
(1988)