Customize your JAMA Network experience by selecting one or more topics from the list below.
Vonberg FW, Bigdeli TB. Genetic Correlation Between Schizophrenia and Epilepsy. JAMA Neurol. 2016;73(1):125–126. doi:10.1001/jamaneurol.2015.3480
Neuropathological, clinical, and epidemiological data suggest that schizophrenia and epilepsy are associated.1 Reported estimates of the prevalence of schizophrenia among people with epilepsy vary, depending on phenotypic definition, but may be around 7%.2 One hypothesis to account for the relationship is that the 2 diseases share a common etiology. Methodological advances now make it possible to test the extent to which genetic predisposition is common to the 2 conditions based on molecular genetic data.3 We sought to do so using publically available genome-wide association study (GWAS) summary statistics from large meta-analyses of schizophrenia4 and epilepsy.5 In this retrospective case-control analysis, we used a technique called linkage disequilibrium (LD) score regression (known as LDSC regression) to estimate the genetic correlation (rG) between these common disorders.3
The International League Against Epilepsy meta-analysis of GWAS included data on 8696 people with epilepsy of all types and 26 157 control individuals. Data were also included on the subtypes of genetic generalized (n = 2606) and focal (n = 5310) epilepsy. The Psychiatric Genetics Consortium meta-analysis of schizophrenia GWAS included 13 833 cases and 18 310 control individuals. The LDSC regression provides an estimation of rG between 2 diseases based on the effect size of each single-nucleotide polymorphism shared by the 2 traits and incorporates the appropriately weighted effect size of all other single-nucleotide polymorphisms with which it is in LD. The calculation also includes the sample size for each study and the degree of sample overlap between the studies, which in this case was zero. Because sample overlap can impair the ability of this method to detect genetic correlation, we did not use the most recent Psychiatric Genetics Consortium meta-analysis of schizophrenia GWAS because this study shared some control individuals with those of the epilepsy GWAS.
Results are shown in the Table. There was a positive genetic correlation between schizophrenia and epilepsy (all subtypes) of 0.22 (SE, 0.07; P = .001). The heritability for schizophrenia was 0.30 (SE, 0.02). All heritability estimates are presented on the liability scale.
In this study, the LDSC regression has revealed a statistically significant positive association between schizophrenia and epilepsy (all subtypes). The individual significant positive rG for schizophrenia with focal epilepsy, although it does not survive Bonferroni correction for multiple comparisons, could be taken to suggest that it is this subtype of epilepsy driving the overall significant positive correlation.
The value for epilepsy heritability of 0.05 calculated by LDSC here is significantly lower than values calculated previously using alternative methods.6 This is likely attributable in part to the genomic control correction applied to each constituent study of the epilepsy meta-analysis data. This biases estimates of heritability downwards without affecting the value for genetic correlation.3 The schizophrenia data set did not undergo genomic control correction and accordingly the heritability reported here is more in keeping with previously published estimates. We would also note that the complete epilepsy data set included both genetic generalized epilepsy and focal epilepsy, and the low heritability estimate could potentially be explained by heterogeneity among these cases. However, neither of these limitations is likely to produce a falsely significant positive result for genetic correlation.
The power of LDSC lies in the fact that it only requires summary statistics, rather than individual-level genotype data, to estimate trait heritability and genetic correlation. Estimations of correlation can provide insights into shared biology at the molecular level and are especially useful where environmental confounders might otherwise be thought to link 2 diseases. A link between schizophrenia and epilepsy has been the subject of interest and controversy since it was noted early in the 20th century that there was some apparent phenotypic overlap between them. However, whether this link represents a shared etiology had not previously been clarified. Here, we have provided an initial demonstration of a significant shared liability to schizophrenia and epilepsy, suggesting that the relationship between the 2 disorders occurs at the level of the genome.
Corresponding Author: Frederick W. Vonberg, MA, MBBS, Boston Children’s Hospital, Harvard Medical School, 300 Longwood Ave, Boston, MA 02115 (firstname.lastname@example.org).
Published Online: November 9, 2015. doi:10.1001/jamaneurol.2015.3480.
Author Contributions: Dr Vonberg had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Both authors.
Acquisition, analysis, or interpretation of data: Both authors.
Drafting of the manuscript: Both authors.
Critical revision of the manuscript for important intellectual content: Both authors.
Statistical analysis: Both authors.
Administrative, technical, or material support: Vonberg.
Study supervision: Bigdeli.
Conflict of Interest Disclosures: None reported.
Funding/Support: The National Health Service of England and Oxford University Clinical Academic Graduate School provided funding for this study.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank Jonathan Flint, BM, BCh (Wellcome Trust Centre for Human Genetics, Oxford University), for review of the manuscript. He did not receive compensation for his contribution.
Create a personal account or sign in to: