Association Between Substance Use Disorder and Polygenic Liability to Schizophrenia

Abstract

Background There are high levels of comorbidity between schizophrenia and substance use disorder, but little is known about the genetic etiology of this comorbidity. Methods Here, we test the hypothesis that shared genetic liability contributes to the high rates of comorbidity between schizophrenia and substance use disorder. To do this, polygenic risk scores for schizophrenia derived from a large meta-analysis by the Psychiatric Genomics Consortium were computed in three substance use disorder datasets: COGEND (ascertained for nicotine dependence n=918 cases, 988 controls), COGA (ascertained for alcohol dependence n=643 cases, 384 controls), and FSCD (ascertained for cocaine dependence n=210 cases, 317 controls). Phenotypes were harmonized across the three datasets and standardized analyses were performed. Genome-wide genotypes were imputed to 1000 Genomes reference panel. Results In each individual dataset and in the mega-analysis, strong associations were observed between any substance use disorder diagnosis and the polygenic risk score for schizophrenia (mega-analysis pseudo R2 range 0.8%–3.7%, minimum p=4×10−23). Conclusions These results suggest that comorbidity between schizophrenia and substance use disorder is partially attributable to shared polygenic liability. This shared liability is most consistent with a general risk for substance use disorder rather than specific risks for individual substance use disorders and adds to increasing evidence of a blurred boundary between schizophrenia and substance use disorder.

Relationships between substance abuse/dependence and psychiatric disorders based on polygenic scores

“…This is in agreement with results from Hartz et al 19 who noted that the percentage of variance in any SUD diagnosis explained by SCZ2…”

“…Hartz et al 19 studied the three datasets included in SAGE, ascertained for nicotine, alcohol or cocaine dependence, in relation to the SCZ PGS. Regarding dependence, they found significant associations between major depressive disorder (MDD) PGS and severe cocaine dependence, and SCZ PGS and severe cannabis and cocaine dependence, after correction for multiple testing.…”

Genome-wide association meta-analysis of cocaine dependence: Shared genetics with comorbid conditions

“…To our knowledge, this is the first report of a shared genetic etiology between cocaine dependence and ADHD, antisocial behavior, risk-taking behavior and children's aggressive behavior based on genome-wide data. Previous studies have reported significant PRS associations between cocaine dependence and SCZ or MDD (Carey et al 2016;Hartz et al 2017;Reginsson et al 2018), and also between SUD and other psychiatric disorders (Du Rietz et al 2017;Gurriarán et al 2018), although our study used the largest sample of cocaine dependence for this type of analysis so far. This correlation can reflect biological pleiotropy, where similar genetic mechanisms influence more than one trait, or mediated pleiotropy, where one phenotype is causally related to a second phenotype, so that the variants associated with this phenotype are indirectly associated with the second one.…”

“…Such comorbidity is associated with an increase of severity for all disorders, although it is unclear whether this relationship is causal or the result of shared genetic and/or environmental risk factors. Some studies have started to inspect these relationships using both genetic correlation and polygenic risk score approaches, supporting the hypothesized role of shared genetic risk factors in the lifetime co-occurrence of several psychiatric disorders and SUD (Carey et al 2016;Hartz et al 2017;Du Rietz et al 2017;Reginsson et al 2018). …”

Exploring the relationship between polygenic risk for cannabis use, peer cannabis use and the longitudinal course of cannabis involvement

“…All the above analyses were conducted in R [43]. The use of ΔR [2] (typically Nagelkerke's pseudo-R 2 for binary traits [20]) as an index of the most predictive PRS relates to its role as an index of predictor efficacy [38,39,44], such that the addition of the PRS to a model improves the fit of the model, thus indicating unique variance attributable to the PRS, over and above covariates. We use the conditional R 2 to select the most predictive P T (see Supporting information), but report both statistics for the most predictive PRS threshold.…”