A multivariable Mendelian randomization to appraise the pleiotropy between intelligence, education, and bipolar disorder in relation to schizophrenia.

Education and intelligence are highly correlated and inversely associated with schizophrenia. Counterintuitively, education genetically associates with an increased risk for the disease. To investigate why, this study applies a multivariable Mendelian randomization of intelligence and education. For those without college degrees, older age of finishing school associates with a decreased likelihood of schizophrenia-independent of intelligence-and, hence, may be entangled with the health inequalities reflecting differences in education. A different picture is observed for schooling years inclusive of college: more years of schooling increases the likelihood of schizophrenia, whereas higher intelligence distinctly and independently decreases it. This implies the pleiotropy between years of schooling and schizophrenia is horizontal and likely confounded by a third trait influencing education. A multivariable Mendelian randomization of schooling years and bipolar disorder reveals that the increased risk of schizophrenia conferred by more schooling years is an artefact of bipolar disorder - not education.