Re-evaluating classical body type theories: genetic correlation between psychiatric disorders and body mass index

Body type theories, proposed by Kretschmer and Sheldon, are historical concepts that attempt to correlate somatotypes and personalities or psychiatric disorders (Kretschmer, 1925; Sheldon, 1940). Accordingly, Kretschmer classified four types of people: (1) the asthenic type who has a slender body (‘leptosome’) and is more prone to schizophrenia (SCZ); (2) the pyknic type who has a round body and is likely to become manic-depressive illness [bipolar disorder (BD)]; (3) the athletic type with a muscular body who may suffer from epilepsy; and (4) the dysplastic type who cannot be classified as any of the other three types (Kretschmer, 1925). Sheldon had a similar insight, classifying people into ectomorphic (corresponding to Kretschmer's asthenic type), endomorphic (corresponding to the pyknic type), and mesomorphic (corresponding to the athletic type). Interestingly, Sheldon also claimed that there was a deeper connection, with genetic components being linked to both the somatotypes and personality (Sheldon, 1940). Such classifications proposed by these celebrated psychiatrists or psychologists were based on their observational surveys in the beginning of the twentieth century; however, scientific evidence thereof is limited (Sorensen et al. 2006; Zammit et al. 2007). Nevertheless, recent genetic studies have re-evaluated their body type theories by conducting genetic correlation analysis, which employs genetic association results from genome-wide association studies (GWASs) to examine the genetic relationship between two independent phenotypes, that is, theoretically, excluding environmental factors. The first study reported a significant correlation between SCZ and low body mass index (BMI) in a European (EUR) ancestry (Bulik-Sullivan et al. 2015); however, no correlation was observed between BD and BMI (Bulik-Sullivan et al. 2015). Following this, we conducted a study in which we replicated a significant negative correlation between SCZ and BMI, but again found no correlation between BMI and BD in the Japanese samples (Akiyama et al. 2017). More recently, the PGC2 major depressive disorder (MDD) group has reported a positive correlation between BMI and MDD (Major Depressive Disorder Working Group of the PGC, 2017), using EUR samples.