Human aggression encompasses a wide range of behaviors and is related to many psychiatric disorders. We introduce the different classification systems of aggression and related disorders as a basis for discussing biochemical biomarkers and then present an overview of studies in humans (published between 1990 and 2015) that reported statistically significant associations of biochemical biomarkers with aggression, DSM-IV disorders involving aggression, and their subtypes. The markers are of different types, including inflammation markers, neurotransmitters, lipoproteins, and hormones from various classes. Most studies focused on only a limited portfolio of biomarkers, frequently a specific class only. When integrating the data, it is clear that compounds from several biological pathways have been found to be associated with aggressive behavior, indicating complexity and the need for a broad approach. In the second part of the paper, using examples from the aggression literature and psychiatric metabolomics studies, we argue that a better understanding of aggression would benefit from a more holistic approach such as provided by metabolomics. Key words: aggression; biomarkers; psychiatry; metabolomics
INTRODUCTIONArguably, "violence and aggression are the most serious problems facing humanity" [Chichinadze et al., 2011]. One approach in the study of aggression and its etiology is at a biochemical level.Biochemical markers are intermediate to the outward phenotype and the underlying biology of aggression. They can aid in elucidating the causes and (patho) physiology of aggressive behavior [James et al., 2006;Scoriels et al., 2015]. One of the promises of
719Neuropsychiatric Genetics biochemical studies of central nervous system diseases is the development of biochemically based biomarkers [KaddurahDaouk and Krishnan, 2009]. Biomarkers can be applied in diagnosis, in assessment of disease stages, as indicators of disease prognosis, and for prediction or monitoring of interventions and treatment responses. Some biomarkers are "surrogate endpoints" which substitute clinical endpoints that reflect "how a patient feels, functions, or survives" [Atkinson et al., 2001]. There are several comprehensive overviews of biochemical compounds as potential biomarkers for aggression [e.g., Siever, 2008;Yanowitch and Coccaro, 2011], including reviews of neurotransmitters [e.g., de Almeida et al., 2005;Seo et al., 2008;Siever, 2008;Wallner and Machatschke, 2009;Chichinadze et al., 2011;Yanowitch and Coccaro, 2011;Haller, 2013;Umukoro et al., 2013;Morrison and Melloni, 2014;Narvaes and Almeida, 2014;Willner, 2015], hormonal networks [Simpson and Hons, 2001;Wingfield et al., 2006;Siever, 2008;Soma et al., 2008;Chichinadze et al., 2011;Eisenegger et al., 2011; Haller, 2014a,c;Soma et al., 2015], and cytokines [e.g., Zalcman and Siegel, 2006]. With a few exceptions [e.g., Siever, 2008], these reviews focus on a single biochemical class, studying biomarkers belonging to various different biochemical pathways in isolation. This makes it...