“…However, meta-analyzes present numerous methodological and statistical force problems. In particular, as far as we are concerned, they cannot discriminate the presence of certain confounding factors such as socio-demographic variables ( Wasek and Endicott, 1983 ; Cloninger, 1992 ; Cloninger et al, 1993 ; Peselow et al, 1995 ; Gutierrez-Zotes et al, 2004 ; Hansenne et al, 2005 ; Mikołajczyk et al, 2008 ; Chen et al, 2013 ) and the number of psychopathological episodes in the arc of life ( Hirschfeld et al, 1989 ; Zaninotto et al, 2015 ; Paolini et al, 2016 ). Finally, meta-analyzes should also include follow-up studies as these can help to disentangle the “state-trait” effect to determine if a specific character vulnerability model exists for the development of affective disorders ( Farmer et al, 2003 ).…”