“…The authors of that study also found that the subjects' thinness was related to PCOS symptoms and to IR (Davies et al 2012; Table 3). These apparent regional differences may result from the limitations of these studies, which include the use of varying definitions of PCOS (Cresswell et al 1997, Laitinen et al 2003, Sadrzadeh et al 2003, Ibáñ ez et al 2008, Hizli et al 2012, Mumm et al 2013, the inclusion of women using hormonal contraceptives (Ibáñ ez et al 2007, Davies et al 2012, Mumm et al 2013, the small numbers of participants (Ibáñez et al 2001, Pandolfi et al 2008, the use of self-reported birth data (Laitinen et al 2003, Sadrzadeh et al 2003, Legro et al 2010, Hizli et al 2012, the inclusion of women with immature HPG axes (Ibáñ ez et al 2001) and the absence of compensatory growth assessments (Cresswell et al 1997, Michelmore et al 2001, Laitinen et al 2003, Sadrzadeh et al 2003, Ibáñez et al 2008, Pandolfi et al 2008, Legro et al 2010, Melo et al 2010, Hizli et al 2012, Mumm et al 2013, Shayeb et al 2014. The use of birth weight classifications as an exposure factor across studies was perhaps the most important limitation, because not all of the studies adjusted the subjects' birth weight for GA. For that reason, some of the studies included preterm newborn infants.…”