BACKGROUND: Youth with conduct disorder (CD) and high callous-unemotional (CU) traits are not a homogenous group and can be disaggregated into primary and secondary subgroups. However, there are inconsistencies in defining primary and secondary subgroups, with some studies using anxiety, others using maltreatment and still others using both features to identify subgroups. There is a paucity of work comparing primary and secondary subgroups with typically developing (TD) youth on experiences of maltreatment and parenting as well as a lack of studies investigating sex differences.

METHODS: In a large sample of TD youth (n = 946, 66% female) and youth with CD (n = 885, 60% female), we used latent profile analysis in youth with CD aged between 9 and 18 years to address four aims: (i) to demonstrate how primary and secondary subgroup membership differs when anxiety, maltreatment, or both are used as continuous indicators, (ii) to compare primary and secondary subgroups with TD youth on abuse and neglect measures, and (iii) to compare primary and secondary subgroups with TD youth on parenting experiences, and (iv) to examine whether the results were consistent across sexes.

RESULTS: Anxiety without maltreatment yielded the best fitting and most theoretically interpretable classification of primary and secondary subgroups across both sexes (Bayesian information criterion = 17832.33, Entropy = 0.75, Lo-Mendell-Rubin: p < 0.01). Compared with TD youth, youth with primary and secondary CU traits experienced greater levels of abuse and neglect (p < 0.001, η 2 p = 0.04-0.16) and maladaptive parenting practices (p < 0.001, η 2 p = 0.04-0.13). Youth with primary and secondary CU traits were equally high on levels of abuse, neglect, and maladaptive parenting (all p values >0.05).

CONCLUSIONS: We provide evidence that anxiety and maltreatment cannot be used interchangeably to identify youth with primary versus secondary CU traits. Anxiey yielded the best fitting and most theoretically interpretable classifications across both sexes. Our results signify the need for researchers and clinicians to adopt a unified approach to defining primary and secondary subgroups of CU traits using anxiety in both sexes.