We also conducted a number of sensitivity analyses. To examine whether our results were sensitive to the age threshold for EAU, we reran the discordant twin analyses using age cutoffs of ≤13 and ≤15 years. Due to power limitations, the SA model for ≤13 years could not be estimated, but for ≤15 years, the adjusted OR (3.45) was significant and highly comparable to the existing results. Adjusted ORs were significant for NSSI using both age cutoffs: ≤13 (OR=10.91) and ≤15 (OR=3.56), which is also consistent with the current findings. Notably, confidence limits around point estimates were extremely wide when ≤13 years was used for the EAU cutoff, suggesting that our choice of ≤14 years optimizes power while also capturing individuals with early onsets. Additional analyses examined whether results from the discordant twin models were sex or cohort specific. Neither the interaction of EAU with sex nor with cohort was significant indicating that the results generalize to both men and women and across both twin samples.
