# paper.pdf ## Metadata - PDFFormatVersion=1.7 - Language=en-GB - IsLinearized=false - IsAcroFormPresent=false - IsXFAPresent=false - IsCollectionPresent=false - IsSignaturesPresent=false - Author=Clara von Wendorff - Creator=Microsoft® Word 2019 - CreationDate=D:20251208163632+01'00' - ModDate=D:20251208163632+01'00' - Producer=Microsoft® Word 2019 - pdf:producer=Microsoft® Word 2019 - dc:creator=Clara von Wendorff - xmp:creatortool=Microsoft® Word 2019 - xmp:createdate=2025-12-08T16:36:32+01:00 - xmp:modifydate=2025-12-08T16:36:32+01:00 - xmpmm:documentid=uuid:67D887D7-B8C5-4E3A-831F-A239988F12DB - xmpmm:instanceid=uuid:67D887D7-B8C5-4E3A-831F-A239988F12DB ## Contents ### Page 1 Supplements for “Redox dysregulation, inflammation and mental healthin adults with high risk of childhood maltreatment” Figure S1: Flow Chart of Study Participants through the study baseline (MAZ. study), follow-up study (JAEL & JAELred study) with final sample sizes included in the current dataanalysis. ### Page 2 Table S1. Comparison included cases (N=115) versus non-included participants (N=482) atbaseline (MAZ.) regarding psychosocial characteristics and mental health burden. Notes. N=number of participants; IQ=intelligence quotient; total N of the tests performed vary between 218-592;1 assessed with the semi-structured clinical interview Kiddie-SADS (Dölitzsch et al., 2014; Seker, Boonmann,d'Huart, et al., 2022) 2 assessed with the semi-structured clinical interview SCID-II (d’Huart, Steppan, et al., 2022;Dölitzsch et al., 2014) 3 assessed with the ‘Essen Trauma-Inventory (ETI-KJ)‘ (Fischer et al., 2016), 4 Dimensionalpsychopathology was self-reported by participants at baseline with the ‘Youth Self Report (YSR)‘ and the ‚YoungAdult Self Report (YASR)‘ of the Achenbach scales, t-values reported (Dölitzsch, Kölch, Fegert, Schmeck, &Schmid, 2016; Seker, Bürgin, et al., 2022); 5 assessed with the Childhood Trauma Questionnaire (CTQ) sum score(Bernstein et al., 2003). 6 assessed with the Standard Progressive Matrices of Raven or the Culture Fair IntelligenceTest (Habersaat et al., 2018). ### Page 3 Table S2: ANOVA group comparison of childhood maltreatment severity and mental healthwith potential dichotomic covariables ### Page 4 Table S3: Pearson’s correlations between mental health, CM, OS markers, andinflammatory markers Note. Between n = 97 and n = 130, no correlation coefficient is statically significant. ### Page 5 Table S4: Pearson’s correlation coefficients between the biomarkers, CTQ, ASEBA andpotential continuous covariates Note. *** p < .001, ** p <.01, * p < .05, BMI= Body mass index,BH = significance after correction for multiple testing using theBenjamini–Hochberg false discovery rate (FDR) procedure.Between n = 86 and n = 131. ### Page 6 Table S5: ANOVA-based group comparison of oxidative stress and inflammationBiomarkers with dichotomous covariates ### Page 7 ### Page 8 ### Page 9 Note: The substance use was measured as use in the last 30 days. All means are z-standardizes. Table S6: Results of the regression model in which we predicted mental health problems(ASEBA total score) by CM (CTQ total score). Note. 1 1 = yes, 0 = no, *** p < .001, n = 96, p (BH) = p-valuesadjusted for multiple testing using the Benjamini–Hochbergfalse discovery rate (FDR) procedure. ### Page 10 Table S7: Results of the regression model in which we predicted GPx/GRed (log) by CM(CTQ total score). Note. 1 1 = yes, 0 = no, * p < .05, n = 87, p (BH) = p-valuesadjusted for multiple testing using the Benjamini–Hochbergfalse discovery rate (FDR) procedure. Table S8: Results of the regression model in which we predicted SOD by CM (CTQ totalscore). Note. 1 1 = yes, 0 = no, n = 114, p (BH) = p-values adjustedfor multiple testing using the Benjamini–Hochberg falsediscovery rate (FDR) procedure. Table S9: Results of the regression model in which we predicted GPx/GRed (log) byemotional abuse (CTQ subscale). Note. 1 1 = yes, 0 = no, * p < .05, n = 87, p (BH) = p-valuesadjusted for multiple testing using the Benjamini–Hochbergfalse discovery rate (FDR) procedure. ### Page 11 Table S10: Results of the regression model in which we predicted IL1RA(log) by CM (CTQtotal score). Note. 1 1 = yes, 0 = no, *** p < .001, n = 96, p (BH) = p-valuesadjusted for multiple testing using the Benjamini–Hochbergfalse discovery rate (FDR) procedure. Table S11: Results of the regression model in which we predicted IL-6(log) by CM (CTQtotal score) Note. 1 1 = yes, 0 = no, * p < .05, n = 96, for substance use 0 =“no use”, 1 = “use of the substance within last 30 days” , p (BH)= p-values adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR) procedure. ### Page 12 Table S12: Results of the regression model in which we predicted IL-10(log) by CM (CTQtotal score) Note. 1 1 = yes, 0 = no, n = 124, p (BH) = p-values adjusted formultiple testing using the Benjamini–Hochberg false discoveryrate (FDR) procedure. Table S13: Results of the regression model in which we predicted CRP(log) by CM (CTQtotal score). Note. 1 1 = yes, 0 = no, *** p < .001, n = 99, for substance use0 = “no use”, 1 = “use of the substance within last 30 days” , p(BH) = p-values adjusted for multiple testing using theBenjamini–Hochberg false discovery rate (FDR) procedure. ### Page 13 Table S14: Results of the regression model in which we predicted TNF𝞪 by CM (CTQ totalscore). Note. 1 1 = yes, 0 = no, n = 124, p (BH) = p-values adjustedfor multiple testing using the Benjamini–Hochberg falsediscovery rate (FDR) procedure. Table S15: Results of the regression model in which we predicted mental health problems(ASEBA total score) by TNF𝞪. Note. 1 1 = yes, 0 = no, * p < .05, ** p <.01, n = 94, p (BH) =p-values adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR) procedure. ### Page 14 Table S16: Results of the regression model in which we predicted mental health problems(ASEBA total score) by SOD. Note. 1 1 = yes, 0 = no, * p < .05, ** p <.01, n = 87, p (BH) =p-values adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR) procedure. TableS17: Results of the regression model in which we predicted mental health problems(ASEBA total score) by GPx/GRed (log). Note. 1 1 = yes, 0 = no, * p < .05, ** p <.01, n = 87, p (BH) =p-values adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR) procedure. ### Page 15 Table S18: Results of the regression model in which we predicted mental health problems(ASEBA total score) by IL1RA(log). Note. 1 1 = yes, 0 = no, ** p <.01, n = 94, p (BH) = p-valuesadjusted for multiple testing using the Benjamini–Hochbergfalse discovery rate (FDR) procedure. Table S19: Results of the regression model in which we predicted mental health problems(ASEBA total score) by IL-6(log). Note. 1 1 = yes, 0 = no, ** p <.01, n = 94, p (BH) = p-valuesadjusted for multiple testing using the Benjamini–Hochbergfalse discovery rate (FDR) procedure. Table S20: Results of the regression model in which we predicted mental health problems(ASEBA total score) by IL-10(log). Note. 1 1 = yes, 0 = no, * p < .05, ** p <.01, n = 94, p (BH) =p-values adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR) procedure. ### Page 16 Table S21: Results of the regression model in which we predicted mental health problems(ASEBA total score) by CRP(log). Note. 1 1 = yes, 0 = no, * p < .05, ** p <.01, n = 96, p (BH) =p-values adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR) procedure. Table S22: Pearson’s correlations between biomarkers and subtypes of CM (CTQsubscales) and mental health problems (ASEBA subscales) Note. n = between 115 and 130, * p<.05., p-value no longer significant after Benjamini–Hochberg FDR correction for multiple testing. ### Page 17 Table S23: Differences in OS and inflammation level between participants with and withoutany psychiatric diagnosis (HITOP). Note. All biomarkers were z-standardized.