The minimum amount of data for factor analyses was not satisfied, with a final sample size of 101 and 117 unique inflammatory analytes, there was less than 1 case per variable. Therefore, due to both low sample size and low levels of correlation between inflammatory variables EFA was not feasible for this sample. Subsequent analyses were completed using the NAPLS2 15-analyte z-score index as well as the individual analytes previously identified to be meaningfully associated with childhood trauma and psychosis symptom severity in existing research. This study sought to examine the associations between childhood trauma, psychosis risk symptoms, functioning, and inflammatory analytes in a sample of 67 CHR subjects and 34 UCs, thus contributing to a growing body of literature examining the effect of early life adversity on later life outcomes in individuals at risk for psychosis and identifying associated biological mechanisms. More specifically, this study sought to examine whether inflammation mediates the relationship between childhood trauma and clinical outcomes in individuals at risk for psychosis. This is one of few studies to explore whether inflammatory analytes mediate the relationship between childhood trauma and psychosis risk symptoms or functioning in CHR subjects, providing important implications for clinical intervention in youth at risk for psychosisby helping to uncover biological mechanisms for therapeutic intervention. However, this is the first study to evaluate this hypothesis using an enhanced sample of CHR subjects. The first primary aim of this study was to confirm known group differences between CHR and UC subjects in experience of childhood trauma, psychosis risk symptoms, and functioning,vertical grow racking system as well as evaluate groups differences across inflammatory analytes known to be associated with childhood trauma. First, we hypothesized that CHR subjects would demonstrate higher levels of pro-inflammatory markers known to be associated with experience of childhood trauma as well as the 15-Analyte Index developed by Perkins et al. relative to UC subjects.
Interestingly, CHR individuals in this sample demonstrated lower levels of TNF-a as compared to UCs, which is inconsistent with research demonstrating significantly elevated baseline blood plasma levels of TNF-a, CRP, and IL-6 in CHR individuals who endorsed a history of trauma . Significant differences were not revealed for Cortisol, CRP, IL-6 or the 15-Analyte Index between UC and CHR groups. The non-significant findings of differences in the 15-Analyte Index is attributable to the index being derived to discriminate between subjects who progress to psychosis, versus those that do not, and unaffected individuals, thus grouping the CHR subjects together resulted in non-significant differences between CHR and UC groups. Further, TNF-a is a pro-inflammatory cytokine, and thus, is involved in the initiation and aggravation of inflammatory responses, including cell apotosis. Interestingly, the biology of TNF in the brain allows for it to both protect neurons, as well as initiate their destruction through different protein activation processes . Although we are unable to evaluate this type of process in the current dataset, the observed decreased levels of TNF-a between CHR and UC subjects may relate to a meaningful narrative of complex inflammatory activation and suppression processes associated with enduring effects of childhood trauma such as increased stress reactivity in individuals at risk for psychosis . For example, Jeffries et al. reports that as compared to CHR-NC, CHR subjects who convert to psychosis demonstrate a striking loss of complexity in analyte correlation networks that could be prognostic, indicating that network imbalance in pro-inflammatory suppression and activation processes is an important feature of in understanding progression to psychosis. Second, we hypothesized that CHR subjects would demonstrate significantly higher incidence of childhood trauma, greater severity of psychosis risk symptoms, as well as lower global, social, and role functioning as compared to UC subjects. Consistent with our hypothesis, this sub-sample of CHR participants demonstrated significantly higher overall psychosis risk symptoms severity, as measured by the SOPS, and lower functioning on the GAF, GFS, and GFR as compared to UC subjects.
By definition, CHR individuals experience more psychosis risk symptoms and lower functioning as compared to unaffected individuals . Thus, the findings that CHR subjects in this sample demonstrated higher psychosis-risk symptoms and lower functioning is to be expected. More importantly, as compared to UC, CHR subjects in this sample demonstrated significantly higher total unique trauma, as well as higher incidence of trauma on most individual trauma sub-types, including, psychological bullying, physical bullying, emotional neglect, psychological abuse, and physical abuse. Three-fourths of CHR subjects in this sample reported history of childhood trauma, which is consistent with previous reports that prevalence of childhood trauma in individuals at risk for psychosis may be up to 90% ; however, this proportion is higher than larger sample from which this data was derived with approximately 60% of all CHR subjects reporting history of trauma. Thus, results from this study replicate previously demonstrated findings from Addington et al. , that CHR subjects experienced greater total number of unique trauma and bullying than UC subjects; however, this sub-sample of participants also demonstrates significant differences in emotional neglect, psychological abuse, and physical abuse, which was not demonstrated in the larger sample. Although only CHR subjects demonstrated a history of sexual abuse, the group differences between CHR and UC subjects was not significant. We can hypothesize that since the current sample of CHR subjects is enriched with a higher proportion of individuals who are known to have converted to psychosis , that increased childhood trauma is associated with poorer clinical outcomes in this CHR sample, which was further explored in Aim 4.
Finally, we hypothesized that CHR subjects who experienced history of childhood trauma would demonstrate higher levels of pro-inflammatory markers known to be associated with experience of childhood trauma as well as the 15-Analyte Index developed by Perkins et al , higher levels of baseline psychosis symptom severity, and lower baseline global/social/role functioning relative to CHR subjects with no history of childhood trauma and that these differences would vary by trauma sub-type. Inconsistent with our hypothesis, no differences were observed between CHRTrauma and CHRNoTrauma subjects in levels of the 15-Analyte Index, Cortisol, CRP, TNF-a, or IL-6. However, when analyzing groups by sub-type of trauma, it was revealed that CHR individuals who endorsed psychological bullying, physical abuse,vertical growing cannabis and sexual trauma also demonstrated higher total incidence of trauma as compared to CHR subjects that did not endorse one of those sub-types of trauma. Further, CHRTrauma subjects who endorsed a history of psychological bullying demonstrated significantly lower Cortisol than CHRTrauma individuals with no history of psychological bullying. While blunted morning salivary Cortisol response in has been observed in first episode psychosis subjects who experienced a higher incidence of childhood trauma , the opposite effect has been seen with blood based Cortisol. In this sample, higher levels of blood based markers of Cortisol were associated with conversion to psychosis in this sample and added as one of the 15-analytes in the Perkins et al. index. Further, reported that higher incidence of psychological bullying was associated with poorer global role functioning in CHR subjects. Thus, while we know increased levels of Cortisol are important to predicting conversion to psychosis and it is possible that the lower levels of Cortisol seen here as associated with higher levels of childhood trauma is associated with a different phenomenon . Put more simply, these results may indicate that the association or difference between blood-based Cortisol and childhood trauma is not clinically relevant for conversion to psychosis and that childhood trauma is independently predictive of clinical or functional outcomes, while higher levels of blood based Cortisol are predictive of conversion status.The second primary aim of this study was to identify highly correlated networks of inflammatory analytes using exploratory factory analysis. However, lack of correlation between inflammatory analytes and the relatively small sample size indicated that the sample was notsuitable for factor analysis. To our knowledge, only one study to date has used EFA to understand the correlation between inflammatory cytokines and severity of psychosis symptoms , demonstrating positive correlations between levels of cytokines and the Positive and Negative Symptoms Scale scores in subjects with schizophrenia. used unweighted co-expression network analyses to identify highly correlated networks of analytes in CHR and HC subjects, providing evidence of marked simplification of networks of correlated proteins that regulate tissue remodeling consistent with a hypothesis of blood-brain-barrier dysregulation in schizophrenia. Thus, the investigation of both clusters of inflammatory analytes and networks of inflammatory analytes is important to improving our understanding of high complex interplay between pro-inflammatory and anti-inflammatory processes in the development of psychosis and clinical outcomes. The third primary aim of this study was to determine the relationship between childhood trauma, psychosis risk symptom severity, and functioning in CHR. First, we hypothesized that there would be a significant positive relationship between childhood trauma, and psychosis risk symptom severity, as well as a significant negative relationship between childhood trauma and global/social/role functioning. Consistent with our hypothesis, partial correlation analyses revealed that total childhood trauma was associated with greater positive psychosis risk symptoms and lower global functioning .
Both findings are novel compared to Addington et al. ; however, we were unable to replicate the finding that total childhood trauma is associated with global role functioning in this smaller sub-sample. The association between childhood trauma and positive psychosis risk symptoms is consistent with existing research that higher incidence of trauma is associated with higher levels of positive symptoms in CHR . An extensive review on the relationship between childhood trauma and schizophrenia , concluded that childhood trauma is strongly related to symptoms of psychosis, specifically hallucinations and that the relationship may be dose-dependent. Second, we hypothesized that there would be a significant positive relationship between inflammation, total childhood trauma, psychosis risk symptom severity, and functioning, as well as a significant negative relationship between global/social/role functioning and inflammatory analytes in CHR subjects. Consistent with our hypothesis and replicating results from Perkins et al. , the 15-Analyte Index was positively correlated with all SOPS domains and negatively correlated with social, role, and global functioning in CHR subjects. However, a novel finding in this sample was revealed, a negative correlation between CRP and role functioning, indicating that higher levels of CRP are associated with lower scores on GFR. As GFR measures the level of impairment in academic, occupational, and homemaking roles, this is consistent with research linking blood levels of CRP to impaired cognitive performance in acute psychosis . Further, demonstrated that higher levels of CRP were associated with significantly worse working memory and inversely correlated with cortical thickness in individuals diagnosed with schizophrenia. Thus, the association between higher CRP and lower GFR may be a marker of impaired cognitive performance in CHR subjects. Inconsistent with our hypothesis, inflammatory analytes were not associated with total childhood trauma in CHR subjects. Previous research demonstrating the association between inflammation and childhood trauma and psychosis used a measure of childhood trauma that captured severity and chronicity of trauma occurrence. For example, demonstrated associations between TNF-a and severity of childhood trauma in first episode subjects using the CTQ which captures not only presence of trauma, but also severity and of the trauma experienced . Further, Hepgul et al. demonstrated associations between childhood trauma and CRP using The childhood experience of care and abuse scale: CECA.Q , which also measures severity of trauma. Thus, the measurement of childhood trauma used in this study did not capture severity or chronicity of the trauma experienced and thus were unable to be examined in the current study. Finally, we hypothesized that inflammation would partially mediate the relationship between childhood trauma and psychosis-risk symptom severity, as well as between childhood trauma and functioning in CHR youth. Using the information gathered from partial correlation analyses, we tested two mediation models. In Model 1, we explored whether the relationship between total childhood trauma and SOPSP was mediated by the 15-Analyte Index. In Model 2, we explored whether the relationship between total childhood trauma and GAF was mediated by the 15-Analyte Index. Inconsistent with our hypothesis, but consistent with the results from the partial correlation analyses, total childhood trauma and the 15-analyte index independently accounted for a significant proportion of variance in SOPS positive symptoms in Model 1. Further, total childhood trauma and the 15-analyte index independently accounted for a significant proportion of variance in GAF in Model 2.