Trauma in childhood is associated with greater pain catastrophizing but not anxiety sensitivity: a cross-sectional study

Supplemental Digital Content is Available in the Text. Trauma during childhood but not adulthood affects pain catastrophizing. Emotional trauma is the most impactful in this association. Trauma has no impact on anxiety sensitivity.


Supplementary materials 2:
From the models tested the following showed potential violations of some assumption: -Anxiety Sensitivity ~ Age The assumption diagnostic for the model can be seen here.

Figure 11: Fitted values vs Standardized residuals
Cook's Distance

Supplementary materials 3: Joint regression results
A joint regression in which the two dependent variables were included was carried out through the GJTM package in R.
Once again, five different models were calculated.
▪ Childhood trauma without control variables model (pain catastrophizing ~ age + childhood trauma, anxiety sensitivity ~ age + childhood trauma): This was a joint regression model with childhood trauma as a predictor and both pain catastrophizing and anxiety sensitivity as outcomes. ▪ Full lifespan trauma without control variables model (pain catastrophizing ~ age + childhood trauma + impact of trauma history, anxiety sensitivity ~ age + childhood trauma + impact of trauma history): This model was a joint multiple regression model with childhood trauma and trauma history impact as predictors and both pain catastrophizing and anxiety sensitivity as outcomes. ▪ Covariate only model (pain catastrophizing ~ age + depression + anxiety, anxiety sensitivity ~ age + anxiety + depression): This model was a joint multiple regression model with the control variables as predictors and both pain catastrophizing and anxiety sensitivity as outcomes. ▪ Childhood trauma model (pain catastrophizing ~ age + depression + anxiety + childhood trauma, anxiety sensitivity ~ age + anxiety + depression + childhood trauma): This model was a joint multiple regression model with depression, anxiety and childhood trauma as predictors and both pain catastrophizing and anxiety sensitivity as outcomes. ▪ Full lifespan trauma model (pain catastrophizing ~ age + depression + anxiety + childhood trauma + impact of trauma history, anxiety sensitivity ~ age + anxiety + depression + childhood trauma + impact of trauma history): This model was a joint multiple regression model with depression, anxiety, childhood trauma and trauma history impact as predictors, and both pain catastrophizing and anxiety sensitivity as outcomes.
In table 2 the BIC (Bayesian Information Criterion) and AIC (Akaike information criterion) associated with each model can be seen. These criteria measure the fit of the model and their ability to predict future data, as well as, penalising the models with higher complexity. A lower value is an indicator of a better model fit.
Results show the same pattern as the one observed with the two independent regressions. In the models without anxiety and depression as covariates, the model was improved by adding childhood trauma but nor by adding impact of trauma history. In the case of models with covariates, the AIC indicated that the model was improved when adding childhood trauma but not when adding the variable impact of trauma history. The BIC did not show such convincing results of the overall benefits of including childhood trauma (supplementary table 2). This could be due to the lack of effect of childhood trauma on anxiety sensitivity. When looking at the effect of each predictor on the different outcome variables this is corroborated. The same way as with the independent models childhood trauma was shown to be a significant predictor of pain catastrophizing even when controlling for depression and anxiety, but impact of lifetime trauma was not. Meanwhile, childhood trauma showed a significant effect on anxiety sensitivity only until we controlled for the effect of lifetime trauma. Furthermore, in the models controlling for anxiety and depression childhood trauma did not show a significant effect (supplementary  table 3).

Anxiety sensitivity
When age was not included in the model as a covariate, results showed that lifetime trauma (but not childhood trauma) was a significant predictor of anxiety sensitivity, but, this effect disappeared when controlling for anxiety. From the covariates, anxiety was a significant predictor but depression was not. A summary of the results can be found in supplementary table 4. The model comparison showed that in the models without covariates adding impact of trauma history improved the model. In the case of models with covariates, it showed that adding neither childhood trauma or impact of trauma history improved the prediction. In other words, when controlling for anxiety and depression, trauma does not show to explain a significant part of the anxiety sensitivity variance. A summary of the results can be seen in