Bidirectional associations between PTSD severity and glycemic control in trauma-exposed women with type 2 diabetes

ObjectivePosttraumatic stress disorder (PTSD) is linked with metabolic disturbance and increased risk of Type 2 diabetes (DM2), yet mechanisms explaining this connection have yet to be defined. Here we examine commonly hypothesized risk factors influencing the strength and directionality of the relationship between PTSD and DM2 severity in trauma-exposed black women with DM2. MethodsWe examined the relationships among PTSD severity (Clinician-Administered PTSD Scale, CAPS), glycemic control (A1c), age, smoking, body mass index (BMI), depression (Beck Depression Inventory, BDI) and medications in trauma-exposed Black women with DM2 recruited from an urban hospital between 2013-2015 (n=95) as a part of the Grady Trauma Project. Missing data were handled with multiple imputation. Relationships were assessed using lasso regression performed on each imputed dataset. Mediation analysis tested whether BMI or depression mediated associations between PTSD severity and glycemic control. ResultsLasso regression identified BDI, ACE inhibitor/ARB/beta blocker use, and CAPS severity as predictors in the final model predicting A1c. In pooled linear regression analyses across imputed datasets, overall model fit was modest, and none of these predictors were statistically significant. Lasso regression in the reverse direction yielded BMI, BDI, ACE inhibitor/ARB use, and A1c for the final model predicting CAPS severity. In pooled analyses, model fit was substantially stronger and BDI emerged as the only statistically significant predictor (p<0.05). Mediation analyses indicated that BDI, but not BMI, significantly mediated the effect of A1c on CAPS severity, as well as to a lesser extent the reverse relationship, the effect of CAPS severity on A1c. ConclusionIn this examination of the relationships between PTSD and DM2 severity, CAPS severity and A1c were found to be bidirectionally correlated in lasso regression models, with BDI emerging as a significant predictor and mediator of these relationships. BMI was not found to be a mediator in either direction. These results suggest the possibility of a reciprocal relationship between PTSD and DM2, where worsening of either condition may influence worsening of the other, where treating one condition in isolation may not be sufficient to prevent increased overall morbidity and mortality. Further research should include physiological interventions to ascertain causality in these relationships.