0945 Sleep Architecture Predicts Hyperarousal Among Trauma-exposed Women

Abstract

Introduction: Traumatic-stress disorders are associated with self-reported hyperarousal and physiological hyperarousal such as augmented startle response. Among trauma-exposed women showing a wide spectrum of post-traumatic symptoms, we sought to predict the magnitude of responses to an acoustic-startle stimulus measured by skin conductance response (SCR), orbicularis blink-startle response (EMGR), and heart rate acceleration (HRA), using both self-reported hyperarousal measures and sleep architecture. Methods: Thirty-three women (ages 18-40, 15 with current PTSD diagnosis) who had experienced a criterion-A trauma within two years completed the Clinician Administered PTSD Scale (CAPS), PTSD Checklist (PCL-5), Hyperarousal Scale, Hypervigilance Scale, SCL- 90, and Depression Anxiety and Stress Scale (DASS). Each of the six hypervigilance items were extracted from CAPS and PCL and the single anxiety subscores from SCL-90 and DASS. Using principal component analysis, we reduced the resulting 16 self-report items to 4 factors that together accounted for 79% of their combined variance. These included: 1) hypervigilance/arousal, 2) subjective sleep and concentration impairments, 3) anxiety, and 4) recklessness. Following an acclimation night, ambulatory polysomnography provided sleep-stage percentages (N1-3, REM) and REM latency (REML). Participants underwent an acoustic-startle paradigm in which mean SCR, EMGR, and HRA responses to 15 loud (102dB) tones were obtained. Separate stepwise multiple regressions then used these four factors and five sleep parameters to predict each startle-response variable. Results: The anxiety factor was positively correlated with the mean EMGR (r=0.537, p=0.003) and HRA (r=.412, p=.029), and a positive correlation trend was observed between this factor and SCR (r=.320, p=.09). However, none of the other self-report factors were associated with SCR, EMGR, or HRA. Among the four self-report factors and five sleep parameters, greater REML predicted greater EMGR (R2=0.212, β=0.501, p=0.021), and greater HRA (R2=0.271, β=0.619, p=0.002). Decreased N2% predicted greater HRA (R2=0.271, β=-0.441, p=0.021). Conclusion: Sleep architecture predicted physiological hyperarousal indexed by elevated startle responses as well as - or better than - factors generated from subjective measures of hyperarousal. Sleep predicts hyperarousal across a spectrum of trauma symptomatology, emphasizing the need to further examine interactions between sleep, hyperarousal, and anxiety disorders.