Date of Award

Spring 5-15-2015

Degree Type


Degree Name

Master of Public Health (MPH)


Public Health

First Advisor

Ashli Owen-Smith

Second Advisor

Matthew Magee

Third Advisor

David Wright


INTRODUCTION: Traumatic brain injury (TBI) is a major public health problem, causing approximately 52,000 deaths from 1.7 million injuries in the United States annually, with a combined direct and indirect economic cost estimated at $60-75 billion per year. Traumatic subarachnoid hemorrhage (tSAH), a subtype of closed head injury, has a high prevalence within TBI—evident in up to two-thirds of moderately and severely brain injured patients. tSAH is also associated with poor clinical outcomes; some research suggests mortality and unfavorable outcome rates are two-to-three times higher in patients with tSAH, based on brain imaging, compared to those without. To date, no pharmacological treatment has been conclusively shown to improve outcomes in humans for either moderate or severe TBI or for specific tSAH injury. The aim of this study was to assess whether the effect of PROG was substantially different in study TBI patients with evidence of tSAH on initial brain imaging compared to those that did not have evidence of tSAH.

METHODS: ProTECT III clinical trial data was used for an exploratory, post hoc subgroup analysis to determine the effect of the hormone progesterone (PROG) on outcome. Study subjects with any abnormality on baseline brain imaging were included in the analysis and two subgroups, tSAH positive (+tSAH) and tSAH negative (–tSAH), were selected. The primary outcome evaluated was a favorable/unfavorable dichotomy derived from the 6-months post-injury Extended Glasgow Outcome Scale (GOSE) assessment, which evaluates both mortality and functional outcomes. Risk ratios (RRs) were calculated for the total sample and each of the two subgroups and used as statistical evidence for interaction between PROG and tSAH.

RESULTS: All subjects from the original ProTECT III trial cohort (N=882) with no abnormalities found on baseline computed tomography (CT) image (n=125) or missing image (n=1) were excluded from this analysis. Subjects with one or more abnormalities noted on CT (+CT, n=756) were then divided into subgroups based on presence (n=582) or absence (n=174) of tSAH. Subjects with +tSAH were more severely injured than –tSAH (mean Rotterdam CT score 3.3 vs. 2.2; 3.1 overall) and had a lesser proportion of favorable outcomes (47.4% vs. 74.3%; 53.6% overall). Compared to placebo, patients treated with progesterone had marginally better likelihood of favorable outcomes (risk ratio among +tSAH 1.06, 95% confidence interval [CI], 0.89 to 1.26; and RR among –tSAH 1.02, 95% CI 0.85 to 1.22). A multivariable model, adjusted for baseline differences in treatment group covariates did not yield substantially different results for the effect of progesterone on favorable outcomes (+tSAH 1.07; 95% confidence interval [CI], 0.84 to 1.36, –tSAH 1.08; 95% CI 0.75 to 1.56, +CT 1.06; 95% CI 0.87 to 1.29).

CONCLUSION: Our study demonstrated that progesterone did not result in different effects in patients with or without tSAH than those without based on initial brain imaging. This investigation supports previous research findings; tSAH is correlated with more severe injury and worsened outcomes. Concomitant injuries found in +tSAH group are likely worsening the outcomes over –tSAH, but this was not evaluated here. More complex statistical modeling should be used on this data to determine if it provides evidence that tSAH is an independent prognosticator of unfavorable outcome or merely associated with more severely injured patients.