Little evidence that posttraumatic stress is associated with diurnal hormone dysregulation in Turkana pastoralists

Abstract

Research in industrialized populations suggests that posttraumatic stress disorder (PTSD) may be associated with decreased cortisol or testosterone sensitivity, resulting in a blunted diurnal rhythm. However, the evolutionary implications of this association are unclear. Studies have primarily been conducted in Western industrialized populations, so we do not know whether hormonal blunting is a reliable physiological response to PTSD or stems from factors unique to industrialized settings. Furthermore, existing studies combine PTSD from diverse types of traumas, and comparison groups with and without PTSD differ along multiple dimensions, making it hard to know if PTSD or other life factors drive the blunted cortisol response. We conducted a study among n = 60 male Turkana pastoralists, aged between about 18-65 years in Kenya, exposed to high levels of lethal inter-ethnic cattle raiding. 28% of men in this area have PTSD symptom severity that would qualify them for a provisional PTSD diagnosis. Saliva samples were collected at three points to compare the cortisol and testosterone profiles of Turkana warriors with and without PTSD. Contrary to existing work, our preregistered analysis found little evidence for a difference in the hormonal profiles of warriors with high versus low PTSD symptom severity. Our results imply that the relationship between PTSD and hormonal diurnal variation may vary across populations and ecologies or that the association documented in Western populations stems from other correlated life factors. Studies in a wider range of populations and ecological contexts are needed to understand the evolutionary underpinnings of hormonal responses to trauma.

Keywords: PTSD; cattle raids; combat stress; cortisol; pastoralists; posttraumatic; testosterone.

Figure 1.

The three sensitivity measures we preregistered for this analysis using a typical diurnal curve from the literature. S₁₂ is the slope of the biomarker concentration from the first to the second sample. S₁₃ is the slope of the biomarker concentration from the first to the third sample. AUC is the area under the biomarker concentration curve.

Figure 2.

Change in concentrations of cortisol (top) and testosterone (bottom) for participants with low PTSD symptom severity (left) and high PTSD symptom severity (right). Contrary to our expectations, the concentration of both biomarkers tended to increase from the second to the third sample.

Figure 3.

Mean concentration for measured cortisol (top) and testosterone (bottom) levels at each sample time for participants with high PTSD symptom severity (red), participants with low PTSD symptom severity (blue), and all participants (black). Concentrations of both hormones decrease, on average, between the first and second sample times and increase between the second and third.

Figure 4.

LOOIC model weights for the models with S₁₂ (slope from T1 to T2) as the cortisol and testosterone sensitivity measures that are summarized in Tables 3 and 4. Areas represent the relative weight of the model. Three of the 12 models dominate the LOOIC Weights for cortisol sensitivity, with the most weight on the intercept model (with no predictor variables). For testosterone sensitivity, two of the 12 models dominate the LOOIC weights, with a model including the provisional diagnosis of PTSD as a predictor having the most weight and the intercept model having the second most weight. The models dominating the LOOIC Weight are not very distinguishable in terms of prediction, so there is not much evidence that any model performs better than the intercept model.

Figure 5.

Waking cortisol and testosterone concentrations taken at waking for the population. The difference in waking cortisol concentration between low and high PTSD symptom severity participants was not significant. However, the waking testosterone concentration for the high PTSD severity group was significantly lower. This analysis was exploratory.