The impact of an exercise training intervention on cortisol levels and post-traumatic stress disorder in juveniles from an Ugandan refugee settlement: study protocol for a randomized control trial

Exercise Training (ET) has been shown to reduce an individual’s perception of stress and to improve overall mental health [10]. It is thereby important to demarcate exercise from physical activity. Both activities can involve large muscle groups that result in substantial increases in heart rate and energy expenditure. The most significant difference between the two is that exercise is planned and structured while physical activity is not [11]. Both can be distinguished as acute and chronic. Acute exercise is the physiological response to a single bout of exercise, while chronic exercise, which can also be called exercise training, refers to the repeated performance of acute exercise [11].

There are several studies describing chronic exercise as a useful intervention for PTSD patients to reduce trauma-related symptoms [12‐14]. After implementing an ET program, Fetzner and Asmundson [14] found significantly reduced PTSD symptoms, with 88.9% of the participants reporting a clinically significant improvement. Furthermore, previous research has shown significantly higher reductions in PTSD, depression, anxiety, and stress-symptom severity in PTSD patients who committed to ET, compared to controls [12, 13]. According to a meta-analysis of Rosenbaum et al. [15], who investigated four intervention studies, a higher physical activity was related to less frequent and less-severe PTSD symptoms. Two of these programs used combined aerobic exercises [12, 13], whereas the other two used yoga sessions [16, 17] in their program.

Exercise is a very effective stimulator of the hypothalamic-pituitary-adrenal (HPA) axis [18]. The hypothalamus produces the corticotropin-releasing hormone (CRH) which stimulates the pituitary gland which in turn releases the adrenocorticotropic hormone (ACTH) [19]. ACTH stimulates the adrenal glands to release cortisol and dehydroepiandrosterone (DHEA). DHEA and its sulfate form, dehydroepiandrosterone sulfate (DHEAS), have been found to possess anabolic properties [20], whereas high cortisol has catabolic properties [21]. Furthermore, DHEA/S has neuroprotective and anti-glucocorticoid effects [22]. Under normal conditions DHEA levels are closely correlated with cortisol; however, an imbalance of cortisol/DHEA secretion may occur when an individual experiences chronic stress [23]. Cortisol and DHEA are often addressed as a ratio that represents the balance between anabolic and catabolic hormones [24]. Budde et al. [25] found an increase in cortisol levels due to acute exercise in adolescents. According to Hill et al. [26], the exercise intensity necessary to provoke a significant increase in circulation cortisol is 60% maximum rate of oxygen consumption (VO_2max), which shows an increase of + 39.9% in circulating cortisol levels. In general, untrained participants had lower VO_2max levels, significantly higher responses to cortisol post exercise, and lower hair-cortisol levels compared to participants who exercised regularly [27‐29]. In regard to the chronic effects of exercise, it turned out that a higher training volume (measured in kilometers run per week) was associated with an increase in cortisol levels in young healthy men [29]. After ET, a decrease [30] as well as an increase [31] in cortisol levels was found. A distinct rise in cortisol has been observed also immediately after waking, typically peaking 30–45 min after waking [32], and has been appropriately termed the cortisol awakening response (CAR). This response is a neuroendocrine manifestation of the hypothalamic-pituitary-adrenal axis (HPA-axis) has been demonstrated to be sensitive to stressors like exercise. Although normative ranges have been developed for several populations [33], it is still unclear what may constitute a “healthy” cortisol response, as both elevated and depressed responses have been related to dysfunctional psychosocial health status [34]. Compared to adults, fewer data are available on hormone responses to exercise in children and adolescents. Regarding the HPA response to the physical activity status in adolescents, cross-sectional findings revealed no significant changes [35]. However, after presenting an acute exercise-related stressor, young individuals at early stages of puberty showed higher cortisol increases and lower DHEAS-to-cortisol ratios than late pubertal participants [36].

In various studies, alterations in the function of the glucocorticoid system have also been reported for PTSD patients. It is commonly accepted that PTSD patients can have a dysfunction in the HPA-axis [37]. Results on differences in baseline cortisol levels between individuals with PTSD and controls are inconsistent, depending on the research paradigm [38]. Several researchers found reduced basal cortisol levels in PTSD patients compared to healthy controls [39, 40]. In addition, cortisol levels turned out to be a significant predictor for PTSD symptoms 6 weeks and 6 months after the traumatic event [39]. At some time points, even higher cortisol levels were found in participants who had PTSD [41]. Similarly, equivocal results exist surrounding the DHEA response, in particular due to the smaller number of systematic studies in this field. Interestingly, there were negative associations between hair-cortisol levels and the number of different lifetime traumatic events, the frequency of traumatization, and the time interval since traumatization [40]. Regarding the cortisol levels of adolescents and children with PTSD, research reports an increase in cortisol concentration in PTSD victims [42]. Some other studies did not find any significant differences in cortisol levels in this age group [43] or even an attenuation in the CAR [44].

Thus, even if cortisol and DHEA are associated with PTSD and provide promising targets for the detection of changes in HPA functioning, functional consequences related to the development of stress-related psychopathologies will have to be investigated in a subsequent step, constituting a future research avenue. However, the latest studies have demonstrated the usefulness of the present research approach to utilize stress hormones as an objective marker to validate the effectiveness of ET.

Kim and colleagues [45] showed a design that is comparable to the present study by demonstrating significant changes in cortisol levels and also an improvement of PTSD symptom severity after the adult participants completed an 8-week ET program. In their study, “Mindfulness-based Stretching” and “Deep Breath Exercises” were used as low-intensity chronic exercise interventions. However, the intensity of ET had positive effects on the improvement of PTSD symptoms [12‐14].

Even if a few studies suggest ET to be beneficial for children and adolescents with PTSD, no randomized controlled trial (RCT) has been reported yet that directly compares the effects of a cardiovascular condition to a sham condition in a longitudinal design. To our knowledge, there is neither a study investigating the relationship between a chronic exercise intervention and the cortisol response in young PTSD victims, nor has it ever been tested in a resource-poor, post-conflict setting.

The purpose of this paper is to introduce a study protocol for a RCT that aims to systematically investigate the impact of ET on PTSD symptoms and associated cortisol levels in adolescents. It is hypothesized that ET in adolescents not only leads to a sustainable decrease in PTSD symptoms in comparison to two control conditions (H1) but that the changes in PTSD symptoms also correspond to changes in cortisol (H2) and DHEA levels (H3) as markers for beneficial alterations in HPA-axis functioning. Furthermore, we expect differences in PTSD symptom severity and cortisol/DHEA ratios between participants who are highly physically active and those who are less physically active at baseline (t₁) (H4).

This study investigates juvenile trauma victims from the Democratic Republic of the Congo (DRC), aged 13–16 years, who are based at the Nakivale refugee settlement, an official refugee camp in Uganda, Africa, which is among the largest refugee camps in the world. The government of Uganda and its United Nations High Commissioner for Refugees (UNHCR) partners provide general health services to the refugees in the settlement with few psychotherapeutic interventions. Therefore, Ainamani, Elbert, Olema, and Hecker [46] propose the ample provision of trauma-focused treatment options particularly in these settings in order to support not only individual traumatized refugees but also their families and the whole community. All participants will have a history of traumatic experiences. Participants eligible for inclusion in the intervention only involve individuals with a diagnosis of PTSD according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-V). Exclusion criteria cover (1) acute suicidality, acute intoxication, or psychotic symptoms, (2) inability of the parent or legal guardian to provide consent, or (3) child protection issues (e.g., acute maltreatment) that are identified during the initial cross-sectional assessment and are judged by a clinician to make trial inclusion inappropriate. Additionally, the participants may not be part of any other psychological therapy or take any medication with psychoactive drugs until the study is completed. We further control for an unchanged physical activity level during the intervention.

A prior sample size calculation was performed using g*power 3 [47] based on (1) an expected medium effect size in the overall model according to cortisol and PTSD changes reported in previous studies [45], (2) the PTSD symptom severity, the trauma load and the number of potential stressors occurring during the intervention as covariates, and (3) the number of time points. The desired statistical power level was set at 0.95. The sample size calculation was corrected for an estimated dropout rate of 20%, according to other intervention studies that have already been performed in the Nakivale refugee settlement [48]. According to the power analysis, 159 participants are sufficient so that a total sample size of 198 participants is targeted to acknowledge potential dropouts.

The present project, which was also registered at www.​drks.​de before the start of data collection, contains two subsequent steps: (1) a PTSD screening in a large and representative sample of juvenile refugees from the DRC that will serve to identify participants eligible for inclusion in the intervention. A cross-sectional analysis comparing the physical activity levels of the refugees according to their PTSD symptom severity will serve to identify and balance potentially confounding variables. With regard to cross-sectional studies, it is difficult to make causal inferences because it is unknown if the parameter itself (e.g., being cardiovascularly fit) caused the benefit (e.g., reduced symptoms) or if other factors might have accounted for the resultant differences [49]. Therefore, (2) a longitudinal, prospective, and RCT will be implemented covering the ET intervention and two control conditions. Group (Exercise Training = ET, Alternative Intervention = AI, and Waiting-list Control = WC) serves as the between-subject factor. Eligible participants will be randomly assigned to one of the three conditions. The participants’ age, their traumatic load indicated by the number of experienced traumatic event types, the baseline PTSD symptom severity, and their physical activity level will serve as matching criteria across groups. The study utilizes a double-blind RCT design.

The ET group is subjected to an ET of 60–80% of the maximal heart rate (HR_max). The AI group serves as the placebo group and receives fine and gross motor body coordination exercises through playful balance, bilateral coordination, hand-eye coordination, and leg-arm coordination exercises [49]. With such a sham-controlled study, it will be examined whether the expected changes in cortisol levels and symptom severity in PTSD patients are attributed to chronic exercise and no other influencing factors [50]. The WC group serves as a no-treatment control and is measured at the same intervals as the treatment group. It will not be implemented until the ET and AI groups have finished their second follow-up (t₄). Introducing a second control group with the alternative intervention is mandatory to differentiate beneficial effects of miscellaneous physical activity from standardized ET. This is not only important to investigate a potential cross-cultural effectiveness of ET but also serves to justify the specific training requirements of ET and the need for sufficient resources in order to build up sustainable intervention programs. The flow diagram is presented in Fig. 1.

To assess the impact of chronic exercise on PTSD symptoms and cortisol profiles, repeated measurements with time (baseline (t₁), interim test (t_1b), post test (t₂), 3-month follow-up (t₃), and 6-month follow-up (t₄)) as the within-subject factor will be conducted on the primary outcome measures PTSD symptom severity and awakening cortisol profiles. At eligibility (t₀), the diagnosis will be verified and co-morbid diseases will be identified. In addition, at baseline (t₁), the PTSD symptom severity, relevant information about the participants’ trauma and flight history, as well as their exposure to early life adversities will be assessed. As physiological measures, cortisol and DHEA levels will be determined. The assignment of participants to one of the three groups will be conducted afterwards. An interim measuring time point (t_1b) at 4 weeks after the respective intervention will serve to adjust the training to the individual participant’s progress in the two exercise groups in order to assure that the ET stays over 60% HR_max after the 8-week interventional phase at the post test (t₂). The baseline measures will be repeatedly assessed at the post test as well as the two follow-up measuring points (t₃ and t₄; see Fig. 2 and Additional file 1).

TRLS-D-18-00245R, 18 June 2018, no recruitment yet.