Saliva cortisol levels and physiological parameter fluctuations in mild traumatic brain injury patients compared to controls

Abstract Background: Evidence suggests that fluctuations of cortisol and physiological parameters can emerge during the course of mild Traumatic Brain Injury (mTBI). Objective: To investigate fluctuations of cortisol and physiological parametersduring the acute phase of mTBI in hospitalized patients. Methods: 30 participants (19 patients with mTBI and 11 controls) were examined for saliva cortisol dynamics, heart rate (HR), systolic arterial pressure (SAP), diastolic arterial pressure (DAP), mean arterial pressure (MAP) and body temperature (BT) fluctuations for four consecutive days. Also, the participants completed the Athens Insomnia Scale and Epworth Sleepiness Scales, in order to check for sleep problems. Results: Patients showed elevated levels of cortisol relative to controls (peak at 8 am and lowest levels at 12 am), as well as for most physiological parameters. MAP was significantly higher for patients throughout the measurement period, and BT was elevated for patients relative to controls at almost all measurements of the first and second day. Mean HR tended to track at non-significantly higher levels for the mTBI group. Patients’ sleepiness and insomnia values (ESS and AIS) were initially significantly higher relative to controls but the difference dissipated by day 4. Conclusion: The increase in absolute values of cortisol and physiological parameters measurements, indicates that in the acute phase of mTBI, a stressful process is activated which may affect sleep quality as well. Supplemental data for this article is available online at at doi: 10.1080/00207454.2021.1951264.


Introduction
Hypothalamic-pituitary-adrenal axis (HPAA) activation is a significant mechanism for responding to stressful events which trigger variation in cortisol saliva values above normal levels. Traumatic Brain Injury (TBI) may affect HPAA in both animal studies [1] and clinical studies in humans [2,3]. The majority of TBIs are categorized as mild (mTBI) (scoring between 13 and 15 on the Glasgow Coma Scale-GCS) [4]. Evidence suggests that diurnal fluctuations can emerge during the acute, subacute, or chronic phases following mTBI [4,5].
Stable 24-hour rhythms in melatonin and cortisol levels, as well as heart rate (HR), blood pressure (BP), and body temperature (ΒΤ) are, not surprisingly, interrupted when an individual experiences brain injury, given that master clock functions are situated in the suprachiasmatic nucleus of the hypothalamus [6].
This study examines the fluctuations of cortisol saliva levels, physiological parameters for four consecutive days, and Sleep Scales emanating from mTBI [7][8][9][10]. The aim of the study was to record possible alterations during the acute phase mTBI, considering that mTBI is a stress factor activating the HPA axis. Therefore, we examined cortisol dynamics alongside with key physiological parameters, HR, systolic arterial pressure (SAP), diastolic arterial pressure (DAP) and mean arterial pressure (MAP) and BT, in both mTBI patients and a comparison group of healthy volunteers. This study additionally deploys scales measuring insomnia with the Athens Insomnia Scale (AIS) and tendency to sleepiness with the Epworth Sleepiness Scale (ESS).

Patients
All patients, admitted to the Emergency Department at "KAT" General Hospital between April 2016 to April 2018 experiencing head injury as the single presenting condition, were recruited to the study at an initial phase. Of these, only patients classified by the neurosurgeon in charge as having mTBI and a GCS [13][14][15] were included in the study. These patients were admitted to the Neurosurgical Clinic requiring hospital admission and monitoring for at least 96 h, at the discretion of the physician in charge. CT scanning was performed within the first 4 h after trauma to estimate the severity and localization of the injury. In addition, complete blood count (CBC), coagulation and biochemistry tests were performed.
To avoid gender effects, only males were included in both the patient and control groups for analysis. To sum up, to ensure a high degree of homogeneity, patients were enrolled according to the following inclusion criteria: males, aged 18-80 presenting with mTBI, GCS 13-15, and requiring admission for at least 96 h. Only participants capable and willing of giving informed consent were included. Additional exclusion criteria were glucocorticoid medication prior to or subsequent to admission to hospital and prior pituitary insufficiency. Totally, 19 patients with mTBI (GCS = 15) followed the above criteria and were included in this prospective observational study.

Control group
The control group was randomly selected by the primary investigator using snowball sampling. Inclusion criteria were male, aged 18-80 with a free medical history and no medication. Exclusion criteria included the following: participation in a clinical study within the previous 30 days; use of glucocorticoid medication within 14 days of study entry; recent use (within the previous seven days) of medication with b-adrenergic receptor blocker activity or an MAO inhibitor characteristics; known preexisting sleep disorders, depression, or insomnia; and significant abuse or dependency on nicotine or alcohol. Parents with young children (age < 2.0) and shift workers of the study were also excluded. Participants who had traveled through more than two time zones during the previous four weeks were also excluded [11].
All subjects agreed to refrain from alcohol and caffeine consumption for the duration of the project and to abstain from any significant physical activity or any stress factor, after 20:00 h during the study period. Baseline CBC, biochemistry tests, coagulation and X-rays were completed for all control participants.

Saliva cortisol sampling
Saliva sampling was performed for 4 consecutive days after injury for the experimental group, and matching analyses were conducted with controls. On the 1 st and 4 th day, saliva sampling was conducted at the following time points: 8 a.m., 12 p.m., 4 p.m., 8 p.m., 12 a.m. and 4 a.m. On the 2 nd and 3 rd days, saliva sampling was conducted every 8 h at the time points of 8 a.m., 4 p.m. and 12 a.m.
Saliva samples were collected in specially designed commercially available plastic tubes (Salivette, Sarstedt AG & Co., Nümbrecht, Germany) previously shown to be a convenient and accurate and convenient technique relative to 'passive drool' approach [12,13]. All patients and volunteers were instructed to have dinner and to brush their teeth before 8 p.m., and to rinse the mouth before every saliva collection [13].
After the demonstration, the patients and the control group were given oral and written instructions for saliva collection on the proper use of Salivette. The control group was contacted half an hour prior to measurement with a telephone reminder. As an additional reminder, each Salivette bore a label indicating the exact date and time of sampling. Participants were reminded prior to sampling that they should not drink liquids, smoke or consume food for at least one hour prior to the sample time.

Assays
The samples were centrifuged at 1000xg for 2 min, and 2 ml of the supernatant was collected and placed in microtubes placed in the refrigerator at −80 °C until measurements. The samples were analyzed at the Clinical Biochemistry Department, "KAT" General Hospital, Kifissia, Athens, Greece, by certified technicians.
Salivary cortisol was determined by competitive electrochemiluminescence immunoassay using Elecsys Cortisol reagent kit (Roche). All saliva samples were run in duplicate with mean values were reported (CV < 6%), with a detection limit of was 1.5 nmol/L.The range of the cortisol assay was 1-1750 nmol/L. The analytical and functional sensitivity of the test fell within the < 1 nmol/L and < 8 nmol/L range [13].
In our study, the reference range followed the following values which were determined using saliva samples from healthy individuals in studies with the Elecsys Cortisol II assay [14], (95th/97.5th percentile) that is:

Physiological parameters
Physiological parameters including HR (beats per minute, bpm), systolic (SAP; mmHg), diastolic (DAP; mmHg) and mean arterial pressure (MAP; mmHg) and body temperature (BT; °C) were recorded manually every 4 h during the first and 4 th day and every 8 h during the 2 nd and 3 rd day during hospitalization for both patients and control group. Physiological parameters were recorded simultaneously with the saliva of cortisol sampling.

Measurement of sleepiness and insomnia
The Epworth Sleepiness Scale (ESS) [15] and Athens Insomnia Scale (AIS) [16] were completed on admission and on the fourth day in the experimental group. Controls completed the scales once on the first day. These scales measure daytime sleepiness and insomnia, respectively. ESS scores from 11 to 16 indicate mild to moderate daytime sleepiness, and scores of 16 and above indicate severe daytime sleepiness. Each item of the AIS can be rated 0-3, (with 0 corresponding to no problem at all and 3 to the very serious problem) with a total score ranging from 0 (absence of any sleep-related problem) to 24 (the most severe degree of insomnia). A cutoff score of ≥ 6 on the AIS is used to establish the diagnosis of insomnia.

Ethical and research approvals
The study was conducted in full accordance with ethical principles, including the World Medical Association Declaration of Helsinki (Version, 2002), and was independently reviewed and approved by the local "KAT" Hospital and the National and Kapodistrian university of Athens ethical committees, respectively. Informed consent forms were signed by all participants.

Statistical analysis
For the comparison of proportions, chi-square (χ 2 ) and Fisher's exact tests were used. Student's t-test tests were used for the comparison of continuous variables between patients and controls and paired t-tests were used to evaluate changes in study variables from one-time measure to another. To longitudinally assess changes in cortisol, SAP, DAP and BT levels, mixed linear regression models were fitted, to evaluate changes of cortisol over time in association with other variables. Regression coefficients (β) with standard errors (SE) were computed from the results of the models. All p values reported are two-tailed. Statistical significance was set at 0.05 and analyses were conducted using SPSS Statistics 22 (IBM, Armonk, New York). Continuous variables are presented with mean and standard deviation (SD). Qualitative variables are presented with absolute and relative frequencies. The two groups were matched in terms of age, nationality, BMI, smoking status, comorbidity, allergy and alcohol consumption.

Data relating to neurological status during the acute phase of mTBI
The sample consisted of 30 participants (11 in the control group and 19 in the patient's group). Demographics, clinical characteristics and laboratory findings of the two study groups are presented in Table 1.

Saliva cortisol
Patients had significantly greater levels of saliva cortisol ( Figure 1) relative to controls. Peak levels were consistently observed at 8 a.m. and lowest levels at 12 a.m. for all days for both controls and patients' groups. As depicted in Figure 1 there was a parallel kinetic of saliva cortisol levels between controls and patients but the levels were significantly higher in the patient's group. For the controls, significant increases in cortisol levels were observed at 8:00 a.m., for all days of measurements.

Physiological parameters
Turning to arterial pressure, MAP was significantly higher for the patients' group at Day 1 at 8 a.m., at 8 p.m., and 4 a.m., Day 2 at 8 a.m., at 4 p.m. and at Day 3 at 8 a.m. (Figure 2). No significant difference between patients and controls was observed with respect to DAP, except at 4 a.m. of the 1 st day. Again, we observed a parallel kinetic of MAP values between controls and patients tending to converge when reaching day 4.
In controls, DAP decreased from 12 a.m. to 4 a.m. for a first and fourth day, while in patients DAP remained unchanged. SAP increased from 12 p.m. to 4 p.m. in controls and from 8 a.m. to 4 p.m. in patients.
For the first three days of the studies, SAP measurements from patients tended to be elevated relative to controls (Table 3; supplementary material).
HR was not significantly different between the two groups at any time-point; however, HR decreased from 12 a.m. to 4 a.m. of the fourth day in patients and from 4 p.m. to 12 a.m. of the second and third day in the control group ( Figure 3; Table 4, supplementary material). We observed actually a smaller difference in HR curves between the two groups.
Finally, BT was significantly elevated in patients compared to controls at almost all measurements of the 1 st and the 2 nd day. Additionally, BT was significantly higher in patients at 12 a.m. of 3 rd day and at 4 p.m.,   Table 4, supplementary material). We observed a parallel kinetic of BT values between controls and patients tending to converge when reaching day 4.
Results from mixed linear models showed the expected time effect on cortisol during the day for both groups. In controls, no differences were found between the four days, but in the patient's group    cortisol level significantly decreased on the final two days as compared to the first day. (Table 5, supplements) Results from mixed linear models for SAP, DAP, MAP, BT and HR effect on cortisol levels ( Table 5, supplements), showed a significant effect of SAP on cortisol variability during follow up, indicating that greater SAP is associated with greater cortisol levels. DAP, MAP, HR, BT, ΒΜΙ, age, smoking, allergy, comorbidity and alcohol consumption were not associated with cortisol changes during follow up.
We observed statistically significant differences in Hematocrit (Ht), Hemoglobin (Hb), White Blood Count (WBC) and differential (Neutrophil, Lymphocyte), PLTs, C-RP and Fibrinogen between patients and the control group (Table 1).

Measurement of sleepiness and insomnia
In Table 2 we present ESS, AIS and insomnia percentage among controls and the patients at baseline and day 4. Insomnia was present in 9,1% of controls and 68,4% of the patients at baseline and 26,3% at day 4. At baseline patients had significantly higher values (p 1 ) compared to controls and this difference was ameliorated (p 2 ) at day 4 to non significance. Also, among patients there was a statistically significant decrease in ESS, AIS and insomnia percentage between baseline and day 4.

Discussion
The main finding of this prospective study during the acute phase of mTBI, is that patients monitored for 4 consecutive days do experience elevated levels of cortisol and physiological parameters. Also, we observed significant differences in saliva cortisol levels, physiological parametersand Sleep Scales recorded as compared to controls.
Overall, patients exhibited significantly greater levels of saliva cortisol compared to healthy controls. In line with both classical and more recent reports, our results suggest that patients with TBI have increased cortisol levels in plasma and saliva respectively [2,17].
The HPA axis status is probably the most extensively investigated endocrine function in acute TBI patients. Such patients commonly present with high cortisol levels. Elevated levels may persist up to 15 days after the trauma and are essential in maintaining vascular tone and endothelial integrity and in potentiating the vasoconstrictor actions of catecholamines. Initially, the increase in cortisol is mediated by ACTH. Later other substances, including cytokines or catecholamines, act as mediators. The diurnal variation in cortisol may be also preserved or abolished [18].
In our study, the diurnal variation of cortisol was observed to be preserved albeit at higher levels, and it seems that there is an association between mTBI and saliva cortisol dynamics [8]. This is also supported by the study of Skoglund et al. [17]. However, for both day 2 and 3, peak at 8:00 am is not surprising, since for both days there are only three points per 24-hours. Cortisol must be increased at the beginning of the active phase if subjects are exposed to regular light-dark cycles, as it is assumed for our participants. On the other hand, in a review, it was suggested that higher cortisol values are associated with a more severe head injury and worst clinical outcome, but also suggested that lower cortisol values are found in the most severely injured patients having brainstem dysfunction or are brain-dead. Other studies found no such correlations [18]. Since our patients experienced mTBI, it is reasonable to assume that their ANS was reactive affecting cortisol values and physiological parameters. A review suggests that even the psychological stress associated with mTBI may activate the ANS [19].
Fluctuations in the physiological parameters, including BP and HR, observed in this study may indicate that the activity of the autonomic nervous system (ANS) was affected by mTBI and especially leading to hyperactivity of ANS. Even in patients in a persistent vegetative state after TBI, ANS dysfunction is present [20]. In that study, including normotensive subjects, SBP, DBP, and HR were characterized by a circadian pattern, with higher values of BP and HR during the daytime and lower pressure and bradycardia during the nighttime. An important physiologic mechanism behind the diurnal BP and HR pattern is the day-night variation in the activity of the autonomic nervous system (ANS), which is under the influence of various intrinsic and extrinsic factors. Our results indicate the diurnal pattern of physiological parameters in mTBI patients was preserved, but all at elevated levels.
In our study, the ANS was possibly affected by the brain injury leading in high hypothalamus stimulation resulting in higher values of physiological parameters [20]. In other words, increased physiological parameters observed in mTBI patients are the expression of the physiological adaptation to stressful stimuli, such as mTBI [21].
The statistically significant difference between controls and mTBI patients regarding Hematocrit possibly reflects mild blood loss or hemodilution due to iv fluid administration. Blood loss may be related to significantly lower PLTs observed in our patients. Also, the significant increase in total WBC and Neutrophil in our patients may be attributed to a stress-induced inflammatory response to trauma which is also compatible with significant differences observed for fibrinogen and CRP levels [22].mTBI injury affected subjective sleep measures, suggesting more sleepiness and insomnia symptoms, indicated by differences in AIS and ESS scores between the two groups at baseline but also by the improvement in those scales at day 4 in mTBI patients. This finding is confirmed in other studies of TBI patients, demonstrating sleep disturbances, manifested by insomnia during night sleep and as consequence sleepiness during the day, which is in accordance with other studies [8,10,23]. Possible differences regarding saliva cortisol level fluctuations with previous studies may be explained at least in part, due to different lab settings and operator/technician, different study protocols, differences normal range and different kit. Also, differences in illness severity since we included only mTBI patients (GCS = 13-15) contrary to Poll's study including only healthy participants, and the study of Skoglund et al incorporating only severe trauma patients [9,17].
In this study, we took advantage of noninvasive salivary cortisol sample collection procedures, which allows study participants to reliably collect saliva samples at home. The process also is a less stressful way of collecting samples in hospital wards, which is a key advantage in cortisol studies [24]. Cortisol is diffused into saliva at rates independent of salivary flow, and cortisol levels in saliva offer the opportunity to assess dynamic hypothalamic-pituitary-adrenal changes [17]. Analysis of saliva accurately and quickly reflect the concentration of plasma cortisol's rhythm [25] and being able to track these changes, considering the rapidity of fluctuation, is important in studies such as this. Moreover, levels of cortisol in saliva provide an estimate of the unbound, "free" cortisol fraction [17], which is preferable.
In conclusion, we found that there was a significant increase in cortisol levels and physiological parameters measurements, possibly due to stress HPA axis and ANS activation.
This study provides evidence that: 1. mTBI patients follow a preserved diurnal rhythm, as supported by saliva cortisol measurements being increased compared to controls. 2. Physiological parameters also follow a preserved diurnal rhythm again in an increased pattern.

mTBI affected sleep as measured AIS and ESS
scales showing an improved pattern a day 4 in patients' group.

Limitations
The results of this study need to be considered in the context of a number of limitations. The study would have benefited from a larger sample size, different illness severity, the collection of plasma ACTH, and better subjective and objective evaluation of stress. It is of our knowledge that the sample is small for correcting confounding factors/covariates (e.g. pain or other stress-related factors including the severity of mTBI). There is evidence that cortisol release in acute illness is also promoted by ACTH-independent mechanisms such as neurones, neuropeptides, cytokines, and gonadal steroids, which regulate HPA axis function at the pituitary or adrenal level [26]. Objective measures of stress load possibly by examining stress factors and measuring catecholamines levels, especially at the submission time in order to provide context to cortisol levels found in the study, would have been valuable. In addition, the study would have been improved by using core BT [27,28]. In addition, the daily excretion of free cortisol (nmol/24 h) measured in urine would have been valuable. Another unavoidable limitation of the present study was that cortisol was not assessed before the brain injury. Finally, a substantial number of patients with even mTBI experience long-term pain, sleep disorders, and mental health conditions, including post-traumatic stress disorder and major depression [29], factors that all can plausibly complicate the interpretation of physiological measurements.

Future implications
Our findings suggest that there is a need to investigate whether changes in saliva cortisol and physiological parameters are due to stress-induced ANS activation or to brain injury or both by measuring melatonin and catecholamines or other mediators. Clearly, it would be useful to measure stress load objectively and subjectively, including catecholamines levels. Also, there is a need to investigate if it is needed therapeutic intervention such as pain control, b-blockade for ANS activation, and anxiolytic administration. Also, there is possibly a need for a longer period of clinical and laboratory monitoring including sleep quality and quantity.
More broadly, taking a more extended longitudinal approach is recommended. With the recovery phase defined as three months after TBI in surviving patients, prolonging the study to include 30 days after the discharge in order to check the adrenal status and wider pituitary function [9] would be valuable. Finally, studies of the consequences of mTBI present a challenge for implementation science. Relatively little is known about circadian profile of cortisol and physiological parameters of mTBI patients in clinical and public health settings [4], which it would be scientifically valuable. It is clear from this study and other research that mTBI has implications on sleep, but the picture complicates when one understands that sleep can in turn loop back and have multiple health implications for the recovering patient.