+ Models
PNEC-1879; No. of Pages 8 Psychoneuroendocrinology (2010) xxx, xxx—xxx
a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / p s y n e u e n
Cortisol response to acute trauma and risk of posttraumatic stress disorder Alexander C. McFarlane a, Christopher A. Barton a,*, Rachel Yehuda b, Gary Wittert c a Centre for Military and Veterans Health, School of Population Health and Clinical Practice, The University of Adelaide, South Australia, Australia b Discipline of Medicine, School of Medicine, The University of Adelaide, South Australia, Australia c Mt Sinai School of Medicine, New York, NY, United States
Received 23 December 2009; received in revised form 6 October 2010; accepted 12 October 2010
KEYWORDS Posttraumatic stress disorder; Cortisol; Dexamethasone; Motor vehicle accident; Depression; Prospective study
Summary This study sought to characterize the variability of the acute cortisol response following trauma and its relationship to posttraumatic stress disorder (PTSD). Forty eight participants were recruited within 24 h of a traumatic accident requiring hospital admission. A saliva sample was collected at 08.00 h and 16.00 h 2 days, 1 month and 6 months after hospital admission, together with 24-h urine collection. Participants completed a dexamethasone suppression test (0.5 mg DEX at 21.00 h) at each follow up, together with self-report questionnaires. The Clinician Administered PTSD Scale (CAPS) was administered at 1 and 6 months to identify PTSD. Prevalence of PTSD was 27% at 1 month and 21% at 6 months. PTSD symptoms at 6 months were negatively correlated with salivary cortisol at 08.00 h on day 2 (r = 0.36, p = 0.04), but positively correlated with 16.00 h cortisols (r = 0.41, p = 0.03). A lower rise in cortisol at 08.00 h on day 2 was associated with an increase in risk of PTSD at both 1 month (OR = 1.411 (1.017, 1.957)) and 6 months (OR = 1.411 (1.066, 1.866)). At 1 month, 70% of participants with PTSD suppressed cortisol to more than 90% of pre-dex levels compared with 25% without PTSD (x2 = 6.77, p = 0.034). Urinary cortisol excretion was not different between groups at any time point. The findings support a hypothesis that sensitization of the HPA axis and enhanced suppression of cortisol following the dexamethasone suppression test are established early in the disease process. # 2010 Elsevier Ltd. All rights reserved.
1. Introduction * Corresponding author at: 2/122 Frome St., Centre for Military and Veterans Health, School of Population Health and Clinical Practice, University of Adelaide, Adelaide, South Australia 5005, Australia. E-mail address:
[email protected] (C.A. Barton).
An increasing body of evidence suggests that mounting an adequate cortisol response at the time of exposure to a traumatic event has a protective effect against posttraumatic stress symptoms (Yehuda, 2002). Yet, practically
0306-4530/$ — see front matter # 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.psyneuen.2010.10.007
Please cite this article in press as: McFarlane, A.C., et al., Cortisol response to acute trauma and risk of posttraumatic stress disorder. Psychoneuroendocrinology (2010), doi:10.1016/j.psyneuen.2010.10.007
+ Models
PNEC-1879; No. of Pages 8
2
A.C. McFarlane et al.
speaking, there are several reasons why it is difficult to study this process in individuals who have been exposed to traumatic events. First, this is a time in which there is substantial disruption in people’s lives and increased chaos as individuals deal with the many details involved in restoring their lives post-trauma. Second, the biological methodologies and instruments for tracking biological changes have not been well standardized for this purpose, and have, in fact, yielded conflicting results. Despite the challenging nature of this work, the research in the aggregate supports the idea that there is an attenuated cortisol response during the peritraumatic and early posttraumatic period that contributes to longer-term disruption of the sympathetic nervous system in those that develop posttraumatic stress disorder (PTSD) (Yehuda, 2002; Ehring et al., 2008; Delahanty and Nugent, 2006). Most recently, Ehring et al. (2008) found that lower levels of salivary cortisol measured in the emergency room within 12 h of a motor vehicle accident predicted greater symptom levels of PTSD and depression 6 months later. Similarly, a study of rape victims that assessed plasma cortisol measured within 51 h of the rape found that those with a prior history of assault had lower cortisol levels and were more likely to develop posttraumatic stress disorder than those without a trauma history (Resnick et al., 1995). Further, another study of motor vehicle accident victims (McFarlane et al., 1997) demonstrated a trend toward lower cortisol levels in those who subsequently showed PTSD compared to those who did not develop the disorder, but cortisol levels were significantly lower than in individuals who developed depression. Another study of 99 accident victims who gave urine samples in the first 15 h following admission showed significantly lower urinary cortisol levels in those diagnosed with acute stress disorder 5 weeks later, compared with those who did not have a disorder in this initial period after the accident (Delahanty et al., 2000). This study also found that the initial cortisol levels were negatively correlated with the subsequent symptoms of intrusion and avoidance. In contrast, a large study of mixed civilian trauma, found no biological predictors of PTSD diagnosed 5 months later (Bonne et al., 2003). In this study, PTSD was diagnosed 5 months following the trauma and survivors with and without PTSD had similar levels of hormones 1 week after the trauma and again at 5 months. Similarly, Shalev et al. (2008) found no relationship between PTSD and plasma cortisol levels assessed at 10 am in the morning, 10 days, 1 month and 5 months after Emergency Department attendance amongst 155 civilians who had experienced mixed traumatic events. The discrepancy in the literature may reflect differences in when cortisol levels are measured, and when PTSD is assessed during the post-trauma period. Although it has generally been assumed that cortisol levels remain stable, this may not be the case as illustrated in a study of 31 UN soldiers who had been exposed to a mine accident (AardalEriksson et al., 2001). A negative correlation was found between morning saliva cortisols and concurrent PTSD at 5 days, but a positive correlation between cortisol level and PTSD symptoms 2 and 9 months later. Indeed, although cortisol levels are easy to measure, they do not always produce stable results in either acute or chronic PTSD, owing to both the fact that cortisol levels might be responsive to environmental perturbations and may
yield different estimates depending on whether integrated (e.g. 24-h urinary samples) or single time point estimates are obtained. A more reliable parameter in studies of hypothalamic—pituitary—adrenal (HPA) axis function in PTSD has been cortisol suppression in response to dexamethasone, which reflects the responsiveness of glucocorticoid receptors mediating negative feedback inhibition. To date, no study has examined the reactivity of the HPA axis to challenge in the immediate aftermath of a traumatic exposure. The present study aimed to prospectively investigate the cortisol response to a traumatic accident in the 24—48 h period of the accident, and then 1 month and 6 months following the trauma to understand the relationship between cortisol production and posttraumatic stress symptoms in the 6 months period of follow up. The degree of HPA axis suppression to low dose dexamethasone 2 days, 1 month and 6 months after the accident was also investigated to study whether super-suppression of cortisol is a vulnerability factor for PTSD or an acquired abnormality. Specifically, it was hypothesized that salivary cortisol level at 08.00 h and 16.00 h would be lower in individuals who developed PTSD 1 month and 6 months after the trauma. The PTSD group was hypothesized to suppress cortisol to a greater extent than those who did not develop PTSD in response to the low dose dexamethasone suppression test.
2. Methods 2.1. Study design The study involved longitudinal follow-up of individuals who experienced acute trauma requiring admission to the Royal Adelaide Hospital, South Australia. Participants completed self-administered questionnaires, structured interviews, and provided saliva and urine samples 2 days, 1 month and 6 months after the trauma. Ethical approval for this study was granted by the University of Adelaide Human Research Ethics Committee and the Royal Adelaide Hospital Ethics Committee.
2.2. Participants Participants were recruited through the accident and trauma department of the Royal Adelaide Hospital by a research nurse between July 1998 and March 2001. All participants were involved in an acute traumatic accident (motor vehicle accident, industrial accident, assault, or domestic accident) and had been admitted. The flow of participants through the study is presented in Fig. 1. All participants were between the ages of 18 and 65, and were required to understand written and spoken English and to have an adequate recall of significant events relating to the trauma. Any individual who was unconscious or amnesic for 15 min or longer immediately prior to or following the accident was excluded from the study. All participants lived within the Adelaide metropolitan area or nearby townships. Patients who were heavily sedated, ventilated in intensive care or unconscious at the time of recruitment were excluded from the study as were patients who were pregnant or planning a pregnancy. Given the likelihood of amphetamines
Please cite this article in press as: McFarlane, A.C., et al., Cortisol response to acute trauma and risk of posttraumatic stress disorder. Psychoneuroendocrinology (2010), doi:10.1016/j.psyneuen.2010.10.007
+ Models
PNEC-1879; No. of Pages 8
The temporal changes in HPA axis sensitization following trauma and risk for PTSD
3
Number Screened Following Acute Trauma (July 1998-March 2001) N=1863
Reason for non-participation 883 - Did not meet the selection criteria 202 - Discharged before seen by the research nurse 174 - Either too old or too young 130 - Lived outside Adelaide Metro area 98 – Admitted to Intensive Care Unit 83 - No memory of the accident 49 – Refused 18 - Too unwell 178 - Other reasons
24 hours
N=48 Recruited N=5 Withdraw by day 2 4 - Not interested to continue 1 - Too unwell to continue
2 days
N=43 Complete follow-up N=6 withdrew by 1 month 5 - Not interested 1 - No reason given
1 month N= 37 Complete follow-up
6 months N= 28 Complete follow-up
Figure 1
N= 9 withdrew by 6 month 3 - Did not attend follow-up 3 - Unable to contact 2 - Not interested 1 - Traveling overseas
Flow diagram for the study.
and opiates interacting with the HPA axis, intravenous drugs users were excluded. After complete description of the study to the participants, written informed consent was obtained. In total, 42% of those recruited did not complete the full protocol. Reasons for non-participation and withdrawal from the study are provided in Fig. 1. There was no difference between those who remained in the study compared to those that withdrew for age ( p = 0.58), gender ( p = 1.0), marital status ( p = 0.45) or employment ( p = 0.79). Individuals who withdraw were no more depressed at 1 month ( p = 0.95) or experiencing more PTSD symptoms (CAPS total score at 1 month p = 0.95).
2.3. Procedure Participants were recruited within 24-h of admission to hospital by the research nurse. The recruitment interview consisted of questions relating to the type of accident, level of pain, and current medication. Details of the nature of the injuries were obtained from hospital records. Blood alcohol levels obtained within 4 h of the accident were also made available to research staff with the permission of the participant. All participants were evaluated within 2 days of the accident, and again 1 month and 6 months later. Follow-up
interviews were performed either at the University of Adelaide or in the participant’s home.
2.4. Data collection The Clinician Administered Posttraumatic Stress Disorder Scale (CAPS-II) (Blake et al., 1995) was used to determine PTSD at the 1 month and 6 month assessments. Standard scoring was used such that a symptom was considered present if it had a frequency of 1 or more and an intensity of 2 or more. PTSD was indicated if there was at least 1 ‘‘B’’ symptom, 3 ‘‘C’’ symptoms and 2 ‘‘D’’ symptoms. Participants also completed the Beck Depression Inventory (BDI) as a continuous measure of symptoms of depression at 1 and 6 months. Current distress associated with PTSD symptoms was examined using the self-report measure, the Impact of Events Scale — Revised (Weiss and Marmar, 1997), at the day 2, 1 month and 6 months assessments. Symptom questions were anchored to the accident in question. Scores were summed to produce an overall total score. Lifetime and 1month point prevalence rates of DSM-IV (American Psychiatric Association, 1994) disorder were assessed using a computerized version of the fully structured, standardized and comprehensive Composite International Diagnostic Interview Version 2.1 (WHO, 1997).
Please cite this article in press as: McFarlane, A.C., et al., Cortisol response to acute trauma and risk of posttraumatic stress disorder. Psychoneuroendocrinology (2010), doi:10.1016/j.psyneuen.2010.10.007
+ Models
PNEC-1879; No. of Pages 8
4
A.C. McFarlane et al.
2.5. Collection of biological data 2.5.1. Saliva samples Saliva samples were collected at day 2, 1 month and 6 months, at 08.00 h, 16.00 h and 08.00 h the next day (post-dexamethasone). Samples were collected using salivette tubes without citric acid (Sarstedt, Australia). Fifteen minutes before saliva was collected, subjects rinsed their mouths with water and did not eat or drink until the sample collection was completed. Samples were spun at 1200 rpm for 3 min to separate saliva from the cotton swab and the saliva was then stored at 20 8C until assay.
Cortisol levels at each assessment were found to be approximately normally distributed and were analysed to identify cross-sectional differences between groups using independent samples t-test. Adjusted p-values were calculated using binominal logistic regression. A bionomial generalised estimating equation (GEE) with log link was used to examine whether salivary cortisol at the day 2 assessments were associated with PTSD at 1 month and 6 months. The GEE was undertaken using SAS (9.1) (SAS Institute Inc., Cary, NC, USA, 2002) with scale parameters held fixed.
3. Results
2.5.2. Dexamethasone Suppression Test Patients were instructed to take a 0.5 mg dose of dexamethasone at 21.00 h on the evening after the 08.00 h and 16.00 h saliva collection at each follow-up assessment. A further saliva sample was then collected at 08.00 h the following morning to determine change in cortisol level. We have shown previously that there is no difference in the suppression of cortisol at 08.00 h when dexamethasone is administered at 21.00 h compared with 23.00 h (Barton et al., 2002).
3.1. Demographic characteristics
2.5.3. Urine samples Subjects were asked to complete a 24-h urine collection, beginning the morning before the research nurse conducted the structured interviews at each of the follow-up visits. This occurred on the day prior to the dexamethasone suppression test. Urine was collected in a 2 L urine collection bottle. The amount of urine passed was measured and urine was stored in a 50 ml urine aliquot at 20 8C until assessed to determine cortisol excretion.
At the 1 month assessment, 10 (27%) participants met criteria for PTSD. The mean total score on the CAPS for those with PTSD at 1 month was 71.6 (SD = 19.8) and for those without PTSD was 19.6 (SD = 13.9). At the 6-month assessment, 28 participants completed the CAPS, of whom 6 (21%) met the criteria for PTSD. The mean score on the CAPS for those with PTSD at 6 months was 67.5 (SD = 19.2), and for those without PTSD was 13.3 (SD = 13.0). Female gender (female sex OR = 8.0, 95% CI 1.4—44.9) was associated with increased risk of PTSD while employment at the time of the accident was protective against PTSD (employed OR = 0.06, 95% CI 0.005—0.6) at 1 month and remained so at 6 months. No other socio-demographic variables were significantly associated with PTSD at 1 month. At the 1 month assessment, 10 participants (27%) scored greater than 14 on the BDI which is indicative of major depression. At 6 months, 5 (19%) scored greater than 14. Of the 10 participants with PTSD at 1 month, 8 (80%) scored greater than 14 on the BDI. At the 6 month follow up, 4 of the 7 (57%) participants with PTSD had BDI score greater than 14. A weak negative correlation was found between BDI total score at 1 month and 6 months and 08.00 h salivary cortisol measured at the day 2 assessment but this did not reach statistical significance at the 0.05 level (r = 0.26, p = 0.068 (1 tailed), r = 0.25, p = 0.11 (1 tailed)). There was no correlation between BDI total score at 1 month and 16.00 h salivary cortisol (r = 0.03, p = 0.445) but a weak positive correlation was found at the 6 month assessment that just failed to reach statistical significance at the 0.05 level (r = 0.33, p = 0.065 (1 tailed)).
2.6. Assay procedures 2.6.1. Determination of salivary cortisol Saliva samples were sent to Mt Sinai Medical School, Bronx New York for analysis of salivary cortisol. Free cortisol in saliva was analysed by commercial radioimmunoassay (Incstarr GammaCoat (125I) — Cortisol RIA kit). Standard curves and cortisol concentration were determined using ICM micromedic system AGC operating software version 2.07. This procedure is associated with an inter-assay and intra-assay coefficient of variation of 6.1% and 2.3%, respectively. 2.6.2. Determination of urinary free cortisol Assays were performed at the Queen Elizabeth Hospital endocrine and metabolism laboratory in South Australia, using a commercial RadioImmunoAssay (RIA) kit (Amerlex). This procedure was associated with an intra-assay coefficient of variation of 5.3%, and an inter-assay coefficient of variation of 8.2%.
2.7. Data analysis Data was entered in to SPSS version 15.0 (SPSS Inc., Chicago, 2006) for analysis. Continuous measures were checked for approximation of the normal distribution. Socio-demographic and other descriptive data were analysed using cross-tabs to identify potential confounding variables.
A total of 48 participants were recruited. The flow of participants through the study is presented in Fig. 1. The mean age of the sample at recruitment was 34 years (SD = 12.72). Most were male (75%) and 46% had never been married.
3.2. Posttraumatic stress disorder (PTSD) following acute trauma
3.3. Relationship between cortisol levels and PTSD Bivariate correlations were conducted between level of cortisol in saliva at 08.00 h and 16.00 h, 24-h urinary free cortisol and PTSD symptoms assessed by the IES-R. A clear correlation was found between cortisol level at day 2 and
Please cite this article in press as: McFarlane, A.C., et al., Cortisol response to acute trauma and risk of posttraumatic stress disorder. Psychoneuroendocrinology (2010), doi:10.1016/j.psyneuen.2010.10.007
+ Models
PNEC-1879; No. of Pages 8
The temporal changes in HPA axis sensitization following trauma and risk for PTSD
5
Table 1 Levels of cortisol in saliva (mg/dl) and 24-h urine excretion (nmol/24 h) of subjects with and those without PTSD, 2 days, 1 month, and 6 months after an acute traumatic event. PTSD positive
PTSD negative
Mean difference
Unadjusted p-value
N
Mean (SD)
N
Mean (SD)
Day 2 assessment 08.00 h saliva cortisol 16.00 h saliva cortisol 24 h UFC
9 9 10
307.9 (221.1) 161.7 (151.1) 182.7 (35.7)
26 22 26
490.1 (249.9) 193.9 (136.1) 192.1 (105.7)
182.1 32.2 9.3
0.061 0.566 0.789
1 month assessment 08.00 h saliva cortisol 16.00 h saliva cortisol 24 h UFC
10 10 9
324.0 (253.5) 201.9 (119.3) 157.2 (120.8)
25 21 24
540.9 (363.7) 222.4 (108.7) 147.1 (96.2)
216.9 20.4 10.1
0.095 0.619 0.803
6 months assessment 08.00 h saliva cortisol 16.00 h saliva cortisol 24 h UFC
6 5 6
315.3 (88.1) 163.6 (61.6) 119.9 (59.0)
22 20 21
496.3 (286.7) 194.5 (216.4) 166.2 (210.8)
181.0 30.9 46.3
0.143 0.758 0.604
a
UFC, urinary free cortisol. a PTSD caseness determined at 1 month assessment.
3.4. Suppression of cortisol following administration of 0.5 mg dexamethasone One third of participants (33.3%) failed to suppress cortisol at 08.00 h in response to 0.5 mg dexamethasone administered at 21.00 h on day 2 (Table 2). By 1 month, most participants with PTSD demonstrated hyper-suppression of cortisol in response to dexamethasone at 21.00 h. Seven of ten participants with PTSD (70%) suppressed cortisol to more than 90% of pre-dex levels, compared with 25% of those who did not develop PTSD (x2 = 6.77, p = 0.034) at the 1 month follow-up. A trend toward hyper-suppression in the PTSD group was still apparent at 6 months (Table 2), but failed to reach statistical significance (x2 = 3.03, p = 0.219).
a
1200
Salivary Cortisol ug/dl
1000
800
600
400
200
0 PTSD +’ve
b
PTSD –‘ve
700 600
Salivary Cortisol ug/dl
PTSD symptoms at 6 months, where lower cortisol excretion in the morning (08.00 h) at day 2 and higher in the afternoon (16.00 h) at day 2 was associated with more symptoms at 6 months, as measured by the IES-R total score (r = 0.36, p = 0.04 and r = 0.41, p = 0.03, respectively). For participants who were PTSD positive at 1 month, salivary cortisol levels on the day 2 assessment at 08.00 h tended to be lower (08.00 h mean difference 182.1 mg/dl, p = 0.061) compared to those who did not have PTSD. The difference in means between those with PTSD and those who did not, approach statistical significance at the 0.05 level at any of the other assessment points (Table 1 and Fig. 2). There was no statistically significant difference in 24-h urinary free cortisol excretion between those with PTSD and those without at each follow-up assessment (Table 1). A binomial GEE with log link was used to examine whether salivary cortisol assessed at day 2 was significantly associated with PTSD status at 1 month and 6 months. Adjusting for gender and age, each 100 mg/dl decrease in salivary cortisol at 08.00 h at day 2 was associated with a 41% increase in the odds of PTSD at 1 month (OR = 1.411, 95% CI 1.017—1.957) and 6 months (OR = 1.411, 95% CI 1.066—1.866).
500 400 300 200 100 0 PTSD +’ve
PTSD -’ve
Figure 2 Salivary cortisol level at (a) 08.00 h at the day 2 assessment and PTSD caseness at 1 month; (b) 16.00 h at the day 2 assessment and PTSD caseness at 1 month.
Please cite this article in press as: McFarlane, A.C., et al., Cortisol response to acute trauma and risk of posttraumatic stress disorder. Psychoneuroendocrinology (2010), doi:10.1016/j.psyneuen.2010.10.007
+ Models
PNEC-1879; No. of Pages 8
6
A.C. McFarlane et al. Table 2 Suppression of salivary cortisol from the dexamethasone suppression test, by PTSD status. Day 2 0—50% suppression 51—89% suppression 90% + suppression
N=2 N=5 N=2
N=9 N=5 N = 10
x2 = 3.7, p = 0.154
1 month 0—50% suppression 51—89% suppression 90% + suppression
N=2 N=1 N=7
N=6 N = 12 N=6
x2 = 6.8, p = 0.034y
6 months 0—50% suppression 51—89% suppression 90% + suppression
N=1 N=1 N=4
N=5 N = 10 N=6
x2 = 3.0, p = 0.219
PTSD diagnosis determined at 1 month assessment. y p < 0.05.
4. Discussion The variability of the acute biological stress response following acute traumatic injuries was studied by measuring 24-h urinary cortisol, salivary cortisol at 08.00 h and 16.00 h and salivary cortisol production in response to administration of 0.5 mg dexamethasone. This was repeated 24—48 h after the accident, and again 1 and 6 months following admission to a major teaching hospital, to determine the relationship between the acute pattern of cortisol reactivity and the emergence of the symptoms of posttraumatic stress disorder. We identified a moderate negative correlation between morning salivary cortisol level and a moderate positive correlation between afternoon salivary cortisol determined 2 days after trauma and PTSD symptoms assessed at 6 month follow-up. Similarly, lower salivary cortisol concentration at 08.00 h was observed 2 days after an acute traumatic event for participants who were PTSD positive on the CAPS 1 and 6 months later, taking into account the effects of gender on risk for PTSD. There was evidence of hyper-suppression of cortisol following the overnight dexamethasone suppression test in participants who developed PTSD. We did not identify any difference in 24-h urinary free cortisol excretion (assessed 24 h, 1 month and 6 months post accident) between participants who developed PTSD and those who did not. Our study is one of an increasing number of studies that have now reported findings of HPA axis dys-function soon after acute trauma and risk of subsequent PTSD. Our results are in line with those of Ehring et al. (2008), McFarlane et al. (1997) and Delahanty et al. (2000), which suggest HPA axis dysfunction is evident shortly after trauma, but are at odds with findings of Bonne et al. (2003) and Shalev et al. (2008) who found no differences in HPA axis function between those who developed PTSD and those who did not. The lack of consistent findings may, in part, be accounted for by the different time windows for the measurement of the cortisol response and also the methods of assessment. One of the negative investigations for example, first examined plasma cortisols 1 week after the trauma in survivors with and without PTSD (Bonne et al., 2003). In general it appears that studies that have looked at the cortisol response within the first 2 days after the trauma have found a negative correla-
tion between PTSD symptoms and cortisol. Similarly, the negative findings of Shalev et al. (2008) occurred when samples were first collected more than a week after the trauma, and furthermore, the morning cortisol sample was collected at 10.00 h. Morning cortisol in our study was determined from a saliva sample collected at 08.00 h, 2 days after the accident. The positive relationship in our data suggest that there may be a window of the first few days after a trauma where there is an effect of cortisol but this has disappeared by the first week. These results also demonstrate the importance of collecting biological samples for determination of cortisol in the afternoon or evening, during the diurnal nadir in cortisol levels, as well as in the morning during the peak in cortisol levels. This is seen from the combination of the negative correlation we found between symptom scores on the Impact of Events Scale and 8 am cortisol level and the contrasting positive correlation with the 4 pm cortisol. Furthermore, this difference may explain the lack of any relationship between the 24-h urinary free cortisol samples and PTSD symptoms in our study. These findings highlight the potential role of some abnormality of the circadian cycle in individuals who are vulnerable to developing PTSD. Our findings should be considered in relation to the proposed role of cortisol on memory consolidation and retrieval (de Quervain et al., 2009) as they provide a different perspective on the role of cortisol in the days after an accident in contrast to levels at the time of the trauma (McFarlane et al., 1997). In particular the beneficial inhibitory effect of higher levels of cortisol on memory retrieval (de Quervain et al., 2009) may explain why there was a positive correlation with the 8 am level but not the 4 pm level due to the inverted U nature of this relationship. de Quervain et al. (2009) emphasise that this effect of cortisol may disrupt the vicious cycle of reliving and reconsolidating aversive memories. This is also indicated by studies involving the administration of stress doses of hydrocortisone following critical illness and major surgery that have been found to reduce the experience of traumatic stress symptoms (Schelling et al., 2006). However, the lack of a significant correlation with the 24 h excretion suggests this is an effect associated with circadian regulation and peak levels rather than total production or lower levels at 4 pm. It may be that the reason why higher levels at 4 pm have a positive correlation with PTSD symptoms is that the amount of cortisol during the afternoon drop is in the range where cortisol has the reverse effect by enhancing consolidation. Our finding of a negative correlation between early morning cortisol and symptoms but a positive correlation between afternoon cortisol and symptoms may be impacted upon by the high level of comorbid depression in the group that developed PTSD. This possibility is indicated from studies that have explored diurnal cortisol variation in twins and vulnerability to depression. One study found that a genetic liability to affective disorders was associated with a high level of evening cortisol but not morning cortisol (Vinberg et al., 2008). Further twin studies suggest that probands with lifetime depression have different diurnal cortisol profiles than those without, suggesting that altered HPA axis functioning is an indicator of depression susceptibility (Wichers et al., 2008). Hence, the higher 16.00 h cortisol levels in our participants may be indicative of this general relationship
Please cite this article in press as: McFarlane, A.C., et al., Cortisol response to acute trauma and risk of posttraumatic stress disorder. Psychoneuroendocrinology (2010), doi:10.1016/j.psyneuen.2010.10.007
+ Models
PNEC-1879; No. of Pages 8
The temporal changes in HPA axis sensitization following trauma and risk for PTSD that indicates a susceptibility to depression and posttraumatic stress disorder. This suggests that the cortisol response in the 48 h after the accident at 4 pm may indicate a specific risk, namely the risk of depression. We found a weak positive correlation between Beck Depression Inventory total score at the 6 months assessment and salivary cortisol level at 16.00 h, however this did not reach statistical significance. As a result of the high level of comorbidity between depression and PTSD in this group, we could not conduct sub-group analyses of those with and without depression. However, a recent meta analysis of HPA axis function in PTSD and depression suggests that comorbid depression in people with PTSD has no influence on the association between PTSD and cortisol levels following traumatic injury, and it is suggested PTSD and depression reflect a shared vulnerability to trauma (Meewisse et al., 2007). A strength of this study is that it is the first investigation to have specifically examined the role of the reactivity of the HPA axis using dexamethasone within 48 h of a traumatic event. Surprisingly, however, there was a significant group of 10 individuals who had greater than 90% suppression of cortisol in response to dexamethasone at the day 2 assessment, who did not go on to develop PTSD. Yet, over time, there was a tendency toward super-suppression of the HPA axis in those individuals who developed PTSD. It remains unclear however, whether super-suppression of the HPA axis to dexamethasone is a vulnerability factor for developing PTSD on exposure, or a consequence of exposure. Some individuals who develop PTSD develop this abnormality after the traumatic exposure, whereas others do not, emphasizing that this abnormality is neither core to the vulnerability to the condition or its psychobiology (Pitman et al., 2006). A further strength of this study was that the subjects were followed for 6 months. While this is longer than most other studies of acute cortisol response, a longer period of followup still would be optimal. A number of studies have demonstrated that cases of PTSD continue to emerge in the 12 months following trauma. Hence, many cases will have been misclassified in most of the cortisol studies to date (Carty et al., 2006). Future studies should include a clinical assessment of PTSD and more frequent collection of biological samples (preferably including pre-trauma function) than has occurred in existing studies, such that changes in HPA axis function can be more accurately described. However our experience would suggest the recruitment and retention of such cohorts from general community settings will be difficult. Investigations involving special groups such as emergency service workers and police or military personnel may be useful as these groups have a much higher probability of exposure to traumatic events than the general population and through their employer are more easily tracked. A final advantage of this study was that recruitment was on the basis of those admitted to a major teaching hospital, some of whom developed PTSD and some of whom did not. Hence, there was a highly comparable, natural control group. This is an important issue because the contradictions that have been found in relation to cortisol in other studies may, in part, arise from the potential bias in the recruitment of control samples. Despite the strengths in the study design, a limitation was the low recruitment rate, resulting in a small number of participants. However, the low recruitment rate is not unex-
7
pected given the timing of the recruitment in the context of serious trauma, the nature of the injuries experienced by eligible participants, the longitudinal and time consuming data collection process, and the makeup of the at risk group being predominately young males. As a result of small numbers, some differences between groups, for example males and females, could not be analysed, and the impact of the stage of the menstrual cycle and alcohol use could not be controlled for. Finally, because the time of morning cortisol collection was standardised at 8 am, and the time of awakening was not recorded, effects of the cortisol awakening rise were not considered. However, the morning cortisol collection was always at least more than 30 min after waking due to the arrival of the research team at the participants home at approximately 07.30 h in order to prepare for the collection of biological samples at 08.00 h. Differences in waking time between participants may have introduced some additional variance in the cortisol levels. In addition, we were unable to reliably obtain samples for assessing cortisol from the emergency room or on admission to the hospital, which has been utilized in a number of other studies and would have provided information about the immediate cortisol response to trauma.
5. Conclusion The findings provide further support of the hypothesis that sensitisation of the HPA axis and enhanced suppression of cortisol following the dexamethasone suppression test are established early in the disease process. Yehuda (2002) has hypothesized that there is a two factor model where the delayed HPA axis response terminates the adrenergic surge immediately following a traumatic event, and via negative feedback, contains the enhancement of emotional recall. The fact that we looked at this in the 24—48 h period after a traumatic accident provides an opportunity to investigate the importance of the dynamics of the HPA axis in a time during which there is potential extinction of any conditioned fear response (Pitman and Delahanty, 2005). Studies with greater sample size and pre-trauma assessment would further clarify the temporal changes in HPA axis sensitization following trauma and risk for PTSD.
Contributors AM and GW devised and gained funding for the study, developed study design and interpreted data. CB was involved in data collection, analysis and interpretation of data and drafted the manuscript together with AM. RY provided critical review of the study design and interpreted data. All authors have read and approved the final manuscript for submission.
Role of funding source The study was funded by a grant from the Australian Government Department of Veterans Affairs. The study sponsor had no further role in the study design, collection, analysis or interpretation of data; in the writing of this manuscript; and in the decision to submit the paper for publication. Professor McFarlane holds a National Health and Medical Research Council Program Grant (300403).
Please cite this article in press as: McFarlane, A.C., et al., Cortisol response to acute trauma and risk of posttraumatic stress disorder. Psychoneuroendocrinology (2010), doi:10.1016/j.psyneuen.2010.10.007
+ Models
PNEC-1879; No. of Pages 8
8
A.C. McFarlane et al.
Conflict of interest The authors have no conflicts of interest to declare.
Acknowledgements The authors wish to thank the study participants, Ms Susan March for study coordination and data collection, and Dr. Nancy Briggs for statistical advice.
References Aardal-Eriksson, E., Eriksson, T., Thorell, L., 2001. Salivary cortisol, posttraumatic stress symptoms, and general health in the acute phase and during 9-month follow-up. Biol. Psychiatry 50, 986—993. American Association Psychiatric, 1994. Diagnostic Criteria from DSM-IV. Amercian Psychiatric Association, Washington. Barton, C., March, S., Wittert, G., 2002. The low dose dexamethasone suppression test: effect of time of administration and dose. J. Endocrinol. Invest. 25, RC10—RC12. Blake, D., Weathers, F., Nagy, L., Kaloupek, D., Gusman, F., Charney D., Keane, T., 1995. The developmet of a clinician-administered PTSD Scale. J. Trauma. Stress 8, 75—90. Bonne, O., Brandes, D., Segman, R., Pitman, R., Yehuda, R., Shalev, A., 2003. Prospective evaluation of plasma cortisol in recent trauma survivors with posttraumatic stress disorder. Psychiatry Res. 119, 171—175. Carty, J., O’Donnell, M., Creamer, M., 2006. Delayed-onset PTSD: a prospective study of injury survivors. J. Affect. Disord. 90, 257—261. de Quervain, D., Aerni, A., Schelling, G., Roozendaal, B., 2009. Glucocorticoids and the regulation of memory in health and disease. Front. Neuroendocrinol. 30, 358—370. Delahanty, D., Nugent, N., 2006. Predicting PTSD prospectively based on prior trauma history and immediate biological responses. Ann. N.Y. Acad. Sci. 1071, 27—40. Delahanty, D., Raimonde, A., Spoonster, E., 2000. Initial posttraumatic urinary cortisol levels predict subsequent PTSD symptoms in motor vehicle accident victims. Biol. Psychiatry 48, 940—947.
Ehring, T., Ehlers, A., Cleare, A., Glucksman, E., 2008. Do acute psychological and psychobiological responses to trauma predict subsequent symptom severieties and PTSD and depression? Psychiatry Res. 161, 67—75. McFarlane, A.C., Atchison, M., Yehuda, R., 1997. The acute stress response following motor vehicle accidents and its relation to PTSD. Ann. N.Y. Acad. Sci. 821, 437—441. Meewisse, M.-L., Reitsma, J., de Vries, G., Gersons, B., Olff, M., 2007. Cortisol and post-traumatic stress disorder in adults. Br. J. Psychiatry 191, 387—392. Pitman, R., Delahanty, D., 2005. Conceptually driven pharmacologic approaches to acute trauma. CNS Spectr. 10, 99—106. Pitman, R., Gilbertson, M., Gurvits, T., May, F., Lasko, N., Metzger, L., Shenton, M., Yehuda, R., Orr, S., 2006. Clarifying the origin of biological abnormalities in PTSD through the study of identical twins discordant for combat exposure. Ann. N.Y. Acad. Sci. 1071, 242—254. Resnick, H., Yehuda, R., Pitman, R., Foy, D., 1995. Effect of previous trauma on acute plasma cortisol level following rape. Am. J. Psychiatry 152, 1675—1677. Schelling, G., Roozendaal, B., Krauseneck, T., Schmolez, M., de Quervain, D., Briegel, J., 2006. Efficacy of hydrocortisone in preventing posttraumatic stress disorder following critical illness and major surgery. Ann. N.Y. Acad. Sci. 1071, 46—53. Shalev, A., Videlock, E., Peleg, T., Segman, R., Pitman, R., Yehuda, R., 2008. Stress hormones and post-traumatic stress disorder in civilian trauma victims: a longitudinal study Part I: HPA axis responses. Int. J. Neuropsychopharmacol. 11, 365—372. Vinberg, M., Bennike, B., Kyvik, K., Andersen, P., Kessing, L., 2008. Salivary cortisol in unaffected twins discordant for affective disorder. Psychiatry Res. 161, 292—301. Weiss, D., Marmar, C., 1997. In: Wilson, J., Keane, T. (Eds.), The Impact of Event Scale-Revised. Assessing Psychological Trauma and PTSD, New York, Guildford. Wichers, M., Myin-Germeys, I., Jacobs, N., Kenis, G., Derom, C., Vlietinck, R., Delespaul, P., Mengelers, R., Peeters, F., Nicolson, N., Vanos, J., 2008. Susceptibility to depression expressed as alterations in cortisol day curve: a cross-twin, cross-trait study. Psychosom. Med. 70, 314—318. Yehuda, R., 2002. Post-traumatic stress disorder. N. Engl. J. Med. 346, 108—114.
Please cite this article in press as: McFarlane, A.C., et al., Cortisol response to acute trauma and risk of posttraumatic stress disorder. Psychoneuroendocrinology (2010), doi:10.1016/j.psyneuen.2010.10.007 View publication stats
