Abnormal diurnal pattern of cortisol secretion in patients after aneurysmal subarachnoid hemorrhage

Abstract

Substantial evidence suggests that impairment of the hypothalamus–pituitary system can occur following an aneurysmal subarachnoid hemorrhage (aSAH). Given that the diurnal cortisol rhythm is primarily controlled by the hypothalamus–pituitary system, this study examined whether changes in diurnal cortisol rhythm occurred after aSAH. Cortisol concentrations were measured in the saliva samples collected from patients after aSAH and other types of cerebral hemorrhage (non-aSAH) in the post-awakening period and at night (21:00 h), and the cortisol awakening response (CAR) and diurnal cortisol decline were determined. The area under the cortisol curve from immediately after to 45 min after awakening (CARauc) in the aSAH patient group was comparable to that in the non-aSAH or healthy control groups. However, an obvious cortisol peak was not found after the awakening period, and the morning/nighttime cortisol ratio in the aSAH patient group was significantly lower than that in other examined groups due to higher nighttime cortisol concentrations. In aSAH patients, the CARauc and nighttime cortisol concentrations were negatively correlated with the Fisher CT grade. These results indicate that the diurnal cortisol rhythm is not regulated normally after aSAH, and cortisol secretory activity decreases as the volume of subarachnoid bleeding increases. Our findings will be helpful to understand altered hypothalamus–pituitary–adrenal axis function after aSAH.

Keywords::

• aSAH
• cortisol awakening response
• diurnal cortisol decline
• HPA axis dysfunction
• hemorrhage severity
• salivary cortisol

Introduction

Aneurysmal subarachnoid hemorrhage (aSAH) is the rupture of a blood vessel into the subarachnoid space, between the arachnoid membrane and the pia mater surrounding the brain. Because the hypothalamus is located adjacent to the circle of Willis, it has been suggested that hypothalamic function might be directly influenced by the rupture of a cerebral aneurysm (Crompton 1963). Vernet et al. (2001) measured levels of pituitary hormones, thyroid and sex steroids several months after aSAH; they found that hypothalamus lesions after aSAH can result in secondary hypopituitarism. In the last few years, other investigators have confirmed the incidence and prevalence of hypopituitarism several months after aSAH or traumatic brain injury; hypopituitarism is often marked by a deficiency in growth hormone combined with a deficiency in either adrenocorticotropic hormone (ACTH) or gonadotropin (Dimopoulou et al. 2004; Kreitschmann-Andermahr et al. 2004).

Additional studies have found high rates of psychosomatic and psychiatric disorders in patients after aSAH; these disorders are often marked by fatigue, lack of initiative, anxiety, and depression and can occur even in patients who had good neurological recovery (Brandt et al. 2002; Powell et al. 2002). Although these psychosomatic and psychiatric disorders are associated with hypopituitarism (Schneider et al. 2007), many of these symptoms are similar to those occurring in patients with adrenal insufficiency (Arlt and Allolio 2003) or altered diurnal rhythm-mediated endogenous cortisol secretion after awakening (Chida and Steptoe 2009). Some studies have found abnormally high nighttime cortisol levels in patients after aSAH (Jenkins 1969; Savaridas et al. 2004). These findings suggest that aSAH patients have abnormal diurnal cortisol secretory rhythms. Overall, little information about the effect of aSAH on HPA axis function is available.

The measurement of cortisol concentrations in saliva has become more common over the past few years, and it is clear that salivary cortisol measurements can reflect the levels of free forms of the steroid in the blood (Vining et al. 1983). The collection of saliva is much easier than the collection of blood through venipuncture, and saliva can be readily collected at frequent intervals. Because of these advantages, saliva can be used to effectively assess the diurnal rhythm-mediated endogenous cortisol secretion that matches an individual's sleep/wake cycle. The diurnal cortisol rhythm is marked by two main characteristics. First, serum and salivary cortisol concentrations peak within 30–45 min after awakening (Wilhelm et al. 2007); this is the cortisol awakening response (CAR). It has been hypothesized that the CAR represents activation of the HPA axis following awakening, because an ACTH peak occurs prior to the initiation of the CAR (Chida and Steptoe 2009). Second, cortisol secretion gradually declines throughout the day, reaching a nadir in the late nighttime (i.e. the diurnal cortisol decline). The CAR and diurnal cortisol decline have been used as indices of subtle changes in HPA axis regulation. The CAR has been studied extensively in healthy populations (Wust et al. 2000), and an altered CAR (i.e. blunted or heightened) and reduced diurnal cortisol decline have been documented in patients with a variety of diseases, including cardiovascular disease, cancer, diabetes, and psychosomatic and psychiatric disorders (Sephton et al. 2000; Edwards et al. 2001; Clow et al. 2004; Fries et al. 2009).

As described above, previous studies have shown hypopituitarism as evidence of impaired hypothalamus function after aSAH. Because the hypothalamus plays a pivotal role in maintenance of the normal diurnal cortisol rhythm, we hypothesized that a ruptured aneurysm in the circle of Willis may disturb the diurnal cortisol rhythm. This study was carried out to examine the CAR and diurnal cortisol decline in patients after aSAH. Cortisol concentrations were determined in repeatedly collected saliva samples from patients after aSAH and from patients after other types of cerebral hemorrhage (non-aSAH). Saliva samples were collected approximately 2–3 weeks after the onset of the hemorrhage and before the patients' discharge from the general neurosurgical ward (GNW). The CAR and nighttime cortisol concentrations in the two patient groups were compared.

Materials and methods

Subjects

This study included patients with cerebral hemorrhage who were admitted between June 2008 and June 2009 to the adult neurosurgery service through the emergency room at the Korea University An-am Hospital (KUAH). Upon admission to the hospital, the clinical severity of the patients was evaluated according to the Glasgow coma scale (GCS; Teasdale and Jennett 1974), Hunt-Hess grade (Hunt and Hess 1968), and Fisher CT grade (Fisher et al. 1980). The diagnosis of cerebral hemorrhage was confirmed by computed tomography (CT) scanning, and angiography was used to diagnose aneurysm and determine its location. Patients were diagnosed and received medical treatment within a few hours of arrival at the hospital. After surgical treatment, patients were transferred to the neurosurgical intensive care unit (NICU). They remained in the NICU for several (4–9) days and then spent an addition 8–14 days in the GNW.

All patients were treated according to the standardized management plan developed by the Neurosurgical Department of Korea University. A calcium channel blocker, nimodipine, was given only to patients with aSAH during their stay in the NICU to prevent cerebral vasospasm; none of the patients received glucocorticoids, anesthetics or other metabolic suppressive agents. Patients were excluded from the study if they had one of the following: (1) a tumor-associated intracerebral hemorrhage and cerebral arteriovenous malformation; (2) a previous history of cerebral hemorrhage; (3) neurological deficits such as paralysis, aphasia, impaired cognition and an altered level of consciousness (i.e. a coma state); (4) other severe comorbid conditions such as cancer or diabetes; and (5) abnormal sleep/wake cycle (i.e. slept all day or frequently woke up at night) while in the GNW.

The control group included healthy subjects, none of whom presented with brain pathologies, diabetes, thyroid disorders or hypo- or hypertension. Individuals taking hormone replacement therapy, hormonal birth control, antidepressant medications or sleeping pills were excluded from the study. However, we did not track other minor psychological or physical health problems in the healthy subjects; for example, problems such as intermittent or minor muscle pain, sleep disorders, gastrointestinal disease and psychological or financial stress were not monitored.

A total of 25 patients with aSAH due to a ruptured aneurysm in the circle of Willis were included in this study, as well as 21 patients with a non-aSAH hemorrhage and 23 healthy subjects (Table I). One aSAH patient was treated by endovascular coiling; the rest of the aSAH patients underwent surgical aneurysm clipping. Two patients with traumatic intracranial hemorrhage (tICH) underwent shunt operations, and the other non-aSAH patients (n = 19) underwent craniotomy for hematoma removal.

Table I.  Characteristics of patients with aSAH and other types of cerebral hemorrhage (non-aSAH).

	aSAH patient(n = 25)	Non-aSAH patient(n = 21)	Healthy control(n = 23)
Age, years (range)	56.0 (24–74)	52.0 (15–81)	44.5 (21–70)
Gender (male/female)	8/17	12/9	11/12
Aneurysm location
ICA	2
MCA	10
AcomA	6
PcomA	7
Type of hemorrhage
tSDH/tSAH		2
sIVH		2
tICH		10
tSAH		3
tSDH		4
Initial Fisher grade
2	12
3	9
4	4
Initial Hunt-Hess grade
1–2	14
3	9
4	2
Initial GCS score
15	9	5
14–13	7	12
12–7	9	4
Duration of stay in NICU/GNW (days)	5.9 ± 1.5/11.5 ± 1.7	5.3 ± 1.4/9.6 ± 2.2
Saliva sampling day after onset of hemorrhage	14.4 ± 2.8	11.9 ± 3.2
Sleep duration (h) at day of saliva sampling	6.4 ± 1.4	6.1 ± 1.3	6.5 ± 0.5

All data represented as mean ± SD. Abbreviations: ICA, internal carotid artery; MCA, middle cerebral artery; AcomA, anterior communicating artery; PcomA, posterior cerebral artery; tSDH, traumatic subdural hemorrhage; tSAH, traumatic subarachnoid hemorrhage; sIVH, spontaneous intraventricular hemorrhage; tICH, traumatic intracranial hemorrhage; GCS, Glasgow coma scale; NICU, neurosurgical intensive care unit; GNW, general neurosurgical ward.

Saliva collection

Saliva samples were collected from patients who had fully recovered consciousness (GCS 15) and had normal nocturnal sleep in the GNW. Saliva samples were collected three days before discharge: 14.4 ± 2.8 (mean ± SD) days after treatment of the hemorrhage in aSAH patients and 11.9 ± 3.2 days after treatment of the hemorrhage in non-aSAH patients.

To avoid the effect of delayed saliva collection on the CAR (Kunz-Ebrecht et al. 2004), each patient, caregiver, and family member who cared for the patient was instructed in the procedures for saliva collection and was asked to alert the medical staff when the patient awoke after nocturnal sleep. The medical staff helped each patient collect saliva samples at the designated times. Each patient, caregiver, and family member was questioned about the patient's sleep duration and the patient's ongoing psychosomatic and psychiatric complaints.

Saliva was collected immediately upon awakening, at 15, 30 and 45 min after awakening and at night (21:00 h) for two consecutive days. Saliva was collected without external stimulation but with muscle movement. Healthy subjects were also instructed to collect saliva samples on two consecutive workdays using the same method as the patients.

All participants were asked not to smoke, eat food or drink any fluids 30 min before collecting the sample; participants were also asked to rinse their mouth with water and refrain from brushing their teeth 30 min before sample collection. Cortisol concentrations in saliva are stable when the sample is stored at − 20 to − 80°C for up to 1 year, and repeated freezing and thawing of samples up to four times before analysis does not affect the measured concentrations of cortisol (Groschl et al. 2001); healthy subjects were therefore asked to keep their saliva samples in their own domestic freezers before submitting samples to the neurology laboratory. The healthy subjects' samples were excluded if they were collected outside of the designated time or were not correctly labeled with the collection time and date.

After debris was removed, the collected samples were stored at − 70°C. To precipitate mucins, the samples were thawed and centrifuged (10,000g, 15 min, 4°C; Gozansky et al. 2005). The supernatant was collected and stored at − 70°C until the assay was performed. All saliva samples were analyzed at the Hormone Research Center at Chonnam National University.

All participants gave informed consent and were given information about their hormone levels. This study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the hospital's institutional review board (KUAH).

Measurement of salivary cortisol

Cortisol concentrations in saliva were determined using a radioimmunoassay as previously described (Ahn et al. 2007; Kim et al. 2010). Exogenous 5.5 and 22.1 nmol/l cortisol added to charcoal-stripped saliva were determined to be 5.4 ± 0.7 nmol/l (n = 20) and 22.9 ± 1.9 nmol/l (n = 20), respectively. The inter-assay coefficients of variation (CVs), as assessed from quality controls with mean cortisol concentrations of 2.8, 5.3, and 13.7 nmol/l were 7.8, 10.2, and 14.0%, respectively (n = 26). Intra-assay CVs for the same pool were less than 10% (n = 15). The analytical sensitivity for cortisol was 1 nmol/l.

Data analysis

In all samples, cortisol values were lower than 75 nmol/l. This indicates that samples were not contaminated with blood and did not have altered pH (Steptoe et al. 2004). As previous studies suggested (Wust et al. 2000; Kunz-Ebrecht et al. 2004), a normal CAR in this study was defined as a mean increase in cortisol concentrations between awakening and 30 min after awakening of at least 2.5 nmol/l.

To further analyze the CAR, the mean increases in cortisol concentrations after the post-awakening period (MnInc) was calculated according to the following formula: [Awakening Cortisol (AC)_{15 min} + AC_{30 min} + AC_{45 min}]/3–AC_{0 min} (Wust et al. 2000), and the overall cortisol secretion in the first 45 min after awakening was also calculated as the area under the cortisol curve from immediately after to 45 min after awakening (CARauc) using GraphPad Prism version 5.01 for Windows (CA, USA). To further analyze the diurnal cortisol decline, the molar morning to nighttime cortisol ratio (molar M/N-F ratio, [(AC_{0 min} + AC_{15 min} + AC_{30 min} + AC_{45 min})/4]/[nighttime cortisol]) was calculated. The declining slope of cortisol concentrations from morning to nighttime in each examined groups was analyzed. To do this, the mean cortisol concentrations after the awakening period [(AC_{0 min} + AC_{15 min} + AC_{30 min} + AC_{45 min})/4] and the mean nighttime cortisol concentrations were calculated, then the slope of cortisol decline in each examined group was calculated by linear regression using GraphPad Prism.

The difference in cortisol concentrations between the first and second days of collection was analyzed using a paired t-test. A two-way repeated-measures analysis of variance (ANOVA) was applied to reveal the possible effects of the different variables (collection time and collection day) on cortisol concentrations at each time point. Differences in cortisol concentrations among examined time points were compared with a one-way AVOVA. A one-way ANOVA was also used to compare differences in cortisol concentrations and auxiliary indices (the MnInc, CARauc, and M/N-F ratio) among the examined groups (aSAH, non-aSAH, and healthy controls). When the one-way ANOVA indicated a significant difference (p < 0.05), a Tukey's post-hoc multiple comparisons test was performed to locate specific group differences.

To assess the stability of cortisol concentrations at each measurement and the stability of auxiliary indices over 2 days, Pearson's correlations were computed between cortisol concentrations across the two consecutive sampling days using GraphPad Prism. To assess the relationship between auxiliary indices and clinical variables (GCS score, Hunt-Hess grade and Fisher CT grade), Spearman's r was computed between auxiliary indices and clinical variables using GraphPad Prism 5.01 for Windows (GraphPad Software, San Diego, CA, USA). Statistical calculations were performed using SAS version 9.1 (SAS Institute, Inc., Cary, NC, USA). All results are presented as the mean ± standard deviation (SD). A p-value < 0.05 was considered significant.

Results

Differences in the CAR among the examined groups

To assess the CAR, cortisol concentrations were determined in the saliva samples collected in the morning immediately upon awakening and at 15, 30, and 45 min after awakening in both patient groups (aSAH and non-aSAH group) and in the healthy control group. Repeated-measures two-way ANOVA revealed that the time of saliva collection had a significant effect on cortisol concentrations in the control (F_3,168 = 22.23, p < 0.01) and non-aSAH patient groups (F_3,128 = 4.43, p < 0.01) but did not have a significant effect in the aSAH patient group (F_3,192 = 0.29, p>0.05). The day of saliva collection did not significantly affect cortisol concentrations at any time point during the awakening period in the control (F_1,168 = 1.04, p>0.05), non-aSAH patient (F_1,128 = 1.82, p>0.05) and aSAH patient group (F_1,192 = 0.04, p>0.05). Therefore, cortisol concentrations at each time point after awakening in each examined group were averaged over the 2 days for further analysis, and mean values are presented in Figure 1.

Figure 1.  Changes in salivary cortisol concentrations after awakening in (A) healthy controls, (B) non-aSAH patients, and (C) aSAH patients. The cortisol concentrations were determined from saliva samples collected immediately upon awakening and 15, 30, and 45 min after awakening from healthy controls (n = 23), non-aSAH patients (n = 21), and aSAH patients (n = 25). Each box represents the interquartile range, whiskers represent range, and horizontal lines and the cross symbol within boxes represent median and mean values of cortisol, respectively. Differences in mean cortisol concentrations in each panel were compared with a one-way ANOVA, and Tukey's post-hoc multiple comparisons test was performed to locate specific group differences. Boxes with different letters in each panel are significantly different from each other (p < 0.05). aSAH: aneurysmal subarachnoid hemorrhage.

As seen in Figure 1A,B, cortisol concentrations increased and reached a peak value 30 min after awakening and then decreased in the control and non-aSAH patient group. Cortisol concentrations 30 min after awakening were significantly higher than those immediately upon awakening in both the control and non-aSAH groups (p < 0.05 for all analyses by Tukey's post-hoc test) (Figure 1A,B). Within 30 min after awakening, the absolute increase was 10.4 ± 4.8 and 10.4 ± 8.3 nmol/l in the healthy control and the non-aSAH patient group, respectively. Specifically, all healthy subjects and most non-aSAH patients (20/21, 95.2%) exhibited a normal CAR (at least a 2.5 nmol/l increase in cortisol concentrations within the first 30–45 min after awakening).

However, cortisol concentrations did not increase in the aSAH patient group (Figure 1C), and no obvious peak was found after the awakening period. One-way ANOVA revealed that the cortisol concentrations at each time point were not significantly different from those at the other time points in the aSAH patient group (F_3,96 = 0.17, p>0.05). Specifically, most aSAH patients (23/25, 92%) did not exhibit a normal CAR.

Cortisol concentrations immediately upon awakening were significantly different among the three groups (F_2,66 = 5.16, p < 0.01 by one-way ANOVA). Post-hoc analysis indicated that cortisol concentrations immediately upon awakening in aSAH patients (15.4 ± 10.7 nmol/l) were comparable to those in the control group (10.8 ± 5.0 nmol/l) (p>0.05) but significantly different from those in the non-aSAH patient group (8.2 ± 5.7 nmol/l) (p < 0.05). Interestingly, cortisol concentrations immediately upon awakening in the aSAH group were not significantly different from those 30 min after awakening in other groups (non-aSAH patient group: 18.0 ± 10.0 nmol/l; control group: 20.7 ± 7.3 nmol/l) (F_2,66 = 0.17, p>0.05 by one-way ANOVA). There was a strong correlation in cortisol concentrations at each time point between the first and second collection days in the control group (Pearson's r = 0.76–0.80, p < 0.001) and both patient groups (Pearson's r = 0.63–0.86, p < 0.001).

The auxiliary indices (the CARauc, MnInc and M/N-F-ratio) and nighttime cortisol concentrations in the examined groups

Table II shows the auxiliary indices (the CARauc, MnInc, and M/N-F-ratio) and nighttime cortisol concentrations in the examined groups and patient subgroups according to the initial severity of each clinical variable.

Table II.  The auxiliary indices and nighttime salivary cortisol concentrations in the examined groups and classified patient subgroups according to initial severity grades for each clinical variable.

	CARauc(nmol/l)	MnInc(nmol/l)	Nighttime SAF(nmol/l)	M/N-F ratio
Healthy controls
n = 23	774.1 ± 251.1	8.9 ± 4.3	3.0 ± 1.6	6.0 ± 2.8
aSAH patients
Overall (n = 25)	640.0 ± 418.4	− 1.3 ± 7.1	12.1 ± 12.3	2.3 ± 2.5
Initial Fisher CT grade
2 (n = 12)	892.2 ± 349.9	− 2.6 ± 9.7	14.2 ± 12.0	3.2 ± 3.2
3 (n = 9)	533.8 ± 339.9	− 0.1 ± 4.2	13.9 ± 13.8	1.1 ± 0.4
4 (n = 4)	125.5 ± 48.1	− 0.4 ± 1.3	1.9 ± 1.3	2.2 ± 2.1
Initial Hunt-Hess grade
1–2 (n = 14)	740.3 ± 406.2	0.4 ± 6.4	11.8 ± 11.5	2.9 ± 3.1
3 (n = 9)	527.9 ± 425.4	− 3.6 ± 6.2	13.6 ± 14.8	1.5 ± 1.4
4 (n = 2)	448.3 ± 531.2	− 3.3 ± 6.2	7.4 ± 9.2	1.5 ± 0.2
Initial GCS score
15 (n = 9)	825.6 ± 452.1	0.1 ± 7.8	15.1 ± 14.2	2.9 ± 3.5
14–13 (n = 7)	572.8 ± 392.0	− 3.4 ± 10.4	13.6 ± 12.5	1.1 ± 0.5
12–7 (n = 9)	511.2 ± 413.4	− 1.3 ± 3.2	9.3 ± 11.3	2.1 ± 1.7
Non-aSAH patients
Overall (n = 21)	603.8 ± 229.0	5.9 ± 5.3	5.1 ± 3.4	3.6 ± 2.9
Initial GCS score
15 (n = 5)	484.0 ± 247.5	3.3 ± 2.9	4.6 ± 2.9	4.2 ± 2.4
14–13 (n = 12)	653.0 ± 317.5	5.3 ± 4.7	6.1 ± 3.6	3.0 ± 2.2
12–7 (n = 4)	581.3 ± 318.6	8.0 ± 8.6	4.2 ± 2.8	4.4 ± 4.7

All data represented as mean ± SD. Abbreviations: aSAH, aneurysmal subarachnoid hemorrhage; CARauc, the area under the time-concentration curve from immediately upon awakening to 45 min after awakening; MnInc, mean increases in cortisol levels after awakening period; SAF, salivary cortisol; M/N-F ratio, morning to nighttime cortisol ratio; GCS, Glasgow coma scale.

The CARauc in the aSAH patient group was comparable to those of the other examined groups (F_2,66 = 1.62, p>0.05 by one-way ANOVA). In aSAH patients, there was a significant difference in the CARauc among Fisher CT grades (F_2,22 = 9.29, p < 0.05). Post-hoc analysis indicated that the CARauc in aSAH patients with Fisher CT grade 4 was significantly lower than that in aSAH patients with Fisher CT grades 2 and 3 (p < 0.05 for all analyses). In contrast, one-way ANOVA revealed that the CARauc did not significantly vary with the Hunt-Hess grades in aSAH patients (F_2,22 = 0.93, p>0.05) or with the GSC scores in either aSAH (F_2,22 = 1.42, p>0.05) or non-aSAH patients (F_2,18 = 0.49, p>0.05).

There was a significant difference in the MnInc among examined groups (F_2,66 = 19.48, p < 0.05 by one-way ANOVA). Post-hoc analysis indicated that the MnInc in the aSAH patient group was significantly lower than that in other examined groups (p < 0.05, for all analyses). However, the MnInc did not differ according to the severity of each clinical variable in either aSAH (all F values ≤ 0.97, df = 2, 22, all p values >0.05) or non-aSAH patients (F_2,18 = 0.53, p>0.05).

There was a significant difference in the nighttime cortisol concentrations among examined groups (F_2,66 = 8.89, p < 0.05 by one-way ANOVA). Post-hoc analysis indicated that nighttime cortisol concentrations in the aSAH patient group were significantly higher than those in other examined groups (p < 0.05 for all analyses). Meanwhile, the determined cortisol concentrations after the awakening period and at nighttime in aSAH patients were comparable to each other (F_4,120 = 0.35, p>0.05 by one-way ANOVA). Due to higher nighttime cortisol concentrations, the morning to nighttime cortisol decline slope in the aSAH patients group (slope = − 2.30) was flatter than that in other examined groups (non-aSAH group = − 7.60, healthy control = − 12.0) (Figure 2). Post-hoc analysis indicated that the M/N-F ratio in the aSAH patient group was comparable to that in the non-aSAH group (p>0.05) but significantly lower than that in the healthy control group (p < 0.05). No difference was found in nighttime cortisol concentrations among clinical grades for each variable in both the aSAH (all F values ≤ 1.74, df = 2, 22, all p values >0.05 by one-way ANOVA) and non-aSAH patients (F_2,18 = 1.31, p>0.05 by one-way ANOVA). The M/N-F ratio was also comparable among clinical grades in aSAH (all F values ≤ 1.89, df = 2, 22, all p values >0.05 by one-way ANOVA) and non-aSAH patients (F_2,18 = 0.54, p>0.05 by one-way ANOVA).

Figure 2.  Decline of salivary cortisol concentrations from morning to nighttime in each examined group. The cortisol concentrations were determined from saliva samples collected immediately upon awakening and 15, 30, and 45 min after awakening from healthy controls (n = 23), non-aSAH patients (n = 21), and aSAH patients (n = 25), and the cortisol concentrations (mean ± SEM) after the awakening period and at nighttime of each examined group are depicted. Each slope was calculated by linear regression (y_control = − 12.9x+28.8, y_non-aSAH = − 7.6x+20.3, y_aSAH = − 2.3x+16.7). aSAH: aneurysmal subarachnoid hemorrhage.

There was a negative correlation between Fisher CT grade and the CARauc (Spearman's r = − 0.66, p < 0.001) and between Fisher CT grade and nighttime cortisol concentration (Spearman's r = − 0.39, p < 0.05). However, a significant correlation was not found between Fisher CT grade and the MnInc or the M/N-F ratio (p>0.05 for all Spearman's correlation analyses). The CARauc and nighttime cortisol concentrations decreased slightly as the Hunt-Hess grade increased; CARauc and nighttime cortisol concentrations also increased as the GCS score increased in aSAH patients. However, no correlation was found among these trends (p>0.05 for all Spearman's correlation analyses). There was no relationship between Hunt-Hess grade and other auxiliary indices (the MnInc and M/N-F ratio) or between GCS score and the auxiliary indices in aSAH patients (p>0.05 for all Spearman's correlation analyses). No relationship was found between GCS score and auxiliary indices in non-aSAH patients.

Discussion

The results of this study show abnormal diurnal rhythm-mediated endogenous cortisol secretion in patients after aSAH due to a ruptured aneurysm in the circle of Willis. Specifically, we have demonstrated that the CAR was eliminated and the diurnal cortisol decline was reduced in the aSAH patient group. Furthermore, the CARauc and nighttime cortisol concentrations were negatively correlated with Fisher CT grade, but no relationship was found between other auxiliary indices and Fisher CT grade in patients after aSAH.

Although we found that the CAR was absent in most patients after aSAH, four potential factors could be responsible for the loss of the CAR in these patients: (1) sleep quality and other factors; (2) medication history before and after admission (antihypertensives and nimodipine); (3) duration of stay in the intensive care unit (ICU); and (4) lack of compliance. Sleep disorders have been demonstrated to induce only small increases or decreases in the CAR (Backhaus et al. 2004), and some other factors, such as quality of sleep, sleep duration, nightly awakenings, nap taking, age, gender, and light exposure, have also been shown to influence the CAR, which may be blunted or heightened (Clow et al. 2004; Fries et al. 2009). However, none of these factors have been demonstrated to cause the CAR to disappear, as observed in this study. It is also known that a calcium channel blocker (nimodipine) and antihypertensives do not have a strong effect on ACTH, cortisol levels or the CAR (Adams et al. 1988; Wolf et al. 2005). However, adrenal insufficiency and a lack of response to ACTH stimulation have frequently been found in patients with a prolonged ICU stay (Dimopoulou et al. 2007). Notably, the patients included in this study stayed in the NICU for 3–9 days, and the duration of stay in the NICU was comparable between aSAH and non-aSAH patients. However, the typical CAR was absent only in patients after aSAH. Therefore, this finding suggests that the duration of stay in the NICU was not a primary cause of the absent CAR in patients after aSAH. Because the determination of the CAR is critically dependent on the collection time after awakening (Kunz-Ebrecht et al. 2004), patients, caregivers, and family members were instructed to alert medical staff immediately the patient awakened after nocturnal sleep, and medical staff helped and guided the collection of saliva samples after the awakening period. Thus, collection of saliva samples occurred at the designated time points. Cortisol concentrations after the awakening period in non-aSAH patients were also measured from saliva collected at the designated time points in the present study.

To the best of our knowledge, an absent CAR has only been reported previously in patients with hippocampus damage (Buchanan et al. 2004), severe global amnesia (Wolf et al. 2005) and Asperger syndrome (Brosnan et al. 2009). Anoxia, encephalitis or severe brain injury was the major cause of hippocampus damage and severe global amnesia in these studies. In all the patient groups of previous studies, the cortisol levels that were determined 0 and 30 min after awakening were comparable to baseline (0 min) levels of each corresponding control group, and cortisol levels during the remainder of the day were decreased. In the present study, the cortisol levels determined at each time point (0, 15, 30, and 45 min) after awakening in the aSAH patient group were comparable to the peak cortisol concentrations (30 min after awakening) of the healthy control group, and cortisol levels after the awakening period and at nighttime in aSAH patients were comparable to each other. Thus, cortisol profiles (i.e. elevated after awakening period or at nighttime) in aSAH patients were distinctly different from those in patients of previous studies.

In addition to absence of the CAR, reduced diurnal cortisol decline due to higher nighttime cortisol concentrations was observed in the distinctive cortisol profiles in aSAH patients. Abnormal diurnal cortisol decline has received attention because absent or abnormal diurnal cortisol slope has been found in patients with an acute and severe stress state (Bornstein et al. 1998; Baker et al. 2000), and is associated with earlier mortality from metastatic breast cancer (Sephton et al. 2000). Although we did not determine the diurnal cortisol decline with additional measurement of the cortisol concentrations in the samples collected in the daytime, we assumed that the aSAH patients had a flattened diurnal cortisol slope because their mean M/N-F ratio was 2.3. Abnormal diurnal cortisol decline is defined as a morning/nighttime molar ratio of cortisol less than 2 and is used to identify patients with Cushing's syndrome (Raff et al. 1998). In this study, 20 out of 25 aSAH patients and 7 out of 21 non-aSAH patients showed this low ratio. The M/N-F ratio and incidence of abnormal cortisol decline in aSAH patients in the present study are consistent with the findings of previous studies of patients within the first month after treatment for aSAH (Jenkins et al. 1969; Savaridas et al. 2004). An absence of the CAR and a reduced diurnal cortisol decline due to higher nighttime cortisol concentrations in most aSAH patients implies that the diurnal rhythm of cortisol secretion is not normally regulated in aSAH patients.

Three specialized scores are used to evaluate the severity of a brain injury. In brief, the GCS or Hunt-Hess grade is a neurosurgical scale of a patient's conscious state or clinical condition, respectively, and the Fisher grade is a scale for the volume of subarachnoid blood on a CT scan. It has been reported that the total- and calculated-free cortisol concentrations at 1–2 weeks after aSAH were not associated with any clinical variables (GCS, Hunt-Hess grade or Fisher CT grade; Bendel et al. 2008; Poll et al. 2010). In the present study, the CARauc and nighttime cortisol levels were slightly decreased as the Hunt-Hess and GCS worsened, but these parameters were not significantly correlated with either of the clinical variables. However, the CARauc and nighttime cortisol concentrations were decreased as the Fisher CT grade worsened, and there was a strong relationship between the CARauc and Fisher CT grade and between the nighttime cortisol levels and Fisher CT grade in aSAH patients. The results indicated that cortisol secretory activity is negatively associated with the volume of subarachnoid bleeding, rather than with consciousness, after aSAH. It is conceivable that a decreased hypothalamic perfusion contributed to this result because patients with hypothalamic hypoperfusion exhibited a lower cortisol secretory activity 14–21 days after aSAH (Poll et al. 2010).

Cortisol concentrations that were consistently elevated after the awakening period or at nighttime were considered to be a primary cause for the absence of the CAR and the reduced diurnal cortisol decline in aSAH patients. Although the secretion of cortisol by the adrenal glands is largely regulated by pituitary ACTH, there is evidence that HPA axis function is modulated by non-ACTH factors such as neurotransmitters, neuropeptides, cytokines, and growth factors (Bornstein and Chrousos 1999). In conditions involving a clinical situation and chronic stress, these non-ACTH factors modulate HPA axis function at the pituitary or adrenal level and induce the hypersecretion of cortisol without increasing the ACTH level: that is, ACTH secretion is dissociated from the cortisol level (Pignatelli et al. 1998; Bornstein et al. 2008). Some studies have reported that non-ACTH factors (cytokines and neurotransmitters) in the cerebrospinal fluid are elevated and that the sympathetic nervous system is activated during the first few days or weeks after aSAH (Kawamata et al. 1994; Naredi et al. 2000, 2006; Fassbender et al. 2001). A dissociation of ACTH from the cortisol levels within the first 1–2 weeks after an aSAH has already been reported in previous studies (Bendel et al. 2008; Poll et al. 2010). Because an ACTH peak occurs prior to the initiation of the CAR (Wilhelm et al. 2007), the consistently elevated cortisol levels after the awakening period indicated that there was no an obvious ACTH peak after the awakening period. It is conceivable that there are consistently higher ACTH levels after the awakening period. However, because ACTH is secreted in a pulsatile manner and has a short half-life (Murphy et al. 1998), it is reasonable to hypothesize that non-ACTH factors might be associated with the elevated cortisol levels of aSAH patients, which result in the disruption of the normal diurnal cortisol rhythm of aSAH patients.

The detailed regulatory mechanism of the CAR has not yet been fully elucidated. The hippocampus, a memory-relevant brain region, has received attention regarding the regulation of the CAR (Clow et al. 2009; Fries et al. 2009) because, as described above, an absence of the CAR was found in patients with a damaged hippocampus (Buchanan et al. 2004). The hippocampus is a target for cortisol and also plays a role in the negative feedback regulation of the HPA axis (Jacobson and Sapolsky 1991). In animal studies, lesions of the hippocampus are associated with elevation of basal glucocorticoid levels (Fischette et al. 1980), and hippocampal function appears to be compromised or deactivated by exposure to elevated glucocorticoid levels, as in intensive stress (Reagan and McEwen 1997; Sung et al. 2009). Thus, impairment of the hippocampus is a result of, as well as a contributory cause of, elevated basal glucocorticoid levels. A smaller hippocampal volume has been reported in clinical patients with conditions such as Cushing's syndrome (Starkman et al. 1992), major depression (Sheline et al. 1996), and schizophrenia (Bogerts et al. 1993); reduced diurnal cortisol decline was commonly found in these patients (Raff et al. 1998; Mittal et al. 2007; Hinkelmann et al. 2009). Unfortunately, the available information regarding the impairment of hippocampal function after aSAH is limited. Reduced volume of the hippocampus has only been reported in patients 1 year after aSAH (Bendel et al. 2006). This finding indicates that histological and physiological changes may occur in the hippocampus after aSAH. Thus, the consistently elevated cortisol levels after the awakening period and the nighttime in aSAH patients are considered to be a potential result of the reduced inhibitory role of the hippocampus in HPA axis function after an aSAH, as well as a contributory cause of the impaired hippocampus function. However, more sophisticated evidence needs to be gathered to elucidate the dysregulation of the cortisol secretory activity in aSAH patients.

This study has some limitations. The first concern is the number of patients. Saliva sampling requires the patients to be alert and able to deposit saliva in a collecting tube. Patients with an altered consciousness or sleep/wake cycle were therefore excluded from this study. For these reasons, the overall number of patients was limited, especially of the severe aSAH patients (Fisher CT grade of 4) and non-aSAH patients (GCS equal to or less than 8). Secondly, psychosomatic and psychiatric disorders such as fatigue, depression, and anxiety are known to be prevalent in patients after aSAH or other brain injury (Brandt et al. 2002; Schneider et al. 2007). Each patient usually had two or more symptoms, and the sample size was relatively small. Therefore, we were unable to demonstrate a relationship between cortisol secretory activities and the patient's minor health issues. Our results must be considered to be preliminary, and further investigation is warranted to better elucidate this issue. Thirdly, most aSAH patients exhibited an absent CAR coupled with a reduced diurnal cortisol decline, and most non-aSAH patients exhibited a typical CAR and a normal diurnal cortisol decline. There were some exceptions in the aSAH patients and non-aSAH patients, as described above. We attempted to determine the common denominator in these exceptional cases, but we could not identify any factors (e.g. aneurysmal site, type of hemorrhage, clinical variables, age or gender) that contributed to these results because the number of exceptional cases was too small for a statistical analysis. Further population-based studies are required to confirm this issue.

In summary, the HPA axis function in patients after aSAH is poorly understood. The present study provides information regarding dysregulation of the diurnal cortisol rhythm in aSAH patients. Our findings must be considered to be preliminary, but they will be helpful in further prospective investigations of the HPA axis function after aSAH.

Acknowledgements

This work was supported by the Grant of the Korean Ministry of Education, Science and Technology (The Regional Core Research Program/Biohousing Research Institute).

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.